文献类型：期刊文献

中文题名：并网模式下基于多代理技术的微电网多目标优化

英文题名：Multi-objective Optimization of Connected Microgrid Based on Multi-agent Technology

作者：王思明[1];牛玉刚[1];祖其武[1]

机构：[1]华东理工大学化工过程先进控制和优化技术教育部重点实验室,上海200237

年份：2017

卷号：43

期号：6

起止页码：829

中文期刊名：华东理工大学学报（自然科学版）

外文期刊名：Journal of East China University of Science and Technology

收录：北大核心:【北大核心2014】；CSCD:【CSCD_E2017_2018】；

基金：国家自然科学基金(61673174);上海市优秀学术/技术带头人计划(16XD1421300)

语种：中文

中文关键词：微网;经济性;多目标;多代理;JADE;粒子群

外文关键词：microgrid; economy; multi-objective; multi-agent; JADE; particle swarm

摘要：微电网的经济运行是一个多目标、多约束问题。传统方法将多目标转化为单目标的求解策略往往难以迅速收敛,且在权重的选取上具有很强的主观性。本文提出了一种基于多代理技术的多目标求解策略,考虑微网的整体经济性问题,分别建立发电侧目标代理(PowerAgent)以及需求侧目标代理(LoadAgent),从而避免了权重的选择。虚拟电价信息作为发电侧代理的协商参数,负荷调度信息作为需求侧代理的协商参数。利用JADE平台提供的Agent通信功能,目标代理之间传递协商参数,优化自身子目标问题,迭代协商,最终达到目标均衡。协商过程保证了微网的整体经济性要求,且两个子目标通过目标代理分布式求解,能够快速收敛。最后通过算例讨论了各代理的行为特征,验证了本文方法的可行性。
The economic operation of the microgrid is a multi-objective and multi-constraint problem. The traditional method is to transform the multi-objective problem onto a single-objective problem,whichoften dif f icult ly converges and has a stronger subjectivity in the selection of weights. In this paper,a multi-objective solving strategy is proposed via multi-agent technology. By considering the whole economy of the microgrid,this paper establishes the generation target agent （PowerAgent） and the demand side agent（LoadAgent） , respectively,so as to avoid the choice of the weights. Moreover, the virtual electricity price information is introduced as the negotiation parameter of the generating side agent and the load scheduling information is the negotiating parameter of the demand side agent. By utilizing the agent communication function in JADE platform,the target agents communicate the negotiation parameters each other,optimizes every subtarget problem and achieves the goal balance via iterative negotiates. The above negotiationprocess guarantees the overall economic requirements of the microgrid such that the two sub-targets canattain quickly convergence by distributedly solving the target agent. F in a l ly, an example is given toillustrate the feasibility of the proposed method.