Single populated genetic algorithm and its application to jobshop scheduling

The authors present an efficient genetic algorithm (GA) called the single populated genetic algorithm (SPGA). The algorithm uses an individual in a generation. Without using crossover, the solution is improved through mutations only. The algorithm is very fast in terms of convergence and the solution quality is excellent. The SPGA was applied in the traveling salesman problem (TSP) and was verified to be efficient through simulations. An application of the SPGA to the jobshop scheduling problem (JSP) is also studied. A representation method for the JSP is described. The genotype represented by the new method is simple and no illegal schedule is produced. The new genetic algorithm together with the representation method can realize flexible scheduling for job shop