NEURAL NETWORK-BASED DYNAMIC PROCESS PLANNING AND ITS INTEGRATION WITH SHOP FLOOR SCHEDULING

摘要

Integration between process planning and production scheduling is regarded as an essential problem in implementing intelligent and flexible automated manufacturing systems. Conventional off-line process planning prevents a shop floor controller from coping with dynamic shop status. The objective of the paper is to propose an integrated process planning and production scheduling approach that can adaptively generate the needed production schedule in real-time based on shop floor status. The process plan generated off-line should be modifiable in an adaptive manner based on shop floor status. To this end, process planning builds both static plan information such as process operations and tolerances, and dynamic planning models as neural network forms for machine selection, process parameter determination. The dynamic planning models are later executed to generate the required production schedule in real-time in the shop floor. The proposed approach will enhance the shop floor controller's capability by producing efficient production schedules under dynamic shop floor environments.