1: \documentclass{arSIH}
2:
3: \newtheorem{theorem}{Theorem}
4: \newtheorem{proposition}[theorem]{Proposition}
5: \newtheorem{lemma}[theorem]{Lemma}
6: \newtheorem{corollary}[theorem]{Corollary}
7:
8: \theoremstyle{definition}
9: \newtheorem{definition}[theorem]{\font\=cmssi10\Definition\bf}
10: \newtheorem{example}[theorem]{\font\=cmssi10\Example\bf}
11: \newtheorem{remark}[theorem]{\font\=cmssi10\Remark\bf}
12:
13: %% Seppo Hiltunen macros
14:
15: \def\Cal{\mathcal}
16: %
17: %
18: % special CHARACTER constructions:
19: %
20: %\def\Omega{\hbox{\font\=cmr10\\char'012}}
21: %\def\Delta{\hbox{\font\=cmr10\\char'001}}
22: \def\Eps{\hbox{\font\=cmmi10 scaled\magstep1\\char'017}\kern0.15mm}% bigger varepsilon
23: \def\Iota{\kern.15mm\hbox{\font\=cmmi10 scaled\magstep1\\char'023}\kern0.2mm}% bigger iota
24: \def\Nu{\hbox{\font\=cmmi10 scaled\magstep1\\char'027}\kern0.25mm}% bigger nu
25: \def\uvarPi{\kern.15mm\underline{\kern-.15mm\varPi\kern-.85mm}\kern.85mm}% underlined varPi
26: \def\uOmega{\kern.3mm\underline{\kern-.3mm\Omega\kern-.3mm}\kern.3mm}% underlined Omega
27: \def\varXii{\char'004}\def\Pii{\char'005}\def\varUpsilonn{\char'007}
28: \def\betaa{\char'014}\def\thetaa{\char'022} \def\kappaa{\char'024}
29: \def\lambdaa{\char'025}\def\pii{\char'031}\def\rhoo{\char'032}\def\sigmaa{\char'033}
30: \def\tauu{\char'034}\def\varphii{\char'047}
31: %
32: %
33: % miscellaneous one-place function symbols:
34: %
35: \def\Rlp{\roman{Re\,}}% Re z
36: \def\Imp{\roman{Im\,}}% Im z
37: \def\fRe{\hbox{\font\=cmr9\f\kern.1mm}\roman{Re}\kern.75mm}% the real part of a function
38: \def\fIm{\hbox{\font\=cmr9\f\kern.1mm}\roman{Im}\kern.65mm}% the imaginary part of a function
39: \def\vecc#1{\kern-.5mm\vec{\kern.5mm#1}}% \vec moved a bit backwords so that in (\vecc x) the arrow would not get too close to ')'
40: %
41: %
42: %mathematical OPERATORS (one-place function symbols) for topological vector spaces
43: %
44: %\providecommand{\dimHa}{\mathop{\rm dim_{_{Ha}}}\nolimits}%
45: \def\TVS{\roman{TVS}\kern0.37mm}%
46: \def\LCS{\roman{LCS}\kern0.37mm}%
47: \def\BaS{\roman{BaS}\kern0.37mm}%
48: \def\dimHa{{\rm dim_{_{\kern.2mm Ha}}}}% the Hamel dimension of a structured vector space
49: \def\rajou{{}^{}{\Cal B}_{s\,}}% the set of bounded sets in a TVS
50: \def\Linb{\Cal L\lower.7mm\hbox{\kern.1mm\font\=cmmi6\b}}% \Linb(E,F) gives \Cal L_b(E,F) = the topological vector space of continuous linear maps E to F with the topology of uniform convergence on bounded sets
51: \def\LL^#1{L\kern0.15mm\raise.4mm\hbox{$^{#1}$}\kern0.15mm}% = L^p with p a bit lifted
52:
53:
54:
55: \def\dualbeta{^{\kern0.4mm\prime}_{\kern-.2mm\raise.95mm\hbox{$_{_\beta}$}}} % E\dualbeta = E'_\beta = strong topological dual of E
56: \def\Nbh{\Cal N_{\font\=cmmi6\lower.15mm\hbox{\kern.1mm\bh\kern.15mm}}}% \Nhb(x,T) = set of T-neighborhoods of x
57: \def\Topma{\roman{{Top_{}}_{\hbox{\font\=cmr6\ma}}\kern.15mm}}
58: \def\co{\hbox{\font\=cmmi12\c}\kern.15mm\lower.15mm\hbox{$_{\rm o}$}}% \co(I,F) = the supremumnormable space of functions I to F with become small outside finite sets
59: \def\prodc{\prod{_{_{\kern-.3mm\bold c\kern.15mm}}}}% cartesian propduct of sets
