An Improved Genetic Algorithm Based on Neighborhood Search for Flexible Job-shop Scheduling problem

To deal with the flexible job-shop scheduling problem (FJSP), an improved genetic algorithm based on neighborhood search is proposed. The algorithm adds the design of neighborhood search compared with the traditional GA, which makes the general individuals in the population approach the neighborhood which tending to the excellent individuals, and accelerates the local search ability of the algorithm. Large-scale mutation is also designed in the algorithm to make the population be redistributed in the solution space when falling into local optimum, and find the next local optimum solution, thus find the global optimum solution in multiple local optimum solutions. Finally, a program was developed with the actual data of a workshop to verify the feasibility and effectiveness of the algorithm. The result shows that the algorithm achieves satisfactory results in all indexes mentioned above.