中文摘要：

建立DSM补偿机制问题是电力经济中典型的半结构化问题，传统数学模型很难解决，该文将智能工程与智能体理论结合起来研究了此问题。文中首先将建立DSM补偿机制问题归纳为智能工程理论中的第2类问题，然后构建了一个多智能体系统，并设计了各智能体的目标、结构以及行为策略，通过智能体之间的交互来确定该问题的最理想智能路径，接着给出了系统的仿真流程，最后在2000年和2002年实际经济数据的基础上作了实例分析。研究结果表明以电费附加方式按照适当的比例提取DSM资金用于DSM补偿在中国切实可行。

英文摘要：

Establishing DSM compensation mechanism is one of the typical uncertain issues in electric power economy, which is difficult to be modeled by traditional mathematic models. The theory of intelligent engineering and agent are adopted to study it. Firstly, the issue of establishing DSM compensation mechanism is defined as the second type issue in IE. Secondly, an open-agent system is designed, which consists of a government agent, 15 sectoral agents and human experts. Thirdly, the goal, structure and behavioral strategy of each agent are analyzed, and the optimum intelligent path of the issue is found out after the evolvements of agents. Then a simulation procedure is designed. Finally, the paper provides a case study based on China economic data of 2000 and 2002. The analysis results show that drawing DSM funds by proper percentage of electricity charge is feasible in China.