A differential evolution algorithm with a superior-inferior mutation scheme

A differential evolution (DE) algorithm with superior-inferior mutation scheme (SIDE) is proposed to solve global optimization problems over continuous space. Firstly, a superior-inferior mutation scheme is introduced based on the superior individual and inferior individual. Secondly, the dynamic adjustment strategy of exploration factor for balancing superior individual and inferior individual is proposed, which indicates how much weight to place on local exploitation and global exploration. They are integrated to DE in different milestones of optimization to balance exploration and exploitation of the search space and can alleviate the premature convergence. In order to verify the performance of SIDE, a set of numerical experiments on 32 benchmark functions are executed for performance comparison with 8 advanced DE variants and 4 non-DE-based algorithms for 30, 50 and 100 variables. The experimental results indicate that SIDE is much better than compared DE algorithms in terms of optimization quality. Furthermore, SIDE has the best adaptability to high dimensional problems.