A Review of Imperfect Information Games:Adversarial Solving Methods and Comparative Analysis
Artificial Intelligence(AI)has emerged as a pivotal force in the latest industrial revo-lution and has become a national strategic priority.The fusion of AI and game theory has given rise to"Game Intelligence"as a leading research domain.Among the diverse facets of game intel-ligence,Imperfect-Information Games(IIGs)stand out for their ability to simulate the strategic decision-making of multiple agents amidst private information an accurate portrayal of many real-world scenarios.Compared to perfect-information games,IIGs offer a more nuanced understand-ing of decision-making processes,making them applicable across various real-world domains such as financial trading,business negotiations,and military operations.Recent strides in 1IG research have led to the emergence of two primary streams of offline solving methods:Regret Minimiza-tion and Best Response.Regret Minimization continually refines its strategy towards equilibrium by learning from past decisions,making it particularly advantageous in scenarios with unknown or uncertain opponent strategies.On the other hand,Best Response fine-tunes its strategy to-wards equilibrium by devising tailored countermeasures against opponents'decisions,proving pivotal in training AI for large-scale real-time strategy games like Starcraft and DOT A.The effi-cacy of the Best Response approach hinges on its ability to anticipate and counteract opponents'moves.Moreover,search-based online solving methods optimize blueprint strategies in real-time,facilitating precise Nash equilibrium solutions,constituting a critical technology in IIG sol-ving.The synergy of offline and online solving methods equips AI with the capability to navigate the intricacies of IIGs and attain optimal solutions.This survey aims to provide a comprehensive exploration of the realm of IIGs.Beginning with an elucidation of IIGs'concept and their distin-guishing features,the survey offers an overview of the methods employed for their resolution.Subsequently,it delves into the fundamental principles and historical context of these methods,alongside delineating advanced techniques to enhance their efficacy.Additionally,the survey con-ducts an exhaustive comparison of the strengths and weaknesses of various methods,while provi-ding insights into future research trajectories.It is our aspiration that through this comprehensive scrutiny of IIGs,this survey will drive advancements in game intelligence technology and contrib-ute to the development of artificial intelligence.
imperfect information gameregret minimizationbest responsesafe searchrein-forcement learning
万方数据
维普
