Pareto control in multi-objective dynamic scheduling of a stepper machine in semiconductor wafer fabrication

Abstract

This paper focuses on Pareto control in multi-objective dynamic scheduling of a stepper machine that is considered as a bottleneck machine in the semiconductor wafer fabrication process. We propose the use of compromise programming method for achieving Pareto control in the needs of conflicting objectives such as mean cycle time, cycle time variance and maximum tardiness. Using conjunctive simulated scheduling, at each decision instance in simulated time, a Pareto job is selected and loaded on the machine for processing. Using the real factory data, we demonstrate the concept of Pareto control in dynamic scheduling and show how a stepper machine can be controlled at specified needs of scheduling objectives. The results obtained from Pareto control approach are superior to the simulated results of actual operating heuristic in the factory.