Decentralized algorithm for online workload partition of multi-agent systems

In this paper, we consider workload partition problems in order to dynamically cover an uncertain region by multiple mobile agents. Distributed algorithm for mobile agents is designed to yield equal workload in each subregion by partitioning the entire region. Due to the uncertainty in the region, the agents do not have the total environment information. Instead, they have to update their trajectories in real time when they move forward in order to guarantee accurate workload partition and minimize the time to complete the dynamic coverage. With the help of input-to-state stability, we give the partitioning error estimation for the given algorithm. Then simulation results verify the feasibility of the distributed algorithm.