1: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

2: %%%%   12/30/2004  ver 1.0%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

3: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

4: 

5: \documentclass[12pt]{article}

6: \usepackage{amsmath}

7: \usepackage{amssymb}

8: \usepackage{graphicx}

9: \usepackage{ifthen}

10: %\usepackage{showkeys}

11: \usepackage{verbatim}

12: %\usepackage{psfrag,graphicx,subfigure,amssymb}

13: %\usepackage{psfrag,graphicx,amssymb}

14: 

15: 

16: 

17: \renewcommand{\baselinestretch}{1.4}

18: \textwidth     170mm

19: \textheight    220mm

20: \topmargin      -5mm

21: \oddsidemargin  -2mm

22: \evensidemargin  0mm

23: 

24: 

25: 

26: \newlength{\captionwidth}

27: \setlength{\captionwidth}{0.7\textwidth}

28: 

29: 

30: 

31: %%% NEWCOMMAND %%%

32: \newsavebox{\tempbox}

33: \newcommand{\mycaption}[2]{%

34: \par\vspace{10pt}\sbox{\tempbox}{Figure #1: #2}%

35: \ifthenelse{\lengthtest{\wd\tempbox>\captionwidth}}%

36: {\sbox{\tempbox}{Figure.#1:\ }%

37: \addtolength{\captionwidth}{-\wd\tempbox}%

38: \mbox{Figure #1:\ }\parbox[t]{\captionwidth}{\small\textit{#2}}}%

39: {Figure #1: {\small\textit{#2}}}}%

40: 

41: 

42: 

43: 

44: 

45: 

46: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%

47: \makeatletter

48:   \renewcommand{\theequation}{\thesection.\arabic{equation}}

49:   \@addtoreset{equation}{section}

50: \makeatother

51: 

52: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%

53: \setcounter{page}{0}

54: 

55: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

56: %%%%%%%%%%%%%%%%%%%%%  BEGIN DOCUMENT  %%%%%%%%%%%%%%%%%%%%%

57: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

58: \allowdisplaybreaks

59: \begin{document}

60: \thispagestyle{empty}

61: %%%%%%%%%%%%%%%%%%%%%%

62: %%%%% TITLE PAGE %%%%%

63: %%%%%%%%%%%%%%%%%%%%%%

64: \begin{flushright}

65: \texttt{hep-th/0412329}\\

66: OU-HET 513 \\

67: December 2004

68: \end{flushright}

69: \bigskip

70: \bigskip

71: %%%%% TITLE %%%%%

72: \begin{center}

73: {\Large \textbf{Free Fermion and Seiberg-Witten Differential}}

74: \end{center}

75: \begin{center}

76: {\Large \textbf{in Random Plane Partitions}}

77: \end{center}

78: \bigskip

79: \bigskip

80: %%%%% AUTHOR %%%%%

81: \renewcommand{\thefootnote}{\fnsymbol{footnote}}

82: \begin{center}

83: Takashi Maeda\footnote{E-mail: 

84: \texttt{maeda@het.phys.sci.osaka-u.ac.jp}}$^1$, 

85: Toshio Nakatsu\footnote{E-mail: 

86: \texttt{nakatsu@het.phys.sci.osaka-u.ac.jp}}$^1$, 

87: Kanehisa Takasaki\footnote{E-mail: 

88: \texttt{takasaki@math.h.kyoto-u.ac.jp}}$^2$

89: and 

90: Takeshi Tamakoshi\footnote{E-mail: 

91: \texttt{tamakoshi@het.phys.sci.osaka-u.ac.jp}}$^1$\\

92: \bigskip

93: {\small 

94: \textit{$^1$Department of Physics, Graduate School of Science, 

95: Osaka University,\\

96: Toyonaka, Osaka 560-0043, Japan\\

97: $^2$Graduate School of Human and Environmental Studies, 

98: Kyoto University, \\

99: Yoshida, Sakyo, Kyoto 606-8501, Japan

100: }}

101: \end{center}

102: \bigskip

103: \bigskip

104: \renewcommand{\thefootnote}{\arabic{footnote}}

105: %%%%% ABSTRACT %%%%%

106: \begin{abstract}

107: A model of random plane partitions which describes  

108: five-dimensional $\mathcal{N}=1$ supersymmetric 

109: $SU(N)$ Yang-Mills is studied. 

110: We compute the wave functions of fermions 

111: in this statistical model and investigate 

112: their thermodynamic limits or the semi-classical behaviors. 

113: These become of the WKB type at the thermodynamic limit. 

114: When the fermions are located at the main diagonal 

115: of the plane partition, their semi-classical wave functions 

116: are obtained in a universal form. 

117: We further show that 

118: by taking the four-dimensional limit 

119: the semi-classical wave functions turn to live 

120: on the Seiberg-Witten curve and that the classical action 

121: becomes precisely the integral of the Seiberg-Witten differential. 

122: When the fermions are located away from the main diagonal, 

123: the semi-classical wave functions depend on 

124: another continuous parameter.  

125: It is argued that they are related with the wave functions 

126: at the main diagonal by the renormalization group 

127: flow of the underlying gauge theory. 

128: \end{abstract}

129: 

130: 

131: 

132: 

133: \setcounter{footnote}{0}

134: \newpage

135: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

136: %

137: %  TEXT

138: %

139: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

140: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

141: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

142: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

143: \section{Introduction}

144: 

145: 

146: 

147: 

148: Recently it becomes possible 

149: \cite{Nekrasov,Nekrasov-Okounkov} to compute 

150: the exact partition functions of four-dimensional 

151: $\mathcal{N}=2$ supersymmetric gauge theories. 

152: The celebrated Seiberg-Witten solutions 

153: \cite{Seiberg-Witten} of the gauge theories emerge 

154: \cite{Nekrasov-Okounkov,Nakajima-Yoshioka} through the statistical 

155: models of random partitions. 

156: In particular, Nekrasov and Okounkov show 

157: \cite{Nekrasov-Okounkov} that the Seiberg-Witten 

158: geometry is realized as the thermodynamic limit  

159: or the semi-classical approximation of the statistical 

160: models. 

161: 

162: 

163: 

164: 

165: 

166: In \cite{MNTT} the authors consider 

167: the statistical model of random plane partitions 

168: relevant to describe five-dimensional 

169: $\mathcal{N}=1$ supersymmetric $SU(N)$ Yang-Mills theory 

170: on $\mathbb{R}^4 \times S^1$ 

171: (more precisely plus the Chern-Simons term). 

172: This model has a smooth four-dimensional limit and 

173: gives rise to the exact partition function for 

174: four-dimensional $\mathcal{N}=2$ supersymmetric Yang-Mills.  

175: 

176: 

177: 

178: 

179: 

180: 

181: In this article we further develop the previous study.  

182: Random plane partitions are known 

183: \cite{Okounkov-Reshetikhin} to be described by two-dimensional 

184: free fermions (2$d$ conformal field theory). 

185: We compute the one-point functions of fermions 

186: in the statistical model and clarify their 

187: thermodynamic limits.

188: We start Section 2 with a brief review of the model. 

189: The interpretation as $q$-deformed random partitions 

190: is emphasized. In Section 3 we introduce one-point functions 

191: of fermions which are located at the main diagonal 

192: of the plane partition. They turn out to have a statistical 

193: interpretation.  

194: They are expressed as the statistical sums 

195: of suitable wave functions associated with partitions. 

196: Asymptotics of these wave functions, which are relevant  

197: to obtain the thermodynamic limit of the one-point functions, 

198: are computed. 

199: 

200: 

201: 

202: 

203: 

204: 

205: 

206: As the $q$-deformed random partitions, 

207: the model can be specified by the Boltzmann weights 

208: attached to partitions rather than plane partitions. 

209: For a very large partition, the Boltzmann weight is 

210: measured by the energy function. 

211: In Section 4 and Appendix A, we compute the energy function 

212: in a way different from \cite{Nekrasov-Okounkov}.  

213: Our computations are based on the regularized density of the 

214: Maya diagram. The comparison with the known description 

215: using the profile is presented in Appendix B. 

216: 

217: 

218: 

219: 

220: The thermodynamic limits or the semi-classical approximations of 

221: the fermion one-point functions at the main diagonal 

222: are proved in Section 5 to have 

223: a universal form in the sense that it is irrespective of detail 

224: of the minimizers of the energy function.  

225: They are given by the WKB type wave functions. 

226: In particular, at the four-dimensional limit, 

227: these semi-classical one-point functions are shown to be living 

228: on the Seiberg-Witten curve of the gauge theory.  

229: They are the WKB type wave functions 

230: whose classical action is precisely 

231: the integral of $dS_{s.w}$ on the curve, 

232: where $dS_{s.w}$ is the Seiberg-Witten differential. 

233:   

234: 

235: 

236: 

237: 

238: The diagonal slices of plane partitions are labelled 

239: by the discretized time $m$. The main diagonal slice is 

240: at $m=0$. The discretized time becomes a continuous time $t$ 

241: at the thermodynamic limit. 

242: It is identified with a coordinate of the limit shape 

243: \cite{Okounkov-Reshetikhin} of random plane partition. 

244: In Section 6 we consider the $U(1)$ theory and 

245: compute the semi-classical wave functions of fermions 

246: which are located away from the main diagonal. 

247: We then argue the role of the parameter $t$ in gauge theories. 

248: 

249: 

250: 

251: 

252: 

253: 

254: 

255: 

256: 

257: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

258: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

259: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

260: \section{A model of random plane partitions}

261: 

262: 

263: 

264: 

265: 

266: A plane partition $\pi$ is an array of non-negative integers 

267: \begin{eqnarray}

268: \begin{array}{cccc}

269: \pi_{11} & \pi_{12} & \pi_{13} & \cdots \\

270: \pi_{21} & \pi_{22} & \pi_{23} & \cdots \\

271: \pi_{31} & \pi_{32} & \pi_{33} & \cdots \\

272: \vdots & \vdots & \vdots & ~

273: \end{array}

274: \label{pi}

275: \end{eqnarray}

276: satisfying 

277: $\pi_{ij}\geq \pi_{i+1 j}$ and 

278: $\pi_{ij}\geq \pi_{i j+1}$ for 

279: all $i,j \geq 1$. 

280: Plane partitions are identified 

281: with the three-dimensional Young diagrams 

282: as depicted in Figure 1-(a). 

283: The three-dimensional diagram $\pi$ 

284: is a set of unit cubes such that $\pi_{ij}$ cubes 

285: are stacked vertically on each $(i,j)$-element of $\pi$. 

286: The size of $\pi$ is $|\pi| \equiv \sum_{i,j \geq 1}\pi_{ij}$, 

287: which is the total number of cubes of the diagram. 

288: Each diagonal slice of $\pi$ becomes a partition, 

289: that is, a sequence of weakly decreasing non-negative integers. 

290: See Figure 1-(b).  

291: Let $\pi(m)$ be a partition along the $m$-th diagonal slice. 

292: \begin{eqnarray}

293: \pi(m)= 

294: \left\{

295: \begin{array}{cl}

296: (\pi_{1~ m+1},\pi_{2~ m+2},\pi_{3~ m+3},\cdots) & 

297: ~~\mbox{for}~m \geq 0 \\[1.5mm]

298: (\pi_{-m+1~ 1},\pi_{-m+2~ 2},\pi_{-m+3~ 3},\cdots) & 

299: ~~\mbox{for}~m \leq -1.\\[1.5mm]

300: \end{array}

301: \right.

302: \end{eqnarray}

303: In particular 

304: $\pi(0)=(\pi_{11},\pi_{22},\pi_{33},\cdots)$ will be called 

305: the main diagonal partition. 

306: The series of partitions $\pi(m)$ satisfies the condition

307: \begin{eqnarray}

308: \cdots \prec \pi(-2) \prec \pi(-1) \prec 

309: \pi(0) \succ \pi(1) \succ \pi(2) \succ \cdots,

310: \end{eqnarray}

311: where $\mu \succ \nu$ means the following interlace relation 

312: between two partitions 

313: $\mu=(\mu_1 \geq \mu_2 \geq \cdots \geq 0)$ 

314: and 

315: $\nu=(\nu_1 \geq \nu_2 \geq \cdots \geq 0)$ 

316: \begin{eqnarray}

317: \mu \succ \nu ~~~

318: \Longleftrightarrow ~~~

319: \mu_1 \geq \nu_1 \geq \mu_2 \geq \nu_2 

320: \geq \mu_3 \geq \cdots.

321: \end{eqnarray} 

322: We have $|\pi|=\sum_{m=-\infty}^{+\infty}|\pi(m)|$,  

323: where the size of a partition $\mu$ is also denoted by  

324: $|\mu| \equiv \sum_{i \geq 1}\mu_i$. 

325: %

326: %

327: \begin{figure}[bt]

328: \begin{center}

329: \includegraphics[scale=0.7]{planepartition.eps}

330: \mycaption{1}{

331: The three-dimensional Young diagram (a) and 

332: the corresponding sequence of partitions 

333: (the two-dimensional Young diagrams) (b).}

334: \end{center}\label{3d Young diagram}

335: \end{figure}

336: %

337: %

338: 

339: 

340: 

341: 

342: 

343: A model of random plane partitions 

344: which describes five-dimensional ${\cal N}=1$ 

345: supersymmetric $SU(N)$ gauge theory is discussed 

346: in \cite{MNTT}. 

347: The partition function of the model is given by  

348: \begin{eqnarray}

349: Z^{SU(N)}_{q}

350: \equiv 

351: \sum_{\pi}\, q^{|\pi|}\,Q^{|\pi(0)|}\,e^{V(\pi(0))}. 

352: \label{Z SU(N)}

353: \end{eqnarray}

354: The Boltzmann weight consists of three parts. 

355: The first contribution comes from the energy of a plane 

356: partition $\pi$.  

357: The second contribution is a chemical potential 

358: for the main diagonal partition $\pi(0)$, and 

359: the third comes from the $N$-periodic potential 

360: \cite{Nekrasov-Okounkov} for the main diagonal partition. 

361: Let $\xi_r \in \mathbb{R}$ ($1\leq r \leq N$) be 

362: such that $\sum_{r=1}^N \xi_r=0$, and 

363: define the $N$-periodic potential $v$ on $\mathbb{Z}$ 

364: by $v(k)=\xi_{k+1 \mbox{ \scriptsize{mod} } N}$. 

365: The potential $V$ for a partition $\mu$ is formally 

366: given \footnote{See Appendix A for more information.} by

367: \begin{eqnarray}

368: V(\mu)=

369: \sum_{i=1}^{+\infty} 

370: v(\mu_i-i).

371: \label{N-periodic potential}

372: \end{eqnarray} 

373: To contact with the $SU(N)$ gauge theory 

374: we also need to identify \cite{MNTT} 

375: the indeterminates $q$ and $Q$ 

376: with the following field theory variables. 

377: \begin{eqnarray}

378: q=e^{-\frac{2R}{N}\hbar},~~~

379: Q=(2R\Lambda)^2, 

380: \label{q and Q}

381: \end{eqnarray}

382: where $R$ is the radius of $S^1$ in the fifth dimension, 

383: and $\Lambda$ is the lambda parameter of the underlying 

384: four-dimensional field theory. 

385: 

386: 

387: 

388: 

389: 

390: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

391: \subsubsection*{Transfer matrix approach}

392: 

393: 

394: 

395: 

396: 

397: The transfer matrix approach \cite{Okounkov-Reshetikhin} 

398: allows us to express the random plane partitions 

399: (\ref{Z SU(N)}) in terms of two-dimensional conformal 

400: field theory (2$d$ free fermion system).  

401: 

402: 

403: 

404: It is well known that partitions 

405: are realized as states of 2$d$ free fermions 

406: by using the Maya diagrams. 

407: Let 

408: $\psi(z)=\sum_{k \in \mathbb{Z}}

409: \psi_k z^{-k-1}$

410: and 

411: $\psi^*(z)=\sum_{k \in \mathbb{Z}}

412: \psi^*_k z^{-k}$

413: be complex fermions with the anti-commutation relations 

414: \begin{eqnarray}  

415: \left\{ \psi_k,\psi^*_l \right\}

416: =\delta_{k+l,0},~~~~

417: \left\{ \psi_k,\psi_l \right\}

418: =\left\{ \psi_k^*,\psi_l^* \right\}

419: =0. 

420: \end{eqnarray}

421: Let 

422: $\mu=(\mu_1, \mu_2, \cdots)$ be a partition. 

423: The Maya diagram $\mu$ is 

424: a series of the strictly decreasing integers 

425: $x_{i}(\mu) 

426: \equiv -i+\mu_i$, 

427: where $i \in \mathbb{Z}_{\geq 1}$. 

428: The correspondence with the Young diagram 

429: is depicted in Figure 2.

430: By using the Maya diagram 

431: the partition can be mapped to 

432: the following fermion state 

433: \begin{eqnarray}

434: |\mu;n \rangle 

435: =

436: \psi_{-x_1(\mu)-1-n}\psi_{-x_2(\mu)-1-n}

437: \cdots \psi_{-x_{l(\mu)}(\mu)-1-n}

438: \psi^*_{-l(\mu)+1+n}\psi^*_{-l(\mu)+2+n}

439: \cdots \psi^*_{n} 

440: |\emptyset;\,n \rangle, 

441: \label{|mu;n>}

442: \end{eqnarray}

443: where $l(\mu)$ is the length of $\mu$, that is, 

444: the number of the non-zero $\mu_i$. 

445: In (\ref{|mu;n>}) the state $|\emptyset;\,n\rangle$ is 

446: the ground state of the charge $n$ sector. 

447: It is defined by the conditions 

448: \begin{eqnarray}

449: &&

450: \psi_{k}|\emptyset;\,n \rangle=0~~~\mbox{for}~k \geq -n,

451: \nonumber \\

452: &&

453: \psi_{k}^*|\emptyset;\,n \rangle=0~~~\mbox{for}~k \geq n+1.

454: \label{|phi;n>}

455: \end{eqnarray}

456: We mainly consider the $n=0$ sector in the below. 

457: %

458: %

459: %

460: \begin{figure}[hbt]

461: \begin{center}

462: \includegraphics[scale=0.5]{fig1.eps}

463: \mycaption{2}{The correspondence between 

464: the Maya diagram and 

465: the Young diagram of $\mu=(7,5,4,2,2,1)$.

