An Improved Self-adaptive Control Parameter of Differential Evolution for Global Optimization

Differential evolution (DE), a fast and robust evolutionary algorithm for global optimization, has been widely used in many areas. However, the success of DE for solving different problems mainly depends on properly choosing the control parameter values. On the other hand, DE is good at exploring the search space and locating the region of global minimum, but it is slow at exploiting the solution. In order to alleviate these drawbacks of DE, this paper proposes an improved self-adaptive control parameter of DE, referred to as ISADE, for global numerical optimization. The proposed approach employs the individual fitness information to adapt the parameter settings. Hence, it can exploit the information of the individual and generate the promising offspring efficiently. To verify the viability of the proposed ISADE, 10 high-dimensional benchmark problems are chosen from literature. Experiment results indicate that this approach is efficient and effective. It is proved that this approach performs better than the original DE in terms of the convergence rate and the quality of the final solutions. Moreover, ISADE obtains faster convergence than the original self-adaptive control parameter of DE (SADE).