中文摘要：

文章应用复合DEA算法分析制约电力企业相对效率的瓶颈资源及优化方向。以沪深两市21家电力上市公司为实证分析样本，结果表明：我国电力行业很可能已步入规模收益递减的阶段，电力企业的规模越大，规模收益递减的可能性越大；电力企业投入产出相对效率最敏感的投入指标是固定资产和主营业务成本，产出指标是发电量和主营业务收入；对于非技术和规模有效的电力企业，投入冗余相对严重的是“大专及以下学历的员工”，产出不足相对严重的是“发电量”和“经营净现金流量”。

英文摘要：

Bottleneeked resources that lay restrict on electric power enterprises and optimization methods are analyzed by using com- pound DEA in this paper. Choosing 21 electric power enterprises listed of the two stock markets as sample, the results show that Chinese electric power industry must have entered the stage of decreasing returns to scale because of the larger scale in the electric lX~Wer enterprise, the more possibility of decreasing returns to scale, that the most sensitive to the input index of input- output relative efficiency of electric power enterprise is fixed assets and main business cost while the most sensitive to the output index is generating capacity and main business cost, and that for the non- technical and the limited size electric power enterprises, relatively serious input redundancy is ＂the employees educated in or below college while output deficiency is ＂the generating capacity＂ and ＂the net operating cash flow＂.