466: Elements of the Maya diagram 

467: are denoted by $\bullet$.}

468: \end{center}\label{fig1}

469: \end{figure}

470: %

471: %

472: %

473: 

474: 

475: 

476: 

477: 

478: The basic ingredient of the transfer matrix approach 

479: is the following evolution operator at a discretized time 

480: $m \in \mathbb{Z}$.

481: \begin{eqnarray}

482: \Gamma(m)

483: &\equiv& 

484: \left\{

485: \begin{array}{cl}

486: \exp\Bigl(

487: \displaystyle{\sum_{k=1}^{+\infty}}

488: \frac{1}{k}q^{k(m+\frac{1}{2})}J_{-k}\Bigr)

489: & 

490: ~~~\mbox{for}~m \geq 0 \\[1.5mm]

491: \exp\Bigl(

492: \displaystyle{\sum_{k=1}^{+\infty}}

493: \frac{1}{k}q^{-k(m+\frac{1}{2})}J_{k}\Bigr)

494: & 

495: ~~~\mbox{for}~m \leq -1,  \\[1.5mm]

496: \end{array}

497: \right.

498: \label{gamma m}

499: \end{eqnarray}

500: where $J_{\pm k}$ are the modes of the standard $U(1)$ current 

501: \begin{eqnarray}

502: :\psi \psi^*:(z)=\sum_{n}J_nz^{-n-1}.

503: \end{eqnarray}

504: Implications of the above operators 

505: in random plane partitions can be 

506: understood from their matrix elements: 

507: For $m \geq 0$, 

508: \begin{eqnarray}

509: \langle \mu;\,0|

510: \Gamma(m)

511: |\nu;\,0 \rangle 

512: =

513: \left\{

514: \begin{array}{cl}

515: q^{(m+\frac{1}{2})(|\mu|-|\nu|)}  &

516: \mu \succ \nu    \\[1.5mm]

517: 0 & 

518: \mbox{otherwise}, \\[1.5mm]

519: \end{array}

520: \right.

521: \label{gamma m on mu 1}

522: \end{eqnarray}

523: and for $m \leq -1$,  

524: \begin{eqnarray}

525: \langle \mu;\,0|

526: \Gamma(m)

527: |\nu;\,0 \rangle 

528: =

529: \left\{

530: \begin{array}{cl}

531: q^{(m+\frac{1}{2})(|\mu|-|\nu|)}  &

532: \mu \prec \nu    \\[1.5mm]

533: 0 & 

534: \mbox{otherwise}. \\[1.5mm]

535: \end{array}

536: \right.

537: \label{gamma m on mu 2}

538: \end{eqnarray}

539: It follows from (\ref{gamma m on mu 1}) 

540: and (\ref{gamma m on mu 2}) that the partition 

541: function is expressed as 

542: \begin{eqnarray}

543: Z^{SU(N)}_q=

544: \langle \emptyset;\,0|

545: \Biggl\{ 

546: \prod_{m \leq -1} \Gamma(m)

547: \Biggr\}

548: Q^{L_0}

549: \exp 

550: \Biggl(

551: \sum_{r=1}^{N}\xi_r J_{0}^{(r)}

552: \Biggr)

553: \Biggl\{

554: \prod_{m \geq 0}  \Gamma(m)

555: \Biggr\}

556: | \emptyset;\, 0\rangle.

557: \label{Z SU(N) transfer matrix}

558: \end{eqnarray}

559: Here, 

560: $J_0^{(r)}$ are the zero modes of 

561: the $U(1)$ currents of $N$-component fermions 

562: $\psi^{(r)}$ and $\psi^{(r)*}$ which are 

563: realized \cite{Miwa-Jimbo} by folding $\psi$ and $\psi^{*}$. 

564: 

565: 

566: 

567: 

568: 

569: 

570: 

571: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

572: \subsubsection*{$q$-deformed random partitions}

573: 

574: 

575: 

576: 

577: 

578: 

579: We can interpret the random plane partitions 

580: (\ref{Z SU(N)}) as a model of random partitions. 

581: It is identified with a $q$-deformation of the 

582: random partitions \cite{Nekrasov-Okounkov}. 

583: (The deformation itself differs from \cite{Nekrasov-Okounkov}.) 

584: To see this, 

585: we rewrite the partition function 

586: by using the Schur functions. 

587: An insertion of the unity 

588: $1=\sum_{\mu}|\mu;\,0 \rangle \langle \mu ;\,0|$ 

589: factorizes (\ref{Z SU(N) transfer matrix}) into 

590: \begin{eqnarray}

591: Z_q^{SU(N)}&=&

592: \sum_{\mu}

593: Q^{|\mu|}e^{V(\mu)}

594: \langle \emptyset;\,0|

595: \prod_{m \leq -1}

596: \Gamma(m)

597: |\mu;\, 0\rangle 

598: \langle \mu;\,0| 

599: \prod_{m \geq 0}

600: \Gamma(m)

601: |\emptyset;\,0\rangle.  

602: \label{factorized Z SU(N) transfer matrix}

603: \end{eqnarray}

604: The matrix elements in the above turn to be 

605: \begin{eqnarray}

606: \langle \emptyset ;\,0|

607: \prod_{m \leq -1}\Gamma(m) 

608: | \mu ;\, 0 \rangle 

609: &=&

610: \langle \emptyset ;\,0 | 

611: \prod_{k=1}^{+\infty}

612: \exp \left( 

613: \frac{1}{k}

614: \sum_{i=1}^{+\infty}

615: q^{k(i-\frac{1}{2})}

616: J_k \right) 

617: | \mu; \,0 \rangle 

618: \nonumber \\[1.5mm]

619: &=&

620: s_{\mu}(q^{\frac{1}{2}},q^{\frac{3}{2}},\cdots), 

621: \label{schur mu 1} \\[1.5mm]

622: \langle \mu ;\,0|

623: \prod_{m \geq 0}\Gamma(m) 

624: | \emptyset ;\, 0 \rangle 

625: &=&

626: \langle \mu;0 | 

627: \prod_{k=1}^{+\infty}

628: \exp \left( 

629: \frac{1}{k}

630: \sum_{i=1}^{+\infty}

631: q^{k(i-\frac{1}{2})}

632: J_{-k} \right) 

633: | \emptyset ;\, 0 \rangle 

634: \nonumber \\[1.5mm]

635: &=&

636: s_{\mu}(q^{\frac{1}{2}},q^{\frac{3}{2}},\cdots), 

637: \label{schur mu 2}

638: \end{eqnarray} 

639: where 

640: $s_{\mu}(q^{\frac{1}{2}},q^{\frac{3}{2}},\cdots)$ 

641: is the Schur function 

642: $s_{\mu}(x_1,x_2,\cdots)$ specialized at 

643: $x_i=q^{i-\frac{1}{2}}$ ($i \geq 1$). 

644: Therefore we obtain 

645: \begin{eqnarray}

646: Z_q^{SU(N)}

647: &=&

648: \sum_{\mu}\,Q^{|\mu|}e^{V(\mu)}

649: s_{\mu}(q^{-\rho})^2, 

650: \label{Z SU(N) schur}

651: \end{eqnarray}

652: where the multiple index 

653: $\rho \equiv (-\frac{1}{2},-\frac{3}{2}, 

654: \cdots, -i+\frac{1}{2},\cdots)$ is used. 

655: The expression (\ref{Z SU(N) schur}) allows us 

656: to interpret (\ref{Z SU(N)}) 

657: as a model of $q$-deformed random partitions. 

658: It is also clear from  

659: (\ref{factorized Z SU(N) transfer matrix}) 

660: that partitions $\mu$ are the main diagonal 

661: partitions $\pi(0)$.  

662: 

663: 

664: 

665: 

666: 

667: The four-dimensional limit of the model  

668: is obtained by letting $R \rightarrow 0$ 

669: under the identification (\ref{q and Q}). 

670: To take the limit the following 

671: product formula of the Schur function 

672: \cite{Macdonald} becomes useful. 

673: \begin{eqnarray}

674: s_{\mu}(q^{-\rho})

675: &=&

676: q^{n(\mu)+\frac{1}{2}|\mu|}

677: \prod_{(i,j)\in \mu}

678: \frac{1}

679: {\left(1-q^{h(i,j)}\right)}, 

680: \label{product formula b}

681: \end{eqnarray}

682: where $h(i,j)$ is the hook length of the box $(i,j)$ 

683: in the Young diagram,  

684: and $n(\mu)\equiv \sum_{i \geq 1}(i-1)\mu_i$. 

685: By using (\ref{product formula b}) 

686: we can see from (\ref{Z SU(N) schur}) that 

687: \begin{eqnarray}

688: \lim_{R \rightarrow 0}Z_q^{SU(N)}

689: =

690: \sum_{\mu}

691: \left(\frac{N\Lambda}{\hbar}\right)^{2|\mu|}

692: e^{V(\mu)}

693: \left(\frac{dim \mu}{|\mu|!}\right)^2. 

694: \label{4d U(1) random partitions}

695: \end{eqnarray}

696: This is the model of random partitions \cite{Nekrasov-Okounkov}

697: which describes four-dimensional 

698: ${\cal N}=2$ supersymmetric $SU(N)$ Yang-Mills 

699: at the thermodynamic limit.

700: 

701: 

702: 

703: 

704: 

705: Thermodynamic limit of the model is achieved  

706: by letting $\hbar \rightarrow 0$. To see this, 

707: we consider the $U(1)$ theory or the $SU(N)$ theory with 

708: the periodic potential turned off. 

709: In such a situation the partition function can be computed 

710: by using the standard technique of 

711: 2$d$ conformal field theory.  

712: It becomes 

713: \begin{eqnarray}

714: Z_q^{U(1)}=\prod_{n=1}^{+\infty}

715: (1-Qq^n)^{-n}.

716: \label{MacMahon}

717: \end{eqnarray}

718: The mean values of $|\pi|$ and $|\pi(0)|$ 

719: are respectively given by $q\frac{\partial}{\partial q}\ln Z$ 

720: and $Q\frac{\partial}{\partial Q}\ln Z$. 

721: It follows from (\ref{MacMahon}) that 

722: they behave 

723: $\left \langle |\pi| \right \rangle =O(\hbar^{-3})$ and 

724: $\left \langle |\pi(0)| \right \rangle =O(\hbar^{-2})$ 

725: as $\hbar \rightarrow 0$ ($q \rightarrow 1$). 

726: Therefore a typical plane partition $\pi$ 

727: near the limit $\hbar \rightarrow 0$ is a plane partition 

728: of order $\hbar^{-3}$, 

729: and its main diagonal partition $\pi(0)$ or $\mu$

730: becomes a partition of order $\hbar^{-2}$. 

731: 

732: 

733: 

734: 

735: 

736: \newpage

737: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

738: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

739: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

740: \section{One-point functions at the main diagonal}

741: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

742: 

743: 

744: 

745: 

746: 

747: In the present section and section 5  

748: we investigate one-point functions of the free fermions 

749: located at the main diagonal of the plane partition $\pi$.  

750: From the transfer matrix viewpoint they are given by 

751: \begin{eqnarray}

752: \Psi_q^*(z) 

753: &\equiv& 

754: \frac{1}{Z^{SU(N)}_q}

755: \langle \emptyset;\, -1 |

756: \Biggl\{

757: \prod_{m \leq -1} \Gamma(m)

758: \Biggr\}

759: \,

760: \psi^*(z)

761: \,

762: q^{\frac{1}{2}L_0}

763: Q^{L_0}

764: \,

765: e^{\sum_{r=1}^{N}\xi_r J_{0}^{(r)}}

766: \Biggl\{

767: \prod_{m \geq 0}\Gamma(m)

768: \Biggr\}

769: | \emptyset;\, 0 \rangle,

770: \nonumber \\*

771: &&

772: \label{Psi(lambda)star}\\

773: \Psi_q(z) 

774: &\equiv& 

775: \frac{1}{Z^{SU(N)}_q}

776: \langle \emptyset;\, +1 |

777: \Biggl\{

778: \prod_{m \leq -1} \Gamma(m)

779: \Biggr\}

780: \,

781: \psi(z)

782: \,

783: q^{-\frac{1}{2}L_0}

784: \,

785: Q^{L_0}

786: \,

787: e^{\sum_{r=1}^{N}\xi_r J_{0}^{(r)}}

788: \Biggl\{

789: \prod_{m \geq 0}\Gamma(m)

790: \Biggr\}

791: | \emptyset ;\, 0 \rangle, 

792: \nonumber \\*

793: &&

794: \label{Psi(lambda)}

795: \end{eqnarray}

796: where $q^{\pm \frac{L_0}{2}}$ are inserted 

797: for the later convenience of normalization.  

798: 

799: 

800: 

801: The goal of our discussions is to clarify

802: the thermodynamic limits of the above one-point functions. 

803: These are treated in section 5 

804: by using the results of this section. 

805: 

806: 

807: 

808: 

809: 

810: 

811: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

812: \subsubsection*{Statistical interpretation of 

813: fermion one-point functions}

814: 

815: 

816: 

817: 

818: We regard the random plane partitions (\ref{Z SU(N)}) 

819: as the $q$-deformed random partitions (\ref{Z SU(N) schur}). 

820: The Boltzmann weight for a partition $\mu$ is given by 

821: $Q^{|\mu|}e^{V(\mu)}s_{\mu}(q^{-\rho})^2$. 

822: The above one-point functions admit to have a statistical 

823: interpretation. As we confirm subsequently, they can be

824: expressed as the following statistical sums. 

825: \begin{eqnarray}

826: \Psi_q^*(z)

827: &=&

828: \frac{1}{Z^{SU(N)}_q}

829: \sum_{\mu}\,

830: \chi_q^*(z \mid \mu)\,\,

831: Q^{|\mu|}

832: e^{V(\mu)}

833: s_{\mu}(q^{-\rho})^2,

834: \label{Psi(lambda)star 2} 

835: \\

836: \Psi_q(z)

837: &=&

838: \frac{1}{Z_q^{SU(N)}}

839: \sum_{\mu}\, 

840: \chi_q(z \mid \mu)\,\, 

841: Q^{|\mu|}

842: e^{V(\mu)}

843: s_{\mu}(q^{-\rho})^2,  

844: \label{Psi(lambda) 2} 

845: \end{eqnarray}

846: where $\chi_q^*(z\mid\mu)$ and 

847: $\chi_q(z\mid\mu)$ are thought as wave functions 

848: associated with each partition $\mu$. 

849: They are defined by  

850: \begin{eqnarray}

851: \chi_q^*(z \mid \mu)

852: &\equiv& 

853: \sum_{i=1}^{+\infty}

854: z^{-x_i(\mu)-1}

855: (-)^{i-1}

856: q^{\frac{1}{2}|\mu|}

857: \frac{s_{(\mu:\, i)}(q^{-\rho})}{s_{\mu}(q^{-\rho})}, 

858: \label{chi star} 

859: \\

860: \chi_q(z \mid \mu)

861: &\equiv& 

862: \sum_{i=1}^{+\infty}

863: z^{-x_i(\widetilde{\mu})-1}

864: (-)^{i-1}

865: q^{-\frac{1}{2}|\mu|}

866: \frac{s_{(\widetilde{\mu}:\, i)}(q^{\rho})}

867: {s_{\widetilde{\mu}}(q^{\rho})}.  

868: \label{chi} 

869: \end{eqnarray}

870: Here $(\mu:i)$ is a partition obtained from 

871: $\mu$ for each $i \in \mathbb{Z}_{\geq 1}$. 

872: It is described 

873: in terms of the Maya diagram as follows.    

874: \begin{eqnarray}

875: x_{k}(\mu:i)=

876: \left \{

877: \begin{array}{cl}

878: x_k(\mu)+1 &~~\mbox{for}~~ 1 \leq k <i \\[1.5mm]

879: x_{k+1}(\mu)+1 &~~\mbox{for}~~ k \geq i. \\

880: \end{array}

881: \right.

882: \label{x(mu:i)}

883: \end{eqnarray}

884: We can see from the above that 

885: $|(\mu:i)|=|\mu|-x_i(\mu)-1$. 

886: The partition conjugate to $\mu$ is denoted 

887: by $\widetilde{\mu}$ in (\ref{chi}).

888: The corresponding Young diagram is 

889: obtained by flipping the Young diagram $\mu$ 

890: over its main diagonal. 

891: One can find a relation between 

892: $\chi_q(z \mid \mu)$ and $\chi_q^*(z \mid \mu)$.

893: By comparing (\ref{chi star}) with (\ref{chi}) 

894: we obtain 

895: \begin{eqnarray}

896: \chi_q(z \mid \mu)=

897: \chi_{q^{-1}}^*

898: (z \mid \widetilde{\mu}). 

899: \label{chi vs chi star}

900: \end{eqnarray}

901: 

902: 

903: 

904: 

905: 

906: %%%%%%%%%%%%%%%  PROOF %%%%%%%%%%%%%%%%%%%

907: 

908: 

909: ~

910: 

911: 

912: Let us verify the descriptions 

913: (\ref{Psi(lambda)star 2}) and (\ref{Psi(lambda) 2}). 

914: We first consider the one-point function $\Psi_q^*(z)$. 

915: By the insertion of the unity 

916: $1=\sum_{\mu}|\mu;\,0 \rangle \langle \mu ;\,0|$ 

917: the following factorization of (\ref{Psi(lambda)star}) 

918: is obtained. 

919: \begin{eqnarray}

920: \frac{1}{Z_q^{SU(N)}}

921: \sum_{\mu}\,

922: Q^{|\mu|}

923: e^{V(\mu)}

924: q^{\frac{1}{2}|\mu|}\,

925: \langle \emptyset ;\,-1 | 

926: \Biggl\{

927: \prod_{m \leq -1}\Gamma(m)

928: \Biggr\}

929: \psi^*(z)\,

930: |\mu;\, 0 \rangle 

931: \langle \mu;\, 0 |

932: \prod_{m \geq 0}\Gamma(m)

933: |\emptyset;\,0 \rangle.   

934: \label{Psi star proof}

935: \end{eqnarray}

936: This can be written down by using the Schur functions. 

