Chaotic-Analysis-Based Optimization of Release Policies for Re-entrant Manufacturing Systems

Re-entrant manufacturing systems have a variety of complicated characteristics, including large-scale, multi-objective, high uncertainty, and unbalanced loading. Due to the ambiguity of the industrial environment and the complexity of processing machinery and process flows, optimizing release policies is a particularly challenging task. In order to examine the global impact of release policies on the performance fluctuation of manufacturing systems, this research proposes a chaotic-analysis-based optimization method for release policies. A practical re-entrant manufacturing system is selected as the research object, a parallel simulation platform is built for generating production data. This research designs an improved reservoir computing with “parameter-aware” to associate the complicated characteristics of manufacturing systems and the parameters of release policies. Experiment demonstrates the feasibility of the proposed method in optimizing release policies. The proposed method can effectively monitor the actual production status and perform analysis on chaotic behavior, and then choose the optimal parameters to maximize the production capacity of manufacturing systems.