Two new approach incorporating centroid based mutation operators for Differential Evolution

In the present article we propose two novel variants of Differential Evolution (DE) namely Centroid Differential Evolution (CDE) and Differential Evolution with local Search (DELS) for solving unconstrained, single objective, optimization problems. Both the algorithms make use of geometric centroid to enhance the performance of basic DE in terms of average fitness function value as well as in terms of convergence rate. In CDE a new mutation operator based on geometric centroid is proposed while in DELS a local search mechanism is incorporated in the basic DE algorithm. The proposed algorithms are tested on a set of fifteen test problems taken from literature and four real life problems. The numerical results when compared with original DE and Trigonometric Differential Evolution (TDE), another previous modified version of DE, show that the proposed modifications help in establishing a better tradeoff between the convergence rate and efficiency.