60: \def\vsprod_#1_#2{\prod\kern-0.3mm{}_{_{\roman{#1}\sp{#2}\,}}} % for example, \vsprod_TVS_\Re produces \prod_{\roman{TVS} I\!\!R}
61: \def\vscoprod_#1_#2{\coprod\kern-0.3mm{}_{_{\roman{#1}\sp{#2}\,}}} % see the preceding
62: \def\expnota^#1]_#2{\,^{#1\,]{_{}}_{\roman{#2}}}} % F\expnota^\Omega_{tvs} = the topological product vector space of all functions Omega to F
63: %
64: %
65: \def\sixemptyset{\hbox{\font\=cmsy6\\char'073}}
66: \def\bold#1{{\bf#1}}
67: \def\roman#1{{\rm#1}}
68: \def\limu_#1{\lim\kern-5.5mm\lower1.5mm\hbox{$_{#1}\ $}}
69: \def\oseoy{\raise1.9mm\hbox{\kern.5mm\font\=cmr5\o}\kern-1.7mm y}% kahdessa kohtaa esiintyy
70: \def\O{{}^{}\Cal O}
71: \def\Univ{\hbox{\font\=cmssbx10\U}} % the class of all sets
72: \def\Pows{\Cal P\kern-.4mm_s\kern.3mm}
73: \def\lei{ {}_{ {}^{\,\downarrow\text{\hskip-2.1mm} } } \cap }
74: \def\lei{\hbox{\kern.45mm$_{^\downarrow}\kern-1.280mm\cap\kern.85mm$}}
75: \def\Ze{Z\!\!\!Z} % kokonaisluvut
76: \def\Zepp{{{{Z\!\!\!Z^{\phantom{l}}}^{{}_{{}^{\!}\!+}}}}} % pos. kokonaisluvut
77: \def\Zep{{{{{{{{Z\!\!\!Z_{}}_{}}_{}}_{}}}_{{}^{\!+}}}}} % ei-neg. kokonaisluvut
78: \def\supp{{\rm supp}{}^{}\,}
79: \def\Card{{}^{}{\rm Card}{}^{}\,}
80: \def\inve{\lower.85mm\hbox{$^{^-}$}\kern-.5mm{}^\iota}
81:
82: \def\fvalue{\hbox{\kern.2mm\font\=cmr10\\char'022\kern-.2mm}} % f\value x=f(x)
83: \def\ffvalue{\hbox{\kern.2mm\font\=cmr7\\char'022\kern-.2mm}} % f\value x=f(x)
84: \def\image{\hbox{\font\=cmr10\\char'022\kern-1mm\char'022}} % f\image A=f[A]
85: \def\iimage{\hbox{\font\=cmr7\\kern.3mm\char'022\kern-.7mm\char'022\kern-.3mm}} % f\image A=f[A]
86: \def\images{\hbox{\font\=cmr10\\char'022\kern-1mm\char'022\kern-1mm\char'022}} % f\images\Cal A={f[A]:A in Cal A}
87: \def\timesn{\kern-.2mm\times\kern-.2mm} % same as times but with smaller spaces before and after
88: \def\ttimes{\hbox{\kern-.2mm${}\times\kern-2.5mm\lower.8mm\hbox{\font\=cmr5\t}\kern1.8mm$}} % product topology = \rist
89: \def\ktimes{\hbox{\kern-.2mm${}\times\kern-2.5mm\lower1mm\hbox{\font\=cmr5\k}\kern1.5mm$}} % compactly generated product topology
90: \def\vstimes{\kern.95mm\raise.45mm\hbox{\font\=cmbsy6\\char'002}\kern-2.3mm\lower.9mm\hbox{\font\=cmr5\vs}\kern1.05mm} % X\vstimes Y = the vector space product of X and Y
91: \def\rist{\,{\times}_{{}_{{}_{\!\!\!\!{\bf t}\ }}}} % tulotopologia
92: \def\risti#1{{}^{}\,{\times}_{{\!}_{{}_{\!\!\!{#1}{}^{\!}\ }}}}
93: \def\tanB{{\tau_{_B}}} % tangent bundle \tanB M
94: \def\Circ{\,\text{\font\=cmbsy10\\char'016}\sp\,}
95: \def\Examplee{{\font\=cmssi10\E\kern.15mmx\kern.15mma\kern.15mmm\kern.14mmp\kern.17mml\kern.15mme}\kern.3mm. }
96: \def\Examples{{\font\=cmssi10\E\kern.15mmx\kern.15mma\kern.15mmm\kern.14mmp\kern.17mml\kern.15mme\kern.15mms}\kern.3mm. }
97: \def\Remarkk{{\font\=cmssi10\R\kern.15mme\kern.15mmm\kern.15mma\kern.15mmr\kern.15mmk\kern.15mm. }}
98: \def\Remarkss{{\font\=cmssi10\R\kern.15mme\kern.15mmm\kern.15mma\kern.15mmr\kern.15mmk\kern.15mms}\kern.3mm. }
99: \def\N{{I\!\!N}} % former \Ne, the numbers 1,2,...
100: \def\No{{I\!\!N\kern-.54mm\lower.15mm\hbox{$_{\rm o}$}}} % former \Neo, the natural numbers 0,1,2,...
