MCDDO: Overcoming Challenges and Enhancing Performance in Search Optimization

Abstract

Abstract The child drawing development optimization (CDDO) algorithm, which falls under the category of Human-based algorithms, is a recent example of a metaheuristic approach. This metaheuristic algorithm draws inspiration from the learning behavior of children in terms of drawing and cognitive development as they progress through different stages based on their age. Unlike many other search optimization algorithms, the CDDO algorithm is relatively simple to implement, requires minimal parameter tuning, and outperforms several existing search optimization algorithms. However, despite these advantages, the CDDO algorithm may encounter challenges such as getting trapped in local optima, exhibiting poor performance during the exploration phase, and experiencing stagnation of the local best solution. To overcome these issues, we propose a modified version of the CDDO algorithm (MCDDO). The MCDDO incorporates four key mechanisms: (1) iterative pattern memory updating during the exploitation phase, where new experiences are compared with the child's current drawings; (2) a change in the primary rule employed during the exploitation phase; (3) parameter tuning to strike a balance between exploration and exploitation phases; and (4) preservation of the best solution obtained in each iteration and comparing new solutions with the best solution during the exploration phase. If a new solution is found to be superior, the child's drawings are updated; otherwise, they remain unchanged. This modification introduces entirely different algorithmic mechanisms to update the conditions during the exploitation phase, resulting in improved performance, and leading to the creation of the MCDDO. The performance of the MCDDO algorithm is evaluated through experiments conducted on two standard benchmark functions: 19 classical test functions and 10 CEC-C06 2019 functions. Additionally, an evaluation is made between the MCDDO algorithm and six others widely used algorithms. Statistical analysis using the Wilcoxon rank-sum test confirms that the MCDDO outperforms the alternative algorithms.