Abstract:

For job shop scheduling problem, we proposed a new particle swarm optimization and simulated annealing based algorithm, in which we made use of the information of the problem itself, added a new neighborhood search strategy (multigranularity) in simulated annealing search, and introduced selection optimization and updating parts into the original particle swarm optimization algorithm. All the adjustments make our algorithm more adaptive, improve the ability of selflearning, and reduce the possibility of trapping in the local best. Our algorithm was tested on different scale benchmarks and compared with recently proposed algorithms. The experimental results show that our algorithm is more adaptive and efficient than the other three algorithms.

Key words: job shop scheduling, particle swarm optimization, adaptive, selflearning