Differential evolution based global best algorithm: an efficient optimizer for solving constrained and unconstrained optimization problems

This study proposes an optimization method called Global Best Algorithm for successful solution of constrained and unconstrained optimization problems. This propounded method uses the manipulation equations of Differential Evolution, dexterously combines them with some of the perturbation schemes of Differential Search algorithm, and takes advantages of the global best solution obtained on the course of the iterations to benefit the productive and feasible in the search span through which the optimum solution can be easily achieved. A set of 16 optimization benchmark functions is then applied on the proposed algorithm as well as some of the cutting edge optimizers. Comparative study between these methods reveals that GBEST has the ability to achieve more competitive results when compared to other algorithms. Effects of algorithm parameters on optimization accuracy have been benchmarked with some high-dimensional unimodal and multimodal optimization test functions. Five real world design problems accompanied with three challenging test functions have been solved and verified against the literature approaches. Optimal solution obtained for economic dispatch problem also proves the applicability of the proposed method on multidimensional constrained problems with having large solution spaces.