Some Conditions for Optimal Deterministic Solutions to Stochastic Dynamic Linear Programs

Many problems that require decisions made over time can be formulated as dynamic linear programs. Complications arise in solving these programs when one allows stochastic elements to alter the state to state transitions. Finding the stochastic linear programming solutions may be very difficult since their formulation often greatly increases the problem size. This paper shows that, under certain conditions, a simple deterministic solution technique obtains the same optimal controls as more complicated stochastic methods