Negotiation of coupled constraints in coordinated vehicles

We derive and establish properties for an algorithm for the distributed optimization of a global constrained optimization problem through the limited communication of data between cooperating subsystems; each subsystem solves a local optimization problem. The structure of the optimization problem is such that it is performed by the distinct subsystems, which have their own objective functions and local constraints, but share a common coupled constraint which limits the behavior of the independent systems. Our interest is in performing a distributed computation of the global solution without the participating subsystems needing explicit knowledge of other members´ dynamics, constraints, cost functions, or state. We propose a set of distributed algorithms which solve the global problem given limited communication and identify cases in which the subsystems must communicate more information than is desired.