1: %%%%%% LaTeX2e %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%
2: % Affine Toda-SutherlandSystems %
3: % Multi-particle Dynamics associated with Root
4: %Systems: \\ A Cross between an (affine) Toda Molecule and %
5: %a Calogero-Moser System %
6: % A. Khare, I. Loris and R.Sasaki %
7: % 15 May --29 Aug %
8: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
9: \documentclass[12pt]{article}
10: \usepackage{amsfonts}
11: \usepackage{graphicx}
12: %\usepackage{amssymb}
13: % Change page dimensions to match standard 8 1/2 X 11 inch size
14: \oddsidemargin=-0.1in
15: \evensidemargin=-0.1in
16: \topmargin=-0.2in
17: \textwidth=6.5in
18: \textheight=8.9in
19:
20: \renewcommand{\theequation}{\arabic{section}.\arabic{equation}}
21:
22: \font\twozero=cmr10 at 20pt
23: \font\oneeight=cmr10 at 18pt
24: \font\onesix=cmr10 at 16pt
25: \font\onefour=cmr10 at 14pt
26: \font\larl=cmr10 at 24pt
27: \newcommand{\vT}{\vphantom{\mbox{\twozero I}}}
28: \newcommand{\vTm}{\vphantom{\mbox{\oneeight I}}}
29: \newcommand{\vTmm}{\vphantom{\mbox{\onesix I}}}
30: \newcommand{\vTs}{\vphantom{\mbox{\onefour I}}}
31: \newcommand{\vTb}{\vphantom{\mbox{\larl I}}}
32:
33: \begin{document}
34:
35: \baselineskip=20pt
36:
37: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
38: % %
39: % Title page %
40: % %
41: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
42: \newfont{\elevenmib}{cmmib10 scaled\magstep1}
43: \newcommand{\preprint}{
44: \begin{flushleft}
45: \elevenmib Yukawa\, Institute\, Kyoto\\
46: \end{flushleft}\vspace{-1.3cm}
47: \begin{flushright}\normalsize \sf
48: YITP-03-64\\
49: IP/BBSR/03-13\\
50: {\tt hep-th/0309077} \\ September 2003
51: \end{flushright}}
52: \newcommand{\Title}[1]{{\baselineskip=26pt
53: \begin{center} \Large \bf #1 \\ \ \\ \end{center}}}
54: \newcommand{\Author}{\begin{center}
55: \large \bf Avinash ~Khare${}^a$, I.~Loris${}^{b,c}$ and R.~Sasaki${}^b$
56: \end{center}}
57: \newcommand{\Address}{\begin{center}
58: $^a$ Institute of Physics, Sachivalaya Marg,\\
59: Bhubaneswar, 751005, Orissa, India\\
60: ${}^b$ Yukawa Institute for Theoretical Physics,\\
61: Kyoto University, Kyoto 606-8502, Japan\\
62: ${}^c$ Dienst Theoretische Natuurkunde,
63: Vrije Universiteit Brussel, \\
64: Pleinlaan 2, B-1050 Brussels, Belgium
65: \end{center}}
66: \newcommand{\Accepted}[1]{\begin{center}
67: {\large \sf #1}\\ \vspace{1mm}{\small \sf Accepted for Publication}
68: \end{center}}
69:
70: \preprint
71: \thispagestyle{empty}
72: \bigskip\bigskip\bigskip
73:
74: \Title{Affine Toda-Sutherland Systems}
75: \Author
76: \Address
77:
78: \bigskip
79: \begin{abstract}
80: A cross between two well-known integrable multi-particle dynamics,
81: an affine Toda molecule and a Sutherland system, is introduced for
82: any affine root system.
83: Though it is not completely integrable but partially integrable, or quasi
84: exactly solvable, it inherits many remarkable properties from the parents.
85: The equilibrium position is algebraic, {\em i.e.\/} proportional to the Weyl vector.
86: The frequencies of small oscillations near equilibrium are proportional
87: to the affine Toda masses, which are essential ingredients of the exact
88: factorisable S-matrices of affine Toda field theories.
89: Some lower lying frequencies are {\em integer\/} times a coupling constant
90: for which the corresponding exact quantum eigenvalues and eigenfunctions are
91: obtained. An affine Toda-Calogero system, with a corresponding rational potential,
92: is also discussed.
93: \end{abstract}
94:
95: \newpage
96: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
97: % %
98: % 1. Introduction %
99: % %
100: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
101: \section{Introduction}
102: \label{intro}
103: \setcounter{equation}{0}
104:
105: Calogero-Moser systems and (affine) Toda molecules%
106: \footnote{
107: In this article we use the terminology `molecule' to emphasise the finite degrees
108: of freedom instead of the more familiar `lattice'
109: which might be misinterpreted as meaning
110: an infinitely or macroscopically large system.
111: } are best known examples of integrable/solvable
112: many-particle dynamics on a line which are based on root systems.
113: The original Toda model \cite{Toda} and the Calogero \cite{Cal} and
114: the Sutherland \cite{Sut} models are based on the $A_r$ root system which
115: correspond to the Lie algebra $\mathfrak{su}(r+1)$.
116: Later integrable Toda \cite{Kost,OPC} and Calogero-Moser (C-M)
117: \cite{CalMo,OPC,OP1,DHoker_Phong,bms} systems are formulated for any root system.
118: The potentials of Toda systems are exponential functions of the coordinates,
119: whereas those of Calogero-Moser systems are rational $1/q^2$, trigonometric
120: $1/\sin^2q$, hyperbolic $1/\sinh^2q$ and elliptic $\wp(q)$ functions,
121: in which $\wp$ is Weierstrass function and $q$ denotes the
122: coordinates generically. In the C-M systems,
123: the elliptic potentials are the most general
124: ones and the rest (trigonometric, hyperbolic and rational) is obtained by
125: various degeneration.
126: In fact, a Toda molecule is obtained from an elliptic C-M system
127: by a special limiting procedure \cite{Ino1,DHPh2,kst}.
128: While the potential of a C-M system depends on {\em all\/} (positive) roots,
129: that of an (affine) Toda system contains (affine) {\em simple\/} roots only.
130: For the $A$-type root systems the above feature is usually referred to that
131: the C-M potential gives a {\em pair-wise\/} interactions and the Toda potential is
132: of the {\em nearest neighbour\/} interaction type and
133: the affine simple root corresponds
134: to a {\em periodic\/} boundary condition.
135:
136: In this paper we will present two new types of
137: multi-particle dynamics related to any root
138: system.
139: Roughly speaking each could be considered as a cross between an (affine)
140: Toda molecule and a C-M system. The first, to be tentatively called an (affine)
141: Toda-Sutherland system, has trigonometric potentials $1/\sin^2q$ and
142: depends on the (affine) {\em simple\/} roots only.
143: The second, to be tentatively referred to as an (affine)
144: Toda-Calogero system, has rational potentials $1/q^2$ plus a harmonic confining
145: potential $q^2$ and
146: depends on the (affine) {\em simple\/} roots only.
147: The former has much richer structure than the latter and in this paper we mainly
148: discuss the affine Toda-Sutherland systems.
149: We do not think that they are integrable,
150: either at the classical or the quantum level.
151: But they have many remarkable features as shown in some detail for the systems
152: based on the $A$-type root systems \cite{JGA,AJK,EGKP}.
153: Their potentials have the {\em nearest\/} and {\em next-to-nearest\/} neighbour
154: interactions, in contrast to the nearest neighbour interactions of the
155: (affine) Toda molecule.
156: These dynamical systems exhibit a behaviour intermediate to regular and chaotic.
157: Like the C-M systems, these multi-particle dynamics are closely related to
158: random matrix theory \cite{JGA}.
159:
160:
161: At the classical level, the frequencies of small oscillations at equilibrium
162: \cite{cs} of an affine Toda-Sutherland system have exactly the same
163: pattern as those of the affine Toda molecule based on the same root system.
164: Let us point out that the pattern of the frequencies of small oscillations at
165: equilibrium
166: of an affine Toda molecule, or the so-called {\em affine Toda masses\/}
167: appearing in the affine Toda field theory in $1+1$ dimensions \cite{bcds},
168: are the essential ingredient for its {\em exact factorisable\/} $S$-matrices.
169: At the quantum level, most (but not all) of the multi-particle systems
170: discussed in this paper
171: are {\em Quasi Exactly Solvable\/} (QES) \cite{turb,ush}.
172: That is, on top of the ground state eigenfunctions, a certain small number of
173: eigenvalues and eigenfunctions are obtained exactly.
174: The mechanism for QES seems very different from that of known ones
175: \cite{turb,ush,st1}.
176:
177:
178: This paper is organised as follows. In section 2, the salient features of
179: affine Toda molecules are reviewed with a brief summary of roots and weights as
180: essential ingredients.
181: Section 3 is the main body of the paper. In section \ref{clequil} we obtain the
182: frequencies of small oscillations for Toda-Sutherland systems based on
183: affine root systems. For these multiparticle systems we present some exact
184: eigenvalues and eigenfunctions in section
185: \ref{atsuteigfun}. They correspond to the low lying {\em integer\/} (times a
186: coupling constant) frequencies of the small oscillations at equilibrium
187: \cite{cs,ls}. In section 4 the affine Toda-Calogero systems are briefly discussed.
