
<p>ANNALES DE L’I. H. P., SECTION A </p><p>´</p><p>JEDRZEJ SNIATYCKI </p><p>Dirac brackets in geometric dynamics </p><p>Annales de l’I. H. P., section A, tome 20, n<sup style="top: -0.2793em;">o </sup>4 (1974), p. 365-372 </p><p><<a href="/goto?url=http://www.numdam.org/item?id=AIHPA_1974__20_4_365_0" target="_blank">http://www.numdam.org/item?id=AIH</a><a href="/goto?url=http://www.numdam.org/item?id=AIHPA_1974__20_4_365_0" target="_blank">P</a><a href="/goto?url=http://www.numdam.org/item?id=AIHPA_1974__20_4_365_0" target="_blank">A</a><a href="/goto?url=http://www.numdam.org/item?id=AIHPA_1974__20_4_365_0" target="_blank">_1974__20_4_365_0</a>> </p><p>© Gauthier-Villars, 1974, tous droits réservés. L’accès aux archives de la revue « Annales de l’I. H. P., section A » implique l’accord avec les conditions générales d’utilisation (<a href="/goto?url=http://www.numdam.org/conditions" target="_blank">http://www.numdam. </a><a href="/goto?url=http://www.numdam.org/conditions" target="_blank">org/conditions</a>). Toute utilisation commerciale ou impression systématique est constitutive d’une infraction pénale. Toute copie ou impression de ce fichier doit contenir la présente mention de copyright. </p><p>Article numérisé dans le cadre du programme Numérisation de documents anciens mathématiques </p><p><a href="/goto?url=http://www.numdam.org/" target="_blank">http://www.numdam.org/ </a></p><p>Ann. Inst. Henri </p><p>Section A : </p><p>Poincaré, </p><p>Vol. </p><p>n° </p><p>365 </p><p>XX, </p><p>4, 1974, </p><p>Physique théorique. </p><p>in </p><p>Dirac brackets </p><p>geometric dynamics </p><p>015ANIATYCKI </p><p>Jedrzej </p><p>of </p><p>The </p><p>Canada. </p><p>Alberta, </p><p>University </p><p>Calgary, </p><p>of Mathematics </p><p>Statistics and </p><p>Science </p><p>Department </p><p>Computing </p><p>is formulated in the </p><p>of of constraints in </p><p>ABSTRACT. </p><p>framework of </p><p>- Theory </p><p>symplectic </p><p>dynamics </p><p>Geometric </p><p>significance </p><p>secondary </p><p>of Dirac </p><p>Global existence </p><p>geometry. </p><p>is </p><p>constraints and of Dirac brackets </p><p>given. </p><p>brackets is </p><p>proved. </p><p>1. </p><p>INTRODUCTION </p><p>The <sup style="top: 0.01em;">successes </sup>of the canonical </p><p>of </p><p>with </p><p>quantization </p><p>the </p><p>dynamical systems </p><p>a</p><p>finite number of </p><p>of </p><p>of </p><p>freedom, </p><p>degrees </p><p>necessity </p><p>quan- </p><p>gravi- </p><p>theory </p><p>experimental </p><p>of </p><p>of the </p><p>field </p><p>tization </p><p>and the </p><p>that </p><p>electrodynamics, </p><p>hopes </p><p>quantization </p><p>could resolve </p><p>in </p><p>quantum </p><p>tational field </p><p>difficulties encountered </p><p>of the </p><p>have </p><p>to </p><p>rise </p><p>canonical structure of </p><p>given </p><p>thorough investigation </p><p>field theories. It has been found that the </p><p>formu- </p><p>cases </p><p>standard Hamiltonian </p><p>most </p><p>lation of </p><p>of </p><p>is </p><p>in the </p><p>dynamics </p><p>inadequate </p><p>and </p><p>physically </p><p>interesting </p><p>dueto existence of </p><p>constraints. Methods </p><p>electrodynamics </p><p>gravitation </p><p>with constraints have been </p><p>of </p><p>with </p><p>several </p><p>by </p><p>developed </p><p>dealing </p><p>dynamics </p><p>authors and </p><p>it </p><p>has been realized that the standard Hamiltonian </p><p>dynamics </p><p>can be formulated in terms of constraints </p><p>(1). </p><p>has </p><p>been </p><p>a</p><p>mathematical </p><p>The aim of </p><p>of constraints </p><p>Hamiltonian </p><p>dynamics </p><p>given </p><p>very elegant </p><p>formulation in the framework of </p><p>a</p><p>(2). </p><p>theory </p><p>symplectic geometry </p><p>this </p><p>in </p><p>is to </p><p>formulation of the </p><p>paper </p><p>give </p><p>symplectic </p><p>As </p><p>a</p><p>result the </p><p>of the classification of </p><p>geometric </p><p>and of Dirac brackets is </p><p>dynamics. </p><p>significance </p><p>Further, </p><p>the </p><p>of </p><p>constraints </p><p>given. </p><p>globalization </p><p>See refs. </p><p>the references </p><p>there. </p><p>and </p><p>[4], [5], [6], </p><p>quoted </p><p>C) </p><p>[2], [3], </p><p>[8] </p><p>See </p><p>refs. </p><p>and </p><p>[1] </p><p>[9]. </p><p>(~) </p><p>eg. </p><p>Henri Poincaré - </p><p>de l’Institut </p><p>Section A - </p><p>n° 4 - </p><p>1974. </p><p>XX, </p><p>Annales </p><p>vol. </p><p>366 </p><p>JEDRZEJ SNIATYCKI </p><p>theresults discussedin theliteraturein termsof local coordinatesis obtained. </p><p>the fundamental </p><p>with </p><p>notion </p><p>constraints </p><p>is that </p><p>As in the case of the Hamiltonian </p><p>dynamics, </p><p>of </p><p>in the </p><p>dynamical systems </p><p>geometric analysis </p><p>of </p><p>a</p><p>manifold. Basic </p><p>of </p><p>manifolds are </p><p>symplectic </p><p>properties </p><p>symplectic </p><p>reviewed in Section 2. </p><p>a</p><p>with <sup style="top: 0.01em;">constraints </sup>can </p><p>be </p><p>constraint </p><p>A</p><p>dynamical system </p><p>represented by </p><p>a</p><p>submanifold of </p><p>is described </p><p>and </p><p>manifold. </p><p>constraint </p><p>Therefore, </p><p>symplectic </p><p>dynamics </p><p>manifold </p><p>a</p><p>a</p><p>where </p><p>is </p><p>M, </p><p>by </p><p>triplet (P, </p><p>P, </p><p>of canonical </p><p>(P, cv) </p><p>symplectic </p><p>M</p><p>is </p><p>a</p><p>submanifold of </p><p>which is called </p><p>a</p><p>canonical </p><p>Ele- </p><p>system [12]. </p><p>their relations to </p><p>and the </p><p>mentary properties </p><p>with </p><p>systems, </p><p>Lagrangian </p><p>a</p><p>relation </p><p>between </p><p>cano- </p><p>systems </p><p>nical </p><p>homogeneous Lagrangians, </p><p>and its reduced </p><p>phase space </p><p>are </p><p>in Section 3. </p><p>discussed </p><p>system </p><p>4</p><p>Section </p><p>contains a</p><p>of </p><p>constraints. The </p><p>discussion </p><p>secondary </p><p>genera- </p><p>lization of Dirac classification of constraints