S. Zhang and R. Jarvis (Eds.): AI 2005, LNAI 3809, pp. 1086 – 1091, 2005. © Springer-Verlag Berlin Heidelberg 2005 Evolutionary Optimisation of Distributed Energy Resources Ying Guo, Jiaming Li, and Geoff James CSIRO Information and Communications Technology Centre, Locked Bag 17, North Ryde, NSW 1670, Australia {firstname.lastname}@csiro.au Abstract. Genetic optimisation is used to minimise operational costs across a system of electrical loads and generators controlled by local intelligent agents and connected to the electricity grid at market rates. Experimental results in a simulated environment show that coordinated market-sensitive behaviours are achieved. A large network of 500 loads and generators, each characterised by different randomly selected parameters, was optimised using a two-stage ge- netic algorithm to achieve scalability. 1 Introduction Many countries including Australia are experiencing a growing gap between electric- ity supply and demand, and distributed electricity generation technologies alongside improved demand-side management techniques have been identified as one set of solutions to this challenge [1]. Significant reductions in greenhouse gases can also be achieved by the large-scale deployment of clean, efficient distributed generation in place of increased investment in centralised generation. We are developing multi-agent technology for the management and control of dis- tributed energy resources [2], aimed at deployment in the Australian National Elec- tricity Market within the next five years, and a component of this work is the devel- opment of intelligent coordination algorithms agents controlling distributed energy resources (DERs). Agent-based coordination is used in a range of fields including computing, manufacturing, and energy [3-5]. The purposes of coordination of DER agents are to achieve local efficiency goals and to aggregate sufficient quantities of distributed capacity to be of strategic value to market participants. Retailers exposed to volatile wholesale prices and network businesses making infrastructure investment decisions would be significant beneficiaries of such aggregates. In this paper we tackle the problem of cost minimisation across a set of loads and generators controlled by local agents and connected to the electricity grid at market rates. In contrast to [3] we wish to investigate non-market algorithms. Our approach permits a cap on the power drawn from the grid that can be a local offering to a large- scale aggregation of distributed capacity. We consider systems of small to moderate scale, comprising up to 500 load and generator agents, and our simulated experimen- tal results show that coordinated market-sensitive behaviours are achieved. The remainder of this paper is organised as follows. Section 2 outlines our ap- proach and Section 3 describes the optimisation process. In Section 4, we present