Active network management using distributed constraint optimisation

Abstract

A fully distributed intelligence and control philosophy is needed for future flexible grids to facilitate the low carbon transition and the adoption of emerging network technologies. Future grids need scalable network management solutions in order to cope with the increase in uncertainty and complexity. Fundamental research in intelligent systems and network control will deliver the next generation of intelligent electricity network. This paper presents a network management function formalised as a Distributed Constraint Optimization (DCOP) problem, in particular power flow management. DCOP is an approach to negotiation and arbitration within decentralised control systems where conflicting control decisions arise. Furthermore, the problem will be visualized and decomposed as a factor graph which is a graphical presentation of factorization of a global function into a product of local functions. Additionally, a message passing algorithm, the max-sum algorithm, will be applied which can provide almost optimal results for decentralised coordination problems and limits the computation and communication problems.