基于蚁群算法和轮盘算法的多Agent Q学习

Multiagent Q-learning based on ant colony algorithm and roulette algorithm

提出了一种新颖的基于Q-学习、蚁群算法和轮盘赌算法的多Agent强化学习。在强化学习算法中,当Agent数量增加到足够大时,就会出现动作空间灾难性问题,即:其学习速度骤然下降。另外,Agent是利用Q值来选择下一步动作的,因此,在学习早期,动作的选择严重束缚于高Q值。把蚁群算法、轮盘赌算法和强化学习三者结合起来,期望解决上述提出的问题。最后,对新算法的理论分析和实验结果都证明了改进的Q学习是可行的,并且可以有效地提高学习效率。

Authors present a novel Multiagent Reinforcement Learning Algorithm based on Q-Learning,ant colony algorithm and roulette algorithm.In reinforcement learning algorithm,when the number of agents is large enough,all of the action selection methods will be failed：the speed of learning is decreased sharply.Besides,as the Agent makes use of the Q value to choose the next action,the next action is restrainted seriously by the high Q value,in the prophase.So,authors combine the ant conlony algorithm,roulette algorithm with Q-learning,hope that the problems will be resolved with the algorithm proposed.At last,the theory analysis and experiment result both demonstrate that the improved Q-learning is feasible and increases the learning efficiency.