101: \def\Nopot#1{I\!\!N\kern-.54mm\lower.15mm\hbox{$_{\rm o}$}\kern-.7mm{}^{#1}} % N_o^{#1}
102: \def\potNo{^{\kern.37mm I\!\!{N_{}}_{\kern-.22mm{\rm o}}}} % ^{IN_o}
103: \def\minus{\kern.2mm\lower1.05mm\hbox{$^-$}}
104: \def\pplus{\raise.22mm\hbox{\font\=cmr5\\char'053}}% 5 point +
105: \def\mminus{\raise.18mm\hbox{\font\=cmsy5\\char'000}}% 5 point -
106: \def\plusinftyy{\raise.18mm\hbox{\font\=cmr5\\char'053}\infty}% +infty for sub- and superscripts with smaller +
107: \def\minusinftyy{\raise.18mm\hbox{\font\=cmsy5\\char'000}\infty}% -infty for sub- and superscripts with smaller -
108: \def\plusinfty{\lower1.05mm\hbox{$^+$}\infty}
109: \def\minusinfty{\lower1.05mm\hbox{$^-$}\infty}
110: \def\Qe{\hbox{$Q\kern-2.6mm\raise.2mm\hbox{\font\=cmssqi8\I}\kern1.7mm$}}% the set of rational numbers
111: \def\Re{{I\!\!R}}
112: \def\Rep{{{I\!\!R^{\phantom{l}}}^{{}_{{}^{\!}\!+}}}}
113: \def\Repp{{{{{{{I\!\!R_{}}_{}}_{}}_{}}}_{{}^{\!+}}}}
114: \def\Ce{{\hbox{$C\kern-2.5mm\raise.2mm\hbox{\font\=cmssqi8\I}\kern1.48mm$}}}
115: \def\imag{\kern.15mm\lower.6mm\hbox{$^{^*}$}\kern-1.8mm\imath\kern.1mm} % the imaginary unit
116: \def\imgn{\roman i} % another imaginary unit
117: \def\Ke{{{}^{}I\!\!K}}
118: \def\bvarSigma{\hbox{\font\=cmmib10\\char'006}} % math-bold-italic varSigma
119: \def\ebiF{\kern.1mm\hbox{\font\=cmmib8\F}\kern.5mm} % 8 point bold italic F
120: \def\ebiT{\kern.1mm\hbox{\font\=cmmib8\T}\kern.6mm} % 8 point bold italic T
121: \def\ebiU{\kern.1mm\hbox{\font\=cmmib8\U}\kern.5mm} % 8 point bold italic U
122: \def\bit#1{{\font\=cmmib10\text{\\char'#1}}} % math-bold-italic, for example $\bit{166}$ is bold italic v in math
123: \def\biit#1{\hbox{\font\=cmmib10\#1}} % math-bold-italic, for example $\biit v$ is bold italic v in math
124: \def\bmi#1#2{\hbox{\font\=cmmib#1\\char'#2}}
125: \def\bmii#1#2{\hbox{\font\=cmmib#1\#2}}
126: \def\zzeta{020}
127: \def\eeta{021}\def\xxi{030}\def\varPii{\char'005}\def\varOmegaa{\char'012}
128: \def\fssi#1{\hbox{\font\=cmssi10\#1}\kern0.15mm} % cmssi in math
129: \def\smb#1{\hbox{\font\=cmmi8\#1\kern.3mm}} % eightpoint (capital) math symbols
130: \def\ssmb#1{\hbox{\font\=cmmi6\#1}} % smaller capital for index
131: \def\bCal#1{\hbox{\font\=cmbsy10\#1}} % math-bold-calligraphic, for example $\bCal C$ is bold calligraphic C in math
132: \def\eCal#1{\kern.1mm\hbox{\font\=cmbsy8\#1\kern.4mm}} % 8point bold calligraphic #1
133: \def\ecal#1{\kern.1mm\hbox{\font\=cmsy8\#1\kern.3mm}} % 8point calligraphic #1
134: \def\ncal#1{\kern.1mm\hbox{\font\=cmsy9\#1\kern.3mm}} % 9point calligraphic #1
135: \def\vcal#1{\kern-.1mm\vec{\kern.2mm\hbox{\font\=cmsy7\#1}\kern.3mm}} % e.g., vector bundle \vcal E
136: \def\conc{\!{\bf{\hat{\phantom w}}}\!}
137: \def\concc{{}^{}{}^{}\!{\bf{\hat{\phantom w}}}\!{}^{}}
138: \def\idv{\hbox{\font\=cmr10\id}\kern.25mm\lower.8mm\hbox{\font\=cmr7\v}\kern.3mm} % id_v E=id(v_s E)
139: \def\Idec{\roman{{Id_{\ssp}}_{c\,}}} % Idec bCal X is the class of identities of a category bCal X
140: \def\seq#1{\langle#1\rangle}
141: \def\Seq#1{\big\langle#1\big\rangle}
142: \def\ymp{{}^{}{\Cal N}_o\,}
143: \def\SemiNor{\Cal S_{_N}\kern0.15mm}% \SemiNor E = set of continuous seminorms in E
144: \def\vecs{\upsilon\kern-0.3mm\lower.15mm\hbox{$_s$}\kern0.2mm} % underlying set of a structured vector space
145: \def\vecss{\hbox{\font\=cmitt10\v}\kern-0.1mm\lower.15mm\hbox{$_s$}\kern0.2mm} % underlying set of a vector space / vector (= linear) struc
146: \def\bnull#1{\hbox{\font\=cmssbx10\0}{}_{\font\=cmmi6\lower.15mm\hbox{\kern-.1mm\#1\kern.15mm}}} % \bnull X = zero in vector space X
147: \def\bzero#1{\hbox{\font\=cmbx10\0}{}_{\font\=cmmi6\lower.15mm\hbox{\kern-.1mm\#1\kern.15mm}}} % \bzero X = zero in vector space X
148: \def\dom{{{}^{}{\rm dom}\,{}_{{}^{}}}}
149: \def\domm{{}^{}{\rm dom}^{\kern.3mm\hbox{\font\=cmr6\2}}\,}
150: \def\Dom{{}^{}{\rm Dom}\,{}_{{}^{}}}
