Research of Assembly Job Shop Scheduling Problem Based on Modified Genetic Programming

摘要

The study of Job Shop Scheduling Problem (JSP) enables effective control of the production process and improves corporate economic profitability. This research extends traditional JSP to include assembly operations, which is called assembly job shop scheduling (AJSSP). AJSSP is often considered multi-objective decision problems just like JSP. This research adopts the commonly used mean total lateness as the objective for optimization. GP (Genetic Programming) is proposed as the solution approach for AJSSP to obtain composite dispatching rules (CDR). Through experimental computation, the optimized CDR is shown to perform better than traditional simple rules including SPT, JDD and FIFO. Besides, as searching of the optimized rule is based on 85% machine utilization, the performance is further tested under 80% and 90% utilization. Results confirm that the combined rule shows perfect robustness.