To avoid premature and sensitivity of operator parameters selecting of Standard genetic algorithm (SGA) and simulated annealing genetic algorithm (SAGA), a parallel hybrid genetic algorithm based on multi-group (hybrid GA) is presented. The algorithm combines the ideas of parallel computation, simulated annealing and genetic algorithm, and uses orthogonal test table selecting operator parameters to improve the efficiency and robust of the algorithm. And benchmark example of job shop scheduling problem (JSP) is used to validate the effectiveness of the algorithm. Results show the hybrid genetic algorithm converges quickly with small impact to operator parameters.
An enhanced genetic algorithm with simulated annealing for job-shop scheduling