151: \def\rng{{}^{}{\rm rng}\,{}_{{}^{}}}
152: \def\CPi#1{C\kern-.2mm\lower.05mm\hbox{$_{_\Pi}$}\kern-1.52mm{}^{#1}}
153: \def\CinftyPi{C\kern.4mm\raise.3mm\hbox{$^\infty$}\kern-3.35mm_{_\Pi}\kern1.45mm}
154: \def\CinftyS{\Cinfty\kern-3.9mm_{_{\Cal S}}\kern1.45mm}
155: \def\Cinfty{C\kern.4mm\raise.3mm\hbox{$^\infty$}\kern.15mm}
156: \def\Cinftyzero{\hbox{$C\kern.4mm\raise.3mm\hbox{$^\infty$}\kern.15mm\kern-3.5mm_{\font\=cmr6\lower.15mm\hbox{\kern.1mm\0}}\kern1.9mm$}}
157: \def\Lip#1{{}^{}{{{{{{\Cal L}{}^{}ip}^{}}^{}}^{}}^{}}^{#1{}^{}}}
158: \def\RHB#1#2{\raise#1mm\hbox{$#2$}} % raised (by #1 mm) horizontal box of #2
159: \def\LHB#1#2{\lower#1mm\hbox{$#2$}} % lowered (by #1 mm) horizontal box of #2
160: \def\fivemath#1{\hbox{\font\=cmmi5\#1}}
161: \def\fiveroman#1{\hbox{\font\=cmr5\#1\kern.1mm}}
162: \def\sixmath#1{\hbox{\font\=cmmi6\#1}}
163: \def\sixroman#1{\hbox{\font\=cmr6\#1\kern.1mm}}
164: \def\eightmath#1{\hbox{\font\=cmmi8\{#1}\kern.1mm}}% 8 point math italic
165: \def\eightroman#1{\hbox{\font\=cmr8\{#1}\kern.1mm}}% 8 point roman
166: \def\erm#1{{\font\=cmr8\#1}}% 8 point roman for text, e.g. \erm{DF\,}--\,space
167: \def\subtext#1{\raise.2mm\hbox{$_{_{\kern0.15mm\roman{#1}}}$}}
168: \def\subtexT#1{\raise.2mm\hbox{$_{_{\kern0.15mm\hbox{\font\=cmr5\#1}}}$}}
169: \def\sNor#1{\kern.25mm\lower.38mm\hbox{$_{#1}$}}
170: \def\sNorr#1{\kern-.2mm\lower.38mm\hbox{$_{#1}$}}
171: \def\sNoreset_#1{\kern.13mm\lower.83mm\hbox{\font\=cmmi6\C}\kern.32mm\lower.1mm\hbox{$_{^{\emptyset,#1}}$}}% \|y\|\sNoreset_i produces \|y\|_{C^{\emptyset,i}}
172: \def\sbi#1{{_{\kern-0.1mm}}_{#1}} % same as _#1 but a little lower
173: \def\ais#1_#2{{}_{\font\=cmmi6\lower.15mm\hbox{\kern-.1mm\#1\kern.15mm}}\lower.3mm\hbox{${_{\kern-0.3mm_{#2}}}$}} %
174: %% D\ais E_2 gives D_{E_2} where E is 6 point
175: \def\aais#1_#2{\kern.1mm{}_{\font\=cmmi6\lower.25mm\hbox{\kern-.1mm\#1\kern.15mm}}\lower.4mm\hbox{${_{\kern-0.3mm_{#2}}}$}} %
176: \def\ai#1{{}_{\font\=cmmi6\lower.15mm\hbox{\kern-.1mm\#1\kern.15mm}}} % 6:n pisteen kirjainalaindeksi
177: \def\yi#1{^{\font\=cmmi6\raise.0mm\hbox{\kern-.1mm\#1\kern.15mm}}} % 6:n pisteen kirjainylindeksi
178: \def\ar#1{{}_{\font\=cmr6\lower.15mm\hbox{\kern.1mm\#1}}} % 6:n pisteen numeroalaindeksi
179: \def\aar#1{_{\font\=cmr6\lower.15mm\hbox{\kern.1mm\#1}}} % 6:n pisteen numeroalaindeksi
180: \def\yr#1{^{\font\=cmr6\raise.0mm\hbox{\kern.3mm\#1}}} % 6:n pisteen numeroylindeksi
181: \def\yrai^#1_#2{^{\kern.4mm\hbox{\font\=cmr6\{#1}}}_{\kern.2mm{#2}}}
182: \def\upparentes#1{^{\kern.2mm\raise.2mm\hbox{\font\=cmr6\\char'050}\kern.1mm{#1}\kern.1mm\raise.2mm\hbox{\font\=cmr6\\char'051}}} % e.g. x\upparentes l gives x^{(l)} with 6point parentheses
183: \def\lupar{\kern.2mm\lower1mm\hbox{$^{^(}$}} % pieni vasen ylsulku
184: \def\rupar{\lower1mm\hbox{$^{^)}$}\kern-.15mm} % pieni oikea ylsulku
185: \def\yyi#1{^{\font\=cmmi6\lower.6mm\hbox{\kern-.25mm\#1\kern-.05mm}}} % pieni kirjainylindeksi, tarkoitettu derivaatan merkintn
186: \def\yyr#1{^{\font\=cmr6\lower.45mm\hbox{\kern-.25mm\#1\kern-.15mm}}} % pieni numeroylindeksi, tarkoitettu derivaatan merkintn
187: \def\yplus{\lower1mm\hbox{$^{^+}$}} % + merkki edelliseen
188: \def\yminus{\lower1mm\hbox{$^{^-}$}} % - merkki edelliseen
189: \def\aminus{{\kern.15mm\raise.3mm\hbox{$_{_-}$}\kern-.1mm}}%
190: \def\yplk{{}^{{}_{{}^{'\!}}}} % isoille kirjaimille sopiva ylpilkku TEXma.tex
191: \def\adot{\kern.2mm\hbox{\font\=cmb10\\char'056}}%
192: \def\ydot{\kern.2mm\raise1.9mm\hbox{\font\=cmb10\\char'056}}% k\ydot = k^. = the integer corresponding to the natural number k
193: \def\yydot{\kern.2mm\raise1.35mm\hbox{\font\=cmb7\\char'056}\kern.2mm}% the above for 7 point mode
194: \def\yydott{\kern.2mm\raise1.35mm\hbox{\font\=cmb6\\char'056}\kern.2mm}% the above for 7 point mode
195: \def\ClT{{\rm Cl}\kern.25mm\lower.4mm\hbox{$_{\Cal T}$}\kern0.2mm} % Cl_Cal T A = closure of A w.r.t. T
