Research on agent optimal decision algorithm improvement and network design based on hierarchical reinforcement learning

Artificial intelligence technology has been developing and growing since its birth, and has achieved good results in practical applications in many fields. Among them, deep reinforcement learning algorithm, which has both deep learning perception ability and reinforcement learning decision-making ability, has gradually become a key research direction. Electronic games are also favoured by more and more researchers because they can provide a research environment to verify various artificial intelligence algorithms. Traditional script-based game AI technology cannot meet the needs of decision-making with certain flexibility and effectiveness. This paper uses the interstellar hegemony game as the research environment of intelligent command decision-making method. Aiming at the problem of how to make the optimal decision in the real-time strategy game, the related algorithm is designed, and the hierarchical reinforcement learning idea is used to process the state space and the strategy space hierarchically. The experimental results show that the Agent using hierarchical task network has achieved good results in solving high-dimensional decision-making problems.