Combining Genetic Algorithms And Or-Tools To Solve Flexible Job-Shop Scheduling Problems

Flexible job shop scheduling problems (FJSPs) require optimally assigning operations to machines and sequencing them to minimize objectives like makespan. Metaheuristics like genetic algorithms (GAs) are well-suited for large FJSPs. This study presents a combination of GAs and Google OR-Tools to solve FJSPs, with GAs generating machine assignments and OR-Tools developing schedules. Specifically, GAs first assign operations to machines, and then OR-Tools schedules the operations to minimize makespan and calculate the solution's characteristics. These characteristics inform the GA's selection, crossover, and mutation to produce improved solutions over iterations. Computational experiments on standard benchmark instances demonstrate solution quality comparable to state-of-the-art tabu search approaches. The feasibility of the algorithm to effectively solve FJSPs is thereby established. Additionally, the study introduces a viable hybrid GA for providing high-quality FJSP solutions.