Decision Making Under Multi Task Based on Priority for Each Task

3.1. Outline of Proposal Technique

In this paper, authors propose the action learning and the action selection method of a robot that has multi task. Proposal method is roughly divided into the action selection part and the action learning part. The system does action learning for each task in the action learning part. In the action learning part, the system does action learning and accumulate action value into learning space. Thereafter, action value that learned is passed to action selection part.