188: The final section is reserved for summary and comments.
189: In this paper we adopt the convention
190: that $\hbar=1$ and do not show the dependence on the Planck's constant.
191:
192:
193: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
194: \section{Affine Toda molecule}
195: \label{affine Toda}
196: \setcounter{equation}{0}
197:
198:
199: The dynamical variables of a classical (quantum)
200: multi-particle system to be discussed
201: in this paper, an (affine) Toda molecule, a C-M system,
202: an (affine) Toda-Sutherland
203: system and an (affine) Toda-Calogero system,
204: are the coordinates
205: \(\{q_{j}|\,j=1,\ldots,r\}\) and their canonically conjugate momenta
206: \(\{p_{j}|\,j=1,\ldots,r\}\), with the Poisson bracket (Heisenberg
207: commutation) relations:
208: \begin{eqnarray*}
209: \{q_{j},p_{k}\}&=&\delta_{j\,k},\qquad \{q_{j},q_{k}\}=
210: \{p_{j},p_{k}\}=0,\\
211: \ [q_{j},p_{k}]&=&i\delta_{j\,k},\qquad [q_{j},q_{k}]=
212: [p_{j},p_{k}]=0.
213: \end{eqnarray*}
214: These will be denoted by vectors in \(\mathbf{R}^{r}\)
215: \[
216: q=(q_{1},\ldots,q_{r}),\qquad p=(p_{1},\ldots,p_{r}),
217: \]
218: in which $r$ is the number of particles and it is also the
219: rank of the underlying root system $\Delta$.
220:
221: \subsection{Roots and weights}
222: Let $\Pi$ be the set of simple roots of $\Delta$:
223: \begin{equation}
224: \Pi=(\alpha_1,\alpha_2,\ldots,\alpha_r).
225: \end{equation}
226: Any positive roots in $\Delta$ can be expressed as a linear combination
227: of the simple roots with non-negative integer coefficients
228: \begin{equation}
229: \alpha=\sum_{j=1}^rm_j\alpha_j,\qquad m_j\in\mathbb{Z}_+,\quad
230: \forall\alpha\in\Delta_+.
231: \end{equation}
232: In the case of {\em simply laced\/} root systems ($A$, $D$, $E$)
233: all the roots have the same length. We adopt the convention
234: $\alpha^2=\alpha\cdot\alpha=2$.
235: In the case of {\em non-simply laced\/} root systems ($B$, $C$, $F_4$, $G_2$),
236: there are long roots and short roots.
237: We adopt the convention
238: $\alpha_L^2=2$ except for the $C$-series of the root system in which we adopt
239: $\alpha_S^2=2$. Since $\Delta$ is a finite set, there exists an element $\alpha_h$
240: for which $\sum_{j=1}^rm_j$ is the maximum in $\Delta_+$.
241: We call it the {\em highest root\/} and write it
242: \begin{equation}
243: \alpha_h=\sum_{j=1}^rn_j\alpha_j,\qquad n_j\in\mathbb{Z}_+.
244: \label{alphah}
245: \end{equation}
246: For the non-simply laced root systems, the highest roots are always long.
247: We also introduce {\em highest short root\/} and denote it in the same way as
248: (\ref{alphah}) to avoid duplicating many formulas.
249:
250:
251: The positive integers $\{n_j\}$ are called {\em Dynkin-Kac labels\/}.
252: We define the affine simple root $\alpha_0$ as the {\em lowest\/} ({\em short\/})
253: {\em root\/}, that is the negative of the highest (short) root:
254: \begin{equation}
255: \alpha_0=-\alpha_h=-(\sum_{j=1}^rn_j\alpha_j).
256: \end{equation}
257: The above relationship can be rewritten in a symmetrical way:
258: \begin{equation}
259: \sum_{j=0}^rn_j\alpha_j=0,\quad n_0\equiv1.
260: \label{nzerosum}
261: \end{equation}
262: We call $\Pi_0$ the set of {\em affine simple roots\/}:
263: \begin{equation}
264: \Pi_0=\alpha_0\cup\Pi=(\alpha_0,\alpha_1,\ldots,\alpha_r),
265: \label{afsimp}
266: \end{equation}
267: which specifies the {\em affine Lie algebra\/}, to be denoted as
268: $A_r^{(1)}$, $E_6^{(2)}$, $D_4^{(3)}$, etc. It has the necessary and sufficient
269: information for defining affine Toda molecule (and its field theory version, the
270: affine Toda field theory
271: \cite{bcds}, too).
272:
273: The fundamental weights $\{\lambda_j\}$ are the dual to the simple roots:
274: \begin{eqnarray}
275: \alpha_j^\vee\cdot\lambda_k&=&\delta_{jk},\quad
276: \alpha_j^\vee\equiv {2\alpha_j\over{\alpha_j^2}},\quad j=1,\ldots,r,
277: \label{lamdef}\\
278: \alpha_j\cdot\lambda_j^\vee&=&\delta_{jk},\quad
279: \lambda_j^\vee\equiv{2\over{\alpha_j^2}}\lambda_j,\quad j=1,\ldots,r.
280: \label{lamveedef}
281: \end{eqnarray}
282: The equation (\ref{lamdef}) defines $\{\lambda_j\}$ and (\ref{lamveedef})
283: defines $\{\lambda_j^\vee\}$ in turn. Next we define $\varrho$:
284: \begin{equation}
285: \varrho\equiv\sum_{j=1}^r\lambda_j^\vee,
286: \label{rhodef}
287: \end{equation}
288: which is essentially the {\em Weyl vector\/} having the following properties
289: \begin{eqnarray}
290: \alpha_j\cdot\varrho&=&1,\qquad j=1,\ldots,r,
291: \label{rhoeq1}\\
292: \alpha_0\cdot\varrho&=&-(\sum_{j=1}^rn_j\alpha_j)\cdot\varrho
293: =-(\sum_{j=1}^rn_j)
294: =-(h-1),
295: \label{rhoeq2}
296: \end{eqnarray}
297: in which $h$ is the (dual) {\em Coxeter number\/}:
298: \begin{equation}
299: h\equiv\sum_{j=0}^rn_j=1+\sum_{j=1}^rn_j.
300: \label{Coxdef}
301: \end{equation}
302:
303:
304: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
305: \subsection{Hamiltonian, equilibrium
306: position and frequencies of small oscillations}
307:
308: The Hamiltonian of the affine Toda molecule based on the set of affine simple
309: roots $\Pi_0$ is
310: \begin{eqnarray}
311: H&=&{1\over2}p^2+V(q),
312: \label{afTodaham}\\
313: V(q)&=&{1\over{\beta^2}}\sum_{j=0}^rn_je^{\beta\alpha_j\cdot q},
314: \label{afTodapot}
315: \end{eqnarray}
316: in which $\beta\in\mathbf{R}$ is the coupling constant.
317: Note that all the particle masses are the same and normalised to unity.
318: The potential $V(q)$ has a minimum ({\em equilibrium point\/}) at $q=0$ as
319: \begin{equation}
320: V(q)={h\over{\beta^2}}+{1\over{\beta}}(\sum_{j=0}^rn_j\alpha_j)\cdot q+
321: {1\over2}\sum_{j=0}^r\sum_{k,l}^rn_j(\alpha_j)_k(\alpha_j)_lq_kq_l+
322: o(q^3).
323: \end{equation}
324: Here, $(\alpha_j)_k$ is the $k$-th component of the
325: (affine) simple root $\alpha_j$.
326: The linear term vanishes due to (\ref{nzerosum}) and the constant term
327: is proportional to the Coxeter number $h$ given by (\ref{Coxdef}).
328:
329: The symmetric matrix $M$
330: \begin{equation}
331: M_{kl}=\sum_{j=0}^rn_j(\alpha_j)_k(\alpha_j)_l,\qquad \mbox{or}
332: \qquad M=\sum_{j=0}^rn_j\alpha_j\otimes\alpha_j,
333: \label{Mdef}
334: \end{equation}
335: is called {\em affine Toda mass matrix\/}. Its eigenvalues
336: \begin{equation}
337: \mbox{Spec}(M)=\left\{m_1^2,m_2^2,\ldots,m_r^2\right\},\quad m_j>0.
338: \end{equation}
339: are called affine Toda masses (squared). The set
340: $\left\{m_1,m_2,\ldots,m_r\right\}$ gives $r$ (angular) frequencies of small
341: oscillations at the equilibrium $q=0$. The above
342: Hamiltonian (\ref{afTodaham})-(\ref{afTodapot}) is {\em completely
343: integrable\/} and {\em classical\/} Lax pair is known for all the affine simple root
344: systems. This is a {\em periodic\/} Toda lattice if $\Pi_0$ is for $A_r^{(1)}$.
345:
346:
347:
348: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
349: \section{Affine Toda-Sutherland systems}
350: \label{atsut}
351: \setcounter{equation}{0}
352: The multi-particle dynamics with nearest and next-to-nearest trigonometric
353: interactions introduced in
354: \cite{JGA, AJK} can be called {\em affine Toda-Sutherland model\/} based on
355: $A_r^{(1)}$. They can be generalised to any root system as follows.
