Explore vs. Exploit: Task Allocation for Multi-robot Foraging

Christopher Baldassano and Naomi Ehrich Leonard

This paper describes two measures of performance that can be used to allocate tasks in a multi-robot foraging problem. These heuristics can be used by a human supervisor or an automated control algorithm to adjust the number of robots exploring the environment versus the number of robots greedily harvesting based on current knowledge. A numerical simulation study is presented that offers preliminary support for the usefulness of these heuristics.