Dynamics on Leibniz manifolds (Q556187): Difference between revisions

With help of the Poisson tensor \(P\) on a manifold \(M\), one associates a Hamiltonian vector field (dynamics) \(X_f=i_{df}P\) with every function \(f\) on the Poisson manifold \((M,P)\). Then, the map \(f\mapsto X_f\) is a Lie algebra homomorphism if one takes the Poisson bracket \(\{ f,g\}_P=\langle P,df\wedge dg\rangle\) on the associative algebra of functions \(C^\infty(M)\) on one hand, and the bracket of vector fields on the second. The Poisson bracket has the obvious Leibniz property: \(\{ f,gh\}_P=\{ f,g\}_Ph+g\{ f,h\}_P\), which just tells us that \(X_f=\{ f,\cdot\}_P\) is a vector field. If we do not assume that \(P\) is Poisson, or even that it is a bi-vector field, so that \(P\) is just a 2-contravariant tensor field on \(M\), then the corresponding bracket of functions \(\{ f,g\}_P\) is no longer Lie nor even skew-symmetric. What survives is only the Leibniz property which, however, should be understood for both arguments separately. This explains the name -- the Leibniz structure -- for such a tensor. The paper is devoted to dynamics associated with Leibniz structures. The gradient flows, some control and dissipative systems, and certain nonholonomic systems turn out to be examples of the Hamiltonian vector fields for Leibniz structures. Symmetries of such systems and the associated reduction procedures are described as well.