937: The second component is just (\ref{schur mu 2}). 

938: In order to express the first component in terms of 

939: the Schur functions, it is convenient to 

940: describe the action of 

941: $\psi^*(z)$ on the partition $\mu$ explicitly. 

942: By using (\ref{|mu;n>}) we can see 

943: that the actions of $\psi^*_k$ on the partition become

944: \begin{eqnarray}

945: \psi^*_k\,\,

946: |\mu;\,0 \rangle 

947: =

948: \left\{

949: \begin{array}{cl}

950: (-)^{i-1}|(\mu:i);\,-1\rangle &~~~ 

951: \mbox{if}~~k=x_{i}(\mu)+1 \\[1.5mm]

952: 0 & ~~~\mbox{otherwise},\\

953: \end{array}

954: \right.

955: \label{psi star mode on mu}

956: \end{eqnarray}

957: where $(\mu:i)$ is the partition given 

958: in (\ref{x(mu:i)}). 

959: The above actions are translated to  

960: \begin{eqnarray}

961: \psi^*(z)\,|\mu;\,0\rangle 

962: =

963: \sum_{i=1}^{+\infty}

964: z^{-x_i(\mu)-1}

965: (-)^{i-1}

966: |(\mu:i);\,-1\rangle. 

967: \label{psi star on mu}

968: \end{eqnarray}

969: By using (\ref{psi star on mu}) and also 

970: taking (\ref{schur mu 1}) into account, 

971: the first component of (\ref{Psi star proof}) becomes  

972: \begin{eqnarray}

973: \langle \emptyset ;\,-1 | 

974: \Biggl\{

975: \prod_{m \leq -1}\Gamma(m)

976: \Biggr\}

977: \psi^*(z)\,

978: |\mu;\, 0 \rangle 

979: =

980: \sum_{i=1}^{+\infty}

981: z^{-x_i(\mu)-1}

982: (-)^{i-1}

983: s_{(\mu:\,i)}(q^{-\rho}). 

984: \label{1st factor}

985: \end{eqnarray}

986: The description (\ref{Psi(lambda)star 2}) is obtained 

987: from (\ref{Psi(lambda)star}) by plugging  

988: (\ref{1st factor}) and (\ref{schur mu 2}) 

989: into (\ref{Psi star proof}) and then 

990: arranging them noting the definition (\ref{chi star}). 

991: 

992: 

993: 

994: 

995:  

996: 

997: 

998: Nextly we consider the case of $\Psi_q(z)$.  

999: The description (\ref{Psi(lambda) 2}) can be obtained 

1000: from (\ref{Psi(lambda)}) by taking the same steps as above. 

1001: We first factorize (\ref{Psi(lambda)}) into 

1002: \begin{eqnarray}

1003: \frac{1}{Z_q^{SU(N)}}

1004: \sum_{\mu}

1005: Q^{|\mu|}

1006: e^{V(\mu)}

1007: q^{-\frac{|\mu|}{2}}

1008: \langle \emptyset;\,+1 | 

1009: \Biggl\{

1010: \prod_{m \leq -1}\Gamma(m)

1011: \Biggr\}

1012: \psi(z)\,

1013: |\mu;\, 0 \rangle 

1014: \langle \mu;\, 0 |

1015: \prod_{m \geq 0}\Gamma(m)

1016: |\emptyset ;\,0 \rangle.   

1017: \label{Psi proof}

1018: \end{eqnarray}

1019: To write down the first component in terms of 

1020: the Schur functions we compute the action of 

1021: $\psi(z)$ on the partition $\mu$. 

1022: The actions of $\psi_k$ on the partition  

1023: can be read as 

1024: \begin{eqnarray}

1025: \psi_k

1026: |\mu;\,0 \rangle 

1027: =

1028: \left\{

1029: \begin{array}{cl}

1030: (-)^{i+x_i(\widetilde{\mu})}

1031: |\widetilde{(\widetilde{\mu}:i)};\,+1\rangle & 

1032: ~~~\mbox{if}~~

1033: k=x_{i}(\widetilde{\mu}) \\[1.5mm]

1034: 0 & 

1035: ~~~\mbox{otherwise}\,. 

1036: \end{array}

1037: \right.

1038: \label{psi mode on mu}

1039: \end{eqnarray}

1040: These are translated to 

1041: \begin{eqnarray}

1042: \psi(z)\,|\mu;\,0\rangle =

1043: \sum_{i=1}^{+\infty}

1044: z^{-x_i(\widetilde{\mu})-1}

1045: (-)^{i+x_i(\widetilde{\mu})}

1046: |\widetilde{(\widetilde{\mu}:i)};\,+1 \rangle. 

1047: \label{psi on mu}

1048: \end{eqnarray}

1049: Therefore the first component of (\ref{Psi proof}) becomes  

1050: \begin{eqnarray}

1051: \langle \emptyset ;\, +1 |

1052: \Biggl\{

1053: \prod_{m \leq -1} \Gamma(m)

1054: \Biggr\}

1055: \psi(z)\,

1056: | \mu;\, 0 \rangle 

1057: =\sum_{i=1}^{+\infty}

1058: z^{-x_i(\widetilde{\mu})-1}

1059: (-)^{i+x_i(\widetilde{\mu})}

1060: s_{\widetilde{(\widetilde{\mu}:\,i)}}

1061: (q^{-\rho}). 

1062: \label{1st factor 2}

1063: \end{eqnarray}

1064: The description (\ref{Psi(lambda) 2}) is obtained 

1065: from (\ref{Psi(lambda)}) by 

1066: plugging (\ref{1st factor 2}) and (\ref{schur mu 2}) 

1067: into (\ref{Psi proof}) and making use of the identity 

1068: $s_{\widetilde{\mu}}(q^{-\rho})=

1069: (-)^{|\mu|}s_{\mu}(q^{\rho})$. 

1070: 

1071: 

1072: 

1073: 

1074: 

1075: 

1076: 

1077: 

1078: 

1079: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1080: \subsubsection*{Large $|\mu|$ behaviors of 

1081: wave functions associated with partitions}

1082: 

1083: 

1084: 

1085: 

1086: 

1087: 

1088: Near the thermodynamic limit, 

1089: which is achieved by letting 

1090: $\hbar \rightarrow 0$ in the model, 

1091: very large partitions dominate. 

1092: Their size becomes of order $\hbar^{-2}$.

1093: We will clarify the asymptotics of 

1094: $\chi_q^*(z\mid \mu)$ and $\chi_q(z\mid \mu)$ 

1095: when $\mu$ is such a large partition. 

1096: The Maya diagram of such a partition 

1097: can be thought as a quantity of order $\hbar^{-1}$. 

1098: For the description we conveniently rescale $x_i(\mu)$. 

1099: Let us introduce the order $\hbar^0$ quantities 

1100: $u$ and $s$ by the following rescalings. 

1101: \begin{eqnarray}

1102: x=\frac{N}{\hbar}u,~~~~~

1103: i=\frac{N}{\hbar}s, 

1104: \label{u and s}

1105: \end{eqnarray}

1106: where $u \in \mathbb{R}$ and $s \in \mathbb{R}_{\geq 0}$. 

1107: Correspondingly, 

1108: the Maya diagram is scaled to a function $u(s\mid \mu)$ as

1109: \begin{eqnarray}

1110: x_{i}(\mu)=\frac{N}{\hbar}u(s\mid \mu)+O(\hbar^0).

1111: \label{u(s)}

1112: \end{eqnarray} 

1113: Notice that the Maya diagram is a strictly decreasing series 

1114: satisfying the conditions that 

1115: $x_{i}(\mu) \geq -i$ for $\forall i \in \mathbb{Z}_{\geq 1}$ 

1116: and that $x_{i}(\mu)=-i$ for sufficiently large $i$. 

1117: Therefore, the function $u(s\mid \mu)$ is monotonically decreasing,  

1118: and satisfies the conditions that 

1119: $u(s \mid\mu) \geq -s$ for $\forall s \in 

1120: \mathbb{R}_{\geq 0}$ and that $u(s\mid \mu) \rightarrow -s$ 

1121: as $s \rightarrow +\infty$. 

1122: 

1123: 

1124: 

1125: 

1126: 

1127: The large $|\mu|$ behaviors of the wave functions 

1128: can be computed from (\ref{chi star}) and (\ref{chi}) 

1129: by scaling $\mu$ as (\ref{u(s)}). As the result, 

1130: they turn out to be  

1131: \begin{eqnarray}

1132: &&

1133: \chi_q^*(z\mid \mu)

1134: \,\approx\,

1135: \frac{N}{\hbar}

1136: \int_0^{+\infty}

1137: d\underline{s}

1138: ~z^{-\frac{N}{\hbar}u(\underline{s}\mid\, \mu)}

1139: \nonumber \\[1.5mm]

1140: &&

1141: ~~~~~~~~~

1142: \times 

1143: \exp \left\{

1144: -\frac{RN}{2\hbar}u(\underline{s}\mid \mu)^2

1145: +\frac{N}{\hbar}

1146: \int_{-\infty}^{+\infty}

1147: dv \,

1148: \frac{ds(v\mid \mu)}{dv}

1149: \ln 

1150: \left(

1151: \frac{\sinh R(u(\underline{s}\mid\mu)-v)}{R\Lambda}

1152: \right)

1153: \right\}, 

1154: \nonumber \\[1.5mm]

1155: \label{asymptotic chi star} \\[1.5mm]

1156: &&

1157: \chi_q(z\mid \mu)

1158: \,\approx\,

1159: \frac{N}{\hbar}

1160: \int_0^{+\infty}

1161: d\underline{s}

1162: ~z^{-\frac{N}{\hbar}u(\underline{s}\mid\, \widetilde{\mu})}

1163: \nonumber \\*[1.5mm]

1164: &&

1165: ~~~~~~~~~

1166: \times 

1167: \exp \left\{

1168: \frac{RN}{2\hbar}u(\underline{s}\mid \widetilde{\mu})^2

1169: -\frac{N}{\hbar}

1170: \int_{-\infty}^{+\infty}

1171: dv\, 

1172: \frac{ds(v\mid \mu)}{dv}

1173: \ln 

1174: \left(

1175: \frac{\sinh R(-u(\underline{s}\mid\widetilde{\mu})-v)}{R\Lambda}

1176: \right)

1177: \right\}. 

1178: \nonumber \\*[1.5mm]

1179: \label{asymptotic chi}

1180: \end{eqnarray}

1181: In the above we have used the density function 

1182: $\frac{d}{du}s(u\mid\mu)$. 

1183: The function $s(u\mid\mu)$ is 

1184: the inverse of $u(s\mid\mu)$, that is,  

1185: $u(s(v|\,\mu)\,|\,\mu)=v$. 

1186: It becomes non-negative and weakly decreasing over $\mathbb{R}$, 

1187: satisfying the condition that 

1188: $s(u\mid\mu) \geq -u$. 

1189: The asymptotic behaviors are 

1190: $s(u\mid\mu) \rightarrow -u$ as $u \rightarrow -\infty$ 

1191: and

1192: $s(u\mid\mu) \rightarrow 0$ as $u \rightarrow +\infty$. 

1193: Therefore, 

1194: the density function 

1195: $\frac{d}{du}s(u\mid\mu)$ 

1196: takes values in $[-1,0]$, 

1197: and satisfies the asymptotic conditions 

1198: that 

1199: $\frac{d}{du}s(u\mid\mu)\rightarrow -1$ 

1200: as 

1201: $u \rightarrow -\infty$ and that 

1202: $\frac{d}{du}s(u\mid\mu)\rightarrow 0$ 

1203: as 

1204: $u \rightarrow +\infty$. 

1205: We should note 

1206: the density function in the proper sense

1207: \footnote{See (\ref{rho vs ds}) in Appendix A.} is 

1208: $-\frac{d}{du}s(u\mid\mu)$ rather than 

1209: $\frac{d}{du}s(u\mid\mu)$. 

1210: However, in this article,  

1211: $\frac{d}{du}s(u\mid\mu)$ is called the density function 

1212: as well, unless it causes any confusion.

1213: 

1214: 

1215: 

1216: 

1217: 

1218: Before confirming the asymptotics 

1219: (\ref{asymptotic chi star}) and (\ref{asymptotic chi}), 

1220: we shall describe their four-dimensional limits.  

1221: The limits are obtained from 

1222: (\ref{asymptotic chi star}) and (\ref{asymptotic chi}) 

1223: by letting $R \rightarrow 0$. 

1224: They become  

1225: \begin{eqnarray}

1226: &&

1227: \chi_{4d}^{*}(z\mid\mu)

1228: \, \equiv \, 

1229: \lim_{R \rightarrow 0}

1230: \chi_q^*(z\mid \mu)

1231: \nonumber \\*[1.5mm]

1232: &&~~~~~~

1233: \, \approx \,

1234: \frac{N}{\hbar}

1235: \int_0^{+\infty}

1236: d\underline{s}

1237: ~z^{-\frac{N}{\hbar}u(\underline{s} \mid \, \mu)}

1238: \exp \left\{

1239: \frac{N}{\hbar}

1240: \int_{-\infty}^{+\infty}

1241: dv \,

1242: \frac{ds(v\mid \mu)}{dv}

1243: \ln 

1244: \left(

1245: \frac{u(\underline{s}\mid\mu)-v}{\Lambda}

1246: \right)

1247: \right\}, 

1248: \nonumber \\* 

1249: \label{4d asymptotic chi star} \\[1.5mm]

1250: &&

1251: \chi_{4d}(z\mid\mu)

1252: \, \equiv \, 

1253: \lim_{R \rightarrow 0}

1254: \chi_q(z\mid \mu)

1255: \nonumber \\*[1.5mm]

1256: &&~~~~~~

1257: \,\approx \, 

1258: \frac{N}{\hbar}

1259: \int_0^{+\infty}

1260: d\underline{s}

1261: ~z^{-\frac{N}{\hbar}u(\underline{s}\mid\, \widetilde{\mu})}

1262: \exp \left\{

1263: -\frac{N}{\hbar}

1264: \int_{-\infty}^{+\infty}

1265: dv \,

1266: \frac{ds(v\mid \mu)}{dv}

1267: \ln 

1268: \left(

1269: \frac{-u(\underline{s}\mid\widetilde{\mu})-v}{\Lambda}

1270: \right)

1271: \right\}. 

1272: \nonumber \\*[1.5mm] 

1273: &&

1274: \label{4d asymptotic chi}

1275: \end{eqnarray}

1276: 

1277: 

1278: 

1279: 

1280: 

1281: 

1282: %%%%%%%%%%% PROOF %%%%%%%%%%%%%%%

1283: 

1284: 

1285: ~

1286: 

1287: 

1288: Let us derive the asymptotics 

1289: (\ref{asymptotic chi star}) and (\ref{asymptotic chi}). 

1290: Consider the case of $\chi^*_q(z\mid\mu)$. 

1291: The following product formula of 

1292: the (specialized) Schur function 

1293: \cite{Macdonald} becomes useful to obtain the asymptotics.

1294: \begin{eqnarray}

1295: s_{\mu}(q^{-\rho})=

1296: q^{-\frac{1}{4}\kappa(\mu)}

1297: \prod_{1 \leq i<j < +\infty}

1298: \left\{

1299: \displaystyle{

1300: \frac{q^{\frac{1}{2}(x_j(\mu)-x_i(\mu))}-

1301:           q^{\frac{1}{2}(x_i(\mu)-x_j(\mu))}}

1302:      {q^{\frac{1}{2}(i-j)}-q^{\frac{1}{2}(j-i)}}}

1303: \right\}, 

1304: \label{product formula a} 

1305: \end{eqnarray}

1306: where $\kappa(\mu)\equiv 2\sum_{(i,j)\in \mu}(j-i)$. 

1307: 

1308: 

1309: 

1310: 

1311: We translate the Schur functions in (\ref{chi star}) 

1312: into the infinite products by using the above formula. 

1313: $x_k(\mu:i)$ which appear in the product representation 

1314: of $s_{(\mu:\,i)}(q^{-\rho})$ can be converted to 

1315: $x_k(\mu)$ by using (\ref{x(mu:i)}). 

1316: It turns out that almost all the products cancel between 

1317: $s_{\mu}(q^{-\rho})$ and $s_{(\mu:\,i)}(q^{-\rho})$.  

1318: Finally the ratio of the Schur functions 

1319: in (\ref{chi star}) becomes as follows;   

1320: \begin{eqnarray}

1321: \frac{s_{(\mu:\,i)}(q^{-\rho})}

1322: {s_{\mu}(q^{-\rho})}

1323: &=&

1324: q^{-\frac{1}{2}|\mu|+

1325: \frac{1}{4}(x_i(\mu)+1)(x_i(\mu)+2)}

1326: \displaystyle{

1327:   \prod_{1 \leq j \leq i-1}

1328:   \left( 

1329:   q^{\frac{1}{2}(j-i)}-q^{\frac{1}{2}(i-j)}

1330:   \right)^{-1}} 

1331: \nonumber \\*[1.5mm]

1332: &&

1333: \times 

1334: \displaystyle{

1335:     \prod_{j \neq i}^{+\infty}

1336:     \left \{

1337:        \frac{q^{\frac{1}{2}(x_j(\mu)-x_i(\mu))}

1338:              -q^{\frac{1}{2}(x_i(\mu)-x_j(\mu))}}

1339:             {q^{\frac{1}{2}(i-j)}-q^{\frac{1}{2}(j-i)}}

1340:              \right \}^{-1}}. 

