Symbiotic simulation for optimisation of tool operations in semiconductor manufacturing

A symbiotic simulation-based problem solver agent is proposed that can be used to automatically solve decision making problems regarding the operations of the various tools of an entire semiconductor manufacturing plant (fab). In comparison with common practice decision making, performed by human operators, the advantage of the symbiotic simulation-based approach is its ability to simulate how decisions will affect operations in different parts of the fab. Previous work has been concerned with the optimization of a single tool group. Here, we show that our approach can also be applied to control an entire fab which typically involves several hundred tools. Unlike other approaches, ours is not limited to a set of pre-defined decision making policies. Instead, the problem solver agent can directly schedule setup changes for an arbitrary number of tools. Experiments show that higher throughput can be achieved by using our approach as compared to common practice decision making.