Intelligent game system based on multi-agent deep reinforcement learning algorithm

Abstract

The invention discloses an intelligent game system based on multi-agent deep reinforcement learning. A decision-making object software model and an intelligent decision-making technology are constructed by using a multi-agent deep reinforcement learning algorithm, software can represent the characteristics of an intelligent game, the game can be carried out, winning and losing can be displayed, and the process of the intelligent game can be displayed.

Description

technical field [0001] The invention relates to an intelligent game system based on a multi-agent deep reinforcement learning algorithm. Background technique [0002] Reinforcement learning is one of the main methods in the field of machine learning and intelligent control in recent years. That is to say, reinforcement learning focuses on how the agent takes a series of actions in the environment, so as to obtain the maximum cumulative return. With reinforcement learning, an agent should know what action to take in what state. Reinforcement learning is the learning of a mapping from an environment state to an action, which we call a policy. [0003] Early reinforcement learning algorithms focused on problems where both states and actions were discrete and finite, and tables could be used to record these probabilities. But in many practical problems, the number of states and actions of some tasks is very large. In order to effectively solve these problems, a complex funct...

AI Technical Summary

Problems solved by technology

[0004] For the intelligent game problem in the space-frequency domain, the scale of the environment is large, including the state space, action space and the number of agents that need to be controlled. It is obviously unreasonable to directly use traditional methods to solve the problem.

Images