Application Of Multi-Agent System In Energy Conservation Generation Scheduling System

Behind the rapid economic development in our country, the problem is associatedwith excessive consumption of energy and the deteriorating environment. Facing thedomestic and international pressure, to promote energy conservation has become along-term and sustainable development in China, which makes the electric powerindustry of large energy consumption and pollution face tremendous pressure. Becauseof China’s energy structure dominated by coal, low carbon energy resource selection islimited, so unit of generation scheduling system is optimized to reduce coalconsumption and pollutant emissions will be the core of our current generationscheduling work.Generation scheduling system is a complex and volatile system with highlynonlinear and dynamic. And Agent theory has inherent advantages in dealing withnonlinear and dynamic complex system, which provides a new theoretical basis forsolving complex systems. Therefore, multi-agent theory provides a new thinking forgeneration scheduling system.Based on the status of domestic and international energy conservation ingeneration scheduling system, the research on unit commitment optimization methodand the theoretical study of Multi-Agent, Multi-Agent approach is proposed in the paper,which is used to build modeling for the hierarchical structure of generation schedulingsystem. According to the function of each part of generation scheduling system, thepaper design the relative function for sub-Agent module and study the mechanism ofcommunication and collaboration between agents at the same time. In this paper, authorestablishes minimum objective function of the lowest energy consumption and pollutionemissions in accordance with the requirement of energy-efficient scheduling generation.Particle swarm optimization (PSO) as unit commitment Agent scheduling algorithm isused to optimize the multi-objective and the simulation result shows the feasibility ofthe method by an example. Finally, author writes software interface for the Agent modelof generation scheduling system using the Java language, which provides a simplescheduling decisions for policy makers.