Article ID: | iaor2014505 |
Volume: | 8 |
Issue: | 2 |
Start Page Number: | 105 |
End Page Number: | 114 |
Publication Date: | May 2014 |
Journal: | Journal of Simulation |
Authors: | Aydt Heiko, Cai Wentong, John Turner Stephen |
Keywords: | decision support, evolutionary algorithms, semiconductor manufacturing |
Simulation‐based what‐if analysis is an important tool that can be used to support operational decision making. Solutions to operational decision problems have to be found in a limited period of time. As a result of these time constraints, not all candidate solutions can be evaluated by means of simulation. Essentially, this problem solving process is an optimization process. Specialization of the optimization method can help to reduce the computing budget needed to find an appropriate solution. However, since operational conditions are constantly changing, the exact problem is not known at design time. Therefore, the optimization method has to be specialized during runtime. This can be done by dynamically incorporating situation‐dependent problem‐specific knowledge into the optimization process. We introduce a method based on an evolutionary algorithm, which supports dynamic specialization. Furthermore, we demonstrate the effectiveness of this method in the context of a semiconductor manufacturing application.