In this paper, the need for developing an aggregate production plan within a hierarchical decision process is discussed. A production planning problem formulated as a sequential stochastic optimal control model with chance constraints is introduced. A global optimal solution is only possible in particular cases with a small number of decision variables. As a consequence, near optimal solutions are usually developed and applied to solve sequential stochastic problems. A suboptimal approach, called Open-Loop Feedback Controller (OLFC), is proposed to solve the aggregate stochastic production planning problem. A simple example is used to evaluate the performance of the OLFC solution computationally in relation to the global (closed-loop) and mean (open-loop) optimal solutions. Quasi-adaptability property of the OLFC procedure is analysed in this example.
Suboptimal control; Probabilistic models; Dynamic programming; Production costs; Inventory control