Efficient composition of good enough dispatching policies for semiconductor manufacturing

In a fab with heterogeneous machine groups, the number of scheduling policies grows in a combinatorial way because each machine group has its specific dispatching rules. In this paper, we introduce a fast simulation methodology, OODOE, which is an innovative combination of ordinal optimization (OO) and design of experiments (DOE), to efficiently select a good enough policy for fab dispatching with a largely reduced number of policies to be evaluated by simulation. Our analysis by simulation show that the coupling effect among the dispatching policies among tool groups (factors) are negligible for frequently used performance measures. This leads to the effectiveness and computation efficiency of OODOE. Application of OODOE to composing short-term dispatching policies is also demonstrated.