Effective and Interpretable Rule Mining for Dynamic Job-Shop Scheduling via Improved Gene Expression Programming with Feature Selection

Gene expression programming (GEP) is frequently used to create intelligent dispatching rules for job-shop scheduling. The proper selection of the terminal set a critical factor success GEP. However, there are various job features and machine that can be included in sets capture different characteristics state. Moreover, importance varies greatly between scenarios. irrelevant redundant may lead high computational requirements increased difficulty interpreting generated rules. Consequently, feature approach evolving with improved GEP has been proposed, so as select dynamic First, adaptive variable neighborhood search algorithm was embedded into obtain diverse good Secondly, based on fitness contribution rules, weighted voting ranking method from set. proposed then compared GEP-based algorithms benchmark conditions scheduling objectives. experimentally obtained results illustrated performance using after process better than both baseline algorithm.