Job scheduling is an important and highly complex process in semiconductor manufacturing. This is routinely carried out by senior planners manually through heuristics. The manual approach provides feasible but not optimum solutions. Increasingly, modern fabs are advancing towards the concept of cyber physical system (CPS) with tightly integrated physical environment, wireless sensor network enabled by powerful computation and real-time data acquisition. Manual planners can no longer cope with the increased complexities. There is a need for a simulation based optimization approach to job scheduling.
There is an immediate demand for such technology in job dispatching in the diffusion area of a semiconductor fab. The technology helps to increase throughput of the existing fab by taking into account the preventive maintenance scheduling at the same time.
The generic optimisation engine can also be extended to other capital intensive industries such as aerospace and precision engineering.
Technology Features, Specifications and Advantages
The technology is a seamless integration of IoT & data analytics, process-oriented simulation, AI & metaheuristic, cloud computing & parallel processing. It supports the decision making and provides intelligent guidance to the job processing sequence, tool maintenance schedules and spare parts inventory control in the production environment.
The approach is designed to tackle problems with complex real world dynamics. Specifically, it can be applied in conjunction with a massively parallel computing strategy to solve large-scale job scheduling challenges in wafer fabrication plants. The optimisation engine uses matheuistics approach to reduce the highly complex, multi-objectives and multi-variables schedule planning problem to a mathematical form that is amenable to a parallel solution finding process. The novel approach has been demonstrated on the scheduling of furnaces for either diffusion, photolithography, etching, deposition or a combination of processes in a wafer fab.
The technology has been demonstrated for job processing sequence, tool maintenance schedules and spare parts inventory control in the semiconductor industry. Currently, the research is extending to applications in the smart pharmacy of Hospitals and aerospace companies.The targeted users are the firms which have the demand on reducing manpower through robotic solutions as well as the firms which have considerable investment on robotics, but with limited return on the productivity improvement, such as the semiconductor industry.
The solutions can boost productivity, drive revenue growth, attract investment, reduce labor requirement and create knowledge-based high pay jobs in Singapore.