COMPLEXITY REDUCTION DURING INTERRUPTION ANALYSIS IN A FLEXIBLE MANUFACTURING SYSTEM USING KNOWLEDGE-BASED ONLINE SIMULATION

One of the major challenges for a supervisory controller in a flexible manufacturing system (FMS) is to react quickly to a problem and provide decisions in the shortest possible time using the most accurate and reliable information obtained from the shop floor. The response time of the controller depends upon the complexity of the control problem, number of alternative policies considered, and the methodology to arrive at the best decision, among several others. In this paper, the complexity during the on-line analysis of interruptions caused by machine breakdowns and rush orders in an FMS are discussed. These interruptions affect the manufacturing plan and require changes in the loading and scheduling decisions. Even if the number of machines and part types in an FMS is small, the complexity is extremely high. An integrated decision support system using a knowledge-based on-line simulation (KBOLS) architecture is developed to perform the interruption analysis. Using an emulated FMS, several experiments were conducted to examine the impact of the KBOLS architecture in reducing the complexity problem. The results from these experiments, along with response time studies are reported.