An improved Differential Evolution Algorithm with Self Adaptive Mutation Strategies for Global Optimization

摘要

To solve real-world global optimization problem differential evolution algorithm is used as one of the best nature influenced algorithm. The use of different effective mutation strategies and proper selection of effective control parameters directly affect the performance and convergent rate of differential evolution method. Although its performance is very good but suffers from population diversity and stagnation.In this paper, new self-adaptive mutation strategies with booster vector to improve global optimization of DE/rand/1/bin and DE/best/1/bin is proposed. Elite archive strategies with dynamic adjustment of control parameter with booster vector added to afford more bandwidth for electing an effective mutant solution. The proposed algorithm is compared with five DE and six non-DE algorithms by using a set of twenty benchmark functions on COCO (comparing Continuous Optimizers) framework. The experimental result verifies that proposed self-adaptive strategies outperformed the competitors.