An Entropy-based Adaptive PBIL Algorithm and Its Application
Hali Pang · Acta Simulata Systematica Sinica · 2003
This paper analyzes the basic principle and disadvantage of population-based increased learning algorithm, and proposes an improved strategy which has adaptive function and mutation capability. Information entropy is introduced to evaluate population evolutionary degree, and learning rate and mutation probability are adaptively adjusted according to information entropy value in the new algorithm. The algorithm is applied to resolve the typical Flow Shop scheduling, problem and the result shows that compared with standard PBIL algorithm and genetic algorithm (GA), the calculation efficiency and local search capability are much improved, the optimum effect is satisfied, and the convergence process is very stable.