1341: \label{maeda formula}

1342: \end{eqnarray}

1343: This gives the following expression 

1344: of $\chi_q^*(z\mid\mu)$; 

1345: \begin{eqnarray}

1346: \chi_q^*(z \mid\mu)=

1347: \sum_{i=1}^{+\infty}\, 

1348: z^{-x_i(\mu)-1}\,

1349: q^{\frac{1}{4}(x_i(\mu)+1)(x_i(\mu)+2)}\,

1350: \exp \Bigl\{\Phi_q(i\mid\mu)\Bigr\}, 

1351: \label{chi star 2} 

1352: \end{eqnarray}

1353: where $\Phi_q(i\mid\mu)$ is defined by 

1354: \begin{eqnarray}

1355: \exp 

1356: \Bigl\{\Phi_q(i\mid\mu) \Bigr\}

1357: \equiv

1358: \displaystyle{ 

1359: \prod_{j=1}^{i-1}

1360: \left( 

1361:   q^{\frac{1}{2}(i-j)}-q^{\frac{1}{2}(j-i)}

1362: \right)^{-1}}

1363: \, 

1364: \displaystyle{ 

1365: \prod_{j \neq i}^{+\infty}

1366: \left \{

1367:    \frac{q^{\frac{1}{2}(x_j(\mu)-x_i(\mu))}

1368:          -q^{\frac{1}{2}(x_i(\mu)-x_j(\mu))}}

1369:         {q^{\frac{1}{2}(i-j)}-q^{\frac{1}{2}(j-i)}}

1370: \right \}^{-1}}. 

1371: \label{Phi}

1372: \end{eqnarray}

1373: The similar expression of $\chi_q(z\mid\mu)$ 

1374: is obtainable from (\ref{chi star 2}) by using 

1375: (\ref{chi vs chi star}). 

1376: 

1377: 

1378: 

1379: 

1380: 

1381: 

1382: 

1383: The large $|\mu|$ behavior of $\chi_q^*(z\mid\mu)$ 

1384: can be extracted from (\ref{chi star 2}) 

1385: by scaling $x_i(\mu)$ as prescribed in (\ref{u(s)}). 

1386: Taking account of $q=e^{-2R\hbar/N}$ 

1387: we first observe   

1388: \begin{eqnarray}

1389: q^{\frac{1}{4}(x_i(\mu)+1)(x_i(\mu)+2)}

1390: =\exp

1391: \left\{

1392: -\frac{RN}{2\hbar}u(\underline{s}\mid\mu)^2

1393: +O(\hbar^0) 

1394: \right\}, 

1395: \label{asymptotic K}

1396: \end{eqnarray}

1397: where we put $i=\frac{N}{\hbar}\underline{s}$. 

1398: The large $|\mu|$ behavior of $\Phi_q(i\mid\mu)$ 

1399: can be computed as follows;  

1400: \begin{eqnarray}

1401: &&

1402: \Phi_q(i\mid \mu)

1403: \nonumber \\[1.5mm]

1404: &&

1405: =

1406: \displaystyle{ 

1407:  \sum_{j=i+1}^{+\infty}\,

1408:  \ln \left(q^{\frac{1}{2}\left(i-j\right)}

1409:            -q^{\frac{1}{2}\left(j-i\right)}\right)

1410:  -\sum_{j \neq i}^{+\infty}\,

1411:  \ln \left(

1412:     q^{\frac{1}{2}\left(x_j(\mu)-x_i(\mu)\right)}

1413:     -q^{\frac{1}{2}\left(x_i(\mu)-x_j(\mu)\right)}

1414:      \right)}

1415: \nonumber \\[1.5mm]

1416: &&

1417: =

1418: \displaystyle{

1419:   \frac{N}{\hbar}

1420:   \int_{0}^{+\infty}dx \,

1421:   \ln \Bigl(e^{Rx}-e^{-Rx}\Bigr)

1422:   -\frac{N}{\hbar}

1423:   \int_{0}^{+\infty}ds \, 

1424:   \ln \Bigl(

1425:       e^{R(u(\underline{s}\mid\,\mu)-u(s\mid\,\mu))}

1426:       -e^{R(u(s\mid\,\mu)-u(\underline{s}\mid\,\mu))}

1427:       \Bigr)}

1428: \nonumber \\*

1429: &&

1430: ~~+O(\hbar^0)

1431: \nonumber \\[1.5mm]

1432: &&

1433: =

1434: \displaystyle{

1435:    \frac{N}{\hbar}

1436:    \int_0^{+\infty}dx\, 

1437:    \ln \Bigl( \frac{\sinh Rx}{R \Lambda}\Bigr)

1438:    -\frac{N}{\hbar}

1439:    \int_0^{+\infty}ds\, 

1440:    \ln \left\{

1441:          \frac{\sinh R(u(\underline{s}\mid\mu)-u(s\mid \mu))}

1442:               {R \Lambda}

1443:        \right\}}

1444: \nonumber \\*

1445: &&

1446: ~~+O(\hbar^0),  

1447: \label{asymptotic Phi}

1448: \end{eqnarray}

1449: where the last expression enables us  

1450: to make contact with the four-dimensional theory.  

1451: By plugging (\ref{asymptotic K}) and (\ref{asymptotic Phi})

1452: into (\ref{chi star 2}), 

1453: the large $|\mu|$ behavior of $\chi_q^*(z\mid\mu)$ can 

1454: be read as 

1455: \begin{eqnarray}

1456: &&

1457: \chi_q^*(z\mid \mu)

1458: \approx

1459: C\, \frac{N}{\hbar}

1460: \int_0^{+\infty} d\underline{s}

1461: \,\,z^{-\frac{N}{\hbar}u(\underline{s}\mid\, \mu)}

1462: \nonumber \\*[1.5mm]

1463: &&

1464: ~~~~~~~

1465: \times 

1466: \exp \left\{

1467: -\frac{RN}{2\hbar}u(\underline{s}\mid \mu)^2

1468: +\frac{N}{\hbar}

1469: \int_{-\infty}^{+\infty}

1470: dv \,

1471: \frac{ds(v\mid \mu)}{dv}

1472: \ln 

1473: \left(

1474: \frac{\sinh R(u(\underline{s}\mid\mu)-v)}{R\Lambda}

1475: \right)

1476: \right\}, 

1477: \nonumber \\*

1478: \label{asymptotic chi star 2}

1479: \end{eqnarray}

1480: where 

1481: $C \equiv 

1482: \exp \left\{ 

1483: \frac{N}{\hbar}

1484: \int_0^{+\infty} dx 

1485: \ln \left(\frac{\sinh Rx}{R\Lambda}\right)

1486: \right\}$. 

1487: The constant $C$ will be dropped out by absorbing it 

1488: into $\psi^*(z)$ as a wave function renormalization. 

1489: The density function appears in the above  

1490: by the change of variable, $v=u(s\mid\mu)$. 

1491: Thus we obtain the asymptotics (\ref{asymptotic chi star}). 

1492: 

1493: 

1494: 

1495: 

1496: 

1497: The large $|\mu|$ behavior of $\chi_q(z\mid\mu)$ 

1498: can be computed in the similar manner as above 

1499: by noting the relation (\ref{chi vs chi star}). 

1500: It becomes 

1501: \begin{eqnarray}

1502: &&

1503: \chi_q(z\mid \mu)

1504: \approx

1505: C^{-1}

1506: \frac{N}{\hbar}

1507: \int_0^{+\infty}

1508: d\underline{s}

1509: ~z^{-\frac{N}{\hbar}u(\underline{s}\mid\, \widetilde{\mu})}

1510: \nonumber \\*[1.5mm]

1511: &&

1512: ~~~~~~~

1513: \times 

1514: \exp \left\{

1515: \frac{RN}{2\hbar}u(\underline{s}\mid \widetilde{\mu})^2

1516: -\frac{N}{\hbar}

1517: \int_{-\infty}^{+\infty}

1518: dv\, 

1519: \frac{ds(v\mid \mu)}{dv}

1520: \ln 

1521: \left(

1522: \frac{\sinh R(-u(\underline{s}\mid\widetilde{\mu})-v)}{R\Lambda}

1523: \right)

1524: \right\}.  

1525: \nonumber \\

1526: \label{asymptotic chi 2}

1527: \end{eqnarray}

1528: The constant $C^{-1}$ will be dropped out. 

1529: In the above, the density function of $\mu$ has been used 

1530: instead of $\widetilde{\mu}$. These two are related by 

1531: \begin{eqnarray}

1532: \frac{ds}{du}(u\mid\widetilde{\mu})=

1533: -1+\frac{ds}{du}(-u\mid\mu). 

1534: \label{density reflection formula}

1535: \end{eqnarray}

1536: 

1537: 

1538: 

1539: 

1540: 

1541: 

1542: 

1543: 

1544: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1545: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1546: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1547: \section{Energy functions of partition} 

1548: 

1549: 

1550: 

1551: 

1552: 

1553: The Boltzmann weight for a partition $\mu$ is 

1554: $Q^{|\mu|}e^{V(\mu)}s_{\mu}(q^{-\rho})^2$. 

1555: The Boltzmann weight for a very large partition 

1556: can be measured by the so-called energy function.  

1557: The technique used in the previous section 

1558: is also relevant to obtain the energy function. 

1559: 

1560: 

1561: 

1562: Let $\mu$ be a partition of order $\hbar^{-2}$. 

1563: We rescale the Maya diagram $x_i(\mu)$ to 

1564: $u(s\mid\mu)$ by (\ref{u(s)}).  

1565: The asymptotics of the Boltzmann weight   

1566: follows from the large $|\mu|$ behaviors of 

1567: $s_{\mu}(q^{-\rho})$ and 

1568: the $N$-periodic potential $V(\mu)$.  

1569: The large $|\mu|$ behavior of 

1570: $s_{\mu}(q^{-\rho})$ 

1571: can be obtained by using (\ref{product formula a}).  

1572: The computations analogous to the previous section 

1573: give rise to  

1574: \begin{eqnarray}

1575: &&

1576: \ln s_{\mu}(q^{-\rho})^2

1577: \nonumber \\*[1.5mm]

1578: &&

1579: =\,

1580: \frac{N^2}{\hbar^2}

1581: \int_0^{+\infty}dx\, \,

1582: x \ln \left(\frac{\sinh Rx}{R \Lambda}\right)^2 

1583: -\frac{N^2}{\hbar^2}

1584: \int_{-\infty}^{+\infty}du \,

1585: \left( 1+2\frac{ds(u\mid\mu)}{du} \right)

1586: \frac{Ru^2}{2}

1587: \nonumber \\*[1.5mm]

1588: &&

1589: ~~~

1590: +\frac{N^2}{\hbar^2}

1591: \int \int_{-\infty <u<v< +\infty} 

1592: dudv \,

1593: \left(1+\frac{ds(u\mid\mu)}{du}\right)

1594: \frac{ds(v\mid\mu)}{dv}

1595: \ln 

1596: \left(

1597: \frac{\sinh R(u-v)}{R \Lambda}

1598: \right)^2 

1599: \nonumber \\*[1.5mm]

1600: &&

1601: ~~~

1602: +O(\hbar^{-1}).

1603: \label{asymptotic schur}

1604: \end{eqnarray}

1605: 

1606: 

1607: 

1608: 

1609: In order to make the asymptotics of 

1610: the $N$-periodic potential be of order $\hbar^{-2}$ 

1611: (the same order as (\ref{asymptotic schur})), 

1612: we should regard $\xi_r$ in the potential 

1613: as the order $\hbar^{-1}$ quantities and  

1614: rescale them as well. 

1615: Let $\zeta_r$ be the order $\hbar^0$ quantities 

1616: defined by the rescalings 

1617: \begin{eqnarray}

1618: \xi_r=\frac{\zeta_r}{\hbar}, 

1619: \label{zeta r}

1620: \end{eqnarray}

1621: where $\sum_{r=1}^N\zeta_r=0$. 

1622: In Appendix A, we compute the asymptotics of 

1623: the $N$-periodic potential 

1624: in a way different from \cite{Nekrasov-Okounkov}. 

1625: Our computations are based on the regularized density. 

1626: To carry out these computations 

1627: we need to make the following assumption; 

1628: the density functions are non-decreasing for 

1629: the partitions dominating near the thermodynamic limit.

1630: As the result, we obtain 

1631: \begin{eqnarray}

1632: V(\mu)=

1633: \frac{N}{\hbar^2}

1634: \sum_{r=1}^{N}

1635: \zeta_r

1636: \int_{u_r}^{u_{r-1}}du\,\, u 

1637: \frac{d^2s(u\mid\mu)}{du^2}

1638: +O(\hbar^{-1}),

1639: \label{asymptotic V}

1640: \end{eqnarray}

1641: where $u_r$ ($0 \leq r \leq N$) 

1642: are defined by the conditions 

1643: \begin{eqnarray}

1644: \left.

1645: \frac{d}{du}s(u\mid\mu)\,

1646: \right|_{u=u_r}

1647: =-\frac{r}{N}.

1648: \label{ur}

1649: \end{eqnarray}

1650: Our assumption ensures that 

1651: $u_r$ exist and are ordered 

1652: $u_r < u_{r-1}$ with 

1653: $u_{0,\,N}$$=\pm \infty$. 

1654: 

1655: 

1656: 

1657: 

1658: 

1659: 

1660: By using (\ref{asymptotic schur}) and (\ref{asymptotic V}) 

1661: we can see that the Boltzmann weight asymptotes to

1662: \footnote{

1663: The statistical model considered in this paper 

1664: is related \cite{quantum calabi-yau} 

1665: to topological string on a non-compact toric 

1666: Calabi-Yau threefold 

1667: \cite{topological vertex}. 

1668: The precise correspondence is described in 

1669: \cite{MNTT}.  

1670: $\hbar$ is identified with string coupling constant $g_{st}$.  

1671: The asymptotic expansion corresponds to 

1672: the perturbative string theory. 

1673: The identification $q=e^{-\frac{2R\hbar}{N}}$

1674: indicates that $\hbar$ appears in a pair with $1/N$ 

1675: as observed in (\ref{asymptotic boltzmann}). 

1676: This seems to indicate the correspondence with 

1677: $1/N$ expansion of the underlying gauge theory.}

1678: \begin{eqnarray}

1679: Q^{|\mu|}e^{V(\mu)}s_{\mu}(q^{-\rho})^2

1680: &\approx&

1681: C'

1682: Q^{|\mu|}

1683: \exp \left\{

1684: -\frac{N^2}{\hbar^2}

1685: E \left[s(\cdot \mid\mu) \right]

1686: \right\}, 

1687: \label{asymptotic boltzmann} 

1688: \end{eqnarray}

1689: where $E\left[s(\cdot \mid\mu) \right]$ 

1690: is the energy function of $\mu$ defined by

1691: \begin{eqnarray}

1692: &&

1693: E\left[ s(\cdot \mid\mu) \right] 

1694: \equiv 

1695: \int_{-\infty}^{+\infty} du\, 

1696: \left( 1+2\frac{ds(u\mid\mu)}{du} \right)

1697: \frac{Ru^2}{2}

1698: \nonumber \\*[1.5mm]

1699: &&~~~~~

1700: -\int \int_{-\infty <u<v< +\infty} 

1701: dudv \,

1702: \left(1+\frac{ds(u\mid\mu)}{du}\right)

1703: \frac{ds(v\mid\mu)}{dv}

1704: \ln 

1705: \left(

1706: \frac{\sinh R(u-v)}{R \Lambda}

1707: \right)^2 

1708: \nonumber \\*[1.5mm]

1709: &&~~~~~

1710: -\frac{1}{N}

1711: \sum_{r=1}^{N}

1712: \zeta_r

1713: \int_{u_r}^{u_{r-1}}

1714: du\, u 

1715: \frac{d^2s(u\mid\mu)}{du^2},

1716: \label{energy function} 

1717: \end{eqnarray} 

1718: and $C'\equiv 

1719: \exp \left\{

1720: \frac{N^2}{\hbar^2}

1721: \int_0^{+\infty}

1722: dx x \ln \left(\frac{\sinh Rx}{R \Lambda}\right)^2 

1723: \right\}$. The constant $C'$ will be dropped out. 

1724: At the four-dimensional limit the above energy function 

1725: becomes 

1726: \begin{eqnarray}

1727: &&

1728: E_{4d}\left[s(\cdot \mid\mu)\right]

1729: \equiv 

1730: \lim_{R \rightarrow 0}

1731: E\left[ s(\cdot \mid\mu) \right] 

1732: \nonumber \\*[1.5mm]

1733: &&

1734: ~~~

1735: =-\int \int_{-\infty <u<v< +\infty} 

1736: dudv \, 

1737: \left(1+\frac{ds(u\mid\mu)}{du}\right)

1738: \frac{ds(v\mid\mu)}{dv}

1739: \ln 

1740: \left(

1741: \frac{u-v}{\Lambda}

1742: \right)^2

1743: \nonumber \\*[1.5mm]

1744: &&

1745: ~~~~~~

1746: -\frac{1}{N}

1747: \sum_{r=1}^{N}

1748: \zeta_r

1749: \int_{u_r}^{u_{r-1}}

1750: du\, u 

1751: \frac{d^2s(u\mid\mu)}{du^2}. 

1752: \label{4d energy function}

1753: \end{eqnarray}

1754: 

1755: 

1756: 

1757: 

1758: 

1759: The energy functions (\ref{energy function}) 

1760: and (\ref{4d energy function}) are the functionals of $s(u\mid\mu)$. 

1761: It is also possible to express them 

1762: by using the so-called profile of a partition $\mu$.  

1763: In Appendix B, we present the descriptions in terms of the profile 

1764: and thereby compare  (\ref{energy function}) 

1765: and (\ref{4d energy function}) with the energy functions 

1766: used in \cite{Nekrasov-Okounkov}.  

1767: The energy function (\ref{energy function}) turns out to be 

1768: different from that used \cite{Nekrasov-Okounkov} 

1769: for a description of five-dimensional supersymmetric 

1770: $SU(N)$ Yang-Mills. 

1771: The difference originates in the existence of 

1772: the Chern-Simons term of five-dimensional gauge theories. 

1773: It is shown \cite{MNTT} that   

1774: the random plane partitions (\ref{Z SU(N)}) describe  

1775: five-dimensional supersymmetric $SU(N)$ Yang-Mills 

1776: with the Chern-Simons term at the thermodynamic limit. 

1777: The Chern-Simons correction appears in (\ref{energy function}) 

1778: as the $u^2$ potential term. 

1779: This term vanishes at the four-dimensional limit. 

1780: Therefore the energy function (\ref{4d energy function}) coincides 

1781: with that used \cite{Nekrasov-Okounkov} for the 

1782: description of four-dimensional $\mathcal{N}=2$ supersymmetric 

1783: $SU(N)$ Yang-Mills.

