Hybrid single operator HHO, SMA and HGS based metaheuristic algorithms

Abstract

There are many population-based meta-heuristic optimization algorithms, but none can outperform all existing algorithms on all existing optimization problems, or solve all optimization problems. This leads some algorithms to use one or more control parameters to adjust the properties of the algorithm depending on the problem to be solved. The optimization behavior of these algorithms can be improved by, among other aspects, incorporating hybridization techniques. Hybrid algorithms are suitable for a wide range of applications, but they are usually intended for specific engineering problems, and they increase the difficulty of correctly adjusting the control parameters. This paper presents hybrid algorithms based on three operators from prevalent configuration parameter-free optimization algorithms. Each hybrid approach uses a different strategy to change the algorithm responsible for generating each new individual. These algorithms are HHO, SMA and HGS. Experimental results show that the proposed algorithms perform better than the original algorithms, which implies that the optimal use of these basic algorithms depends on the problem to be solved. Another advantage of the hybrid algorithms is that there is no need for a prior process of adjusting the control parameters.