According to the scheduling problems combining the characteristics of Flexible Job Shop Sch in the multi-variety

small-batch complex production

and eduling Problem and Compute Unified Device Architecture

an island model genetic algorithm for Flexible Job Shop Scheduling Problem was proposed in this paper. First

a storage model and execution model which are well suited for Compute Unified Device Architecture was designed. Then the mode of chromosome encoding and decoding was shown and choosing

crossing

variation

transfer opera- tion of genetic algorithm are all improved. The algorithm raises parallel processing efficiency of genetic algorithm on computing unified device architecture. Comparing the results and efficiency with other algorithms using benchmark data

the experimental result shows that the algorithm is effective and feasible.