Large-scale task/target assignment for UAV fleets using a distributed branch and price optimization

In this work we consider the large-scale distributed task/target assignment problem across a fleet of autonomous UAVs. By using delayed column generation approach on the most primitive non-convex supply-demand formulation, a computationally tractable distributed coordination structure (i.e. a market created by the UAV fleet for tasks/targets) is exploited. This particular structure is solved via a fleet-optimal dual simplex ascent in which each UAV updates its respective flight plan costs with a linear update of way-point task values as evaluated by the market. We show synchronized and asynchronous distributed implementations of this approximation algorithm for dynamically changing scenarios with random pop-up targets. The tests performed on an in-house built network mission simulator provides numerical verification of the algorithm on a) bounded polynomial-time computational complexity and b) hard real-time performance for problem sizes on the order of hundred waypoints per UAV.