Algorithms for Optimal Control of Stochastic Switching Systems

Optimal c ontrol problems of swit hing type with linear state dynamic s are ubiquitous in appli ations of sto hasti c optimization. For high-dimensional problems of this type, solutions whi h utilize some c onvexity related properties are useful. For su ch problems, we present novel algorithmic solutions whi h require minimal assumptions while demonstrating remarkable computational eff icienc y. Furthermore, we devise pro edures of the primal-dual kind to assess the distan e to optimality of these approximate solutions.