Improved differential evolution algorithm for nonlinear programming and engineering design problems

An improved differential evolution algorithm (IDEA) is proposed to solve nonlinear programming and engineering design problems. The proposed IDEA combines the Taguchi method with sliding levels and a differential evolution algorithm (DEA). The DEA has a powerful global exploration capability on macrospace and uses fewer control parameters. The systematic reasoning ability of the orthogonal array with sliding level and response table is used to exploit the better individuals on microspace to be potential offspring. Therefore, the proposed IDEA is well enhanced and balanced on exploration and exploitation. In this study, the sensitivity of evolutionary parameters for the performance of the IDEA is explored, and the IDEA shows its effectiveness and robustness compared with both the DEA and the real-coded genetic algorithm. The engineering design problems usually encounter a large number of design variables, a mix type of both discrete and continuous design variables, and many design constraints. The proposed IDEA is used to solve these engineering design optimization problems, and demonstrates its capability, feasibility, and robustness. From the computational experiments, the introduced IDEA can obtain better results and more prominent performance than the methods presented in the literatures.