356:
357: Given an affine root system $\Pi_0$, let us introduce
358: a {\em prepotential\/} $W$
359: \begin{equation}
360: W(q)=\beta\sum_{j=0}^rn_j\log|\sin(\alpha_j\cdot q)|,
361: \label{prep}
362: \end{equation}
363: in which $\beta\in\mathbf{R}_+$ is a positive coupling constant
364: and $\{n_j\}$ are the Dynkin-Kac labels for $\Pi_0$.
365: This leads to the Hamiltonians with the
366: classical and quantum potentials $V_C$ and
367: $V_Q$ as \cite{bms}
368: \begin{eqnarray}
369: H_C&=&{1\over2}p^2+V_C(q),\qquad V_C(q)={1\over2}\sum_{j=1}^r\left({\partial
370: W\over{\partial q_j}}\right)^2,
371: \label{casham}\\
372: H_Q&=&{1\over2}p^2+V_Q(q),
373: \qquad
374: V_Q(q)={1\over2}\sum_{j=1}^r\left[\left({\partial W\over{\partial
375: q_j}}\right)^2 +{\partial^2 W\over{\partial q_j^2}}\right].
376: \label{qasham}
377: \end{eqnarray}
378: Again note that all the particle masses are the same and normalised to unity.
379: Explicitly $V_Q$ reads
380: \begin{eqnarray}
381: V_Q&=&{1\over2}\sum_{j=0}^r{\beta n_j(\beta
382: n_j-1)\alpha_j^2\over{\sin^2(\alpha_j\cdot q)}}+
383: \beta^2\sum_{j<k}n_j n_k\alpha_j\cdot\alpha_k\cot(\alpha_j\cdot q)
384: \cot(\alpha_k\cdot q)-E_0,\\
385: E_0&=&{\beta^2\over2}\sum_{j=0}^rn_j^2\alpha_j^2,
386: \label{hamexpl}
387: \end{eqnarray}
388: in which the constant part $E_0$ can be considered as the {\em ground state
389: energy\/}. The extended Dynkin diagram of $\Pi_0$ encodes all the necessary
390: information $\{\alpha_j^2\}$, $\{\alpha_j\cdot\alpha_k\}$
391: and $\{n_j\}$ to determine
392: $V_Q$.
393: See \cite{bms,cs,ls} for the formulation of Hamiltonian dynamics in terms of a
394: prepotential and the frequencies of small oscillations at equilibrium.
395: The
396: corresponding ground state wavefunction is
397: \begin{equation}
398: H_Q\psi_0=0,\qquad \psi_0(q)=e^{W(q)}=
399: \prod_{j=0}^r|\sin (\alpha_j\cdot q)|^{\beta
400: n_j}.
401: \end{equation}
402: In contrast to the Calogero-Moser systems \cite{OP1,DHoker_Phong}, the prepotential
403: (\ref{prep}), potential (\ref{qasham}) and thus the Hamiltonian are not
404: Weyl-invariant. For simplicity we consider
405: the configuration space in
406: the {\em principal Weyl alcove\/}:
407: \begin{equation}
408: PW_T=\{q\in{\bf R}^r|\ \alpha\cdot q>0,\quad \alpha\in\Pi,
409: \quad \alpha_h\cdot q<\pi\},
410: \label{PWT}
411: \end{equation}
412: where \(\alpha_h\) is the highest root.
413: (Due the non-invariance under the Weyl group, theories with
414: different configuration spaces are physically different.
415: For example, they have different (non-equivalent) equilibrium positions.)
416:
417: For the simplest affine Lie algebra of $A_r^{(1)}$
418: the quantum Hamiltonian reads%
419: \footnote{For $A_r$ models, it is customary to
420: introduce one more degree of freedom,
421: $q_{r+1}$ and $p_{r+1}$ and embed
422: all of the roots in ${\bf R}^{r+1}$. Here we also adopt the `periodic' convention,
423: $q_{r+1}\equiv q_0$, $q_{r+2}\equiv q_1$, etc.
424: \label{embedding}}
425: \begin{eqnarray}
426: H_Q&=&{1\over2}p^2+\beta(\beta-1)\sum_{j=1}^{r+1}{1\over{\sin^2(q_j-q_{j+1})}}
427: \nonumber\\
428: &&\hspace{20mm}
429: -\beta^2\sum_{j=1}^{r+1}\cot(q_{j-1}-q_j)\cot(q_{j}-q_{j+1})-\beta^2(r+1).
430: \end{eqnarray}
431: This has the {\em nearest\/} and {\em next-to-nearest\/} neighbour
432: interactions \cite{JGA,AJK}.
433: The $B$, $BC$ and $D$ models in
434: \cite{JGA,AJK,EGKP} are different from those in this
435: paper.
436: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
437: \subsection{Classical equilibrium}
438: \label{clequil}
439: The equilibrium point ($\bar{q}$) of the classical Hamiltonian
440: of the affine Toda-Sutherland system
441: \begin{eqnarray}
442: H&=&{1\over2}p^2+V_C(q),\quad V_C(q)={1\over2}\sum_{j=1}^r\left({\partial
443: W\over{\partial q_j}}\right)^2,\\ {\partial W\over{\partial
444: q_j}}&=&\beta\sum_{k=0}^rn_k(\alpha_k)_j\cot[\alpha_k\cdot q],\quad
445: j=1,\ldots,r.
446: \end{eqnarray}
447: has a very intuitive characterisation. It is proportional
448: to the Weyl vector $\varrho$ (\ref{rhodef}), $\bar{q}\propto\varrho$, the
449: fundamental quantity of the Lie algebra. This is much simpler than
450: the cases in the
451: Calogero as well as Sutherland systems in which $\bar{q}$ correspond to the zeros
452: of certain polynomials, {\em i.e.\/} the Hermite, Laguerre, Chebyshev and Jacobi
453: polynomials for classical root systems
454: \cite{zero,cs}. Since \cite{cs,ls}
455: \begin{equation}
456: {\partial
457: W(\bar{q})\over{\partial q_j}}=0,\quad
458: j=1,\ldots,r \quad
459: \Rightarrow\quad
460: {\partial V_C(\bar{q})\over{\partial q_l}}
461: =\sum_{j=1}^r{\partial^2
462: W(\bar{q})\over{\partial q_j\partial q_l}}{\partial
463: W(\bar{q})\over{\partial q_j}}=0,
464: \end{equation}
465: the equilibrium is achieved at
466: the point $\bar{q}$ where all $\partial W/\partial q_j$ vanish,
467: {\em i.e.\/} at the maximum of the ground state wavefunction.
468: It is easy to see that
469: \begin{equation}
470: \bar{q}=c\varrho,\quad c:const.,
471: \end{equation}
472: gives a solution.
473: Using (\ref{rhodef})--(\ref{rhoeq2}),
474: \begin{equation}
475: \alpha_k\cdot\bar{q}=\left\{
476: \begin{array}
477: {cl}
478: c&k=1,\ldots,r,\\
479: -(h-1)c&k=0,
480: \end{array}
481: \right.
482: \end{equation}
483: we obtain
484: \begin{equation}
485: {\partial
486: W(\bar{q})\over{\partial q_j}}=
487: \beta\left(\cot(c)\sum_{k=1}^rn_k(\alpha_k)_j-\cot[(h-1)c](\alpha_0)_j\right).
488: \end{equation}
489: For
490: \begin{equation}
491: c={\pi\over h},
492: \end{equation}
493: $c\varrho$ is in the principal Weyl alcove (\ref{PWT}) and
494: \begin{equation}
495: \cot[(h-1)c]=\cot(\pi-c)=-\cot(c).
496: \end{equation}
497: Thus we find $\bar{q}=\pi\varrho/h$ is the equilibrium
498: \begin{equation}
499: {\partial
500: W(\bar{q})\over{\partial q_j}}=
501: \beta\cot[{\pi\over h}]\left(\sum_{k=0}^rn_k\alpha_k\right)_j=0.
502: \end{equation}
503: The equilibrium points are {\em equally spaced\/}
504: for all the classical root systems. The situation is different for the
505: exceptional root systems.
506: The equilibrium point $\bar{q}=\pi\varrho/h$ is unique in the
507: principle Weyl alcove (\ref{PWT}).
508:
509: \bigskip
510: The squared frequencies of small oscillations at equilibrium $\bar{q}$ are
511: given by the eigenvalues of the matrix
512: \begin{equation}
513: \left.{\partial^2
514: V_C(q)\over{\partial q_j \partial q_k}}\right|_{\bar{q}}=
515: \sum_{j=1}^r\left.{\partial^2
516: W(q)\over{\partial q_j\partial q_l}}\right|_{\bar{q}}\left.{\partial^2
517: W(q)\over{\partial q_l\partial q_k}}\right|_{\bar{q}}=(\widetilde{W}^2)_{jk}.
518: \end{equation}
519: Thus the frequencies of small oscillations at equilibrium $\bar{q}$ are
520: given by the eigenvalues of
521: a symmetric matrix $\widetilde{W}$ defined by
522: \begin{equation}
523: \widetilde{W}_{jk}=-\left.{\partial^2
524: W(q)\over{\partial q_j\partial
525: q_k}}\right|_{\bar{q}}={\beta\over{\sin^2{\pi\over h}}}
526: \sum_{l=0}^rn_l(\alpha_l)_j(\alpha_l)_k={\beta\over{\sin^2{\pi\over h}}}M_{jk},
527: \end{equation}
528: in which matrix $M$ is the mass square matrix of the affine Toda molecule
529: associated with the affine root system $\Pi_0$ defined in (\ref{Mdef}).
