Aiming at the problems of high input dimensionality and long training time in traditional game intelligent algorithm models,we proposed a novel deep reinforcement learning game intelligent guidance algorithm that integrated state information transformation and reward function shaping techniques.Firstly,using the interface provided by the Unity engine to directly read game backend information effectively compressed the dimensionality of the state space and reduced the amount of input data.Secondly,by finely designing the reward mechanism,the convergence process of the model was accelerated.Finally,we conducted comparative experiments between the proposed algorithm model and existing methods from both subjective qualitative and objective quantitative perspectives.The experimental results show that this algorithm not only significantly improves the training efficiency of the model,but also markedly enhances the performance of the agent.
关键词
深度强化学习/游戏智能体/奖励函数塑形/近端策略优化算法
Key words
deep reinforcement learning/game agent/reward function shaping/proximal policy optimization algorithm
维普
万方数据