An Evolutionary Simulation-Optimization Approach to Product-Driven Manufacturing Control

Abstract

Abstract. The presently reported research proposes an adaptive manufacturing scheduling and control framework that exploits the challenging combination of the main capabilities of product-driven control paradigm and online simulationoptimization approaches. Mainly, the proposed approach employs a scheduling rule-based evolutionary simulation-optimization strategy to dynamically select the most appropriate local decision policies to be used by the agentified manufacturing system components. In addition, this approach addresses products and machines agents' local decisional efficiency issues by dynamically adapting their behaviour to the fluctuations of the manufacturing system state. The main motivation of the developed hybrid intelligent system framework is the realization of an effective and efficient distributed dynamic scheduling and control strategy, that enhances manufacturing system reactivity, flexibility and faulttolerance, as well as maintaining global behavioural stability and optimality. In order to assess the significance of the proposed approach, a proof of proposal prototype implementation is presented and a series of numerical benchmarking experiments are discussed.

Intelligent Products through SOHOMA Prism

“…A state of the art was produced in this respect [17] and several works of the SOHOMA community address this issue. [18] propose a PDS employing a scheduling rule-based evolutionary simulation-optimization strategy to dynamically select the most appropriate local decision policies to be used by the products, when disturbances appear. The originality is to choose decision policies (i.e dispatch rules) instead of fixed schedules, allowing more flexibility on the product level.…”