530:
531: \bigskip
532: The frequencies (not frequencies squared) of small oscillations at equilibrium
533: of affine Toda-Sutherland model are given up to the coupling constant $\beta$ by
534: \begin{equation}
535: {1\over{\sin^2{\pi\over h}}}\left\{m_1^2,m_2^2,\ldots,m_r^2\right\},
536: \label{atsspectrum}
537: \end{equation}
538: in which $m_j^2$ are the affine Toda masses. In \cite{bcds} it is shown that
539: the vector $\mathbf{m}=(m_1,\ldots,m_r)$, if ordered properly, is the {\em
540: Perron-Frobenius\/} eigenvector of the incidence matrix (the Cartan matrix)
541: of the corresponding root system.
542: Therefore there exists a one-to-one correspondence between the mass $m_j$ and
543: a vertex (or the fundamental weight) of the Dynkin diagram.
544: This fact will be important in the next subsection for the explicit construction
545: of exact eigenvalues and eigenfunctions.
546: In Table I
547: we list the affine Toda masses and the Coxeter number $h$ for the classical {\em
548: untwisted\/} affine Lie algebras, $A_r^{(1)}$,
549: $B_r^{(1)}$,
550: $C_r^{(1)}$,
551: $D_r^{(1)}$, see \cite{bcds}:
552: \begin{center}
553: \begin{tabular}{||c|c|l||}
554: \hline
555: \(\Pi_0\)&\(h\)&\mbox{affine Toda masses}\\
556: \hline
557: \vTb\(A_r^{(1)}\)&\(r+1\)& \(m_j^2=4\sin^2({j\pi\over h}),\quad
558: j=1,\ldots,r,\)\\[3pt]
559: \hline
560: \vTb\(B_r^{(1)}\)&\(2r\)&\(m_j^2=
561: 8\sin^2({j\pi\over h}),\quad j=1,\ldots,r-1,\quad
562: m_r^2=2,\)\\[3pt]
563: \hline
564: \vTb\(C_r^{(1)}\)&\(2r\)&\(m_j^2=
565: 8\sin^2({j\pi\over h}),\quad j=1,\ldots,r,\)\\[3pt]
566: \hline
567: \vTb\(D_r^{(1)}\)&\(2(r-1)\)&\(m_j^2=
568: 8\sin^2({j\pi\over h}),\quad j=1,\ldots,r-2,\quad
569: m_{r-1}^2=m_r^2=2.\)\\[3pt]
570: \hline
571: \end{tabular}\\
572: \bigskip
573: Table I: The Coxeter number \(h\) and the affine Toda masses \(m_j^2\)\\ for
574: classical untwisted affine Lie algebras.
575: \end{center}
576: Those for the exceptional affine Lie algebras $E_r^{(1)}$, $F_4^{(1)}$
577: and $G_2^{(1)}$ we refer to \cite{bcds}.
578: (The affine Toda masses for $E_8$ reported
579: there need a factor 2.) The {\em twisted affine Lie algebras\/},
580: for example
581: $D_{r+1}^{(2)}$, $E_6^{(2)}$,
582: $D_4^{(3)}$, etc., which are characterised by the highest short roots,
583: can also be obtained
584: from untwisted affine Lie algebras by {\em folding\/} \cite{bcds}.
585: The affine Toda masses for the
586: {twisted affine Lie algebra} are closely related to those of the original untwisted
587: affine Lie algebra.
588:
589:
590: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
591: \subsection{Quantum eigenfunctions}
592: \label{atsuteigfun}
593: Here we demonstrate that some of the quantum affine Toda-Sutherland
594: (\ref{qasham}) systems have a number of exact eigenvalues and eigenfunctions
595: and thus they are partially integrable or {\em quasi exactly solvable\/}
596: \cite{turb,ush}.
597: These are usually a small number of lowest lying excited states.
598: The occurrence of such exact states is strongly correlated with the appearance
599: of the {\em integer eigenvalues\/} in the spectrum of the small oscillations near
600: the {\em classical equilibrium\/}, as shown in the recent general theorems by
601: Loris-Sasaki
602: \cite{ls}. Let us express the eigenfunctions in product forms
603: \begin{equation}
604: \psi_n(q)=\phi_n(q)\psi_0(q),\quad n=0,1,\ldots,
605: \qquad \phi_0\equiv1,
606: \label{phipsi}
607: \end{equation}
608: in which $\phi_n$ obeys a simplified equation with the similarity transformed
609: Hamiltonian $\tilde{H}$ \cite{bms}:
610: \begin{eqnarray}
611: \tilde{H}\phi_n&=&E_n\phi_n,\label{sthameq}\\
612: \tilde{H}=e^{-W}H_Q e^{W}
613: &=&-{1\over2}\triangle-
614: \sum_{j=1}^r{\partial W\over{\partial q_j}}{\partial \over{\partial q_j}},
615: \quad \triangle\equiv \sum_{j=1}^r{\partial^2\over{\partial q_j^2}}.
616: \label{htilform}
617: \end{eqnarray}
618:
619: %%%%%%%%%%%%%%%%%%%
620: \subsubsection{$A_r^{(1)}$}
621: In this case the spectrum of the small oscillations, up to the coupling
622: constant $\beta$
623: is easily read from Table I:
624: \begin{equation}
625: 4\left\{1,\ldots,\sin^2(j\pi/(r+1))/\sin^2(\pi/(r+1)),\ldots,1\right\}.
626: \label{aasspec}
627: \end{equation}
628: Reflecting the left-right mirror symmetry
629: $j\leftrightarrow r+1-j$ of the Dynkin diagram Fig.\ref{fig:ar}, the spectrum is
630: doubly degenerate except for the possible singlet at the middle point
631: $j=(r+1)/2$ for odd $r$.
632:
633: The doubly degenerate integer eigenvalues 4 correspond to the two end points of the
634: $A_r^{(1)}$ Dynkin diagram, Fig.\ref{fig:ar}.
635: They correspond to the fundamental
636: {\em vector} and {\em conjugate vector\/}
637: representations and to the eigenfunctions:
638: \begin{equation}
639: \mathbf{v}=\sum_{j=1}^{r+1}e^{2iq_j},\quad
640: \bar\mathbf{v}=\sum_{j=1}^{r+1}e^{-2iq_j}.
641: \label{arvav}
642: \end{equation}
643: \begin{figure}%[tbp]
644: \centering
645: \includegraphics{ardynk.eps}
646: \caption{$A_r^{(1)}$ Dynkin diagram with the numbers $n_j$
647: attached. The black spot is the affine simple root. }
648: \label{fig:ar}
649: \end{figure}
650: It is easy to verify
651: \begin{equation}
652: \tilde{H}\mathbf{v}=(4\beta+2)\mathbf{v},\quad
653: \tilde{H}\bar\mathbf{v}=(4\beta+2)\bar\mathbf{v},\quad
654: -{1\over2}\triangle\mathbf{v}=2\mathbf{v},\quad
655: -{1\over2}\triangle\bar\mathbf{v}=2\bar\mathbf{v}.
656: \label{arvecs}
657: \end{equation}
658: The affine simple root corresponds to the adjoint representation. Let us define
659: \begin{equation}
660: \phi_a=\phi_{v\bar{v}}+2\beta/(1+2\beta),\quad
661: \phi_{v\bar{v}}=\sum_{j\neq k}e^{2i(q_j-q_k)}.
662: \label{aradj}
663: \end{equation}
664: It is easy to show
665: \begin{equation}
666: \tilde{H}\phi_a=(8\beta+4)\phi_a,
667: \end{equation}
668: in which $8\beta$ is simply a sum of $4\beta$ for $\mathbf{v}$ and another $4\beta$
669: for $\bar\mathbf{v}$ in (\ref{arvav}).
670:
671: For $A_2^{(1)}$, the system is identical with the $A_2$ Sutherland model.
672: For the special case of $A_3^{(1)}$,
673: the above spectrum (\ref{aasspec}) is $\{4,8,4\}$.
674: We find another complex eigenfunction with the classical eigenvalue $8\beta$
675: \begin{equation}
676: \phi_t=\sum_{j=1}^4e^{4iq_j}-e^{2i(q_1+q_2+q_3+q_4)}\sum_{j=1}^4e^{-4iq_j},
677: \quad \tilde{H}\phi_t=(8\beta+8)\phi_t.
678: \label{a3add}
679: \end{equation}
680:
681: %%%%%%%%%%%%%%%%%%%
682: \subsubsection{$D_{r}^{(1)}$}
683: \label{drone}
684: The spectrum of the small oscillations, up to the coupling
685: constant $\beta$ is easily read from Table I:
686: \begin{equation}
687: 8\left\{1,\ldots,\sin^2(j\pi/2(r-1))/\sin^2(\pi/2(r-1)),\ldots\right\},\quad
688: \mbox{and}\quad 2/\sin^2(\pi/2(r-1))[2],
689: \label{drasspec}
690: \end{equation}
691: in which the two degenerate frequencies at the end correspond to
692: the spinor and anti-spinor weights at the
693: right end of the $D_r^{(1)}$ Dynkin diagram
694: in Fig.\ref{fig:dr}.
