An improved genetic algorithm for the re-entrant and flexible job-shop scheduling problem

To solve a class of complex job shop scheduling problem with characteristics of re-entrant and flexible, which is a NP-hard problem, an improved genetic algorithm (GA) is proposed in this paper. In order to effectively overcome the contradiction between "convergence rate" and "convergence accuracy" of GA, a comprehensive search mechanism is presented. In this search mechanism, a strategy with excluding close relative and an elitist selection have been applied into the select operator. Meanwhile, a partition crossover operator with competition between parents and children has also been used in GA. In addition, considering diversity of population, random perturbation is added into the evolution process of GA. Finally, the improved algorithm is applied into the optimization scheduling problem of automatic fluorescence analyzer which has characteristics of re-entrant and flexible. And its experimental results compared with other GA indicate the improved GA with comprehensive search mechanism can effectively overcome the contradiction between "convergence rate" and "convergence accuracy".