1784: 

1785: 

1786: 

1787: 

1788: 

1789: 

1790: 

1791: 

1792: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1793: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1794: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1795: \section{Semi-classical wave function and 

1796: Seiberg-Witten differential}

1797: 

1798: 

1799: 

1800: 

1801: 

1802: 

1803: In this section we investigate the thermodynamic limits 

1804: of the one-point functions 

1805: (\ref{Psi(lambda)star}) and (\ref{Psi(lambda)}). 

1806: We use the asymptotics obtained in the previous sections. 

1807: Near the thermodynamic limit 

1808: (the $\hbar \rightarrow 0$ limit), 

1809: partitions of order $\hbar^{-2}$ dominate 

1810: in the statistical model.   

1811: In order to describe the one-point functions at the 

1812: thermodynamic limit, 

1813: taking account of the expressions 

1814: (\ref{Psi(lambda)star 2}) and (\ref{Psi(lambda) 2}), 

1815: we need to find out the classical configurations 

1816: from among these dominant partitions. 

1817: As described in (\ref{asymptotic boltzmann}) 

1818: their Boltzmann weights are measured 

1819: by the energy function $E\left[s(\cdot\mid\mu)\right]$. 

1820: This means that the classical configurations 

1821: are the minimizers of $E\left[s(\cdot)\right]$.

1822: Let us suppose $s_{\star}(u)$ be such a minimizer. 

1823: 

1824: 

1825: 

1826: 

1827: The thermodynamic limits or 

1828: the semi-classical approximations  

1829: of the fermion one-point functions prove to  

1830: have a universal form in the sense that 

1831: it is irrespective of detail of the minimizers. 

1832: They turn out to be given by the following 

1833: WKB type wave functions. 

1834: \begin{eqnarray}

1835: \Psi_q^*(z)&\approx& 

1836: \exp \left\{

1837: -\frac{N}{\hbar}

1838: \int^{z}u(z)\,d \ln z 

1839: \right\}, 

1840: \label{Psi(lambda)star wkb}

1841: \\[1.5mm]

1842: \Psi_q(z)&\approx& 

1843: \exp \left\{

1844: \frac{N}{\hbar}

1845: \int^{z}u(z)\,d \ln z 

1846: \right\}.  

1847: \label{Psi(lambda) wkb}

1848: \end{eqnarray}

1849: In the above, 

1850: the function $u(z)$ is determined by the equation

1851: \begin{eqnarray}

1852: \frac{dS(u)}{du}=

1853: N \ln z \,, 

1854: \label{critical eq for 5d diagonal}

1855: \end{eqnarray}

1856: where 

1857: \begin{eqnarray}

1858: S(u)\equiv

1859: -\frac{RN}{2}u^2

1860: +N\int_{-\infty}^{+\infty}

1861: dv\, \frac{ds_{\star}(v)}{dv}

1862: \ln 

1863: \left(

1864: \frac{\sinh R(u-v)}{R\Lambda}

1865: \right). 

1866: \label{explicit classical action} 

1867: \end{eqnarray}

1868: 

1869: 

1870: 

1871: 

1872: %%%%%%%%%%%%% PROOF %%%%%%%%%%%%%%%%%

1873: 

1874: 

1875: 

1876: ~

1877: 

1878: 

1879: Let us confirm (\ref{Psi(lambda)star wkb}) and 

1880: (\ref{Psi(lambda) wkb}). 

1881: By following the mode expansions of 

1882: the complex fermions, 

1883: we first write the fermion one-point functions as 

1884: \begin{eqnarray}

1885: \Psi_q^*(z)= 

1886: \sum_{j=-\infty}^{+\infty}

1887: \widehat{\Psi}_q^*(j)\,z^{-j},

1888: ~~~~~~

1889: \Psi_q(j)= 

1890: \sum_{j=-\infty}^{+\infty}

1891: \widehat{\Psi}_q(j)\,z^{j-1}\,, 

1892: \label{Psi(lambda) mode}

1893: \end{eqnarray}

1894: where 

1895: $\widehat{\Psi}_q^*(j)$ and $\widehat{\Psi}_q(j)$ 

1896: are respectively the one-point functions of 

1897: $\psi^*_{j}$ and $\psi_{-j}$. 

1898: In other words 

1899: they are given by the contour integrals  

1900: \begin{eqnarray}

1901: \widehat{\Psi}_q^*(j)= 

1902: \oint \frac{dz}{2\pi i}\,

1903: z^{j-1} \, \Psi_q^*(z),

1904: ~~~~~~~

1905: \widehat{\Psi}_q(j)=

1906: \oint \frac{dz}{2\pi i}\,

1907: z^{-j}\, \Psi_q(z)\, , 

1908: \label{Psi(j)}

1909: \end{eqnarray}

1910: for each $j \in \mathbb{Z}$.

1911: 

1912: 

1913: 

1914: 

1915: 

1916: 

1917: Tracing the representations 

1918: (\ref{Psi(lambda)star 2}) and (\ref{Psi(lambda) 2})  

1919: these wave functions can be expressed as the following 

1920: statistical sums; 

1921: \begin{eqnarray}

1922: \widehat{\Psi}_q^*(j)&=&

1923: \frac{1}{Z_q^{SU(N)}}

1924: \sum_{\mu}

1925: Q^{|\mu|}

1926: e^{V(\mu)}

1927: s_{\mu}(q^{-\rho})^2\,

1928: \widehat{\psi}_q^*(j\mid \mu), 

1929: \label{Psi(j)star 2}\\[1.5mm]

1930: \widehat{\Psi}_q(j)&=&

1931: \frac{1}{Z^{SU(N)}_q}

1932: \sum_{\mu}

1933: Q^{|\mu|}

1934: e^{V(\mu)}

1935: s_{\mu}(q^{-\rho})^2\,

1936: \widehat{\psi}_q(j\mid \mu), 

1937: \label{Psi(j) 2}

1938: \end{eqnarray}

1939: where $\widehat{\psi}^*_q(j\mid\mu)$ 

1940: and $\widehat{\psi}_q(j\mid\mu)$ are obtained 

1941: from $\chi_q^*(z\mid\mu)$ and 

1942: $\chi_q(z\mid \mu)$ in 

1943: (\ref{chi star}) and (\ref{chi}) by   

1944: \begin{eqnarray}

1945: \widehat{\psi}_q^*(j\mid\mu)

1946: &=& 

1947: \oint 

1948: \frac{dz}{2\pi i}\,

1949: z^{j-1}\,

1950: \chi_q^*(z\mid\mu),

1951: \label{hat psi star j}\\[1.5mm]

1952: \widehat{\psi}_q(j\mid\mu)

1953: &=& 

1954: \oint 

1955: \frac{dz}{2\pi i}\,

1956: z^{-j}\,

1957: \chi_q(z\mid\mu). 

1958: \label{hat psi j}

1959: \end{eqnarray}

1960: 

1961: 

1962: 

1963: 

1964: 

1965: 

1966: The main contributions to $\chi_q^*(z\mid\mu)$ 

1967: and $\chi_q(z\mid \mu)$  

1968: come from $\widehat{\psi}_q^*(j\mid\mu)$ 

1969: and $\widehat{\psi}_q(j\mid\mu)$ with $j \sim \hbar^{-1}$

1970: when $\mu$ is a partition of order $\hbar^{-2}$. 

1971: These are seen from 

1972: (\ref{psi star mode on mu}) and (\ref{psi mode on mu}). 

1973: It follows that $\psi_j^*$ generically 

1974: subtracts $j$ boxes from the Young diagram 

1975: and $\psi_{-j}$ adds $j$ boxes to it, 

1976: when they act on the partition. 

1977: In particular,  

1978: $\widehat{\psi}_q^*(j\mid\mu)$ with $j \sim \hbar^{-1}$ 

1979: subtracts the same number of boxes from the partition. 

1980: It is also similar to 

1981: $\widehat{\psi}_q(j\mid\mu)$.  

1982: We rescale $j$ 

1983: in (\ref{hat psi star j}) and (\ref{hat psi j}) 

1984: to the order $\hbar^0$ quantity $u$ by  

1985: \begin{eqnarray}

1986: j=\frac{N}{\hbar}u. 

1987: \label{j u}

1988: \end{eqnarray}

1989: 

1990: 

1991: 

1992: 

1993: 

1994: Let us first consider the $\hbar$-expansions of 

1995: $\widehat{\psi}_q^*(j\mid\mu)$ and $\widehat{\psi}_q(j\mid\mu)$, 

1996: where we regard $j=O(\hbar^{-1})$ and $|\mu|=O(\hbar^{-2})$. 

1997: The semi-classical terms of the $\hbar$-expansions 

1998: can be obtained by plugging 

1999: (\ref{asymptotic chi star}) and (\ref{asymptotic chi}) 

2000: respectively into (\ref{hat psi star j}) and (\ref{hat psi j}) 

2001: with putting $j=Nu/\hbar$. 

2002: They turn out to be 

2003: \begin{eqnarray}

2004: \widehat{\psi}^*_q(j\mid\mu) 

2005: &\approx& 

2006: \exp\left\{

2007: \frac{1}{\hbar}

2008: S \left[\,u;s(\cdot\mid\mu)\right]\right\}, 

2009: \label{asymptotic hat star psi(j:mu)} 

2010: \\[1.5mm]

2011: \widehat{\psi}_q(j\mid\mu) 

2012: &\approx&

2013: \exp\left\{

2014: \frac{-1}{\hbar}

2015: S \left[\,u;s(\cdot\mid\mu)\right]\right\},  

2016: \label{asymptotic hat psi(j:mu)}

2017: \end{eqnarray}

2018: where 

2019: \begin{eqnarray}

2020: S \left[\,u;s(\cdot\mid\mu)\right]=

2021: -\frac{RN}{2}u^2

2022: +N\int_{-\infty}^{+\infty}

2023: dv\, \frac{ds(v\mid\mu)}{dv}

2024: \ln 

2025: \left(

2026: \frac{\sinh R(u-v)}{R\Lambda}

2027: \right). 

2028: \label{S[u;s]}

2029: \end{eqnarray}

2030: 

2031: 

2032: 

2033: 

2034: 

2035: 

2036: The semi-classical approximations of 

2037: $\widehat{\Psi}_q^*(j)$ and $\widehat{\Psi}_q(j)$ 

2038: can be obtained as follows. 

2039: Let $s_{\star}(u)$ be a minimizer 

2040: of the energy function $E\left[s(\cdot)\right]$.

2041: The representations (\ref{Psi(j)star 2})

2042: and (\ref{Psi(j) 2}) will be helpful. 

2043: Taking account of these statistical representations, 

2044: their semi-classical approximations are given by 

2045: (\ref{asymptotic hat star psi(j:mu)}) and  

2046: (\ref{asymptotic hat psi(j:mu)}) 

2047: evaluated at the minimizer $s_{\star}(u)$.  

2048: Thus we obtain 

2049: \begin{eqnarray}

2050: \widehat{\Psi}_q^*(j)

2051: &\approx& 

2052: \exp 

2053: \left\{ \frac{1}{\hbar}S(u) \right\}, 

2054: \label{Psi(u)star wkb}

2055: \\[1.5mm]

2056: \widehat{\Psi}_q(j)

2057: &\approx& 

2058: \exp 

2059: \left\{\frac{-1}{\hbar}S(u) \right\}, 

2060: \label{Psi(u) wkb}

2061: \end{eqnarray}

2062: where the classical action $S(u)$ is read from 

2063: (\ref{asymptotic hat star psi(j:mu)}) and 

2064: (\ref{asymptotic hat psi(j:mu)}) as 

2065: \begin{eqnarray}

2066: S(u)= 

2067: S \left[\,u; s_{\star}(\cdot) \right] 

2068: \label{classical action}

2069: \end{eqnarray}

2070: and given by (\ref{explicit classical action}).

2071: 

2072: 

2073: 

2074: 

2075: 

2076: The semi-classical approximations or 

2077: the thermodynamic limits of $\Psi_q^*(z)$ and $\Psi_q(z)$ 

2078: can be obtained from (\ref{Psi(u)star wkb})

2079: and (\ref{Psi(u) wkb}). 

2080: Taking account of the scaling (\ref{j u}) 

2081: we can replace the expressions (\ref{Psi(lambda) mode})  

2082: with the following integrals at the thermodynamic limit. 

2083: \begin{eqnarray}

2084: \Psi_q^*(z)&\approx&

2085: \frac{N}{\hbar}

2086: \int_{-\infty}^{+\infty}du \,

2087: \exp \left\{

2088: \frac{1}{\hbar}

2089: \left(

2090: -u \ln z^N+S(u) 

2091: \right)

2092: \right\}, 

2093: \label{Psi(lambda) star int}

2094: \\[1.5mm]

2095: \Psi_q(z)&\approx& 

2096: \frac{N}{\hbar}

2097: \int_{-\infty}^{+\infty}du \,

2098: \exp \left\{

2099: \frac{1}{\hbar}

2100: \left(

2101: u \ln z^N-S(u) 

2102: \right)

2103: \right\}.  

2104: \label{Psi(lambda) int}

2105: \end{eqnarray}

2106: Since $\hbar$ is very small, 

2107: we can evaluate the above integrations  

2108: by applying the saddle point method. 

2109: Critical points of the exponents become  

2110: solutions of the equation 

2111: (\ref{critical eq for 5d diagonal}).

2112: Let $u(z)$ be the critical point. 

2113: The critical value can be obtained as follows 

2114: \begin{eqnarray}

2115: u(z)\ln z^N-S(u(z))

2116: &=&

2117: \int^{u(z)}

2118: d \left(u\frac{dS(u)}{du}\right) 

2119: -\int^{u(z)}

2120: du \frac{dS(u)}{du}

2121: \nonumber \\[1.5mm]

2122: &=&

2123: \int^{u(z)}

2124: u d \left(\frac{dS(u)}{du}\right)

2125: \nonumber \\[1.5mm]

2126: &=&

2127: N\int^{z}

2128: u(z)\,d\ln z. 

2129: \label{critical value for 5d diagonal}

2130: \end{eqnarray}

2131: Therefore, the saddle point method gives rise 

2132: to the semi-classical wave functions 

2133: of the one-point functions at the main diagonal 

2134: as described in (\ref{Psi(lambda)star wkb}) and 

2135: (\ref{Psi(lambda) wkb}). 

2136: 

2137:  

2138: 

2139: 

2140: 

2141: 

2142: 

2143: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

2144: \subsubsection*{Semi-classical wave function 

2145: and Seiberg-Witten differential}

2146: 

2147: 

2148: 

2149: 

2150: 

2151: 

2152: Roles of the semi-classical wave functions 

2153: (\ref{Psi(lambda)star wkb}) and (\ref{Psi(lambda) wkb})

2154: in the Seiberg-Witten geometry are not obvious. 

2155: In order to get some insight we look 

2156: at the four-dimensional case.  

2157: Let us first supplement the previous discussions by 

2158: providing the four-dimensional limits of the relevant 

2159: quantities. The four-dimensional limits of 

2160: (\ref{asymptotic hat star psi(j:mu)}) and 

2161: (\ref{asymptotic hat psi(j:mu)}) become  

2162: \begin{eqnarray}

2163: \widehat{\psi}_{4d}^{*}(j\mid\mu) 

2164: \,\equiv\, 

2165: \lim_{R \rightarrow 0}

2166: \widehat{\psi}_q^*(j\mid\mu)

2167: &\approx& 

2168: \exp\left\{

2169: \frac{1}{\hbar}

2170: S_{4d} \left[\,u;s(\cdot\mid\mu)\right]\right\},

2171: \label{asymptotic 4d hat star psi(j:mu)}

2172: \\[1.5mm]

2173: \widehat{\psi}_{4d}(j\mid\mu) 

2174: \,\equiv\, 

2175: \lim_{R \rightarrow 0}

2176: \widehat{\psi}_q^*(j\mid\mu)

2177: &\approx& 

2178: \exp\left\{

2179: \frac{-1}{\hbar}

2180: S_{4d} \left[\,u;s(\cdot\mid\mu)\right]\right\}, 

2181: \label{asymptotic 4d hat psi(j:mu)}

2182: \end{eqnarray}

2183: where 

2184: \begin{eqnarray}

2185: S_{4d}\left[\,u;s(\cdot\mid\mu) \right]=

2186: N\int_{-\infty}^{+\infty}

2187: dv\, \frac{ds(v\mid\mu)}{dv}

2188: \ln 

2189: \left(

2190: \frac{u-v}{\Lambda}

2191: \right). 

2192: \label{4d S[u;s]}

2193: \end{eqnarray}

2194: Thus the minimizer $s_{\star}(u)$ of 

2195: $E_{4d}\left[s(\cdot)\right]$ leads to  

2196: the following semi-classical wave functions.   

2197: \begin{eqnarray}

2198: \widehat{\Psi}_{4d}^*(j)\,\equiv\, 

2199: \lim_{R \rightarrow 0}

2200: \widehat{\Psi}_q^*(j)

2201: &\approx& 

2202: \exp 

2203: \left\{ \frac{1}{\hbar}S_{4d}(u) \right\}, 

2204: \label{4d Psi(u)star wkb}

2205: \\[1.5mm]

2206: \widehat{\Psi}_{4d}(j)\,\equiv\, 

2207: \lim_{R\rightarrow 0}

2208: \widehat{\Psi}_q(j)

2209: &\approx& 

2210: \exp 

2211: \left\{\frac{-1}{\hbar}S_{4d}(u) \right\}, 

2212: \label{4d Psi(u) wkb}

2213: \end{eqnarray}

2214: where 

2215: \begin{eqnarray}

2216: S_{4d}(u)=

2217: S_{4d} \left[\,u; s_{\star}(\cdot) \right]. 

2218: \label{4d classical action}

2219: \end{eqnarray}

2220: 

2221: 

2222: 

2223: 

2224: 

2225: 

2226: We use the solution obtained by Nekrasov and Okounkov

2227: \cite{Nekrasov-Okounkov}. 

2228: Let ${\cal C}_{N}$ be the spectral curve of the $N$-periodic 

2229: Toda chain 

2230: \begin{eqnarray}

2231: h+\frac{1}{h}=\frac{P(x)}{\Lambda^N},  

2232: \label{spectral curve N-toda}

2233: \end{eqnarray}

2234: where $P(x)$ is the $N$-th order monic polynomial. 