196: \def\IntT{\sp{\rm Int}\kern.2mm\lower.4mm\hbox{$_{\Cal T}$}\kern0.2mm} % Int_Cal T A = interior of A w.r.t. T
197: \def\Cl_taurd#1{\roman{Cl_{}}_{\kern0.37mm\hbox{\font\=cmmi8\\char'034}\kern-0.15mm{_{}}_{rd}\kern0.2mm#1\,}}% Cl_{tau_{rd}#1}S = closure of S in the topology of the topologized vector space E (= #1)
198: \def\inc{\subseteq}
199: \def\iinc{\supseteq}
200: \def\exi#1{\exists\,#1\kern.2mm\,;}
201: \def\all#1{\forall\,#1\kern.2mm\,;}
202: \def\imply{\Rightarrow}
203: \def\equivv{\Leftrightarrow}
204: \def\embed{\hookrightarrow}
205: \def\spp{\kern0.07mm} % a horizontal very small positive space
206: \def\sp{\kern0.15mm} % a horizontal small positive space
207: \def\ssp{\kern0.37mm} % a bigger positive space
208: \def\snn{\kern-0.2mm} % a very small horizontal negative space
209: \def\sn{\kern-0.3mm} % a horizontal small negative space
210: \def\ssn{\kern-0.63mm} % a bigger negative space
211: \def\biggerlineskip#1 {\linebreak\nopagebreak\vskip-4.2mm\vskip.#1mm\nopagebreak\noindent}%
212: \def\Biggerlineskip#1 {\linebreak\nopagebreak\vskip-4.2mm\vskip#1pt\nopagebreak\noindent}%
213: \def\KP#1{\kern#1mm} % posive kern of #1 mm
214: \def\KN#1{\kern-#1mm} % negave kern of #1 mm
215: \def\nhskip#1mm{$\null$\kern#1mm}
216: \def\hyppy#1{$\phantom{}$\hskip#1}
217: \def\mhyppy#1{\null\kern#1mm}
218: \def\mathbreak{$\linebreak $}
219: \def\text#1{\hbox{\rm#1}}
220: \def\NS{\vskip1.7mm}
221: \def\NSN{\vskip1.7mm\noindent}
222: \def\VBOX/#1/#2/HEREend{\vbox{#2\vskip-#1mm}\vfill\null\eject}
223: %% use the above to force a vertical box on one page : #1 = number , #2 = text
224: \def\$#1${\hbox{$#1$}} % text math which is not compressed or stretched
225: \def\{\"a} \def\{\"o}
226: \def\q#1{``\kern0.37mm#1\kern0.37mm"}
227: \def\newProCla#1\par#2\par{\vskip1.7mm\noindent\bf#1\it#2\vskip1.7mm}
228: \def\Prooff{{\font\=cmssi10\P\kern.37mmr\kern.37mmo\kern.37mmo\kern.37mmf\kern.37mm. }\rm}
229: \def\newProof{Proof. \rm}
230: \def\QED{\hfill\hbox{$\ \sqcap\kern-2.45mm\sqcup$}}
231: \def\newQED{\hfill\hbox{$\ \sqcap$\hskip-2.45mm$\sqcup$}\vskip1.7mm}
232: \def\noin{\noindent}
233: \def\Newline{\kern-10mm\newline}
234: \font\rp=cmr8
235: \def\eps{\varepsilon}
236:
237: % ----------------------------------------------------------------------------
238:
239: \begin{document}
240:
241: \title[$ \text{\sc On Nash\,--\,Moser applications}$]%
242: {On an assertion about Nash\,--\,Moser applications}
243:
244: \author[S. Hiltunen]{Seppo\ I\. Hiltunen}
245: \address{Helsinki University of Technology \vskip0mm$\hspace{2mm}$
246: Institute of Mathematics, U311 \vskip0mm$\hspace{2mm}$
247: P.O.\ Box 1100 \vskip0mm$\hspace{2mm}$
248: FIN-02015 HUT\vskip0mm
249: FINLAND}
250: \email{shiltune\,@\,cc.hut.fi}
251:
252: \subjclass[2000]{46A61}
253:
254: \keywords{Directional differentiability, P-norm, Fr\'echet space, almost
255: M-tame map, Nash\,--\,Moser application, inverse\sp/\sp implicit function theorem,
256: too strong premise.}
257:
258: \begin{abstract}
259:
260: By an example we show that Olaf M\"uller's assertion about his new theorems
261: being able to give anew some classical results previously obtained via
262: applications of Nash\,--\,Moser type theorems is unfounded. We also give
263: another example indicating some limitations in possible applications of
264: related new inverse function theorems. \end{abstract}
265:
266: \maketitle
267:
268: % ----------------------------------------------------------------------------
269:
270: \noin Below, a {\it space\ssp} will mean a complete metrizable real locally
271: convex space $E\,$, hence a Fr\'echet space. By a {\it P-norm\ssp} for $E$ we
272: mean any function $\varrho:E\owns x\mapsto\varrho(x)=\|x\|$ such that $(x,y)
273: \mapsto\|x-y\|$ is a metric defining the topology of $E\,$.
