Application Of Improved Pseudo-Parallel Genetic Algorithms For Job-Shop Scheduling Problem

Present is the era that the competition fierceness and information technology develop at full speed, the market competition is fiercer and fiercer, and the enterprises face more challenges. China, as big country is emerging rapidly in the capacity of manufacturing industry. Emergency pace, lower costs, and the higher production capacity and efficiency become the most important competitive element for enterprise survive and develop. Furthermore, the effective scheduling algorithm has already become foundation and factor of advanced manufacturing technology, so the study on it has very important theory and practical meaning.Job-Shop Scheduling problem (for short as JSP ) is the simple models of many actual Job-Shop Scheduling problem, it is a typical NP-hard question, so its research has important theory meaning and project value, it is also the most extensively studies of typical Job-Shop Scheduling problem. The complexity of this problem determined that the search space enlarged with the scale of JSP broadens, and calculation quantum increase exponentially. This need guarantee variety of individual in the course of issue in solving JSP, Besides implementing the overall optimization of problem, can also improving the odds of finding the fit value, shortening the search time of the fitness.This paper studied a Improved Pseudo-Parallel Genetic Algorithms based on JSP, which mixed with Simple Genetic Algorithms parallel, this can ensure the variety, and avoid early convergence; the Algorithms improved some genetic operator, apply Iterative-Crossover, multi-parent produced multi-children, made the good gene-segment combination. We also introduced the best individual sequences, to avoid missing the best value.Introduce the Chaos theory in the course of making a variation, a small disturbance is added to the more excellent variation operators, and by using the chaos variable, the disturbance amplitude is adjusted little by little as the search goes on, get the suitable fitness of the problems.