Game decision-making method, system and agent in incomplete information environment

Abstract

The invention discloses a game decision-making method and system in an incomplete information environment. The game decision-making method comprises the following steps of reasoning a current unknownstate of an environment according to the current known state of the environment to obtain an unknown reasoning state; exploring according to the unknown reasoning state to obtain an unknown actual state corresponding to the current unknown state, and obtaining a state for decision according to the unknown actual state and a historical known state; and making a decision according to the unknown reasoning state, the state for decision, the current known state and the feedback of the environment at the previous moment. The system comprises an unknown reasoning module, a state exploration module and a decision module. According to the technical scheme, the known states are utilized for prediction reasoning, the reasoned states are explored and integrated, decision is made by comprehensively exploiting the states and reasoning the states, the decision-making capability is high, and the long-term planning is good.

Description

technical field [0001] The invention belongs to the technical field of artificial intelligence, and in particular relates to a game decision-making method, system and intelligent body in an incomplete information environment. Background technique [0002] Behaviors with the nature of competition or confrontation are called game behaviors. In this type of activity, the parties involved in the competition each have different goals or interests. In order to achieve their respective goals and interests, all parties must consider various possible action plans of the opponent, and try to choose the most beneficial or reasonable plan for themselves, such as chess, cards, and games in daily life, and business Negotiations, bidding, and even a battle are all games. [0003] According to the degree of understanding of the participants to other participants, the game is divided into: complete information game and incomplete information game. Complete information game means that in t...

AI Technical Summary

Problems solved by technology

For a complete information game, it is enough to deduce from the current situation downwards to find a better strategy, but for an incomplete information game, it is not enough to only rely on the current situation information. At this time, the current situation information in the incomplete information environment Compared with the current situation information in a complete information environment, the current situation information is incomplete, that is, the current situation information is not enough. If you rely on the plan in a complete information environment when making a decision, the decision you make is very likely not to be beneficial to you.

Images