Chaotic mutation strategy based single candidate optimizer algorithm

Abstract

Heuristic search algorithms are widely used methods for solving complex optimization problems. These algorithms, usually based on swarm or population-based approaches, try to approach the best solution by moving multiple candidate solutions through the search space. However, these approaches have drawbacks such as requiring a large number of parameters, high computational cost, early convergence or loss of diversity. In this paper, we consider the Single Candidate Optimizer (SCO) algorithm, a new heuristic search strategy that, unlike swarm-based methodologies, uses only a single candidate solution throughout the optimization. SCO uses a two-stage approach to update the position of the candidate solution. In the first stage, the candidate searches different regions of the search space based on its own knowledge. In the second stage, the candidate searches for the best solution in its region using local search methods. In this way, the SCO algorithm balances both exploration and exploitation capabilities. The advantages of the SCO algorithm are simplicity, small number of parameters, low computational cost and high performance. However, the SCO algorithm also has some potential disadvantages. These are the limited exploration capability of a single candidate solution, the risk of getting caught in local optima, and the possibility of getting stuck in suboptimal regions. SCO's exploration mechanism can quickly lead the candidate solution to zero, but this can slow down SCO's convergence to the solution for problems with non-zero solutions. To overcome these drawbacks, various methods can be proposed to improve and refine the SCO algorithm. In this study, a new mutation technique based on chaotic functions such as Chaucy, Gaussian and Levy is used to improve the performance of the SCO algorithm. This mutation operator enhances the exploration capability by allowing the candidate solution to jump to different regions of the search space while updating its position. The proposed Chaotic mutation strategy based Single Candidate Optimizer (CSCO) algorithm is compared with the original SCO algorithm for 23 different benchmark functions taken from the literature. The results show that different mutation techniques significantly improve the performance of the SCO algorithm.