695: \begin{figure}%[tbp]
696: \centering
697: \includegraphics{drdynk.eps}
698: \caption{$D_{r}^{(1)}$ Dynkin diagram with the numbers $n_j$
699: attached. The black spot is the affine simple root. }
700: \label{fig:dr}
701: \end{figure}
702:
703: For $r\ge5$ these eigenvalues are greater than 8,
704: which belongs to the vector weights at the left end of the Dynkin diagram
705: Fig.\ref{fig:dr}. The set of vector weights is
706: $\mathbf{V}=\{\pm\mathbf{e}_j|j=1,\ldots,r\}$. Let us introduce the corresponding
707: wavefunctions
708: \begin{equation}
709: \phi_{\mathbf{V}}=\sum_{\mu\in\mathbf{V}}e^{2i\mu\cdot q}=2\sum_{j=1}^r\cos 2q_j.
710: \label{dreig}
711: \end{equation}
712: However, it is not an eigenfunction
713: \begin{equation}
714: \tilde{H}\phi_{\mathbf{V}}=
715: (8\beta+2)\phi_{\mathbf{V}}-8\beta(\cos2q_1+\cos2q_r),\quad
716: -{1\over2}\triangle\phi_{\mathbf{V}}=2\phi_{\mathbf{V}}.
717: \label{drnon}
718: \end{equation}
719: This would give an eigenfunction in a theory if $q_1$ and $q_r$ are constrained to
720: 0;
721: $q_1\equiv0\equiv q_r$. If this restriction is made in the prepotential
722: $W$ of $D_r^{(1)}$ theory together with $2\beta\to\beta$ (and $r\to
723: r+2$), it gives the prepotential of the $D_{r+1}^{(2)}$ to be discussed shortly
724: in section \ref{subdr2}. The corresponding eigenfunction is (\ref{dr2eig}).
725: The formula (\ref{drnon}) also `explains' the non-existence of the corresponding
726: eigenfunction in
727: $B_r^{(1)}$ theory, which is obtained by restriction $q_r\equiv0$ (together with $r\to
728: r+1$).
729:
730:
731: \bigskip
732: For the special case of $r=3$ the eigenvalues for the spinor and anti-spinor weights
733: are lower than that of the vector weights. We find several lower lying eigenstates:
734: \begin{eqnarray}
735: D_3^{(1)}:\hspace*{-3mm}&&\nonumber\\
736: \phi_{s1}&=&\sin q_1\sin q_2\sin q_3,
737: \qquad\qquad\ \, \tilde{H}\phi_{s1}=(4\beta+3/2)\phi_{s1},\\
738: \phi_{s2}&=&\cos q_1\cos q_2\cos q_3,
739: \qquad\qquad \tilde{H}\phi_{s2}=(4\beta+3/2)\phi_{s2},\\
740: \phi_{ss}&=&\sin 2q_1\sin 2q_2\sin 2q_3,
741: \qquad\quad \tilde{H}\phi_{ss}=(8\beta+6)\phi_{ss},\\
742: \phi_{2}&=&\cos 2q_1\cos 2q_2+\cos 2q_1\cos 2q_3
743: +\cos 2q_2\cos
744: 2q_3+2\beta/(1+2\beta),\nonumber\\
745: && \hspace*{49mm}\tilde{H}\phi_{2}=(8\beta+4)\phi_{2}.
746: \end{eqnarray}
747: These are closely related to the eigenfunctions of the
748: $A_r^{(1)}$, (\ref{arvecs}), (\ref{aradj}) and (\ref{a3add}) since
749: $A_3^{(1)}\cong D_3^{(1)}$.
750: %%%%%%%%%%%%%%%%%%%
751: \subsubsection{$D_{r+1}^{(2)}$}
752: \label{subdr2}
753: The extended Dynkin diagram of $D_{r+1}^{(2)}$, Fig.\ref{fig:dr2}, can be obtained
754: from that of
755: $B_{r+1}^{(1)}$ by folding the left `fish tail' containing the affine simple root.
756: Then $B_{r+1}^{(1)}$ is obtained from $D_{r+2}^{(1)}$ by folding the right `fish
757: tail' corresponding to the spinor and anti-spinor weights.
758: The affine simple root of
759: $D_{r+1}^{(2)}$ is the `lowest short root' of $B_r$.
760: %, see the Dynkin diagram Fig.\ref{fig:dr2}.
761: In this case the spectrum of the small oscillations, up to the
762: coupling constant
763: $\beta$ is:
764: \begin{equation}
765: 4\left\{1,\ldots,\sin^2(j\pi/(r+1))/\sin^2(\pi/(r+1)),\ldots\right\}.
766: \label{dr2asspec}
767: \end{equation}
768: \begin{figure}%[tbp]
769: \centering
770: \includegraphics{dr2dynk.eps}
771: \caption{$D_{r+1}^{(2)}$ Dynkin diagram with the numbers $n_j$
772: attached. The black spot is the affine simple root. }
773: \label{fig:dr2}
774: \end{figure}
775: The lowest eigenvalue is an integer 4 (times $\beta$) which corresponds to the
776: vector weights of $B_r$, the leftmost white vertex in Fig.\ref{fig:dr2}.
777: The set of vector weights is
778: $\mathbf{V}=\{\pm\mathbf{e}_j|j=1,\ldots,r\}$. Let us introduce the corresponding
779: wavefunctions
780: \begin{equation}
781: \phi=\phi_{\mathbf{V}}+{2\beta/(1+2\beta)},\quad
782: \phi_{\mathbf{V}}=\sum_{\mu\in\mathbf{V}}e^{2i\mu\cdot q}=2\sum_{j=1}^r\cos 2q_j.
783: \label{dr2eig}
784: \end{equation}
785: It is easy to see
786: \begin{equation}
787: \tilde{H}\phi=(4\beta+2)\phi,\quad
788: -{1\over2}\triangle\phi_{\mathbf{V}}=2\phi_{\mathbf{V}}.
789: \end{equation}
790:
791:
792: %%%%%%%%%%%%%%%%%%%
793: \subsubsection{$C_{r}^{(1)}$}
794: The spectrum of the small oscillations, up to the coupling
795: constant $\beta$ is easily read from Table I:
796: \begin{equation}
797: 8\left\{1,\ldots,\sin^2(j\pi/(2r))/\sin^2(\pi/(2r)),\ldots\right\}.
798: \label{crasspec}
799: \end{equation}
800: \begin{figure}%[tbp]
801: \centering
802: \includegraphics{crdynk.eps}
803: \caption{$C_{r}^{(1)}$ Dynkin diagram with the numbers $n_j$
804: attached. The black spot is the affine simple root. }
805: \label{fig:cr}
806: \end{figure}
807: The lowest eigenvalue is an integer 8 (times $\beta$) which belongs to the
808: vector weight of $C_r$, corresponding to the leftmost white vertex of
809: Fig.\ref{fig:cr}. The set of vector weights is
810: $\mathbf{V}=\{\pm\mathbf{e}_j|j=1,\ldots,r\}$. Let us introduce the corresponding
811: wavefunctions
812: \begin{equation}
813: \phi_{\mathbf{V}}=\sum_{\mu\in\mathbf{V}}e^{2i\mu\cdot q}=2\sum_{j=1}^r\cos 2q_j.
814: \label{creig}
815: \end{equation}
816: It is easy to see
817: \begin{equation}
818: \tilde{H}\phi_{\mathbf{V}}=(8\beta+2)\phi,\quad
819: -{1\over2}\triangle\phi_{\mathbf{V}}=2\phi_{\mathbf{V}}.
820: \end{equation}
821: As is well known $C_r^{(1)}$ is obtained from $A_{2r-1}^{(1)}$ by folding.
822: The above eigenfunction originates from (\ref{arvav}).
823: %%%%%%%%%%%%%%%%%%%
824: \subsubsection{$A_{2r}^{(2)}$}
825: This is also called a $BC_r$ root system, which is obtained by adding the affine
826: root of
827: $C_r^{(1)}$ to the set of simple roots of $B_r$.
828: The spectrum of the small oscillations, up to the coupling
829: constant $\beta$ is:
830: \begin{equation}
831: 8\left\{1,\ldots,\sin^2(j\pi/(2r+1))/\sin^2(\pi/(2r+1)),\ldots\right\}.
832: \label{ar2asspec}
833: \end{equation}
834: \begin{figure}%[tbp]
835: \centering
836: \includegraphics{ar2dynk.eps}
837: \caption{$A_{2r}^{(2)}$ Dynkin diagram with the numbers $n_j$
838: attached. The black spot is the affine simple root. }
839: \label{fig:ar2}
840: \end{figure}
841: The lowest eigenvalue is an integer 8 (times $\beta$) which corresponds to the
842: vector weight of $B_r$, $\mathbf{V}=\{\pm\mathbf{e}_j|j=1,\ldots,r\}$.
843: Let us introduce the corresponding wavefunctions
844: \begin{equation}
845: \phi=\phi_{\mathbf{V}}+{4\beta/(1+4\beta)},\quad
846: \phi_{\mathbf{V}}=\sum_{\mu\in\mathbf{V}}e^{2i\mu\cdot q}=2\sum_{j=1}^r\cos 2q_j.
