Cooperative dynamic task-allocation through iterative agents´ costs permutations

We propose a multi-robot task-allocation strategy in which individuals´ task assignments are carried out through the application of iterative permutations of agents´ cost values that are associated with completing given tasks. Different possible permutations of agent-to-task assignments are calculated based on profile matrix that comprises all robotic agents´ costs with regards to available tasks. The permutations that are calculated off-line are next used at runtime to dynamically and as per states of agents and available tasks ascertain the optimal agent-task allocations at every decision cycle. Proof of optimality of the adapted allocation strategy is presented. Performance of the proposed approach in multi-robot dynamic multi-task allocation scenario is demonstrated.