An expanded distributed algorithm for dynamic resource allocation over strongly connected topologies

This paper studies a distributed dynamic resource allocation problem for a multi-agent network with a strongly connected digraph, which has increased an equality constraint for individual node requirement from the point of dynamic prediction horizon. Each agent at each prediction horizon in the network is associated with a local resource and a quadratic convex cost function. The goal is to minimize the total cost in a distributed manner and subject to state constraints and collective equality constraints as well as an extra individual requirement simultaneously. With a fully distributed iteration setup, each node individual requirement is ensured and the globally optimal solution is obtained. The convergence of the proposed algorithm is verified theoretically according to eigenvalue perturbation theory and illustrated by simulation results.