Distributed optimization by Newton consensus over undirected graphs
Martin Guay · IFAC-PapersOnLine · 2021
This manuscript proposes a distributed optimization technique based on a Newton consensus approach. The approach implements a Newton step for both the primal and dual problems that can be implemented in a completely decentralized fashion. Unlike existing techniques, no exchange of derivative information between agents is required. In addition, no explicit inversion of the Hessian information is required to generate the required Newton step. The proposed Newton consensus is shown to reduce the dependency of the transient performance on the graph structure. A simulation study demonstrates the effectiveness of the technique to achieve uniform performance.