A rule-based, forward-chaining expert system was developed to schedule a multi-pass glassing and furnacing operation for glass-lined vessels. Due to the nature of the manufacturing process and unexpected disturbances, such as rework, fabrication time of each part cannot be tightly controlled. This results in the need to schedule the glassing and furnacing operation manually. However, because of the large number of parts, allowed part mix, and product constraints that have to be considered and evaluated to generate an acceptable schedule, manual scheduling is very tedious and time-consuming.
The Furnace Scheduling Advisory System (FSAS) captures the expertise of the shop floor supervisor and uses heuristics to account for all resource constraints such as availability and capacity of the furnaces and availability of firing tools, and operational constraints including allowed part mix, firing temperature, and allowed thickness difference. The dynamic schedule generated by the expert system is in a heuristic sense the best solution to meet delivery date requirements, optimize utilization of multiple furnaces and minimize energy consumption.
The expert system is written in OPS5 and implemented on the MicroVAX II minicomputer. Comparison of the actual manual furnacing schedule with the furnacing schedule generated by the expert system using a month's worth of data reveals that significant improvement in furnace utilization can be achieved with the furnace scheduling advisory system. In addition, we expect significant improvement in on-time delivery, energy savings, and manpower savings. The heuristic optimization and scheduling strategies used in the furnace scheduling expert system can be extended to scheduling of any batch manufacturing production process.