is </p><p>in Section 5. Sec- </p><p>given </p><p>tion </p><p>6</p><p>is devoted to an </p><p>ofthe </p><p>of </p><p>Dirac brac- </p><p>analysis </p><p>of </p><p>their </p><p>geometric significance </p><p>a</p><p>kets and </p><p>existence. </p><p>proof </p><p>global </p><p>2. </p><p>SYMPLECTIC MANIFOLDS </p><p>P</p><p>is </p><p>a</p><p>manifold </p><p>and <sup style="top: 0.01em;">OJ </sup></p><p>manifold is </p><p>a</p><p>where </p><p>A</p><p>a</p><p>(3) </p><p>pair (P, OJ) </p><p>symplectic symplectic </p><p>form on <sup style="top: 0.02em;">P. </sup>The most </p><p>of </p><p>a</p><p>is </p><p>important example </p><p>symplectic </p><p>of </p><p>is furnished </p><p>in this case </p><p>the structure of the </p><p>manifold in </p><p>by </p><p>space </p><p>dynamics </p><p>system, </p><p>bracket. </p><p>phase </p><p>P</p><p>the </p><p>and <sup style="top: 0.01em;">cv </sup>is the </p><p>a</p><p>represents </p><p>phase space </p><p>dynamical </p><p>Lagrange </p><p>be </p><p>a</p><p>manifold </p><p>a</p><p>function on <sup style="top: 0.02em;">P. </sup>There exists </p><p>Let </p><p>(P, OJ) </p><p>symplectic </p><p>and </p><p>f</p><p>such </p><p>form </p><p>a</p><p>vector field </p><p>called the Hamiltonian vector field </p><p>unique </p><p>that </p><p>of f, </p><p>J</p><p>~ = 2014 </p><p>d f, </p><p>where </p><p>and </p><p>J</p><p>denotes </p><p>the left interior </p><p>are Hamiltonian vector fields </p><p>of a</p><p>product </p><p>v f </p><p>avector field. It </p><p>by </p><p>correspond- </p><p>v f </p><p>vg </p><p>and </p><p>then their Lie brackets </p><p>is </p><p>to functions </p><p>ing </p><p>f</p><p>g, respectively, </p><p>[v f’ ~] </p><p>the Hamiltonian vector field </p><p>to the Poisson bracket </p><p>and the Jacobi </p><p>corresponding </p><p>( f, g) </p><p>for </p><p>Further, </p><p>( f, </p><p>g) </p><p>an </p><p>identity </p><p>of f and <sup style="top: -0.01em;">g. </sup></p><p>~(~) = 2014 vg( f ), </p><p>the Poisson brackets is </p><p>immediate </p><p>of </p><p>the Jacobi </p><p>that <sup style="top: 0.01em;">6D </sup>is </p><p>consequence </p><p>identity </p><p>of </p><p>fields and the </p><p>closed. </p><p>for the Lie bracket </p><p>vector </p><p>assumption </p><p>3. </p><p>CANONICAL SYSTEMS </p><p>- Acanonical </p><p>is </p><p>a</p><p>where </p><p>DEFINITION </p><p>symplectic </p><p>3 . l. </p><p>M, </p><p>(P, OJ) </p><p>(P, </p><p>triplet </p><p>cv) </p><p>system </p><p>a</p><p>submanifold of P. </p><p>M</p><p>is </p><p>a</p><p>manifold and </p><p>is </p><p>one </p><p>from the </p><p>if </p><p>Canonical </p><p>Lagrangian dynamics </p><p>and </p><p>passes </p><p>systems appear </p><p>All </p><p>manifolds </p><p>considered in this </p><p>are finite </p><p>dimensional, </p><p>(3) </p><p>paper </p><p>paracompact </p><p>of class C ". </p><p>Annales </p><p>de l’Institut Henri </p><p>Poincaré - Section </p><p>A</p><p>DIRAC BRACKETS IN </p><p>DYNAMICS </p><p>367 </p><p>GEOMETRIC </p><p>with </p><p>to the </p><p>canonical </p><p>of </p><p>systems </p><p>dynamics </p><p>homogeneous Lagrangians. </p><p>Let </p><p>L</p><p>be </p><p>a</p><p>one </p><p>a</p><p>on </p><p>a</p><p>conical </p><p>X. function of </p><p>defined </p><p>homogeneous </p><p>degree </p><p>domain </p><p>We denote </p><p>D</p><p>of the </p><p>TX of </p><p>bundle </p><p>tangent </p><p>space </p><p>configuration space </p><p>transformation </p><p>FL : D - T*X the </p><p>by </p><p>Legendre </p><p>given by </p><p>for </p><p>the fibre derivative of </p><p>L</p><p>and </p><p>0</p><p>the Liouville form on T*X </p><p>defined, </p><p>is the </p><p>by </p><p>E</p><p>where n : T*X -~ </p><p>X</p><p>a</p><p>each v </p><p>bundle </p><p>by </p><p>p(Tn(v)) </p><p>cotangent </p><p>canonical </p><p>TpT*X, </p><p>The </p><p>is </p><p>FL, </p><p>(T*X, </p><p>de) </p><p>system </p><p>projection. </p><p>triplet </p><p>range </p><p>submanifold ofT*X. </p><p>canonical There are two subsets </p><p>FL </p><p>be </p><p>is </p><p>a</p><p>range </p><p>provided </p><p>a</p><p>Let </p><p>(P, </p><p>TPM </p><p>K</p><p>and </p><p>N</p><p>M, </p><p>co) </p><p>system. </p><p>of </p><p>associated to </p><p>as </p><p>follows: </p><p>M, </p><p>(P, </p><p>co) </p><p>and </p><p>N=KnTM. </p><p>N</p><p>is called the characteristic set </p><p>for </p><p>canonical </p><p>of </p><p>M. The </p><p>The set </p><p>w</p><p>dynamical signifi- </p><p>cance of </p><p>N</p><p>a</p><p>obtained from </p><p>a</p><p>FL, </p><p>(T*X, range </p><p>homogeneous Lagrangian </p><p>d0) </p><p>system </p><p>with </p><p>on </p><p>X</p><p>a</p><p>L</p><p>is </p><p>given by </p><p>Lagrangian system </p><p>the </p><p>following. </p><p>the </p><p>3 . 2. </p><p>- A curve </p><p>in </p><p>X</p><p>satisfies </p><p>if </p><p>PROPOSITION </p><p>Lagrange-Euler equa- </p><p>where <sub style="top: 0.15em;">y </sub>denotes </p><p>only FL. , </p><p>y</p><p>L</p><p>and </p><p>tions </p><p>the </p><p>to the </p><p>corresponding </p><p>prolongation </p><p>Lagrangian </p><p><sup style="top: 0.01em;">is </sup>an </p><p>this </p><p>curve of N. Proof of </p><p>of to </p><p>TX, </p><p>integral </p><p>propo- </p><p>sition is </p><p>in the reference </p><p>and it will be omitted. </p><p>given </p><p>[11] </p><p>system </p><p>for </p><p>a</p><p>of </p><p>M</p><p>which can be </p><p>canonical </p><p>Thus, </p><p>connected </p><p>M, </p><p>(P, </p><p>points </p><p>curves of </p><p>N</p><p>are related in </p><p>a</p><p>by integral </p><p>be </p><p>physically meaningfull </p><p>related </p><p>time </p><p>evolution as </p><p>a</p><p>in </p><p>3.2 or </p><p>manner; </p><p>they may </p><p>by </p><p>Proposition </p><p>the maximal </p><p>can be </p><p>a</p><p>transformation. </p><p>E</p><p>by </p><p>gauge </p><p>Therefore, </p><p>M, </p><p>given </p><p>point </p><p>p</p><p>N</p><p>manifold of </p><p>it </p><p>exists, </p><p>integral </p><p>as the </p><p>p, </p><p>through </p><p>provided </p><p>interpreted </p><p>of </p><p>history </p><p>p. </p><p>if </p><p>N</p><p>is </p><p>a</p><p>DEFINITION </p><p>3 . 