274:
275: \begin{definition}
276:
277: A directionally differentiable (see \cite[Section 3]{Hic}\sp) map $\tilde f=
278: (E,F,f)$ of spaces $E,F$ where $f$ is a function $E\iinc\dom f\to F$, we call
279: {\it almost M-tame\ssp} at $x$ iff there are P-norms $\varrho_1$ for $E$ and $
280: \varrho_2$ for $F$ such that for any $z,u$ with $\varrho_1(z)\le 1$ and $u\in
281: E$ we have $x+z\in\dom f$ and $\varrho_2(\delta f(x+z,u)-\delta f(x,u))\le
282: \varrho_1(u)\,$.
283:
284: \end{definition}
285:
286: \begin{example}\label{Mul-cou}
287:
288: Let $E$ be the Fr\'echet space of 1--periodic smooth functions $x:\Re\to\Re\,
289: $, and let the smooth $\varphi:\Re\to\Re$ be 1--periodic. Letting $\Iota:t
290: \mapsto n\,t$ and $\bold n:t\mapsto n$ for some fixed $n\in\Ze\setminus\{0\}\,
291: $, we consider the map $\tilde f\subtext{Ex\,2}=(E,E,f)$ where $f$
292: is defined by $U\owns x\mapsto\varphi\circ(\Iota+x)\cdot(\bold n+x')$ with $U$
293: being the set of $x$ in $E$ having $
294: 0 < \inf\ssp\{\,|\,n+x'(s)\ssp|:s\in\Re\sp\,\}\,$. \end{example}
295:
296: \begin{proposition}\label{xx}
297:
298: If $\tilde f\subtext{Ex\,2}$ is almost M-tame at any $x\,$, then $\rng\varphi$
299: is a singleton. \end{proposition}
300:
301: \begin{proof} To proceed by reductio ad absurdum, we assume that $
302: \tilde f\subtext{Ex\,2}$ is almost M-tame at some $x$ and that $\varphi$ is
303: not constant, and we derive a contradiction. Now, there is some $t_0$ such
304: that $\varphi'(t_0)\not=0\,$, and we can also choose $s_0$ so that $
305: n\,s_0+x(s_0)=t_0\,$. There also are P-norms $u\mapsto\|u\|_1$ and $
306: v\mapsto\|v\|_2$ for $E$ such that for $\|z\|_1\le 1$ and for $u\in E$ we have
307: $\|\delta f(x+z,u)-\delta f(x,u)\|_2\le\|u\|_1\,$.
308:
309: The topology of $E$ also being defined by the sequence of norms $\seq{p_i:i\in
310: \No}$ where $p_i(u)=\sup\{|u^{(l)}(s)|:s\in\Re$ and $l\le i\}\,$, there is an
311: odd $k\in\N$ such that $p_{k-1}(z)\le k^{-1}\imply\|z\|_1\le 1\,$. Then there
312: is $\delta_1\in\Rep$ such that $\|v\|_2\le\delta_1\imply p_{k-1}(v)\le 1\,$,
313: and there is $l\in\N$ such that $p_l(u)\le l^{-1}\imply\|u\|_1\le\delta_1\,$.
314:
315: Taking $\eps\ar 0=l^{-1}$ and $m\in\N$ to be fixed below, we let $u:s\mapsto
316: \eps_0 $ and $z:s\mapsto(2\pi\,m)^{-k+\frac12}\sin(m\,2\pi\,(s-s_0))\,$. Then
317: putting $v=\delta f(x+z,u)-\delta f(x,u)\,$, we have $\|v\|_2\le\|u\|_1$ if $
318: p_{k-1}(z)\le k^{-1}\,$, and
319:
320: $\mhyppy{11} v = \varphi'\circ(\Iota+x+z)\cdot u\cdot(\bold n+x'+z')
321: + \varphi\circ(\Iota+x+z)\cdot u'$
322:
323: $\mhyppy{33} - \varphi'\circ(\Iota+x)\cdot u\cdot(\bold n+x')
324: - \varphi\circ(\Iota+x)\cdot u'$
325:
326: \centerline{$\phantom v = \eps_0(\varphi'\circ(\Iota+x+z)\cdot(\bold n+x'+z')
327: - \varphi'\circ(\Iota+x)\cdot(\bold n+x'))\,$,}
328:
329: \noin and hence $v^{(k-1)}=\eps_0(\varphi'\circ(\Iota+x+z)\cdot z^{(k)}+\roman T_z)$
330: where $\roman T_z$ denotes a sum of terms where only derivatives
331: $z^{(i)}$ of $z$ with $i<k$ occur. Consequently, there is $M\in\Rep$ with
332: [ $p_{k-1}(z)\le k^{-1}$ and $s\in\Re\imply |\roman T_z(s)|\le M$ ] .
