Polynomial-Time Algorithms for Multiagent Minimal-Capacity Planning

In this article, we study the problem of minimizing resource capacity autonomous agents that cooperate to achieve a shared task. More specifically, consider high-level planning for team homogeneous operate under constraints in stochastic environments and share common goal: given set target locations, ensure each location is visited infinitely often by some almost surely. We formalize dynamics so-called consumption Markov decision processes. process, agent has limited capacity. Each action may consume amount resource. To avoid exhaustion, can replenish its full designated reload states. The restricts capabilities agent. objective assign locations agents, only responsible visiting assigned subset repeatedly. Moreover, assignment must carry out their tasks with minimal reduce an equivalent graph-theoretical problem. develop algorithm solves graph time polynomial number size process. demonstrate applicability scalability scenario where hundreds unmanned underwater vehicles monitor ocean currents.