3. </p><p>A</p><p>canonical </p><p>is </p><p>M, </p><p>(P, </p><p>w) </p><p>regular </p><p>system </p><p>of TM. </p><p>be </p><p>and since cv is closed </p><p>subbundle </p><p>a</p><p>subbundle </p><p>Frobenius </p><p>a</p><p>canonical </p><p>Then </p><p>N</p><p>is </p><p>Let </p><p>M, </p><p>(P, </p><p>cv) </p><p>regular </p><p>system. </p><p>is also involutive. </p><p>of </p><p>N</p><p>TM, </p><p>theorem </p><p>Hence, </p><p>unique </p><p>by </p><p>each </p><p>E</p><p>M</p><p>is contained in </p><p>a</p><p>maximal </p><p>(4), </p><p>point </p><p>p</p><p>integral </p><p>manifold of N. Let P~ denote the </p><p>set of </p><p>M</p><p>the </p><p>quotient </p><p>by </p><p>equivalence </p><p>relation defined </p><p>maximal </p><p>manifolds of </p><p>that is each element </p><p>N, </p><p>described </p><p>dynamical </p><p>system </p><p>by </p><p>a</p><p>integral </p><p>of the </p><p>in P~ </p><p>our </p><p>cano- </p><p>represents </p><p>history </p><p>by </p><p>M ~ </p><p>Pr denote the canonical </p><p>nical </p><p>and let </p><p>M, </p><p>(P, </p><p>system </p><p>p : </p><p>projec- </p><p>submersion </p><p>M</p><p>P~ admits </p><p>a</p><p>a</p><p>tion. If </p><p>then there exists </p><p>differentiable structure such that </p><p>is </p><p>p</p><p>a</p><p>formwr on <sup style="top: 0.01em;">P~ </sup>such that </p><p>unique symplectic </p><p>w</p><p>The </p><p>manifold </p><p>called the </p><p>to </p><p>of </p><p>is </p><p>reduced </p><p>(Pr, </p><p>symplectic </p><p>wr) </p><p>phase space </p><p>on Pr </p><p>constants of </p><p>Functions </p><p>M, </p><p>(P, </p><p>CD). </p><p>correspond </p><p>(gauge invariant) </p><p>in <sup style="top: 0.01em;">the </sup>manner <sup style="top: 0.01em;">described </sup></p><p>and </p><p>cvr </p><p>rise to their Poisson </p><p>motion, </p><p>algebra </p><p>gives </p><p>in </p><p>2</p><p>Section </p><p>(5). </p><p>All theorems on <sup style="top: 0.01em;">differentiable </sup>manifolds used in this </p><p>can be found </p><p>in ref. </p><p>(~) (~) </p><p>paper </p><p>[7]. </p><p>This Poisson </p><p>was <sup style="top: 0.01em;">first </sup>studied </p><p>in ref. </p><p>[2]. </p><p>algebra </p><p>Vol. </p><p>n° 4 - 1974. </p><p>XX, </p><p>JEDRZEJ SNIATYCKI </p><p>368 </p><p>4. </p><p>CONSTRAINTS </p><p>SECONDARY </p><p>on </p><p>In </p><p>If </p><p>is not </p><p>points </p><p>then </p><p>dim </p><p>on which dim </p><p>e M. </p><p>open </p><p>(P, M, </p><p>regular </p><p>p</p><p>p</p><p>cv) </p><p>depends </p><p>particular </p><p>Np </p><p>the <sup style="top: 0.01em;">set </sup>S° of </p><p>E</p><p>M</p><p>0</p><p><sup style="top: 0.01em;">is </sup>an </p><p>submanifold </p><p>inadmissible </p><p>Np </p><p>S° as </p><p>of </p><p>if it is not </p><p>this set </p><p>M, </p><p>Discarding </p><p>physically </p><p>constraints </p><p>here. Consider the class ofall manifolds </p><p>empty. </p><p>[3] </p><p>have lead Dirac </p><p>to the notion of </p><p>secondary </p><p>generalization </p><p>M</p><p>ofwhich is </p><p>given </p><p>n<br>Y</p><p>dim </p><p>contained in </p><p>and </p><p>Y’ </p><p>such that TY </p><p>N</p><p>is </p><p>a</p><p>subbundle </p><p>of </p><p>TY </p><p>allow here </p><p>(we </p><p>order in this class defined as follows </p><p>Y</p><p>0), </p><p>Y</p><p>let </p><p>denote the </p><p>partial </p><p>if </p><p>and </p><p>if </p><p>Y</p><p>is </p><p>a</p><p>of Y’. </p><p>follows from Zorn’s lemma that </p><p>submanifold </p><p>It </p><p>only </p><p>there </p><p>to </p><p>the </p><p>exist maximal manifolds in </p><p>class. </p><p>we are lead </p><p>this </p><p>Thus, </p><p>following. </p><p>DEFINITION </p><p>4.1. </p><p>A</p><p>constraint manifold of </p><p>a</p><p>canonical </p><p>secondary </p><p>maximal manifold </p><p>is </p><p>a<br>S</p><p>contained in </p><p>M</p><p>such that </p><p>N</p><p>n</p><p>TS </p><p>M, </p><p>(P, </p><p>cv) </p><p>system </p><p>a</p><p>is </p><p>subbundle of TS </p><p>(6). </p><p>S</p><p>be </p><p>submanifold </p><p>cannot use Frobeniustheoremto define the reduced </p><p>a</p><p>constraint manifold of </p><p>If </p><p>S</p><p>is not an </p><p>Let </p><p>M, </p><p>secondary </p><p>of </p><p>(P, </p><p>be involutive. </p><p>open </p><p>one </p><p>case </p><p>M</p><p>then </p><p>n</p><p>TS need not </p><p>N</p><p>In this </p><p>If </p><p>N</p><p>n</p><p>TS </p><p>phase space. </p><p>relation, </p><p>equivalence </p><p>maximal </p><p>manifolds <sup style="top: 0.01em;">define </sup>an </p><p>is involutive its </p><p>and let </p><p>integral </p><p>set of </p><p>denote the </p><p>S</p><p>relation. </p><p>that the </p><p>that there </p><p>this </p><p>by </p><p>such </p><p>N</p><p>quotient </p><p>differentiable structure on </p><p>Ps </p><p>Suppose </p><p>a</p><p>exists </p><p>tion </p><p>canonical </p><p>Ps </p><p>projec- </p><p>S - </p><p>is </p><p>then </p><p>a</p><p>submersion. </p><p>Since </p><p>of<sub style="top: 0.16em;">OJ </sub>M </p><p>is the characteristic set </p><p>TS is contained in the characteristic set </p><p>ps : </p><p>Ps </p><p>n</p><p>and </p><p>N</p><p>TS c </p><p>TM, </p><p>Ns </p><p></p><ul style="display: flex;"><li style="flex:1">of </li><li style="flex:1">defined </li></ul><p></p><p>there exists </p><p>non- </p><p>Therefore </p><p>The form </p><p>the reduced </p><p>OJ </p><p>unique </p><p>S</p><p>by </p><p>03C9rS </p><p>only </p><p>constraint </p><p>and </p><p>a</p><p>2-form </p><p>on </p><p></p><ul style="display: flex;"><li style="flex:1">such that </li><li style="flex:1">S</li></ul><p></p><p>is </p><p>Ps </p><p>N</p><p>03C1*S03C9rS. </p><p>03C9rS </p><p>if </p><p>and </p><p>if </p><p>n</p><p>TS </p><p>manifold </p><p>Thus, </p><p>a</p><p>degenerate </p><p>Ns. </p><p>S</p><p>phase space </p><p>Ps </p><p>ofthe </p><p>has </p><p>natural </p><p>structure </p><p>secondary </p><p>of a symplec- </p><p>if </p><p>if </p><p>TS </p><p>n</p><p>N</p><p>tic manifold </p><p>If this condition </p><p>and </p><p>if </p><p></p><ul style="display: flex;"><li style="flex:1">S</li><li style="flex:1">holds </li></ul><p></p><p>only </p><p>then the structure </p><p>Ns. </p><p>is </p><p>a</p><p>submanifold of </p><p>Sa</p><p>of </p><p>as </p><p>all </p><p>as </p><p>constraint </p><p>canonical </p><p>P, </p><p>OJ) </p><p>secondary </p><p>manifold of </p><p>is </p><p>the </p><p>same </p><p>that of </p><p>cases </p><p>a</p><p>M, </p><p>(P, </p><p>cv). </p><p>exactly </p><p>regular </p><p>This situation </p><p>we </p><p>in </p><p>of </p><p>in </p><p>interest </p><p>S, </p><p>system (P, </p><p>appears </p><p>physics, </p><p>regular </p><p>therefore in the </p><p>shall limit </p><p>our considerations to </p><p>following </p><p>canonical </p><p>systems. </p><p>5. </p><p>A</p><p>CLASS OF </p><p>CANONICAL </p><p>SYSTEM </p><p>on </p><p>a</p><p>first </p><p>M</p><p>be a</p><p>canonical </p><p>A</p><p>function </p><p>P</p><p>is </p><p>called </p><p>Let </p><p>M, </p><p>(P, </p><p>OJ) </p><p>system. </p><p>f</p><p>on </p><p>function constant </p><p>if its Poisson bracket with </p><p>zero on M. This condition </p><p>class function </p><p>every </p><p>can </p><p>be reformulated as follows: </p><p>is </p><p>identically </p><p>This definition is </p><p>a</p><p>of the </p><p>in ref. </p><p>definition </p><p>(~) </p><p>globalization </p><p>[10]. </p><p>given </p><p>Annales de I’Institut Henri </p><p>A</p><p>Poincaré - Section </p><p>369 </p><p>DIRAC BRACKETS IN </p><p>for each v </p><p>GEOMETRIC DYNAMICS </p><p>which </p><p>P</p><p>if </p><p>and </p><p>E</p><p>on </p><p>0. Functions </p><p>first class </p><p>if, </p><p>K, </p><p>v(f) </p><p>only </p><p>f is </p><p>are </p><p>constraint </p><p>not first class are called second class functions. Since the </p><p>a</p><p>M<sub style="top: 0.02em;">can </sub></p><p>P - dim </p><p>be </p><p>described </p><p>of </p><p>system </p><p>submanifold </p><p>0, </p><p>apply </p><p>locally </p><p>by </p><p>equations </p><p>the notion ofa <sup style="top: 0.02em;">class </sup>can be </p><p>extended </p><p>f (p) </p><p>to </p><p>i</p><p>M, </p><p>1, ..., dim </p><p>a</p><p>canonical </p><p>system. regular </p><p>a</p><p>canonical </p><p>is </p><p>the </p><p>5.1. - Class of </p><p>DEFINITION </p><p>of </p><p>M, </p><p>by </p><p>(P, </p><p>regular </p><p>system </p><p>dim K - dim </p><p>N, </p><p>(dim </p><p>N) </p><p><sub style="top: 0.17em;">locally </sub>M </p><p>(’). </p><p>integers </p><p>pair </p><p>If </p><p>a</p><p>is ofclass </p><p>then </p><p>..., </p><p>can </p><p>be described </p><p>system </p><p>where all </p><p>(P, M,cv) </p><p>(n, k) </p><p>1, </p><p>of </p><p>i</p><p>n and </p><p>0, </p><p>1, ..., k, </p><p>equationsf (p) </p><p>0, j </p><p>are first class andall </p><p>are second </p><p>and </p><p>class, </p><p>it </p><p>cannot </p><p>functions £. </p><p>functions gj </p><p>be described </p><p>tions. If k </p><p>of </p><p>with more than n first class func- </p><p>by any system </p><p>equations </p><p>0</p><p>we <sup style="top: 0.01em;">call </sup></p><p>M</p><p>a</p><p>first class submanifold of </p><p>as <sup style="top: 0.01em;">it </sup>can be </p><p>(P, OJ) </p><p>a</p><p>of </p><p>i</p><p>where all </p><p>second </p><p>class </p><p>0, </p><p>1, ..., n, </p><p>locally