The Application of Peer to Peer SAP-based Q-Learning in Task Assignment to Multiple Robots

This paper proposes a method of solving the task assignment of high-speed handling system to multiple robots. In order to reduce the state space and improve the efficiency of learning, an improved Q-learning algorithm based on peer to peer SAP is designed. In the proposed algorithm, all the Q value of the peer to peer SAP are updated after take an action, and these Q value are required only one memory unit. Compare to the standard Q-learning algorithm, the results demonstrate the efficiency of the proposed algorithm.