Exponential Scale-Factor based Differential Evolution Algorithm

摘要

Differential Evolution (DE) Algorithm is a familiar evolutionary and straightforward optimization approach to deal with nonlinear and composite problems. Crossover rate (CR) and scale factor (V) are two control parameters which play a crucial role to retain the proper equilibrium betwixt exploitation and exploration capabilities of DE algorithm. In DE, for a greater value of CR and V, there is an ample possibility to caper the true solution due to huge step length in the search area. So in this paper, we presented another alternative of DE algorithm named as Exponential Scale-Factor based Differential Evolution (ESFDE) for minimizing the step length. In the introduced approach, the scale factor V is exponentially reduced to keep a proper equilibrium betwixt exploitation and exploitation abilities and in this paper, DE/best/2 approach is used. The propounded algorithm is exerted on 15 familiar test problems of various difficulties. A comparison is done using the propounded ESFDE algorithm, DE/best/2, DE/rand/1, DE/rand/2, Particle Swarm Optimization (PSO) and Gbest-guided Differential Evolution Algorithm (Gbest DE).