抄録
This paper describes an efficient solution to a multi-agent path planning problem where a relatively small number of agents track many persons for surveillance purposes. Since search space for path planning is enormous, we developed a meta-heuristics-based planner where the optimization process changes according to the Kullback-Leibler Divergence between the current and the previous prediction of the state transition. We conducted simulated experiments and shows our planning method is considered to be effective.
