Optimizing the evaluation parameters of Amazon chess with parallel genetic algorithm

• Published in: 2023 35th Chinese Control and Decision Conference (CCDC) pp. 2298 - 2302
• Main Authors: Qiu, Shengran, Ma, Yiyun, Xu, Yutao, Ji, Yucheng, Deng, Hongbo
• Format: Conference Proceeding
• Language: English
• Published: IEEE 20-05-2023
• Subjects: Amazons; Artificial intelligence; computer games; evaluation; Games; genetic algorithm; Optimization; Parallel algorithms; Search engines; Servers; Video games
• ISSN: 1948-9447
• DOI: 10.1109/CCDC58219.2023.10326586

In the process of computer game, in order to improve the accuracy of the situation evaluation of Amazon AI program, an optimization scheme based on parallel genetic algorithm is proposed. This scheme uses a multi-node distributed parallel algorithm to optimize the weight of the parameters of Amazon's chess game evaluation function. The whole optimization system is mainly composed of server and client. Each client uses multithread parallel computing and transmits the result data to the server. Through experiments, the optimization results of the parameters of the situation evaluation function can be obtained in 5 hours. It is proved that the situation assessment is more accurate and the capability of the AI program is significantly improved after the optimization of this scheme.