Sand cat swarm optimization combining Adam and Monte Carlo tree search is used to solve complex optimization problems

Sand Cat Swarm Optimization (SCSO), as a metaheuristic algorithm with excellent performance, still faces some inherent challenges. To address the issues of decreasing population diversity, unbalanced exploration and exploitation, and the tendency to fall into local optima in the original algorithm, this study introduces a pioneering variant: Adam Novelty Monte Carlo Tree Search and Cat Swarm Optimization (ANMSCSO). First, by using the Adam algorithm and the Novelty search idea, two types of populations with different functions are constructed: the Adam population focuses on exploitation tasks, while the novelty hybrid population is dedicated to exploration tasks. This dual-population mechanism improves the performance of both exploration and exploitation, effectively enhancing the diversity of the population. Secondly, using the idea of the Monte Carlo tree search strategy, the control of the exploration and exploitation phases is managed adaptively, effectively avoiding imbalances caused by parameter settings. By combining the mutation strategy of the DE algorithm and the crossover concept of the BSA algorithm, the performance of the algorithm is improved, and the integration of the dual-population mechanism with the Monte Carlo tree search strategy effectively prevents the algorithm from falling into local optima. To verify its performance, comparisons were made with mainstream algorithms using problems from CEC2017 and CEC2020, demonstrating the superior ability of ANMSCSO to handle difficult optimization problems. Additionally, experiments on Real world problems from CEC2020 further reveal the strong practical application potential of ANMSCSO.