333:
334: Now fixing $m$ so that $(2\pi\,m)^{-\frac12}\le k^{-1}$ and $
335: l+M < (2\pi\,m)^{\frac12}|\varphi'(t_0)|\,$, we have $p_{k-1}(z)\le k^{-1}$
336: and $p_l(u)\le l^{-1}\,$, hence $\|v\|_2\le\|u\|_1\le\delta_1\,$, and
337: consequently $p_{k-1}(v)\le 1<\eps_0((2\pi\,m)^{\frac12}|\varphi'(t_0)|-M)\le
338: \eps_0|\varphi'(n\,s_0+x(s_0))\cdot z^{(k)}(s_0)+\roman T_z(s_0)|
339: = |v^{(k-1)}(s_0)|\le p_{k-1}(v)\,$, a contradiction. \end{proof}
340:
341: In \cite[Section 5, Applications 1\,--\,5, pp.\ 25\,--\,26]{Mul} it is claimed
342: that certain previous results in \cite{Ha} can be obtained anew by the
343: approach in \cite{Mul}\ssp. In four of the asserted applications it is needed
344: to know that for a fixed tensor field $\omega$ on a manifold $N$ the pullback
345: map $\chi\mapsto\chi^*\omega$ by immersions $\chi:M\to N$ is
346: \q{bounded-differentiable}. It is obvious that if this indeed is the case,
347: then any local representation of the pullback map also is everywhere almost
348: M-tame.
349:
350: It follows from our Proposition \ref{xx} above that this generally fails in
351: the case where $M=N=\mathbb S^1$ since our space $E$ is both a model space for
352: the manifold of immersions and linearly homeomorphic to the space of smooth
353: 1--forms. Namely, in a local representation the pullback map for 1--forms has
354: precisely the form $x\mapsto\varphi\circ(\Iota+x)\cdot(\bold n+x')\,$. Similar
355: results can be obtained for other pairs of manifolds $M\sp,N$ and types of
356: tensors relevant to \cite[pp.\ 25\,--\,26]{Mul} by suitably adapting the idea
357: in the proof of Proposition \ref{xx} above.
358:
359: \begin{example}\label{Hel-cou}
360:
361: Let $E$ be the Fr\'echet space of smooth functions $x:[0,1]=I\to\Re\,$, and
362: let $\varphi:\Re\to\Re$ be a smooth diffeomorphism. We consider the
363: diffeomorphism $\tilde f\subtext{Ex\,4}=(E,E,f)$ where $f$ is defined by $
364: x\mapsto\varphi\circ x\,$. \end{example}
365:
366: \begin{proposition}\label{Helge B}
367:
368: If $\tilde f\subtext{Ex\,4}$ is almost M-tame at $x\,$, then $\varphi''(t)=0$
369: for all $t\in\rng x\,$. \end{proposition}
370:
371: \begin{proof} Letting $\tilde f\subtext{Ex\,4}$ be almost M-tame at $x\,$, we
372: assume that $\varphi''(t_0)\not=0$ for some $t_0\in\rng x\,$, and show that a
373: contradiction follows. Now, there is some $s_0\in I$ with $x(s_0)=t_0\,$.
374: There also are P-norms $u\mapsto\|u\|_1$ and $v\mapsto\|v\|_2$ for $E$ such
375: that for $\|z\|_1\le 1$ and for $u\in E$ we have $
376: \|\delta f(x+z,u)-\delta f(x,u)\|_2\le\|u\|_1\,$.
377:
378: The topology of $E$ also being defined by the sequence of norms \$\seq{\,\sp
379: p\sbi i:i\in\No\ssp}$ where $p_i(u)=\sup\{|u^{(l)}(s)|:s\in\Re$ and $l\le i\}\,
380: $, there is an
381: odd $k\in\N$ such that $p_{k-1}(z)\le k^{-1}\imply\|z\|_1\le 1\,$. Then there
382: is $\delta_1\in\Rep$ such that $\|v\|_2\le\delta_1\imply p_k(v)\le 1\,$, and
383: there is $l\in\N$ such that $p_l(u)\le l^{-1}\imply\|u\|_1\le\delta_1\,$.
