Analysis and Comparison of Monte Carlo Tree Search versus ACO Algorithms in Distribution of Resources and Environment

Abstract

Swarm intelligence (CI) has been widely studied in the past decades. The most famous CI algorithm is ant colony algorithm (ACO), which is used to solve complex path search problems. AlphaZero program, through self-reinforcement learning of Monte Carlo tree algorithm from scratch, has achieved transcendental results in go, chess and general chess. By analyzing and comparing the Monte Carlo Tree Search (MCTS) and ACO algorithms on Distribution of resources and environment, the reasons for AlphaZero's success are revealed. It is not only because of deep neural networks and intensive learning, but also because the algorithm is essentially an evolutionary algorithm of CI emergence.