六子棋博弈系统中机器学习算法设计与研究

Machine learning research and design of algorithm Connect6 game system

本文针对我校参加全国大学生博弈大赛完成的六子棋博弈系统,首先简要介绍本系统的一些关键构成.并重点论述了对现有的六子棋系统一个更进一步的优化,在MTD(f)算法上增加自适应宽度调整和在评估函数上增加TD_BP算法自我学习方法 ,使系统"思考"性的能力更加强.

This article is aimed to analyze Shaoguan University′s students who participatd in the national Com-puter Games finished. It briefly analyzed the system structure and some key opponents with a focus on the Connect6 for a further optimization. It also increases adaptive width adjustment on the MTD （f） algorithm and e-valuation function to increase TD_BP on self-learning algorithm which would strengthen system′s ＂thinking＂ability.