by </p><p>system </p><p>equations </p><p>f (p) </p><p>given </p><p>if n </p><p>M</p><p>is called </p><p>a</p><p>a</p><p>are first class. </p><p>0, </p><p>Similarly, </p><p>functions f </p><p>of </p><p>In </p><p><sup style="top: 0.01em;">this </sup>case </p><p>is </p><p>submanifold </p><p>manifold. </p><p>(P, </p><p>(M, </p><p>OJM) </p><p>symplectic </p><p>two theorems which we </p><p>hold </p><p>shall </p><p>later. </p><p>need </p><p>There </p><p>For </p><p>PROPOSITION 5.2. </p><p>canonical </p><p>M, </p><p>any regular </p><p>(P, </p><p>co), </p><p>system </p><p>N</p><p>is an even number. </p><p>dim K - dim </p><p>M -dim N. </p><p>even </p><p>P~ </p><p>dim </p><p>K</p><p>dim P -dim </p><p>M</p><p>and dim </p><p>dim </p><p>forms </p><p>dim </p><p>Proof. </p><p>dim P~ are </p><p>and </p><p>P</p><p>Since both </p><p>P</p><p>and P~ admit </p><p>symplectic </p><p>is even. </p><p>P - dim P~ </p><p>numbers. </p><p>K - dim </p><p>N</p><p>dim </p><p>dim </p><p>Hence, </p><p>a</p><p>5 . 3. </p><p>Let </p><p>be </p><p>canonical </p><p>system </p><p>j’ and g </p><p>of </p><p>PROPOSITION </p><p>M, </p><p>regular </p><p>(P, </p><p>class </p><p>with </p><p>k</p><p>&#x3E; </p><p>0. </p><p>there exist two functions </p><p>such </p><p>that </p><p>Then, </p><p>(n, lc) </p><p>For each </p><p><sub style="top: 0.01em;">E </sub>M </p><p>there exists </p><p>on </p><p>a</p><p>of in </p><p>P</p><p>and </p><p>0</p><p>M’ </p><p>Proof - </p><p>p</p><p>neighbourhood </p><p>p</p><p>Up </p><p>two functions </p><p>and </p><p>such that </p><p>n</p><p><sub style="top: 0.25em;">gp</sub>Mn </p><p>M</p><p>fp </p><p>fp </p><p>of </p><p>Up </p><p>gp </p><p>Up </p><p>Up </p><p>and </p><p>n</p><p>we </p><p>1. </p><p>associate to the </p><p>Further, </p><p>M</p><p>complement </p><p>( fp, </p><p>gp) </p><p>Up </p><p>of the </p><p>M</p><p>in </p><p>P</p><p>and </p><p>to </p><p>closure </p><p>functions<sub style="top: 0.16em;">f ’ </sub></p><p>covering </p><p>zero. This </p><p>there </p><p>g’ equal identically </p><p>we </p><p>and since </p><p>of </p><p>P</p><p>have obtained </p><p>a</p><p>of </p><p>is </p><p>paracompact, </p><p>P, </p><p>way </p><p>a</p><p>this </p><p>exists </p><p>finite </p><p>and </p><p>For each </p><p>a</p><p>of </p><p>locally </p><p>covering </p><p>}. </p><p>partition </p><p>refinement { </p><p>to the </p><p>such </p><p>we have two </p><p>subordinated </p><p>on </p><p>Ua </p><p>Ua </p><p>unity { </p><p>functions </p><p>covering { </p><p>that </p><p>and g03B1 </p><p>U (1 </p><p><sup style="top: -0.17em;">Let</sup>f </p><p>g</p><p>follows, </p><p>and </p><p>be </p><p>P</p><p>defined </p><p>as </p><p><sup style="top: 0.01em;">functions </sup>on </p><p>for each </p><p>Then, </p><p>is </p><p>to W. M. </p><p>due </p><p>This definition </p><p>(7) </p><p>Tulczyjew </p><p>(unpublished). </p><p>Vol. </p><p>n° 4 - 1974. </p><p>XX, </p><p>JEDRZEJ SNIATYCKI </p>
Details
-
File Typepdf
-
Upload Time-
-
Content LanguagesEnglish
-
Upload UserAnonymous/Not logged-in
-
File Pages9 Page
-
File Size-