Multilevel intelligent scheduling and control system for an automated flow shop manufacturing environment

Current manufacturing scheduling and control systems are incapable of coping with complex system dynamics inherent in real-world situations and, hence, human intervention is required to maintain real-time adaptation and optimization. A unique feature of biological intelligent systems is that they build and improve over their communication, decision-making and control structures in real time autonomously. A challenge is now emerging in the design of manufacturing systems where on-line adaptation and optimization become increasingly important. This paper reports on the development of a new integrated intelligent scheduling and control system for an automated manufacturing environment using a multilevel approach. At the first level, a conventional scheduling and control system is considered, then at the second level, a new fuzzy logic mechanism is developed to enable the conventional system to improve and perceive the changes of system parameters adaptively. A new perturbation mechanism is embedded in the third level to implement on-line optimization for coping with the more complex structural changes of system dynamics. The final level is composed of artificial neural networks that can learn from experiences provided by the perturbation mechanism. The approach is designed to improve system intelligence gradually to cope with various forms of systems dynamics. A fully automated flow shop manufacturing system is taken to demonstrate this approach.