847: \label{ar2eig}
848: \end{equation}
849: It is easy to see
850: \begin{equation}
851: \tilde{H}\phi=(8\beta+2)\phi,\quad
852: -{1\over2}\triangle\phi_{\mathbf{V}}=2\phi_{\mathbf{V}}.
853: \end{equation}
854: As in the $D_{r+1}^{(2)}$ case (\ref{dr2eig}), the eigenfunction (\ref{ar2eig})
855: has a constant part. This is related to the fact that the vector
856: representation of $B_r$ contains a zero weight.
857: In contrast the vector representation of $C_r$ does not contain a zero weight and
858: the corresponding eigenfunction (\ref{creig}) does not have a constant part.
859: This also explains that the eigenfunctions
860: corresponding to the vector and conjugate
861: vector representations (\ref{arvecs}) do not have a constant part,
862: whereas that corresponding to the adjoint representation
863: (\ref{aradj}) has a constant part. The adjoint representation has a rank number of
864: zero weights.
865:
866: %%%%%%%%%%%%%%%%%%%
867: \subsubsection{Exceptional affine Lie algebras}
868: For $E_6^{(1)}$, $E_7^{(1)}$, $E_8^{(1)}$ and $F_4^{(1)}$ none of
869: the frequencies of (\ref{atsspectrum}) are integers. The $G_2^{(1)}$ case, which
870: is obtained from $D_4^{(1)}$ by three-fold folding, has two integer eigenvalues
871: $\left\{8,24\right\}$ inherited from $D_4^{(1)}$. As shown in \ref{drone},
872: we found no exact eigenfunctions
873: for $D_r^{(1)}$ and $D_4^{(1)}$.
874: Therefore we do not expect any exact eigenfunctions
875: for the exceptional affine Toda-Sutherland systems and we have got none.
876: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
877: \subsection{Comments on non-affine Toda-Sutherland systems}
878: \label{natsut}
879: In the Toda molecule (Toda field theory) interactions, the affine simple root
880: $\alpha_0$ plays
881: an essential role for the existence of an equilibrium.
882: However, with $1/\sin^2 q$ type interactions,
883: an equilibrium is achieved without the
884: affine simple root $\alpha_0$. This opens a way to consider (non-affine)
885: Toda-Sutherland systems characterised by a prepotential
886: \begin{equation}
887: W(q)=\beta\sum_{\alpha\in \Pi}\log|\sin(\alpha\cdot
888: q)|=\beta\sum_{j=1}^r\log|\sin(\alpha_j\cdot q)|.
889: \label{nonaffprep}
890: \end{equation}
891: Note that it does not contain the affine simple root $\alpha_0$ nor the Dynkin-Kac
892: labels
893: $\{n_j\}$.
894: Since the highest root is not contained in the prepotential,
895: the configuration space
896: now is
897: \begin{equation}
898: PW_N=\{q\in{\bf R}^r|\ 0<\alpha\cdot q<\pi,\quad \alpha\in\Pi
899: \}.
900: \label{PWN}
901: \end{equation}
902:
903: Finding the equilibrium position $\bar{q}$ of the classical potential is easy.
904: It is again proportional to the Weyl vector $\varrho$ (\ref{rhodef})
905: \begin{eqnarray}
906: \bar{q}={\pi\over2}\varrho,\quad
907: \alpha_j\cdot\bar{q}={\pi\over2},\quad
908: \cot[\alpha_j\cdot\bar{q}]=0\quad \Rightarrow
909: \left.{\partial W(q)\over{\partial q_j}}\right|_{\bar{q}}=0,
910: \quad j=1,\ldots,r.
911: \end{eqnarray}
912: Due to the linear independence of the simple roots,
913: this equilibrium is unique in the
914: configuration space (\ref{PWN}).
915: The frequencies of small oscillations near the
916: equilibrium are the eigenvalues of the
917: matrix
918: \begin{eqnarray}
919: \widetilde{W}_{jk}&=&-\left.{\partial^2
920: W(q)\over{\partial q_j\partial
921: q_k}}\right|_{\bar{q}}=\beta
922: \sum_{l=1}^r(\alpha_l)_j(\alpha_l)_k=\beta\tilde{M}_{jk},\\
923: \tilde{M}&=&\sum_{j=1}^r\alpha_j\otimes\alpha_j.
924: \end{eqnarray}
925: For the simply laced root systems ($A$, $D$, $E$) the spectrum of $\tilde{M}$
926: is the same as the spectrum of the Cartan matrix
927: $C_{jk}=2\alpha_j\cdot\alpha_k/\alpha_k^2$, $j,k\in\Pi$.
928: There is a universal formula for the spectrum of $\tilde{M}$ for the
929: $A$, $D$, $E$ series:
930: \begin{equation}
931: \mbox{Spec}(\tilde{M})=\{4\sin^2(e_1/2h),\ldots,4\sin^2(e_r/2h)\},
932: \quad e_1,\ldots,e_r:\mbox{exponents}.
933: \label{toda-sutuni}
934: \end{equation}
935: The exponents of simply laced root systems are:
936: \begin{center}
937: \begin{tabular}{||c|c|l||c|c|l||}
938: \hline
939: \(\Delta\)&\(h\)&\mbox{exponents},\quad $e_1$,\ldots, $e_r$
940: & \(\Delta\)&\(h\)&\mbox{exponents},\quad $e_1$,\ldots, $e_r$\\
941: \hline
942: \vT\(A_r\)&\(r+1\)& 1, 2, 3, \ldots, $r$&
943: \(E_6\)&\(12\)&1, 4, 5, 7, 8, 11\\[3pt]
944: \hline
945: \vT\(D_r\)&\(2r\)&1, 3, 5,\ldots, $2r-1$;\ $r-1$
946: &\(E_7\)&\(18\)&1, 5, 7, 9, 11, 13, 17\\[3pt]
947: \hline
948: \vT&&
949: &\(E_8\)&\(30\)&1, 7, 11, 13, 17, 19, 23, 29\\[3pt]
950: \hline
951: \end{tabular}\\
952: \bigskip
953: Table II: The exponents
954: \(e_j\) for simply laced root systems.
955: \end{center}
956: For the $B$ series and $G_2$ we have
957: \begin{eqnarray}
958: B_r:\
959: \mbox{Spec}(\tilde{M})&=&\{
960: 4\sin^2(2j-1/2(2r+1))|j=1,\ldots,r\},
961: \label{toda-sutbr}\\
962: G_2:\
963: \mbox{Spec}(\tilde{M})&=&\{(4-\sqrt{13})/3, (4+\sqrt{13})/3\},
964: \end{eqnarray}
965: and analytical formulas are not known for the entire spectrum of $\tilde{M}$
966: in $C_r$ and $F_4$.
967:
968: Although some of the eigenfrequencies of the small oscillations
969: near the classical equilibrium (\ref{toda-sutuni}), (\ref{toda-sutbr}) are
970: integers, they are definitely not the lowest lying ones.
971: According to the quantum-classical correspondence \cite{ls}, we do not expect
972: to find
973: the exact eigenfunctions for the lowest lying states, which have non-integer
974: eigenvalues.
975: Thus it is highly unlikely that the eigenfunctions for the higher excited states,
976: being orthogonal to all the lower lying ones, could be obtained exactly,
977: even for the ones belonging to integer classical eigenvalues.
978: In fact we have not been able to find any exact eigenfunctions
979: for the (non-affine) Toda-Sutherland systems (\ref{nonaffprep}).
980: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
981: \section{Affine Toda-Calogero systems}
982: \label{atcal}
983: \setcounter{equation}{0}
984: Like the affine Toda-Sutherland system, the affine Toda-Calogero system
985: can be defined for any affine root system $\Pi_0$ (\ref{afsimp}).
986: However, in many respects the affine Toda-Calogero systems have less
987: remarkable properties than the affine Toda-Sutherland systems discussed in the
988: preceding section.
989: The equilibrium position $\bar{q}$ does not have a simple characterisation.
990: Except for the systems based on the $A^{(1)}$ series, the small oscillations
991: near the equilibrium do not have {\em integer\/} (times the coupling constant)
992: eigenvalues other than 2, which is universal for all the potentials
993: with quadratic plus inverse quadratic dependence on the coordinate $q$
994: \cite{Gamb}.
995:
996: The prepotential of the affine Toda-Calogero system is obtained from
997: that of affine Toda-Sutherland system (\ref{prep}) by changing $\sin\alpha_j\cdot q
998: \to \alpha_j\cdot q$ and adding a harmonic confining potential with
999: (angular) frequency $\omega>0$:
1000: \begin{equation}
1001: W(q)=\beta\sum_{j=0}^rn_j\log|\alpha_j\cdot q|-{\omega\over2}q^2.
1002: \label{calprep}
1003: \end{equation}
1004: Because of the singularity of the potential we restrict
1005: the configuration space to
1006: the {\em principal Weyl chamber\/} for simplicity:
1007: \begin{equation}
1008: PW=\{q\in{\bf R}^r|\ \alpha\cdot q>0,\quad \alpha\in\Pi
1009: \}.
1010: \label{PW}
1011: \end{equation}
1012: (Due the non-invariance under the Weyl group, theories with
1013: different configuration spaces are physically different.
1014: For example, they have different (non-equivalent) equilibrium positions.)
