Classical BRST cohomology and invariant functions on constraint manifolds. I (Q1189182)

The problem in this very interesting paper is to provide a natural framework for a unified description of BRST cohomology both for Lagrangian and Hamiltonian systems. The mentioned cohomology is constructed via the sophisticated techniques of differential complexes and spectral sequences in algebraic topology. The construction is independent of any additional structure on the underlying manifold, like symplectic one etc. The motivation is that at least some classes of models can be considered as finite-dimensional (mechanical) analogues of gauge theories of Yang-Mills type. For instance, one expects that one must enlarge the configuration space by introducing a Lagrange multiplier which replaces the Lie-algebra-valued gauge potential. More precisely, one is dealing with a manifold \(M\) and a submanifold \(C\) defined by some set of constraints \(\phi=(\phi_ i)\), i.e. \(C=\phi^{- 1}(0)\), where \(\phi: M\to V^*\), \(V^*\) being the dual of a finite- dimensional vector space \(V\). Requiring \(C\) to be a submanifold of \(M\) we obtain that for every point \(m\in M\) the inclusion \(T_ mC\subset \text{ker }T_ m\phi\) holds. The authors introduce the notion of weakly regularity of the origin \(O\) of \(V^*\) as a value of \(\phi\). This is the assumption that \(C\) is a submanifold of \(M\) and that the equality \(T_ mC=\text{ker }T_ m\phi\) holds for every point \(m\in C\). The stronger condition of the regularity of \(O\in V^*\) corresponds to the requirement that \(T_ m\phi: T_ mC\to V^*\) is surjective in every point \(m\in C\). The set of constraints is called irreducible if the origin is a regular value of \(\phi\), and respectively, reducible if the origin is a weakly regular but not regular value of \(\phi\). The setting of the authors' treatment is based on group-theoretical methods: 1. One supposes \(M\) to carry an action of a Lie group \(G\), and 2. \(V^*\) to carry a representation of \(G\) such that the map \(\phi\) becomes \(G\)- covariant. The main result is that the constructed BRST cohomology (in fact, the cohomology of the constructed BRST operator in degree 0) reproduces the \(G\)-invariant functions on \(C\). In this paper the authors deal with an irreducible set of constraints and they make a remark that a generalization to the case of a reducible set of constraints is possible. Applications of the proposed general construction to Lagrangian and Hamiltonian systems are discussed. Especially in some typical cases it is shown how gauge invariance can be replaced by an invariance with respect to some finite-dimensional Lie group.