Competitive hybrid Jaya and β-Hill Climbing algorithm for high-cost optimization problems

Jaya algorithm is a population-based parameter-less optimization algorithm. It is frequently used in high-cost industrial and engineering problems. However, Jaya algorithm gets stuck in local optimum in some problems with constrained conditions. In this paper, we introduce a hybrid variant of Jaya to overcome the problem of getting stuck at the local optimum, the β-Hill Climbing local search algorithm is added to the Jaya algorithm and a hybrid Jaya algorithm (CT-JAYA-BH) is presented. Before deciding to use the β-Hill Climbing algorithm, Jaya was also combined with the Quasi Newton and Nelder-Mead local search algorithms. These variants were tested on the CEC 2015 benchmark set provided by IEEE. According to the results, the Jaya-β-Hill Climbing variant (CT-JAYA-BH) obtained the best results. A parameter analysis of CT-JAYA-BH was also performed to determine at which parameter values this variant achieved the best results. Moreover CT-JAYA-BH was compared with 14 different optimization algorithms using the CEC 2015 benchmark set. According to the results, the proposed CT-JAYA-BH algorithm outperforms the other algorithms with an average rank value of 1.87 in both 10 and 30 dimensions. The results show that CT-JAYA-BH is a highly competitive.