A distributed control algorithm via saddle point dynamics for optimal resource allocation problem over netwoked systems
Phuong H. Hoang, Chương V. Nguyen, Hong-Kyong Kim, Hyo‐Sung Ahn · 2017
This paper proposes a distributed algorithm to tackle the widely-studied problem: Optimal resource allocation over interconnected systems. We construct a dynamical system, which is inspired from a consensus protocol and a saddle point dynamics method, to seek the optimal solutions for the problem. The provably correct strategy considers both practical equality and inequality constraints for the resource allocation problem and guarantees the private information of agents during interaction in the networked system. Moreover, our designed algorithm is beneficial in terms of communication cost that each agent in the interconnected system shares only information of one auxiliary variable. A rigorous analysis is provided in this work to guarantee the correctness of the discussed algorithm. We also validate our proposed solution by a simulation.