1015: The classical and quantum Hamiltonians are given in terms of the prepotential
1016: $W$ by the
1017: same formulas (\ref{casham}) and (\ref{qasham}).
1018: The classical equilibrium position $\bar{q}$ is determined by
1019: \begin{equation}
1020: {\partial
1021: W(\bar{q})\over{\partial q_k}}=0,\quad
1022: k=1,\ldots,r \quad
1023: \Longleftrightarrow\quad
1024: \beta \sum_{j=0}^r{n_j\alpha_j\over{\alpha_j\cdot \bar{q}}}=\omega\bar{q}.
1025: \end{equation}
1026: In
1027: contrast to the Calogero systems in which $\bar{q}$ corresponds to the zeros of
1028: classical polynomials, {\em i.e.\/} the Hermite and the Laguerre polynomials
1029: for the classical root systems \cite{zero,cs}, the present case does not have
1030: such simple characterisation.
1031: The frequencies of small oscillations near the equilibrium are given by the
1032: eigenvalues of the matrix
1033: \[
1034: \widetilde{W}=\mbox{Matrix}
1035: \left(-{\partial^2 W(\bar{q})\over{\partial q_j\partial q_k}}\right).
1036: \]
1037: We have evaluated $\bar{q}$ and $\widetilde{W}$ numerically for various affine root
1038: systems.
1039: We will discuss the systems based on the $A^{(1)}$ series in the section
1040: \ref{nonafar}. In all the other cases the only integer (times $\omega$)
1041: eigenvalues of
1042: $\widetilde{W}$ is 2, which exists in all the cases based on any root system.
1043: In fact it is more
1044: universal and exists for all the potentials with quadratic ($q^2$) plus inverse
1045: quadratic dependence on the coordinate $q$
1046: \cite{Gamb,bms} without any root or weight structure. This eigenvalue 2 gives rise
1047: to exact quantum eigenfunctions
1048: $\phi_n(q)$ which is proportional to the Laguerre polynomial \cite{Gamb,bms}
1049: in $q^2$:
1050: \begin{equation}
1051: \tilde{H}\phi_n(q)=2\omega n\phi_n(q),\qquad \phi_n(q)\propto L_n^{(E_0/\omega
1052: -1)}(\omega q^2),\quad n=1,2,\ldots,
1053: \label{lageig}
1054: \end{equation}
1055: in which $E_0=(\beta h+r/2) \omega$ is the ground state energy and $h$ is the
1056: Coxeter number (\ref{Coxdef}).
1057: Let us emphasise that these quantum eigenfunctions are also universal in the above
1058: sense.
1059:
1060: Here the similarity transformed Hamiltonian
1061: $\tilde{H}$ and the eigenfunctions $\{\phi_n(q)\}$ are defined in terms of
1062: the ground state wavefunction
1063: $\psi_0=e^W$ in the same formulas as before (\ref{phipsi})--(\ref{htilform}).
1064:
1065: %%%%%%%%%%%%%
1066: \subsection{$A_r^{(1)}$}
1067: \label{nonafar}
1068: This theory and its possible generalisation have been
1069: discussed rather extensively by Khare and collaborators \cite{JGA,AJK,EGKP} with
1070: explicit forms of quantum eigenfunctions.
1071: These multi-particle dynamics have nearest and next-to-nearest interactions with
1072: rational $1/q^2$ plus $q^2$ potentials.
1073: Here we discuss the relationship between the exact eigenfunctions and their
1074: classical
1075: counterparts \cite{ls}.
1076: The $A_2^{(1)}$ affine Toda-Calogero system is identical with $A_2$ Calogero system.
1077: The spectrum of $\widetilde{W}$ for $A_r^{(1)}$, $r\ge3$ has a form
1078: \begin{equation}
1079: \mbox{Spec}(\widetilde{W})=\omega\{1,2,3,*,\ldots\},
1080: \label{araftocal}
1081: \end{equation}
1082: in which $*,\ldots$ denote non-integers greater than 3.
1083:
1084: The interpretation of these
1085: three integer eigenvalues is quite clear. The lowest one corresponds to the
1086: elementary excitation of the {\em center of mass\/} coordinates
1087: $Q=q_1+\ldots+q_{r+1}$ and the quantum eigenfunction belonging to the eigenvalue
1088: $n\omega$ is essentially the Hermite polynomial of degree $n$ in $Q$.
1089: The eigenfunctions corresponding to the eigenvalue 2 are the Laguerre polynomials
1090: (\ref{lageig}) mentioned above. Let us introduce the elementary symmetric polynomial
1091: of degree $k$ in $q_1, \ldots, q_{r+1}$ \cite{ls}:
1092: \begin{equation}
1093: \prod_{j=1}^{r+1}(x+q_j)=\sum_{k=0}^{r+1}S_k x^{r+1-k},\quad S_0=1,\quad
1094: S_1=q_1+\cdots+q_{r+1}\equiv Q.
1095: \end{equation}
1096: Since $S_k$ is annihilated by the Laplacian, $\triangle S_k=0$, one finds easily
1097: the exact quantum eigenfunction $\phi_3$
1098: corresponding to the integer eigenvalue 3 in (\ref{araftocal}):
1099: \begin{equation}
1100: \tilde{H}S_3=3\omega S_3+\beta(r-1)Q,\quad
1101: \tilde{H}\phi_3=3\omega \phi_3,\quad \phi_3=S_3+{\beta(r-1)\over{2\omega}}Q.
1102: \end{equation}
1103:
1104: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1105: \section{Summary and comments}
1106: \label{comments}
1107: \setcounter{equation}{0}
1108:
1109: The affine Toda-Sutherland system is introduced for any affine root system
1110: as a cross between the affine Toda molecule and the Sutherland system.
1111: That is, the potential is trigonometric, $1/\sin^2q$, and the multi-particle
1112: interactions are governed by the affine simple roots only,
1113: in contrast to the entire set of roots in the Sutherland system.
1114: It has remarkable universal features.
1115: The classical equilibrium point is $\pi\varrho/h$ ($\varrho$: Weyl vector,
1116: $h$: Coxeter number) and the frequencies of small oscillations near the
1117: equilibrium are proportional to the corresponding affine Toda masses.
1118: In most cases based on classical affine Lie algebras, some low lying frequencies
1119: are integers (times a coupling constant).
1120: They give rise to exact quantum eigenvalues and eigenfunctions. The ground state
1121: eigenfunctions are always given explicitly.
1122: Thus the affine Toda-Sutherland systems provide examples of a new type of
1123: {\em quasi exactly solvable\/} multi-particle dynamics.
1124:
1125: Affine Toda-Calogero systems with rational ($1/q^2$ plus $q^2$) potentials
1126: are found to be less remarkable than their trigonometric counterparts.
1127: They possess an infinite number of exact eigenvalues and eigenfunctions which are
1128: well known.
1129: We have shown that the affine Toda-Calogero systems based on $A^{(1)}$ series
1130: have three lowest frequencies $\omega$, $2\omega$ and $3\omega$ of small
1131: oscillations near the classical equilibrium.
1132: They all correspond to exact quantum eigenvalues and eigenfunctions.
1133:
1134: It would be interesting to understand these `partially integrable'
1135: affine Toda-Sutherland-Calogero systems from various points of view:
1136: relationship with the random matrix models, analysis from the regular and chaotic
1137: dynamics, etc.
1138:
1139: \bigskip
1140: In \cite{JGA,AJK,EGKP} many interesting multi-particle dynamics, rational
1141: and trigonometric, related to the root systems of $B_r$, $C_r$, $BC_r$ and $D_r$
1142: were introduced.
1143: They resemble to our affine Toda-Sutherland and affine Toda-Calogero systems
1144: but they cannot be characterised in terms of affine simple roots.
1145: Unified understanding of these systems is wanted.
1146:
1147:
1148:
1149:
1150:
1151: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1152: % %
1153: % Acknowledgments %
1154: % %
1155: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1156: \section*{Acknowledgements}
1157: I.L. is a post-doctoral fellow with the F.W.O.-Vlaanderen (Belgium).
1158: This work
1159: was initiated when one of us (R.S.) visited Institute of Physics,
1160: Bhubaneswar as a part of JSPS-INSA Exchange Programme, arranged by J.\,Maharana.
1161:
1162:
1163:
1164: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1165: % %
1166: % References %
1167: % %
1168: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1169: \begin{thebibliography}{99}
1170:
1171: \bibitem{Toda}
1172: M.~Toda,
1173: ``Vibration of a chain with nonlinear interaction",
1174: J. Phys. Soc. Jpn. {\bf 22} (1967) 431-436.
1175:
1176: \bibitem{Cal}
1177: F.~Calogero,
1178: ``Solution of the one-dimensional \(N\)-body problem with quadratic
1179: and/or inversely quadratic pair potentials",
1180: J. Math. Phys. {\bf 12} (1971) 419-436.
1181:
1182: \bibitem{Sut}
1183: B.~Sutherland,
1184: ``Exact results for a quantum many-body problem in one-dimension. II'',
1185: Phys. Rev. {\bf A5} (1972) 1372-1376.
1186:
1187: \bibitem{Kost}
1188: B.~ Kostant, ``The solution to a generalized Toda lattice and representation theory",
1189: Adv. in Math. {\bf 34} (1979) 195--338.
