A job-priority based soft scheduling approach for uncertain work area scheduling in Semiconductor Manufacturing

Abstract

This paper studies the uncertain scheduling problem of the oxidation/deposition/diffusion work area in semiconductor manufacturing, where fluctuations in arrival times and processing times are considered. To overcome the shortcomings of the conventional schedule, which has a fixed processing machine and starting time for each operation, we design a new form of schedule called the job-priority based soft schedule. Its basic idea is to assign global priority to each job in advance, and to make remaining decisions (including batch formation, machine assignment and real-time dispatching) in the process of implementing the initial soft schedule. Then we develop a job-priority based soft scheduling approach that involves two layers: (1) in the offline optimisation layer, an initial soft schedule is obtained by a proposed hybrid harmony search algorithm; (2) in the online dispatching layer, the remaining decisions are made in real time by a designed online heuristic rule. The comparative experimental results demonstrate that the two-layer soft scheduling mechanism can effectively adapt to an uncertain production environment due to its integration of the global optimisation perspective and local flexibility.