Improved Evolutionary Programming Algorithm and Its Application Research on the Optimization of Ordering Plan

Abstract

Abstract Evolutionary programming (EP) with Gauss mutation operator has premature convergence. The main reason is that mutation value produced by Gauss mutation operator is so small that every variable in individual and individual itself may not be mutation. This research improved EP algorithm in three aspects of mutation operator, computation of individual opponent value, and search space. First, Gauss mutation operator is replaced with the improved discrete cosine-transformation operator which can produce a large value of mutation. Application of the formula of dynamic and proportional mutation can adjust every component value dynamically in individual, and the strategy of multiindividual competition enlarges the number of searches greatly within the solution space. Second, a model of complicate ordering business is proposed. Finally, the ordering plan is optimized by using the improved EP algorithm, EP with Gauss mutation operator and random mutation operator in Matlab. The result of simulated experiment shows that precision of the solution using the improved algorithm is demonstrated better than other algorithms. As a result, the improved algorithm has effectively solved the problem of premature convergence.