Decentralized task allocation for heterogeneous agent systems with constraints on agent capacity and critical tasks

This paper considers the task allocation problem for a network of heterogeneous agents under the additional assumptions that each agent can execute a limited number of tasks due to its physical limitations and that some tasks considered critical (i.e., mandatory) for the assignment problem must be necessarily assigned. In order to solve this problem in a decentralized way, we propose an algorithm that builds an initial assignment through a market-based approach and resolves conflicts with a consensus procedure based on local communications between neighboring agents. Numerical simulations are performed to evaluate the effectiveness of the proposed approach.