2235: The curve ${\cal C}_N$ is identified 

2236: \cite{Toda curve} with the Seiberg-Witten curve 

2237: of ${\cal N}=2$ supersymmetric 

2238: $SU(N)$ Yang-Mills \cite{ Seiberg-Witten,SU(N) SW curve}. 

2239: The solution of \cite{Nekrasov-Okounkov} is expressed as 

2240: \begin{eqnarray}

2241: \int_{-\infty}^{+\infty}

2242: dv \,

2243: \frac{d^2s_{\star}(v)}{dv^2}

2244: \frac{1}{x-v}=

2245: \frac{1}{N}\frac{d \ln h}{dx}. 

2246: \label{4d solution}

2247: \end{eqnarray}

2248: It follows from (\ref{4d solution}) that

2249: the differential $dS_{4d}(x)$ equals to 

2250: \begin{eqnarray}

2251: dS_{4d}(x)

2252: =dx \ln h. 

2253: \label{dS}

2254: \end{eqnarray}

2255: Therefore, 

2256: the classical action (\ref{4d classical action}) 

2257: can be written as the following integral on ${\cal C}_N$; 

2258: \begin{eqnarray}

2259: S_{4d}(u)

2260: =\int^u dx\,  \ln h. 

2261: \label{4d classical action integral}

2262: \end{eqnarray}

2263: 

2264: 

2265: 

2266: 

2267: 

2268: The thermodynamic limits of the four-dimensional 

2269: counterparts of $\Psi_{q}^*(z)$ 

2270: and $\Psi_{q}^*(z)$ are described by 

2271: \begin{eqnarray}

2272: \Psi_{4d}^*(z)

2273: \,\equiv\,

2274: \lim_{R \rightarrow 0}

2275: \Psi_{q}^*(z)

2276: &\approx&

2277: \frac{N}{\hbar}

2278: \int_{-\infty}^{+\infty}du \, 

2279: \exp \left\{

2280: \frac{1}{\hbar}

2281: \left(

2282: -u \ln z^N+S_{4d}(u) 

2283: \right)

2284: \right\}, 

2285: \label{Psi(lambda) 4d star int}

2286: \\[1.5mm]

2287: \Psi_{4d}(z)

2288: \,\equiv\, 

2289: \lim_{R \rightarrow 0}

2290: \Psi_{q}(z)

2291: &\approx& 

2292: \frac{N}{\hbar}

2293: \int_{-\infty}^{+\infty}du \, 

2294: \exp \left\{

2295: \frac{1}{\hbar}

2296: \left(

2297: u \ln z^N-S_{4d}(u) 

2298: \right)

2299: \right\}.  

2300: \label{Psi(lambda) 4d int}

2301: \end{eqnarray}

2302: We can also apply the saddle point method  

2303: to the above integrations. 

2304: In the present case 

2305: equation for the critical point becomes 

2306: \begin{eqnarray}

2307: h=z^N. 

2308: \label{critical eq for 4d diagonal}

2309: \end{eqnarray}

2310: We then obtain the following semi-classical 

2311: wave functions.

2312: \begin{eqnarray}

2313: \Psi_{4d}^*(z)&\approx& 

2314: \exp \left\{

2315: -\frac{1}{\hbar}

2316: \int^{z}x \,d \ln h 

2317: \right\}, 

2318: \label{Psi(lambda) 4d star wkb}

2319: \\

2320: \Psi_{4d}(z)&\approx& 

2321: \exp \left\{

2322: \frac{1}{\hbar}

2323: \int^{z}x \,d \ln h 

2324: \right\}.  

2325: \label{Psi(lambda) 4d wkb}

2326: \end{eqnarray}

2327: Recall that $x d \ln h$ is nothing but 

2328: the Seiberg-Witten differential $dS_{s.w}$

2329: for the $SU(N)$ Yang-Mills. 

2330: Therefore we have shown above 

2331: that the semi-classical approximations 

2332: or the thermodynamic limits of the fermion one-point 

2333: functions are given by the WKB type wave functions 

2334: whose classical action is the integral of $dS_{s.w}$ 

2335: on the curve ${\cal C}_N$

2336: \begin{eqnarray}

2337: \Psi_{4d}^*(z)&\approx& 

2338: \exp \left\{

2339: -\frac{1}{\hbar}

2340: \int^{z}dS_{s.w}

2341: \right\}, 

2342: \label{Psi(lambda) 4d star sw}

2343: \\

2344: \Psi_{4d}(z)&\approx& 

2345: \exp \left\{

2346: \frac{1}{\hbar}

2347: \int^{z}dS_{s.w}

2348: \right\}.  

2349: \label{Psi(lambda) 4d sw}

2350: \end{eqnarray}

2351: 

2352: 

2353: 

2354: 

2355: Meromorphic function $h$ is the Floquet multiplier 

2356: in the spectral analysis of the 

2357: periodic Toda chain. 

2358: Equation (\ref{critical eq for 4d diagonal}) shows that 

2359: $z$ is nothing but the spectral parameter of the 

2360: associated linear problem. 

2361: This implies a hidden relation between 

2362: the above fermion wave functions and the Baker-Akhiezer 

2363: functions of the Toda hierarchy. 

2364: It is argued in \cite{Takasaki-Nakatsu} that 

2365: the isospectral problem (the $N$-band solitons) 

2366: is converted to an isomonodoromy problem 

2367: by imposing a homogeneity condition 

2368: analogous to the renormalization group equation.  

2369: The associated linear system turns to admit 

2370: a WKB analysis (multiscale analysis of 

2371: the isomonodoromy problem), 

2372: which gives the semi-classical wave functions 

2373: (\ref{Psi(lambda) 4d star sw}) and 

2374: (\ref{Psi(lambda) 4d sw}). 

2375: 

2376: 

2377: 

2378: 

2379: 

2380: 

2381: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

2382: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

2383: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

2384: \section{One-point functions away from the main diagonal} 

2385: 

2386: 

2387: 

2388: 

2389: 

2390: 

2391: $\Gamma(m)$ in the transfer matrix  

2392: is the hamiltonian operator at the discretized time $m$. 

2393: The discretized time becomes a continuous time $t$ 

2394: at the $\hbar \rightarrow 0$ limit. 

2395: In fact, 

2396: $t$ is identified with a coordinate of the limit 

2397: shape \cite{Okounkov-Reshetikhin} 

2398: of random plane partition. 

2399: The limit shape is interpreted 

2400: \cite{quantum calabi-yau} 

2401: as the mirror of the toric $\mathbb{C}^3$. 

2402: Free fermions  

2403: $\psi^*(z)$ and $\psi(z)$ 

2404: have been proposed \cite{brane probe} as probe branes 

2405: in the mirror. In this section we compute 

2406: the semi-classical wave functions of the fermions 

2407: located away from the main diagonal. 

2408: We restrict to the $U(1)$ case. 

2409: In particular, we put $q=e^{-2R\hbar}$. 

2410: We compute these wave functions by applying 

2411: the method developed in \cite{Okounkov-Reshetikhin}.  

2412: It turns out that $t$ is identical to the scale parameter 

2413: of the gauge theory. 

2414: 

2415: 

2416: 

2417: 

2418:  

2419: One-point functions at the discretized time 

2420: $m_0 \in \mathbb{Z}_{\leq 0}$ are given by 

2421: \begin{eqnarray}

2422: \Psi^*_q(z;\,m_0)

2423: &\equiv& 

2424: \frac{1}{Z^{U(1)}_q}

2425: \langle \emptyset ;\, -1|

2426: \Biggl\{

2427: \prod_{m=-\infty}^{m_0-1}\Gamma(m)

2428: \Biggr\}\,

2429: \psi^*(z)\,

2430: \Biggl\{

2431: \prod_{m=m_0}^{-1}\Gamma(m)

2432: \Biggr\}\,

2433: Q^{L_0}\,

2434: \Biggl\{

2435: \prod_{m=0}^{+\infty}\Gamma(m)

2436: \Biggr\}

2437: |\emptyset ;\,0\rangle ,

2438: \nonumber \\

2439: &&

2440: \label{Psi m0 star}

2441: \\[1.5mm]

2442: \Psi_q(z;\,m_0)

2443: &\equiv&

2444: \frac{1}{Z^{U(1)}_q}

2445: \langle \emptyset;\, +1|

2446: \Biggl\{

2447: \prod_{m=-\infty}^{m_0-1}\Gamma(m)

2448: \Biggr\}\,

2449: \psi(z)\,

2450: \Biggl\{

2451: \prod_{m=m_0}^{-1}\Gamma(m)

2452: \Biggr\}\,

2453: Q^{L_0}\,

2454: \Biggl\{

2455: \prod_{m=0}^{+\infty}\Gamma(m)

2456: \Biggr\}

2457: |\emptyset;\,0\rangle .

2458: \nonumber \\

2459: &&

2460: \label{Psi m0}

2461: \end{eqnarray}

2462: Note that one-point functions at a positive 

2463: discretized time are defined in the same way as above,     

2464: and that their computations become similar to those  

2465: at the negative time presented below.     

2466: 

2467: 

2468: 

2469: 

2470: 

2471: The above one-point functions can be calculated exactly. 

2472: Let us define the adjoint actions of $\Gamma(m)$ 

2473: on the fermions as  

2474: \begin{eqnarray}

2475: Ad\Bigl(\Gamma(m)\Bigr)\,

2476: \psi^*(z)=

2477: \Gamma(m)\,\psi^*(z)\,\Gamma(m)^{-1},~~~~

2478: Ad\Bigl(\Gamma(m)\Bigr)\,

2479: \psi(z)=

2480: \Gamma(m)\,\psi(z)\,\Gamma(m)^{-1}. 

2481: \label{def adjoint}

2482: \end{eqnarray}

2483: By using (\ref{def adjoint}) 

2484: we can evaluate 

2485: (\ref{Psi m0 star}) and (\ref{Psi m0}) as follows; 

2486: \begin{eqnarray}  

2487: \Psi^*_q(z;\,m_0)&=&

2488: \langle \emptyset;\, -1|\,

2489: Ad

2490: \Bigl( 

2491: \prod_{m=0}^{+\infty}\Gamma(m)^{-1}

2492: \Bigr)

2493: Ad

2494: \left(

2495: Q^{-L_0}

2496: \right)

2497: Ad

2498: \Bigl( 

2499: \prod_{m=-\infty}^{m_0-1}\Gamma(m)

2500: \Bigr)

2501: \psi^*(z)\,

2502: |\emptyset;\,0 \rangle, 

2503: \nonumber 

2504: \\

2505: \label{Psi m0 star 2}

2506: \\[1.5mm]

2507: \Psi_q(z;\,m_0)&=&

2508: \langle \emptyset;\, +1|\,

2509: Ad

2510: \Bigl( 

2511: \prod_{m=0}^{+\infty}\Gamma(m)^{-1}

2512: \Bigr) 

2513: Ad

2514: \left( 

2515: Q^{-L_0}

2516: \right)

2517: Ad

2518: \Bigl( 

2519: \prod_{m=-\infty}^{m_0-1}\Gamma(m)

2520: \Bigr)

2521: \psi(z)\,

2522: |\emptyset;\,0 \rangle.   

2523: \nonumber \\ 

2524: \label{Psi m0 2}

2525: \end{eqnarray}

2526: The adjoint actions (\ref{def adjoint}) 

2527: are computed from (\ref{gamma m}). 

2528: They become 

2529: \begin{eqnarray}

2530: Ad \Bigl(\Gamma(m)\Bigr) 

2531: \psi^*(z)

2532: &=&

2533: \left\{

2534: \begin{array}{cl}

2535:  \left(1-q^{m+\frac{1}{2}}z^{-1}\right) 

2536:  \psi^*(z) 

2537: &

2538: ~~~~~

2539: \mbox{for}~~m \geq 0 \\[1.5mm]

2540:  \left(1-q^{-m-\frac{1}{2}}z \right) 

2541:  \psi^*(z) 

2542: &

2543: ~~~~~

2544: \mbox{for}~~m \leq -1,

2545: \end{array}

2546: \right.

2547: \label{ad m psi star} \\[1.5mm]

2548: Ad \Bigl(\Gamma(m)\Bigr) 

2549: \psi(z)

2550: &=&

2551: \left\{

2552: \begin{array}{cl}

2553:   \left(1-q^{m+\frac{1}{2}}z^{-1} \right)^{-1}

2554:   \psi(z) 

2555: &

2556: ~~~~~

2557: \mbox{for}~~m \geq 0 \\[1.5mm]

2558:    \left(1-q^{-m-\frac{1}{2}}z \right)^{-1}

2559:    \psi(z) 

2560: &

2561: ~~~~~

2562: \mbox{for}~~m \leq -1.

2563: \end{array}

2564: \right.

2565: \label{ad m psi}

2566: \end{eqnarray}

2567: By plugging (\ref{ad m psi star}) and (\ref{ad m psi}) 

2568: into (\ref{Psi m0 star 2}) and (\ref{Psi m0 2}) 

2569: we obtain 

2570: \begin{eqnarray}

2571: \Psi^*_q(z;\,m_0)\,=\,

2572: \frac{ 

2573:    \Bigl(q^{-m_0+\frac{1}{2}}z;q \Bigr)_{\infty}}

2574: {\Bigl(q^{\frac{1}{2}}Qz^{-1};q\Bigr)_{\infty}},

2575: ~~~~~~

2576: \Psi_q(z;m_0)

2577: \,=\,

2578: \Psi^*_q(z;\,m_0)^{-1},  

2579: \label{exact form}

2580: \end{eqnarray}

2581: where the infinite product 

2582: $(a;q)_{\infty}\equiv \prod_{n=0}^{+\infty}(1-aq^n)$ 

2583: is used.

2584: 

2585: 

2586: 

2587: 

2588: 

2589: 

2590: 

2591: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

2592: \subsubsection*{Semi-classical wave functions}

2593: 

2594: 

2595: 

2596: 

2597: 

2598: 

2599: Asymptotics of the higher modes of 

2600: $\Psi_q^*(z;\,m_0)$ and $\Psi_q(z;\,m_0)$ 

2601: can be studied as in accord with the previous discussion 

2602: at the main diagonal. 

2603: Let $j \in \mathbb{Z}$. 

2604: We introduce 

2605: \begin{eqnarray}

2606: \widehat{\Psi}_q^*(j;\,m_0)

2607: &\equiv&

2608: \oint \frac{dz}{2\pi i}\,

2609: z^{j-1}\, 

2610: \Psi^*_q(z;m_0), ~~~~~

2611: \label{psi star j m0} \\

2612: \widehat{\Psi}_q(j;\,m_0)

2613: &\equiv&

2614: \oint \frac{dz}{2\pi i}\,

2615: z^{-j}\,

2616: \Psi_q(z;m_0)  . 

2617: \label{psi j m0}

2618: \end{eqnarray}

2619: 

2620: 

2621: 

2622: 

2623: To find out their asymptotics 

2624: we need to rescale the integers $j$ and $m_0$ appropriately. 

2625: Guided by (\ref{j u}), 

2626: we regard $m_0$ also of order $\hbar^{-1}$ 

2627: and introduce the order $\hbar^0$ quantities $u$ and $t$ by 

2628: \begin{eqnarray}

2629: j=\frac{u}{\hbar},~~~~~~~~

2630: m_0=\frac{t}{\hbar}, 

2631: \label{u and t}

2632: \end{eqnarray}

2633: where $u \in \mathbb{R}$ and $t \in \mathbb{R}_{\leq 0}$. 

2634: Let us consider the $\hbar$-expansions of 

2635: $\widehat{\Psi}^*_q\left(\frac{u}{\hbar};\frac{t}{\hbar}\right)$

2636: and 

2637: $\widehat{\Psi}_q\left(\frac{u}{\hbar};\frac{t}{\hbar}\right)$ 

2638: ($q=e^{-2R\hbar}$). 

2639: They turn out to be written in the following form.  

2640: \begin{eqnarray}

2641: \widehat{\Psi}_q^*

2642: \left(\frac{u}{\hbar}\,;\,\frac{t}{\hbar}\right)

2643: &=&

2644: \oint \frac{dz}{2\pi i}\,

2645: \exp \left\{ 

2646: \frac{1}{\hbar}S^{(0)}(z;(u,t))

2647: +

2648: O(\hbar^0)

2649: \right\}, 

2650: \label{psi(u,t) star}

2651: \\[1.5mm]

2652: \widehat{\Psi}_q

2653: \left(\frac{u}{\hbar}\,;\,\frac{t}{\hbar}\right)

2654: &=&

2655: \oint \frac{dz}{2\pi i}\,

2656: \exp \left\{ 

2657: -\frac{1}{\hbar}S^{(0)}(z;(u,t))

2658: +

2659: O(\hbar^0)

2660: \right\}.  

2661: \label{psi(u,t)}

2662: \end{eqnarray}

2663: The classical action $S^{(0)}(z;(u,t))$ 

2664: in the above can be computed as follows. 

2665: We first notice that the $\hbar$-expansion of 

2666: $\ln (a;q)_{\infty}$ reads as 

2667: \begin{eqnarray}

2668: \ln (a;q)_{\infty}=

2669: \frac{1}{2R\hbar}

2670: \int_0^a dx \,

2671: \frac{\ln (1-x)}{x}

2672: +O(\hbar^0). 

2673: \label{hbar expansion ln(a;q)}

2674: \end{eqnarray}

2675: Then, (\ref{exact form}) leads to  

2676: \begin{eqnarray}

2677: &&

2678: \ln z^{j-1} \,\Psi_q^*(z;\,m_0) 

2679: \nonumber \\*[1.5mm]

2680: &&

2681: ~~~

2682: =(j-1) \ln z 

2683: +\ln 

2684: \left(q^{-m_0+\frac{1}{2}}z;q \right)_{\infty}

2685: -\ln 

2686: \left(q^{\frac{1}{2}}Qz^{-1};q\right)_{\infty} 

2687: \nonumber \\[1.5mm]

2688: &&

2689: ~~~

2690: =

2691: \frac{1}{\hbar}

2692: \left\{ 

2693: u \ln z 

2694: +\frac{1}{2R}\int_0^{z}dx \,  

2695: \frac{\ln \left( 1-e^{2Rt}x \right)}{x}

2696: -\frac{1}{2R}\int_0^{z^{-1}}dx \,  

2697: \frac{\ln \left(1-Qx \right)}{x}

2698: \right\} 

2699: \nonumber \\

2700: &&

2701: ~~~~~

2702: +

2703: O(\hbar^0). 

