An iterative learning method for multi-cycle flexible production/inventory control under random demands

Abstract

Proposed in this paper is a new economic production lot size (EPLS) model, i.e., a multi-cycle flexible production/inventory model under random demands, which considers the holding and shortage costs of finished products as well as the setup costs of production, and employs an (s,Q) policy for production/inventory control. We have theoretically proved that the total expected cost of the model is a convex function with respect to variables s and Q. On this basis, an iterative learning algorithm is designed to speed up the process by using the model convexity, which enables us to obtain the optimal production preparation point and the optimal production quantity. Results from the comparison between the proposed algorithm and the genetic algorithm (GA) show that the control quantities secured by the two algorithms are consistent but the former has higher solution speed, which verifies the accuracy and effectiveness of the proposed model and algorithm.