Existence of an Optimal Control for Stochastic Systems with Nonlinear Cost Functional

TL;DR

This paper proves the existence of an optimal control for a class of stochastic systems with nonlinear cost functionals, using approximation and convexity assumptions to establish convergence and optimality.

Contribution

It introduces a method to establish the existence of optimal controls in stochastic systems with nonlinear costs under convexity conditions, via approximation and convergence techniques.

Findings

Existence of an optimal control under convexity assumptions.

Convergence of approximate controls to an optimal control.

Construction of an admissible optimal control process.

Abstract

We consider a stochastic control problem which is composed of a controlled stochastic differential equation, and whose associated cost functional is defined through a controlled backward stochastic differential equation. Under appropriate convexity assumptions on the coefficients of the forward and the backward equations we prove the existence of an optimal control on a suitable reference stochastic system. The proof is based on an approximation of the stochastic control problem by a sequence of control problems with smooth coefficients, admitting an optimal feedback control. The quadruplet formed by this optimal feedback control and the associated solution of the forward and the backward equations is shown to converge in law, at least along a subsequence. The convexity assumptions on the coefficients then allow to construct from this limit an admissible control process which, on an appropriate reference stochastic system, is optimal for our stochastic control problem.