Hierarchical Varying Receding-Horizon Policy: A New Production Control Policy for Hybrid Stochastic Production Systems

To simplify the complexities of the evaluation of the optimal production control policy for hybrid stochastic production systems, a new method, called hierarchical varying receding-horizon (HVRH) policy, is initiated. By defining one-times-production of the system, the discounted cost stochastic control proposition over the infinite horizon is simplified to a varying finite horizon discounted cost stochastic control proposition, and the associated HJB equation is also gotten according to the setup sequence and stochastic breakdown of the equipment. Based on evaluating the trajectory of surplus under expectation meaning and the idea of just-in-time, the HVRH policy is actualized by varying receding-horizon optimization sliding on one dimension, which can avoid solving partial differential equations directly and large dynamic programming problems, and makes the policy tractable. Two examples illustrate and confirm the method