More Information
Summary:The invention discloses a battle decision-making method based on progressive evolution, and the method comprises the steps: collecting expert battle data information in a battle game simulation environment, and obtaining situation information; and performing imitation learning on the expert battle data information to obtain an imitation learning agent with the same level as the expert data. According to the battle decision-making method based on progressive evolution, a basic battle rule and a decision-making method are firstly extracted from expert experience through imitation learning to obtain an imitation learning agent with the same level as expert data, and then reinforcement learning is used for performing reinforcement learning training on the imitation learning agent to improve the agent level. And finally, continuously perfecting the decision-making ability by using alliance learning to obtain an optimal decision-making agent, and solving the problems that in the prior art, the agent training and ex
Bibliography:Application Number: CN202310831012