Department of Industrial Engineering Chung-Yuan Christian University 22, Pu-Jen, Chung Li, Taiwan, 320, R.O.C.
Department of Electrical Engineering Chung-Yuan Christian University
The goal of this paper is to explore an effective approach for identifying chaotic types of demands and to develop a production control method for the corresponding chaotic demands. Chaos phenomena is a set of unpredictable behaviors of nonlinear deterministic models. Being distinct from probabilistic types of demands, chaotic demands have a significant impact on production planning due to its butterfly effect. The experiments conducted in the research show that the maximal Lyapunov exponent is very effective in classifying chaos and non-chaos demands. We develop a modified Wagner-Whitin algorithm to facilitate production planning for chaotic demands. The proposed approach is capable of making economical production plans in terms of cost for unpredictable, chaotic demands. The algorithm is tested under a variety of scenarios, such as chaotic demand types, initial conditions, set-up costs, holding costs, and entropy types. The conducted experiment indicates that this algorithm is robust to all of the observed situations, and the minimal total production cost of the proposed method approaches to the optimal one which is produced by the naive Wagner-Whitin algorithm for static demand environments.
Keywords:chaos, production planning, wagner-whitin algorithm
Cite this article as: Kung-Jeng Wang, Hui -Ming Wee, Shin-Feng Gao and Shin-Hsiung Twu, "A Look-Ahead Wagner-Whitin Algorithm for Chaotic Demand," Journal of the Chinese Institute of Industrial Engineers, 18, 1 -12 (2001).