1190:
1191: \bibitem{OPC}
1192: M.\,A.~Olshanetsky and A.\,M.~Perelomov,
1193: ``Classical integrable finite-dimensional systems related to Lie algebras'',
1194: Phys. Rep. {\bf C71} (1981), 314-400.
1195:
1196:
1197: \bibitem{CalMo}
1198: J.~Moser,
1199: ``Three integrable Hamiltonian systems connected with isospectral
1200: deformations'',
1201: Adv. Math. {\bf 16} (1975) 197-220;\
1202: J.~Moser,
1203: ``Integrable systems of non-linear evolution equations",
1204: in {\it Dynamical Systems, Theory and Applications\/};\
1205: J. Moser, ed., Lecture Notes in Physics {\bf 38} (1975), Springer-Verlag;\
1206: F.~Calogero, C.~Marchioro and O.~Ragnisco,
1207: ``Exact solution of the classical and quantal one-dimensional many body
1208: problems with the two body potential \(V_{a}(x)=g^2a^2/\sinh^2\,ax\)'',
1209: Lett. Nuovo Cim. {\bf 13} (1975) 383-387;\
1210: F.~Calogero,
1211: ``Exactly solvable one-dimensional many body problems'',
1212: Lett. Nuovo Cim. {\bf 13} (1975) 411-416.
1213:
1214:
1215: \bibitem{OP1}
1216: M.\,A.~Olshanetsky and A.\,M.~Perelomov,
1217: ``Completely integrable Hamiltonian systems connected with semisimple
1218: Lie algebras",
1219: Inventions Math. {\bf 37} (1976), 93-108.
1220:
1221:
1222: \bibitem{DHoker_Phong}
1223: E.~D'Hoker and D.\,H.~Phong,
1224: ``Calogero-Moser Lax pairs with spectral parameter for general Lie algebras'',
1225: Nucl. Phys. {\bf B530} (1998) 537-610, {\tt hep-th/9804124};
1226: A.\,J.~Bordner, E.~Corrigan and R.~Sasaki,
1227: ``Calogero-Moser models I: a new formulation'',
1228: Prog. Theor. Phys. {\bf 100} (1998) 1107-1129, {\tt hep-th/9805106};
1229: ``Generalized Calogero-Moser models and universal Lax pair operators'',
1230: Prog. Theor. Phys. {\bf 102} (1999) 499-529,
1231: {\tt hep-th/9905011}.
1232:
1233: \bibitem{bms}
1234: A.\,J.~Bordner, N.\,S.~Manton and R.~Sasaki,
1235: ``Calogero-Moser models V: Supersymmetry and Quantum Lax Pair",
1236: Prog. Theor. Phys. {\bf 103} (2000) 463-487, {\tt hep-th/9910033};
1237: S.\, P.~Khastgir, A.\, J.~Pocklington and R.~Sasaki,
1238: ``Quantum Calogero-Moser Models: Integrability for all Root Systems'',
1239: J.\ Phys. {\bf A33} (2000) 9033-9064, {\tt hep-th/0005277}.
1240:
1241:
1242: \bibitem{Ino1} V.\, I.\, Inozemtsev, ``The finite Toda lattices",
1243: Comm. Math. Phys. {\bf 121}
1244: (1989) 628-638.
1245:
1246:
1247:
1248: \bibitem{DHPh2} E.\, D'Hoker and D.\,H.\, Phong,
1249: ``Calogero-Moser and Toda systems for
1250: twisted and untwisted affine Lie Algebras'',
1251: Nucl. Phys. {\bf B530} (1998) 611-640, {\tt
1252: hep-th/9804125}.
1253:
1254: \bibitem{kst}
1255: S.\, P.\, Khastgir, R.\, Sasaki and K.\, Takasaki,
1256: ``Calogero-Moser Models IV: Limits to Toda theory",
1257: Prog. Theor. Phys. {\bf 102} (1999) 749-776, {\tt hep-th/9907102}.
1258:
1259:
1260:
1261:
1262:
1263: \bibitem{JGA}
1264: S.\,R.~Jain and A.~Khare,
1265: ``An exactly solvable many-body problem in one dimension",
1266: Phys. Lett. A262 (1999) 35-39;
1267: S.\,R.~Jain, B.~Gr\'emaud and A.~Khare,
1268: ``Quantum modes on chaotic motion: Analytically exact results",
1269: Phys. Rev. {\bf E66} (2002) 016216.
1270:
1271: \bibitem{AJK}
1272: G.~Auberson, S.\,R.~Jain and A.~Khare,
1273: ``Off-diagonal long-range order in one-dimensional many-body problem",
1274: Phys. Lett. {\bf A267} (2000) 293-295;
1275: ``A class of N-body problems with nearest- and next-to-nearest neighbour interactions",
1276: J. Phys. {\bf A34} (2001) 695-724, {\tt cond-mat/0004012}.
1277:
1278: \bibitem{EGKP}
1279: M.~Ezung, N.~Gurappa, A.~Khare and P.\,K.~Panigrahi,
1280: ``Algebraic study of quantum many-body systems with nearest and next-to-nearest neighbour
1281: long-range interactions,'' {\tt cond-mat/0007005}.
1282: %%CITATION = COND-MAT 0007005;%%
1283:
1284:
1285: %\bibitem{bcdsa}
1286: %H.\,W.~Braden, E.~Corrigan, P.\,E.~Dorey and R.~Sasaki,
1287: %``Extended Toda Field Theory and Exact S-Matrices,''
1288: %Phys.\ Lett.\ B {\bf 227} (1989) 411-416.
1289: %%%CITATION = PHLTA,B227,411;%%
1290:
1291:
1292:
1293:
1294: \bibitem{cs}
1295: E.\,Corrigan and R.\,Sasaki,
1296: ``Quantum vs classical integrability in Calogero-Moser systems",
1297: J.\ Phys.\ A {\bf 35} (2002) 7017-7062, {\tt hep-th/0204039};
1298: S.~Odake and R.~Sasaki,
1299: ``Polynomials associated with equilibrium positions in Calogero-Moser systems,''
1300: J.\ Phys.\ A {\bf 35} (2002) 8283-8314,
1301: {\tt hep-th/0206172};
1302: O.~Ragnisco and R.~Sasaki,
1303: ``Quantum vs classical integrability in Ruijsenaars-Schneider
1304: systems",
1305: Preprint YITP-03-09, {\tt hepth/0305120}.
1306:
1307: \bibitem{bcds}
1308: H.\,W.~Braden, E.~Corrigan, P.\,E.~Dorey and R.~Sasaki,
1309: ``Affine Toda Field Theory and Exact S-Matrices,''
1310: Nucl.\ Phys.\ B {\bf 338} (1990) 689-746.
1311: %%CITATION = NUPHA,B338,689;%%
1312:
1313: \bibitem{turb}
1314: A.\,V.~Turbiner, ``Quasi-exactly-soluble problems and sl(2,R) algebra", Comm.
1315: Math. Phys. {\bf 118} (1988) 467-474.
1316:
1317: \bibitem{ush}
1318: A.\,G.\,~Ushveridze,
1319: ``Qusi-exactly solvable models in quantum mechanics",
1320: IOP Publishing, Bristol, (1994).
1321:
1322: \bibitem{st1}
1323: R.\,Sasaki and K.\,Takasaki,
1324: ``Quantum Inozemtsev model, quasi-exact solvability and ${\cal
1325: N}$-fold supersymmetry",
1326: J.\ Phys.\ {\bf A34} (2001) 9533-9553,
1327: [Erratum-ibid.\ {\bf A34} (2001) 10335],
1328: {\tt hep-th/0109008}.
1329:
1330: \bibitem{ls}
1331: I.\,Loris and R.\,Sasaki,
1332: ``Quantum vs classical mechanics:
1333: role of elementary excitations", Kyoto preprint YITP-03-50,
1334: {\tt quant-ph/0308040};
1335: ``Quantum \& classical eigenfunctions in
1336: Calogero \& Sutherland systems",
1337: Kyoto preprint YITP-03-51,
1338: {\tt hep-th/0308052}.
1339:
1340: \bibitem{zero}
1341: F.~Calogero, ``On the zeros of the classical polynomials'', Lett. Nuovo
1342: Cim. {\bf 19} (1977) 505-507;
1343: ``Equilibrium configuration of one-dimensional many-body problems
1344: with quadratic and inverse quadratic pair potentials",
1345: Lett. Nuovo Cim. {\bf 22} (1977) 251-253;
1346: ``Eigenvectors of a matrix related to the zeros of Hermite polynomials",
1347: Lett. Nuovo Cim. {\bf 24} (1979) 601-604;
1348: ``Matrices, differential operators and polynomials'',
1349: J. Math. Phys. {\bf 22} (1981) 919-934.
1350:
1351: \bibitem{Gamb}
1352: P.\,J.\, Gambardella, ``Exact results in quantum many-body systems of
1353: interacting particles in many dimensions with \(\overline{SU(1,1)}\) as the
1354: dynamical group",
1355: J. Math. Phys. {\bf 16} 1172-1187 (1975).
1356: \end{thebibliography}
1357:
1358:
1359:
1360: \end{document}
1361: %%%%%%%%%%%%%%%%%%%%%%%%%%%%
1362: % End %
1363: %%%%%%%%%%%%%%%%%%%%%%%%%%%%
1364: