Study of League Championship Algorithm Efficiency for Global Optimization Problem

The article objective is to study a new League Championship Algorithm (LCA) algorithm efficiency by its comparing with the efficiency of the Particle Swarm optimization (PSO) algorithm. The article presents a brief description of the terms used in the League Championship algorithm, describes the basic rules of the algorithm, on the basis of which the iterative process for solving the global optimization problem is built. Gives a detailed description of the League Championship algorithm, which comprises a flowchart of the algorithm, as well as a formalization of all its main steps. Depicts an exhaustive description of the software developed to implement the League Championship algorithm to solve global optimization problems. Briefly describes the modified particle swarm algorithm. Presents the values of all free parameters of the algorithm and the algorithm modifications, which make it different from the classical version, as well. The main part of the article shows the results of a great deal of computational experiments using two abovementioned algorithms. All the performance criteria, used for assessment of the algorithms efficiency, are given. Computational experiments were performed using the spherical function, as well as the Rosenbrock, Rastrigin, and Ackley functions. The results of the experiments are summarized in Tables, and also illustrated in Figures. Experiments were performed for the vector dimension of the variable parameters that is equal to 2, 4, 8, 16, 32, and 64. An analysis of the results of computational experiments involves a full assessment of the efficiency of the League Championship algorithm, and also provides an answer about expediency for further algorithm development. It is shown that the League Championship algorithm presented in the article has a high development potential and needs further work for its study.