An enhanced colliding bodies optimization and its application

Abstract

Colliding bodies optimization (CBO) is a recently proposed algorithm, and there are no algorithm-specific parameters that should be previously determined in updating equations of bodies. CBO has been used to solve various optimization problems because of its simple structure. However, CBO suffers from low convergence speed and premature convergence. To enhance CBO’s performance, a new variant named learning strategy based colliding bodies optimization (LSCBO), which is based on the learning strategy of the Teaching–learning-based optimization algorithm (TLBO), is proposed in this paper. In this method, a hybrid strategy combining the colliding process of CBO and the learning process of TLBO is proposed to generate new positions of the bodies. Compared with some other CBO variants, the guidance of the best individual is introduced to improve the convergence speed of CBO, and a random mutation method based on the historic information is designed to help bodies escape from local optima. Moreover, a new method for determining the mass of bodies is designed to avoid computation overflow. To evaluate the effectiveness of LSCBO, 47 benchmark functions and three real-world structural design problems are tested in the simulation experiments, and the results are compared with those of other well-known meta-heuristic algorithms. The statistical simulation results indicate that the performance of CBO is obviously improved by the developed method.