Genetic synthesis of production-control systems for unreliable manufacturing systems with variable demands

The control of manufacturing systems with variable demands has attracted much research attention over the years. However, only limited results have been obtained due to the difficulty of this production-control problem. In this paper, genetically optimized short-run hedging points are used to construct gain-scheduled adaptive controllers for unreliable manufacturing systems with variable demands. The performance of such adaptive controllers is illustrated for unreliable systems subjected to piecewise-constant demands. It is demonstrated that the performance of these adaptive controllers is superior, in general, to that of genetically optimized non-adaptive controllers. However, such gain-scheduled adaptive controllers are designed for variable demands that are piecewise-constant. Therefore, in order to deal with more general classes of variable demands, a genetic rule-induction design methodology is used to synthesize robust fuzzy-logic controllers to provide automatic closed-loop control for unreliable manufacturing systems. Such robust fuzzy-logic controllers are shown to provide effective control for unreliable manufacturing systems with various kinds of variable demands.