On solutions of the distributional Bellman equation
This work provides theoretical foundations for distributional reinforcement learning, addressing a core mathematical problem for researchers in the field, though it appears incremental as it builds on existing distributional Bellman equation concepts.
The paper tackles the problem of existence, uniqueness, and tail properties of solutions to distributional Bellman equations in reinforcement learning, establishing necessary and sufficient conditions and linking them to multivariate affine distributional equations.
In distributional reinforcement learning not only expected returns but the complete return distributions of a policy are taken into account. The return distribution for a fixed policy is given as the solution of an associated distributional Bellman equation. In this note we consider general distributional Bellman equations and study existence and uniqueness of their solutions as well as tail properties of return distributions. We give necessary and sufficient conditions for existence and uniqueness of return distributions and identify cases of regular variation. We link distributional Bellman equations to multivariate affine distributional equations. We show that any solution of a distributional Bellman equation can be obtained as the vector of marginal laws of a solution to a multivariate affine distributional equation. This makes the general theory of such equations applicable to the distributional reinforcement learning setting.