2704: \label{asymptotic ln psi star}

2705: \end{eqnarray}

2706: Therefore we obtain 

2707: \begin{eqnarray}

2708: &&

2709: S^{(0)}(z;(u,t))

2710: \nonumber \\*[1.5mm]

2711: &&

2712: ~~~

2713: =

2714: u\ln z 

2715: +\frac{1}{2R}\int_0^{z}dx \,

2716: \frac{\ln \left( 1-e^{2Rt}x \right)}{x}

2717: -\frac{1}{2R}\int_0^{z^{-1}}dx \,  

2718: \frac{\ln \left(1-Qx \right)}{x}. 

2719: \label{S(lambda;(u,t))}

2720: \end{eqnarray}

2721: 

2722: 

2723: 

2724: 

2725: 

2726: 

2727: Since $\hbar$ is very small,  

2728: the saddle point method becomes applicable  

2729: to the contour integrals 

2730: in (\ref{psi(u,t) star}) and (\ref{psi(u,t)}). 

2731: The critical points of $S^{(0)}(z;(u,t))$ 

2732: turn to be the solutions $\alpha, \bar{\alpha}$ 

2733: of the quadratic equation 

2734: \begin{eqnarray}

2735: e^{2Rt}z^2

2736: -(1+e^{2Rt}Q-e^{-2Rt})z

2737: +Q=0. 

2738: \label{quadratic eq}

2739: \end{eqnarray}

2740: This allows us to write the critical points as    

2741: \begin{eqnarray}

2742: \alpha,\bar{\alpha}=

2743: Q^{\frac{1}{2}} e^{-Rt\pm i \theta_{\star}(u,t)}.  

2744: \label{alpha}

2745: \end{eqnarray}

2746: The phase $\theta_{\star}(u,t)$ 

2747: in (\ref{alpha}) satisfies the equation    

2748: \begin{eqnarray}

2749: \cos \theta_{\star}(u,t)=

2750: R\Lambda_{eff}(t)+

2751: \frac{1}{2\Lambda_{eff}(t)}

2752: \frac{1-e^{-2Ru}}{2R}, 

2753: \label{theta star equation}

2754: \end{eqnarray}

2755: where 

2756: $\Lambda_{eff}(t)

2757: \equiv  

2758: \Lambda e^{Rt}$. 

2759: 

2760: 

2761: 

2762: 

2763: 

2764: 

2765: 

2766: The saddle point method gives rise to 

2767: the following semi-classical wave functions.  

2768: \begin{eqnarray}

2769: \widehat{\Psi}_q^*

2770: \left(\frac{u}{\hbar}\,;\,\frac{t}{\hbar}\right)

2771: &\approx& 

2772: \exp \left\{

2773: \frac{1}{\hbar}

2774: S^{(0)}

2775: (\alpha;(u,t)) \right\},~~~~~~

2776: \label{psi star wkb} 

2777: \\

2778: \widehat{\Psi}_q 

2779: \left(\frac{u}{\hbar}\,;\,\frac{t}{\hbar}\right)

2780: &\approx&

2781: \exp \left\{

2782: \frac{-1}{\hbar}

2783: S^{(0)}

2784: (\alpha;(u,t)) \right\},

2785: \label{psi wkb}

2786: \end{eqnarray}

2787: where the critical point $\alpha$ is taken 

2788: for simplicity. The critical value

2789: $S^{(0)}(\alpha;(u,t))$ can be read easily 

2790: up to the $u$-independent term since we have 

2791: \begin{eqnarray}

2792: \frac{d}{du}S^{(0)}(\alpha;(u,t))=\ln \alpha.

2793: \label{dS(0)/du}

2794: \end{eqnarray}

2795: The integration of (\ref{dS(0)/du}) gives rise to 

2796: \begin{eqnarray}

2797: S^{(0)}(\alpha;(u,t))

2798: &=&

2799: -Rtu

2800: +\frac{u}{2}\ln Q  

2801: +i\int^u dx\, \theta_{\star}(x,t)

2802: \nonumber \\

2803: &&

2804: + u\mbox{-independent term}.

2805: \label{critical value}

2806: \end{eqnarray}

2807: 

2808: 

2809: 

2810: 

2811: 

2812: 

2813: 

2814: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

2815: \subsubsection*{Family of curves} 

2816: 

2817: 

2818: 

2819: 

2820: 

2821: 

2822: 

2823: The above semi-classical analysis has its interpretation 

2824: in complex geometry. To explain this, 

2825: it is convenient to start with the four-dimensional case. 

2826: It follows from (\ref{theta star equation}) that 

2827: the phase $\theta_{\star}(u,t)$ becomes independent of 

2828: $t$ at the limit $R \rightarrow 0$. 

2829: Equation (\ref{theta star equation}) is translated to 

2830: \begin{eqnarray}

2831: \cos \theta_{\star}^{4d}(u)=\frac{u}{2\Lambda}, 

2832: \label{4d theta star equation}

2833: \end{eqnarray}

2834: where  

2835: $\theta_{\star}^{4d}(u) 

2836: \equiv 

2837: \lim_{R \rightarrow 0}

2838: \theta_{\star}(u,t)$. 

2839: The phase $\theta_{\star}^{4d}$ can be thought 

2840: as a meromorphic function on a Riemann surface. 

2841: Let $\mathcal{C}$ be the curve ($\mathbb{P}^1$) 

2842: \begin{eqnarray}

2843: y^2=x^2-4\Lambda^2,   

2844: \label{4d U(1) curve}

2845: \end{eqnarray}

2846: and $h$ be the meromorphic function on 

2847: $\mathcal{C}$ 

2848: \begin{eqnarray}

2849: h=\frac{x+y}{2\Lambda}. 

2850: \label{4d U(1) h}

2851: \end{eqnarray}

2852: Then we can write equation (\ref{4d theta star equation}) as  

2853: \begin{eqnarray} 

2854: \theta_{\star}^{4d}(u)=

2855: \left. \frac{1}{i}\ln h \, \right|_{x=u+i0}. 

2856: \label{4d theta vs h}

2857: \end{eqnarray}

2858: The classical action becomes an integral 

2859: of $dx \ln h$ on the curve $\mathcal{C}$. 

2860: \begin{eqnarray}

2861: i\int^udx \,\theta^{4d}_{\star}(x)=

2862: \int^u dx \,\ln h. 

2863: \label{4d action integral}

2864: \end{eqnarray}

2865: It is tempting to think $x$ and $\ln h$ in the above 

2866: as the action-angle variables of the one-dimensional Toda chain.  

2867: Note that the curve $\mathcal{C}$ can be expressed as 

2868: \begin{eqnarray}

2869: h+\frac{1}{h}=\frac{x}{\Lambda}. 

2870: \label{spectral curve of infinite Toda}

2871: \end{eqnarray} 

2872: This is the spectral curve of the infinite Toda chain. 

2873: 

2874: 

2875: 

2876: 

2877: 

2878: 

2879: 

2880: We now move on to the five-dimensional case. 

2881: For each $t \in \mathbb{R}_{\leq 0}$, we associate 

2882: the curve $\mathcal{C}_t$  

2883: \begin{eqnarray}

2884: y^2=z^2-4\Lambda_{eff}^2(t), 

2885: \label{5d U(1) curve t}

2886: \end{eqnarray}

2887: and the meromorphic function $h_t$ 

2888: \begin{eqnarray}

2889: h_t\equiv \frac{y+z}{2\Lambda_{eff}(t)}. 

2890: \label{5d U(1) ht}

2891: \end{eqnarray} 

2892: The identification of $u$ with $z$ gets 

2893: slightly involved in the five-dimensional theory. 

2894: Let $\mathbb{C}/\mathbb{Z}$ be 

2895: the cylinder obtained by identifying $x$ 

2896: with $x+ \frac{\pi i}{R}n$. 

2897: We will regard $u$ as the real line of the cylinder. 

2898: Note that the cylinder becomes $\mathbb{C}$ as $R \rightarrow 0$. 

2899: We introduce the holomorphic function on $\mathbb{C}/\mathbb{Z}$ by 

2900: \begin{eqnarray}

2901: z_t(x)\equiv 

2902: 2R\Lambda_{eff}^2(t)+

2903: \frac{1-e^{-2Rx}}{2R}. 

2904: \label{z(x)t}

2905: \end{eqnarray}

2906: Then the equation (\ref{theta star equation}) 

2907: can be written as 

2908: \begin{eqnarray}

2909: \theta_{\star}(u,t)=

2910: \left.

2911: \frac{1}{i}

2912: \ln h_t \, 

2913: \right|_{z=z_t(u+i0)}. 

2914: \label{theta vs h}

2915: \end{eqnarray}

2916: This shows that $\theta_{\star}$ has the same form 

2917: as the four-dimensional counterpart. 

2918: Effect of the fifth dimension is encoded in $z_t(x)$ 

2919: such that (\ref{theta vs h}) smoothly reduces to 

2920: (\ref{4d theta vs h}) as $R \rightarrow 0$. 

2921: In fact, we have 

2922: $\lim_{R \rightarrow 0}z_t(x)=x$ and 

2923: $\lim_{R \rightarrow 0}\Lambda_{eff}(t)=\Lambda$.  

2924: Therefore (\ref{theta vs h}) becomes 

2925: (\ref{4d theta vs h}) at the four-dimensional limit. 

2926: The classical action becomes the following integral 

2927: on the curve ${\cal C}_t$. 

2928: \begin{eqnarray}

2929: i\int^u dx \, \theta_{\star}(x,t) 

2930: &=&

2931: \int^u dx \, \ln h_t(z_t(x)) 

2932: \nonumber \\

2933: &=&

2934: \int^{z_t(u+i0)}

2935: dz 

2936: \left(\frac{dz_t}{dx}\right)^{-1} \ln h_t. 

2937: \label{classical action t}

2938: \end{eqnarray}

2939: 

2940: 

2941: 

2942: 

2943: 

2944: 

2945: 

2946: 

2947: An exact solution of the five-dimensional theory can be 

2948: obtained from the geometrical data 

2949: (\ref{5d U(1) curve t}), (\ref{5d U(1) ht}) 

2950: and (\ref{z(x)t}) at $t=0$. 

2951: It should be noted that the time dependence of 

2952: these data is only via $\Lambda_{eff}(t)$. 

2953: Recall that $\Lambda=\Lambda_{eff}(0)$ 

2954: should be identified with the lambda parameter of the gauge theory. 

2955: The standard dimensional argument shows that 

2956: the renormalization group flow is realized effectively 

2957: by scaling the lambda parameter  

2958: although it is originally a RG invariant. 

2959: We conjecturally identify the above time evolution 

2960: of the system with the RG flow. 

2961: Since $\Lambda_{eff}(t)=\Lambda e^{2Rt}$ ($t \leq 0$), 

2962: it becomes zero at $t=-\infty$ and 

2963: the curve (\ref{5d U(1) curve t}) 

2964: gets degenerate to $y^2=z^2$. 

2965: We expect that 

2966: the geometrical data near $t=\pm \infty$, 

2967: reflecting the holographic principle, 

2968: describes the gauge theory in the perturbative regime  

2969: \footnote{

2970: The analysis of the one-point functions at a

2971: positive time leads to the same geometrical data 

2972: with a slight change of $\Lambda_{eff}(t)$. 

2973: Namely, $\Lambda_{eff}(t)=\Lambda e^{-2Rt}$ 

2974: for $t \geq 0$. 

2975: We have $\lim_{t \rightarrow +\infty}\Lambda_{eff}(t)=0$. 

2976: Therefore $\mathcal{C}_t$ becomes the same degenerate curve 

2977: at $t=+\infty$. }. 

2978: This issue will be reported elsewhere 

2979: \cite{MNTT in progress} from the viewpoint of 

2980: integrable systems. 

2981: 

2982: 

2983: 

2984: 

2985: 

2986: 

2987: 

2988: 

2989: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

2990: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

2991: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

2992: \appendix 

2993: \section{Proof of equation (\ref{asymptotic V})}

2994: 

2995: 

2996: 

2997: 

2998: 

2999: 

3000: We first express 

3001: the $N$-periodic potential (\ref{N-periodic potential}) 

3002: in a form relevant to study its asymptotics. 

3003: It is convenient to consider partitions 

3004: paired with the $U(1)$ charges. 

3005: Let $(\mu,n)$ be the charged partition, 

3006: where $\mu$ is a partition and $n$ is the $U(1)$ charge. 

3007: The states $|\mu ; n \rangle$ constitute bases of the Fock space 

3008: of a single complex fermion.  

3009: System of $N$ component fermions is realized 

3010: \cite{Miwa-Jimbo} on this Fock space. 

3011: Thereby any charged partition $(\mu,n)$ can 

3012: be expressed in a unique way 

3013: as a set of $N$ charged partitions $(\lambda^{(r)},p_r)$ 

3014: and vice versa. In terms of the Maya diagrams 

3015: the correspondence can be read as follows;  

3016: \begin{eqnarray}

3017: \Bigl\{ n+x_{i}(\mu) \Bigr\}_{i \geq 1}=

3018: \bigcup_{r=1}^N~

3019: \Bigl\{ N(p_r+x_{i_r}(\lambda^{(r)}))+r-1 

3020: \Bigr\}_{i_r \geq 1}. 

3021: \label{mu vs lambda(r)}

3022: \end{eqnarray}

3023: The periodic potential $V(\mu)$ is shown 

3024: \cite{Nekrasov-Okounkov} to be $\sum_{r=1}^N\xi_rp_r$,  

3025: where $p_r$ are the $U(1)$ charges of 

3026: the $N$ charged partitions corresponding to $(\mu,0)$.  

3027: 

3028: 

3029: 

3030: 

3031: For a charged partition $(\mu,n)$, 

3032: we introduce the density 

3033: $\rho(x \mid \mu \,; n)$ by 

3034: \begin{eqnarray}

3035: \rho(x \mid \mu \,; n)\equiv

3036: \sum_{i=1}^{+\infty}

3037: \delta(x-n-x_{i}(\mu)). 

3038: \label{rho(mu:n)}

3039: \end{eqnarray}

3040: In the neutral case we simply denote 

3041: $\rho(x\mid\mu) \equiv \rho(x\mid \mu\,; 0)$. 

3042: The density (\ref{rho(mu:n)}) is not sensitive to the 

3043: $U(1)$ charge since $n$ could be absorbed into a shift 

3044: of $x$. It is convenient to modify the above density as   

3045: \begin{eqnarray}

3046: \rho_{reg}(x\mid \mu\,; n)=

3047: \sum_{i=1}^{+\infty}

3048: \delta(x-n-x_{i}(\mu))

3049: -\sum_{i=1}^{+\infty}\delta(x+i), 

3050: \label{rho reg (mu,n)}

3051: \end{eqnarray}

3052: where the subtraction is prescribed so that it satisfies 

3053: \begin{eqnarray}

3054: \int_{-\infty}^{+\infty}dx\, \rho_{reg}(x\mid \mu\,; n)=n.

3055: \label{reg condition}

3056: \end{eqnarray}

3057: The following expression of the $N$-periodic potential 

3058: becomes important in the subsequent discussion. 

3059: \begin{eqnarray}

3060: V(\mu)=\sum_{r=1}^{N}\xi_r 

3061: \int_{-\infty}^{+\infty}dx \,

3062: \rho_{reg}(x\mid \lambda^{(r)}\,; p_r), 

3063: \label{N-periodic potential 2}

3064: \end{eqnarray}

3065: where $(\lambda^{(r)},p_r)$ are the charged partitions 

3066: which describe $(\mu,0)$. Note that $\sum_{r=1}^Np_r=0$.

3067: 

3068: 

3069: 

3070: 

3071: When $\lambda^{(r)}$ are of order $\hbar^{-2}$ and  

3072: $p_r$ are of order $\hbar^{-1}$, 

3073: we rescale the variables as 

3074: \begin{eqnarray}

3075: x=\frac{u}{\hbar},~~~

3076: i_r=\frac{s}{\hbar},~~~

3077: p_r=\frac{\eta_r}{\hbar},~~~

3078: x_{i_r}(\lambda^{(r)})

3079: =\frac{u(s\mid\lambda^{(r)})}{\hbar}+O(\hbar^0).   

3080: \label{rescale N pairs}

3081: \end{eqnarray}

3082: The asymptotics of the densities for $(\lambda^{(r)},p_r)$ 

3083: as $\hbar \rightarrow 0$ can be computed 

3084: in the similar manner as in the text. They become 

3085: \begin{eqnarray}

3086: \rho(x\mid \lambda^{(r)}\,; p_r)&=&

3087: -\frac{ds(u-\eta_r\mid\lambda^{(r)})}{du}+O(\hbar^1), 

3088: \label{rho vs ds}

3089: \\

3090: \rho_{reg}(x\mid \lambda^{(r)}\,; p_r)&=&

3091: -\left\{

3092: \frac{ds(u-\eta_r\mid\lambda^{(r)})}{du} 

3093: +\theta(-u)

3094: \right \}

3095: +O(\hbar^1),  

3096: \label{rho reg vs ds}

3097: \end{eqnarray}

3098: where $\theta(u)$ is the step function, that is, 

3099: $\theta(u)=1$ for $u>0$, and $0$ for $u<0$. 

3100: 

3101: 

3102: 

3103: 

3104: 

3105: 

3106: 

3107: Thanks to the correspondence (\ref{mu vs lambda(r)})  

3108: we can write $\rho(x\mid\mu)$ and $\rho_{reg}(x\mid\mu)$ 

3109: as the superpositions of 

3110: the densities for the $N$ charged partitions 

3111: in the following manner; 

3112: \begin{eqnarray}

3113: \rho(x\mid\mu)&=&

3114: \frac{1}{N}\sum_{r=1}^N

3115: \rho 

3116: \Bigl(

3117: \Bigl.