384:
385: Taking $\eps\ar 0=l^{-1}$ and $m\in\N$ to be fixed below, we let $u=I\times
386: \{\eps_0\}$ and \$z=\Seq{\,(\sp 2\,\pi\,m\sp)^{\sp\mminus k\ssp+
387: \ssp\frac12}\ssp\sin\,(\sp 2\,\pi\,m\,(\sp s-s\ar 0)):s\in I\sp\,}\,$. Putting
388: $v=\delta f(x+z,u)-\delta f(x,u)\,$, the implication $p_{k-1}(z)\le k^{-1}
389: \imply\|v\|_2\le\|u\|_1$ holds, and we have\vskip.4mm
390:
391: \centerline{$v = \varphi'\circ(x+z)\cdot u - \varphi'\circ x\cdot u
392: = \eps_0(\varphi'\circ(x+z) - \varphi'\circ x)\,$,}\vskip.4mm
393:
394: \noin and hence $v^{(k)}=\eps_0(\varphi''\circ(x+z)\cdot z^{(k)}+\roman T_z)$
395: where $\roman T_z$ denotes a sum of terms where only derivatives
396: $z^{(i)}$ of $z$ with $i<k$ occur. Consequently, there is $M\in\Rep$ such that
397: the implication [ $p_{k-1}(z)\le k^{-1}$ and $s\in I\imply |\roman T_z(s)|\le M$ ]
398: holds.
399:
400: Now fixing $m>(2\pi)^{-1}\cdot\max\,\{k^2,|\varphi''(t_0)|^{-2}(l+M)^2\}\,$,
401: we have $p_{k-1}(z)\le k^{-1}$ and $p_l(u)\le l^{-1}\,$,
402: hence $\|v\|_2\le\|u\|_1\le\delta_1\,$, and consequently
403: $p_k(v)\le 1<\eps_0((2\pi\,m)^{\frac12}|\varphi''(t_0)|-M)\le
404: \eps_0|\varphi''((x+z)(s_0))\cdot z^{(k)}(s_0)
405: +\roman T_z(s_0)|=|v^{(k)}(s_0)|\le p_k(v)\,$, a contradiction. \end{proof}
406:
407: From Proposition \ref{Helge B} it follows that if $\tilde f\subtext{Ex\,4}$
408: satisfies the premise in \cite[Theorem B, p.\ 3]{Gl} at $x\,$, then $
409: \varphi\,|\,\rng x$ is affine. Furthermore, if there is a set $A$ of functions
410: $x$ such that at every $x\in A$ the premise holds, and if in addition the set
411: $\,\rng\sn\bigcup\ssp A=\{\,x\ssp(t):x\in A$ and $t\in I\sp\,\}\,$ is dense in
412: the real line, then $\varphi$ itself is affine.
413:
414: Noting that the idea in the proof of Proposition \ref{Helge B} can be
415: naturally adapted to show more generally that for a fixed smooth \$\varphi:
416: \varPi\spp\sqcap\varPi\aar 1\iinc\Omega\,|\,Q\to\varPi\aar 2$ only locally
417: affine maps \$\sp\tilde f\subtext{Re\,6}=(\sp\Cinfty(\spp Q\sbi{\varPi}\ssp,
418: \varPi\aar 1)\ssp,\Cinfty(\spp Q\sbi{\varPi}\ssp,\varPi\aar 2)\ssp,f\sp)$ with
419: $f$ given by \$x\mapsto$ \$\varphi\circ[\,\sp\roman{id\,},x\,]$ can
420: generically satisfy the premise in \cite[Theorem B, p.\ 3]{Gl}\ssp, we thus
421: see that problems involving this kind of maps $\sp\tilde f\subtext{Re\,6}$ are
422: practically ruled outside the domain of possible applications.
423:
424: \begin{remark}
425:
426: Our preceding observation, of course, does not preclude the possibility that
427: the theorems in the carefully written exposition \cite{Gl} may have nontrivial
428: applications where maps like $\sp\tilde f\subtext{Re\,6}$ do not play an
429: essential role.
430:
431: There also are some historical implicit and inverse function theorems beyond
432: Banach spaces sharing this property of having premises too strong in order for
433: these theorems to be able to have serious applications involving maps of the
434: type $\sp\tilde f\subtext{Re\,6}$ above. We shall soon publish one such result
435: concerning Seip's various inverse and implicit function theorems in
436: \cite{Se72}\ssp. \end{remark}
437:
438:
439: % ----------------------------------------------------------------------------
440:
441: \begin{thebibliography}{99} \def\{\font\=cmcsc7\}
442:
443: \bibitem{Gl}\bibname{H. G{\l\ckner}}\ssp: `Implicit functions from
444: topological vector spaces to Fr\'echet spaces in the presence of metric
445: estimates' (v4) {\em
446: preprint}\ssp{\rm, http://arXiv.org/abs/math.FA/0612673}.
447: \bibitem{Ha}\bibname{R. S. H{\amilton}}\ssp: `The inverse function theorem of
448: Nash and Moser' {\it Bull.\ Amer.\ Math.\ Soc\ssp}.\ {\bf7} (1982) no 1,
449: 65--222.
450: \bibitem{Hic}\bibname{S. H{\iltunen}}\ssp: `Differentiation, Implicit
451: functions, and applications to generalized well-posed\-ness' {\em
452: preprint}\ssp{\rm, http://arXiv.org/abs/math.FA/0504268}.
453: \bibitem{Mul}\bibname{O. M{\\"uller}}\ssp: `Bounded Fr\'echet geometry' {\em
454: preprint}\ssp{\rm, http://arXiv.org/abs/math.DG/0612379}.
455: \bibitem{Se72}\bibname{U. S{\eip}}\ssp: {\em Kompakt erzeugte Vektorr\ume und
456: Analysis}\sp, Lecture Notes in Math. 273, Springer, Berlin - Heidelberg -
457: New York 1972.
458:
459: \end{thebibliography}
460:
461: \end{document}
462: