FUZZY-SYSTEMS FOR CONTROL OF FLEXIBLE MACHINES OPERATING UNDER INFORMATION DELAYS

The performance of a manufacturing system with a defined level of flex ibility is determined by the effectiveness of the control strategy emp loyed. The success of the latter is critically dependent upon informat ion intensive activities including information collection, transfer an d processing. Each of these activities consumes time and thus causes d elays. We refer to these delays as information delays. Most real-world manufacturing systems operate under conditions that entail significan t information delays. Thus, there is a need to model, analyse and evol ve control strategies that can perform well under such delays. This pa per focuses on the design of a suitable control strategy for a simple system operating in a stochastic environment with information delays. System stochasticity coupled with information delay has a compounding effect on the uncertainty of the environment within which decisions mu st be taken thus providing motivation to explore the development of co ntrol strategies based on fuzzy logic. We introduce a novel fuzzy asso ciative memory based control strategy (FCS) to cope with information d elays. Fuzzy associative memories embody a bank of fuzzy rules that re flect expert knowledge in linguistic form. In demonstrating the use of FCS for the one machine, two queue dynamic sequencing problem wherein information delays manifest in the form of machine setup times, this paper identifies suitable input and output control variables and sugge sts their appropriate fuzzification. We define the relative opportunit y gain and the relative work-in-process as two fuzzy control variables . The output fuzzy variable is the switching confidence level. A compa rison of FCS with the alternating priority heuristic is presented usin g average job flowtime as a performance measure. Simulation results sh ow the efficacy and potential of using fuzzy control in situations whe re information delays are significant.