3118: \frac{x-r+1}{N} \Bigr| \lambda^{(r)}\,; p_r 

3119: \Bigr), 

3120: \label{rho(mu) vs rho(lambda(r))} 

3121: \\

3122: \rho_{reg}(x\mid\mu)&=&

3123: \frac{1}{N}\sum_{r=1}^N

3124: \rho_{reg} 

3125: \Bigl( \Bigl.

3126: \frac{x-r+1}{N} \Bigr| \lambda^{(r)}\,; p_r

3127: \Bigr). 

3128: \label{rho(mu) vs rho(lambda(r)) reg} 

3129: \end{eqnarray} 

3130: (\ref{mu vs lambda(r)}) also shows that 

3131: the rescalings (\ref{u and s}) and (\ref{u(s)}) 

3132: are consistent with (\ref{rescale N pairs}). 

3133: Therefore, at the thermodynamic limit, 

3134: the above relations turn to be  

3135: \begin{eqnarray}

3136: \frac{ds(u\mid\mu)}{du}=

3137: \frac{1}{N}\sum_{r=1}^N

3138: \frac{ds(u-\eta_r\mid\lambda_r)}{du},  

3139: \label{ds(mu) vs ds(lambda(r))}

3140: \end{eqnarray}

3141: where $\sum_{r=1}^N\eta_r=0$. 

3142: 

3143: 

3144: 

3145: 

3146: 

3147: $\frac{d}{du}s(u\mid\mu)$ takes values in $[-1,0]$ 

3148: and asymptotes respectively to $0$ 

3149: as $u \rightarrow +\infty$ and $-1$ 

3150: as $u \rightarrow -\infty$. 

3151: At this stage we impose a condition on partitions. 

3152: In the below our consideration is restricted to 

3153: a class of partitions satisfying the condition 

3154: that $\frac{d}{du}s(u\mid\mu)$ is non-decreasing. 

3155: As shown in Appendix B, 

3156: this is equivalent to say that the profile 

3157: $f(u\mid\mu)$ is convex. 

3158: For such a partition $\mu$, 

3159: it follows from (\ref{reg condition}) 

3160: and (\ref{rho reg vs ds}) that 

3161: $\frac{d}{du}s(u-\eta\mid\mu)+\theta(-u)$ 

3162: has a compact support. 

3163: 

3164: 

3165: 

3166: 

3167: Let us compute the asymptotics of $V(\mu)$ 

3168: based on the expression (\ref{N-periodic potential 2}). 

3169: We rescale $\xi_r$ in the potential 

3170: to $\zeta_r$ by (\ref{zeta r}). 

3171: The asymptotics can be computed 

3172: by using (\ref{rho reg vs ds}) as 

3173: \begin{eqnarray}

3174: &&

3175: \sum_{r=1}^{N}\xi_r 

3176: \int_{-\infty}^{+\infty}dx \,

3177: \rho_{reg}(x\mid \lambda^{(r)}\,; p_r) 

3178: \nonumber \\[1.5mm]

3179: &&

3180: =\,

3181: \frac{-1}{\hbar^2}

3182: \sum_{r=1}^N

3183: \zeta_r

3184: \int_{-\infty}^{+\infty}

3185: du \left\{

3186: \frac{ds(u-\eta_r\mid\lambda^{(r)})}{du}

3187: +\theta(-u) \right\}

3188: +O(\hbar^{-1})

3189: \nonumber \\[1.5mm]

3190: &&

3191: =\,

3192: \frac{1}{\hbar^2}

3193: \sum_{r=1}^N

3194: \zeta_r

3195: \int_{-\infty}^{+\infty}

3196: du\,  u

3197: \frac{d^2s(u-\eta_r\mid\lambda^{(r)})}{du^2}

3198: +O(\hbar^{-1}), 

3199: \label{N-periodic potential 3}

3200: \end{eqnarray}

3201: where the last equality follows 

3202: by the partial integration. 

3203: 

3204: 

3205: 

3206: 

3207: %

3208: %

3209: %

3210: \begin{figure}[ht]

3211: \begin{center}

3212: \includegraphics[scale=0.63]{fig2.eps}

3213: \mycaption{3}{The graph of $\frac{d}{du}s(u\mid\mu)$ 

3214: for $\eta_1>\eta_2>\cdots>\eta_N$, 

3215: where $\eta_r$ are sufficiently separated from one another. 

3216: $u_r$ ($0\leq r\leq N$) are determined by (\ref{ur}).}

3217: \end{center}\label{fig2}

3218: \end{figure}

3219: %

3220: %

3221: %

3222: Without losing generality it is enough to consider 

3223: the case of $\eta_1>\eta_2> \cdots > \eta_N$. 

3224: In addition, we suppose that $\eta_r$ are 

3225: sufficiently separated from one another. 

3226: The graph of $\frac{d}{du}s(u\mid\mu)$ is depicted in Figure 3. 

3227: The relation (\ref{ds(mu) vs ds(lambda(r))}) leads to 

3228: the following equalities; 

3229: \begin{eqnarray}

3230: \int_{-\infty}^{+\infty}

3231: du\, u 

3232: \frac{d^2s(u-\eta_r\mid\lambda^{(r)})}{du^2}=

3233: N\int_{u_r}^{u_{r-1}}

3234: du\,  u 

3235: \frac{d^2s(u\mid\mu)}{du^2}, 

3236: \label{equality}

3237: \end{eqnarray}

3238: where $u_r$ ($0 \leq r \leq N$) are determined 

3239: by the condition (\ref{ur}).  

3240: By plugging (\ref{equality}) into 

3241: (\ref{N-periodic potential 3}) we obtain 

3242: (\ref{asymptotic V}). 

3243: 

3244: 

3245: 

3246: 

3247: 

3248: 

3249: 

3250: 

3251: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

3252: \section{Comparison between the energy functions of partitions}

3253: 

3254: 

3255: 

3256: 

3257: 

3258: We express the energy functions 

3259: (\ref{energy function}) and (\ref{4d energy function}) 

3260: by using the (rescaled) profile of partition. 

3261: Let $\mu$ be a partition. The profile function 

3262: $f_{\mu}(x)$ $(x \in \mathbb{R})$ 

3263: is defined \cite{Nekrasov-Okounkov} by 

3264: \begin{eqnarray}

3265: f_{\mu}(x)

3266: \equiv 

3267: |x|+

3268: \sum_{i=1}^{+\infty}

3269: \Bigl \{

3270: |x-x_i(\mu)-1|-|x-x_i(\mu)|

3271: -|x+i-1|+|x+i|

3272: \Bigr \}. 

3273: \label{profile mu}

3274: \end{eqnarray}

3275: The profile function becomes a quantity of order $\hbar^{-1}$ 

3276: when $\mu$ is a partition of order $\hbar^{-2}$.  

3277: By the rescalings (\ref{u and s}) and (\ref{u(s)}) 

3278: it is translated to 

3279: \begin{eqnarray}

3280: f_{\mu}(x)=

3281: \frac{1}{\hbar}

3282: f(u\mid\mu)+O(\hbar^0), 

3283: \label{rescaled profile mu} 

3284: \end{eqnarray} 

3285: where the rescaled function $f(u\mid\mu)$ is described by 

3286: \begin{eqnarray}

3287: f(u\mid\mu)=

3288: N|u|-2N

3289: \int_0^{+\infty}

3290: ds \Bigl \{ 

3291: \theta(u-u(s\mid\mu))-\theta(u+s) 

3292: \Bigr \}. 

3293: \label{f vs u}

3294: \end{eqnarray}

3295: (\ref{f vs u}) leads to the following relation 

3296: with the density function. 

3297: \begin{eqnarray}

3298: \frac{df(u\mid\mu)}{du}=

3299: N 

3300: \left( 

3301: 1+

3302: 2\frac{ds(u\mid\mu)}{du}

3303: \right). 

3304: \label{df vs ds}

3305: \end{eqnarray}

3306: We can rephrase the assumption made in Appendix A 

3307: such that the rescaled profile functions for partitions 

3308: dominating near the thermodynamic limit are convex. 

3309: 

3310: 

3311: 

3312: 

3313: 

3314: 

3315: The asymptotics (\ref{asymptotic V}) turns out to be 

3316: the surface tension \cite{Nekrasov-Okounkov}. 

3317: Let $\sigma(y)$ be the concave and piecewise-linear function 

3318: on $[-N,N]$ defined by $\frac{d}{dy}\sigma(y)=\zeta_r$ for 

3319: $y \in [N-2r,N-2(r-1)]$. 

3320: It is a straightforward computation to see 

3321: \begin{eqnarray}

3322: N

3323: \sum_{r=1}^{N}

3324: \zeta_r

3325: \int_{u_r}^{u_{r-1}}

3326: du\, u 

3327: \frac{d^2s(u\mid\mu)}{du^2}=

3328: -\frac{1}{2}

3329: \int_{-\infty}^{+\infty}du \,

3330: \sigma \left( 

3331: \frac{df(u\mid\mu)}{du} \right).

3332: \label{surface tension}

3333: \end{eqnarray}

3334: 

3335: 

3336: 

3337: 

3338: 

3339:  

3340: The energy function 

3341: (\ref{energy function}) 

3342: can be rewritten as a functional of 

3343: $f(u\mid\mu)$ by using (\ref{df vs ds}). 

3344: It becomes  

3345: \begin{eqnarray}

3346: && 

3347: N^2

3348: E \left[ s(\cdot \mid\mu) \right]

3349: \,=\,

3350: \int_{-\infty}^{+\infty} du\, 

3351: \frac{df(u\mid\mu)}{du}

3352: \frac{NRu^2}{2}

3353: \nonumber \\*[1.5mm]

3354: &&~~~~~

3355: +

3356: \frac{1}{4}

3357: \int \int_{-\infty <u<v< +\infty} 

3358: dudv \,

3359: \left(N+\frac{df(u\mid\mu)}{du}\right)

3360: \left(N-\frac{df(v\mid\mu)}{dv}\right) 

3361: \ln 

3362: \left(

3363: \frac{\sinh R(u-v)}{R \Lambda}

3364: \right)^2 

3365: \nonumber \\*[1.5mm]

3366: &&~~~~~

3367: +\frac{1}{2}

3368: \int_{-\infty}^{+\infty}du \,

3369: \sigma \left( 

3370: \frac{df(u\mid\mu)}{du} \right).

3371: \label{energy function f}

3372: \end{eqnarray}

3373: Note that the above energy function is different from 

3374: that used in \cite{Nekrasov-Okounkov} since  

3375: the $u^2$ potential term does not appear there. 

3376: While this, the four-dimensional limit 

3377: (\ref{4d energy function}) becomes  

3378: \begin{eqnarray}

3379: && 

3380: N^2

3381: E_{4d} \left[ s(\cdot\mid\mu) \right]

3382: \nonumber \\*[1,5mm]

3383: &&

3384: ~~~

3385: =

3386: \frac{1}{4}

3387: \int \int_{-\infty <u<v< +\infty} 

3388: dudv \,

3389: \left(N+\frac{df(u\mid\mu)}{du}\right)

3390: \left(N-\frac{df(v\mid\mu)}{dv}\right) 

3391: \ln 

3392: \left(

3393: \frac{u-v}{\Lambda}

3394: \right)^2

3395: \nonumber \\*[1.5mm]

3396: &&

3397: ~~~~~~

3398: +\frac{1}{2}

3399: \int_{-\infty}^{+\infty}du \,

3400: \sigma \left( 

3401: \frac{df(u\mid\mu)}{du} \right).

3402: \label{4d energy function f}

3403: \end{eqnarray}

3404: This coincides with the energy function used  

3405: in \cite{Nekrasov-Okounkov}.  

3406: It is shown there that the minimizer of 

3407: (\ref{4d energy function f}) 

3408: is described by the Seiberg-Witten geometry 

3409: of four-dimensional ${\cal N}=2$ supersymmetric $SU(N)$ 

3410: Yang-Mills.   

3411: 

3412: 

3413: 

3414: 

3415: 

3416: 

3417: 

3418: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

3419: %%%%%%%%%%%%%%% ACKNOWLEDGEMENTS %%%%%%%%%%%%%%%

3420: \subsection*{Acknowledgements}

3421: T.N. is supported in part by Grant-in-Aid for 

3422: Scientific Research 15540273. 

3423: K.T. is supported in part by Grant-in-Aid for 

3424: Scientific Research 16340040. 

3425: 

3426: 

3427: 

3428: 

3429: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

3430: \begin{thebibliography}{99}

3431: 

3432: 

3433: 

3434: \bibitem{Nekrasov}

3435: N.~Nekrasov, 

3436: \textit{``Seiberg-Witten prepotential from instanton counting,''}

3437: Adv. Theor. Math. Phys. {\bf 7} (2004) 831, 

3438: \texttt{hep-th/0206161.}

3439: 

3440: 

3441: 

3442: 

3443: \bibitem{Nekrasov-Okounkov}

3444: N.~Nekrasov and A.~Okounkov,

3445: \textit{``Seiberg-Witten theory and random partitions,''}

3446: \texttt{hep-th/0306238.}

3447: 

3448: 

3449: 

3450: \bibitem{Seiberg-Witten}

3451: N.~Seiberg and E.~Witten, 

3452: Nucl. Phys. {\bf B426} (1994) 19, 

3453: \texttt{hep-th/9407087}; 

3454: errata, {\it ibid.} {\bf B430} (1994) 485;  

3455: {\it ibid.} {\bf B431} (1994) 484, 

3456: \texttt{hep-th/9408099.}

3457: 

3458: 

3459: 

3460: 

3461: \bibitem{Nakajima-Yoshioka}

3462: H.~Nakajima and K.~Yoshioka, 

3463: \textit{``Instanton counting on blowup I,''}

3464: \texttt{math.AG/0306198.}

3465: 

3466: 

3467: 

3468: 

3469: 

3470: \bibitem{MNTT}

3471: T.~Maeda, T.~Nakatsu, K.~Takasaki and T.~Tamakoshi, 

3472: \textit{``Five-Dimensional Supersymmetric Yang-Mills Theories 

3473: and Random Plane Partitions,''}

3474: \texttt{hep-th/0412327.}

3475: 

3476: 

3477: 

3478: 

3479: \bibitem{Okounkov-Reshetikhin}

3480: A.~Okounkov and N.~Reshetikhin,  

3481: \textit{``Correlation Function of Schur Process with Application 

3482: to Local Geometry of a Random 3-Dimensional Young Diagram,''} 

3483: J. Amer. Math. Soc. \textbf{16} (2003) no.3 581, 

3484: \texttt{math.CO/0107056}. 

3485: 

3486: 

3487: 

3488: \bibitem{Miwa-Jimbo}

3489: M.~Jimbo and T.~Miwa, 

3490: \textit{``Solitons and Infinite Dimensional Lie Algebras,''}

3491: Publ. RIMS {\bf 19} (1983) 943. 

3492: 

3493: 

3494: 

3495: 

3496: 

3497: \bibitem{Macdonald}

3498: I.~G.~Macdonald,

3499: \textit{``Symmetric functions and Hall polynomials,''}

3500: Clarendon Press, 1995.

3501: 

3502: 

3503: 

3504: 

3505: 

3506: 

3507: \bibitem{quantum calabi-yau}

3508: A.~Okounkov, N.~Reshetikhin and C.~Vafa,  

3509: \textit{``Quantum Calabi-Yau and Classical Crystals,''} 

3510: \texttt{hep-th/0309208}. 

3511: 

3512: 

3513: 

3514: 

3515: 

3516: 

3517: \bibitem{topological vertex}

3518: A.~Iqbal, 

3519: \textit{``All Genus Topological String Amplitudes 

3520: and 5-brane Webs as Feynman Diagrams,''}

3521: \texttt{hep-th/0207114}.\\

3522: M.~Aganagic, A.~Klemm, M.~Marino and C.~Vafa, 

3523: \textit{``The Topological Vertex,''}

3524: \texttt{hep-th/0305132}.

3525: 

3526: 

3527: 

3528: \bibitem{Toda curve}

3529: A.~Gorsky, I.~Krichever, A~.Marshakov, A.~Mironov and A.~Morozov, 

3530: Phys. Lett. {\bf 335B} (1995) 466, 

3531: \texttt{hep-th/9505035};  

3532: E.~Martinec and N.~Warner, 

3533: Nucl. Phys. {\bf B459} (1996) 97 

3534: \texttt{hep-th/9509161}; 

3535: T.~Nakatsu and K.~Takasaki, 

3536: Mod. Phys. Lett. {\bf A11} (1996) 157, 

3537: \texttt{hep-th/9509162.}

3538: 

3539: 

3540: 

3541: \bibitem{SU(N) SW curve} 

3542: A.~Klemm, W.~Lerche, S.~Theisen and S.~Yankielowicz, 

3543: Phys. Lett. {\bf 344B} (1995) 169 

3544: \texttt{hep-th/9411048}; 

3545: P.~Argyres and A.~Faraggi, 

3546: Phys. Rev. Lett. {\bf 73} (1995) 3931,  

3547: \texttt{hep-th/9411057.}

3548: 

3549: 

3550: 

3551: 

3552: 

3553: \bibitem{Takasaki-Nakatsu}

3554: K.~Takasaki and T.~Nakatsu, 

3555: \textit{``Isomonodromic Deformations and 

3556: Supersymmetric Gauge Theories,''} 

3557: Inter. Jour. Mod. Phys. {\bf A11} (1996) 5505, 

3558: \texttt{hep-th/9603129}. 

3559: 

3560: 

3561: 

3562: \bibitem{brane probe}

3563: M.~Aganagic, R.~Dijkgraaf, A.~Klemm, M.~Marino 

3564: and C.~Vafa, 

3565: \textit{``Topological Strings and Integrable Hierarchies,''}

3566: \texttt{hep-th/0312085.}

3567: 

3568: 

3569: 

3570: 

3571: 

3572: \bibitem{MNTT in progress}

3573: T.~Maeda, T.~Nakatsu, K.~Takasaki and T.~Tamakoshi, 

3574: work in progress. 

3575: 

3576: 

3577: 

3578: 

3579: \end{thebibliography}

3580: 

3581: 

3582: 

3583: \end{document}

3584: 

3585: