Investigation of modified heuristic algorithms for simulation-based optimization

For optimization of manufacturing processes in electronics and semiconductor production, simulation-based optimization algorithms become more and more important. In addition to the usual dispatching rules we use iterative local search algorithms for scheduling. In this paper we focus on highly complex manufacturing models with difficult technological operational sequences. In particular, these contain cost extensive machine setups with a limited number of tools. So it is our aim to optimize the manufacturing workflow by minimizing the setup and machine idle times. This was realized by modifying the generally used control variables (job sequences, priorities, etc..) to model-specific product mixture control variables. Furthermore we describe some extensions of known heuristic search algorithms to increase their efficiency. We also developed a Matlab-based visualization tool to better evaluate the effectiveness of the investigated optimization processes.