With the further increase in investment in power distribution network construction, traditional project management methods are confronted with the challenge of quantifying and identifying the bottleneck of efficiency in each project execution stage from "reservation, project initiation, implementation and commissioning". Therefore, this paper first constructs a comprehensive efficiency evaluation index system that includes multi-dimensional evaluation indicators such as basic project attributes, process sequence, and resource consumption. Secondly, this paper proposes a comprehensive evaluation model combining Data Envelopment Analysis (DEA) and entropy-weighted TOPSIS (Technique for Order Preference by Similarity to Ideal Solution). Finally, graph computing and association rule mining methods are applied in the process of efficiency diagnosis. Through analyzing the complex correlation network between evaluation indicators and efficiency results, the key influencing paths and hidden bottleneck factors are discovered. The empirical research indicates that, compared with traditional methods, the comprehensive evaluation model combining above methods can significantly improve the discriminative ability of efficiency scores, and there exists an obvious negative correlation between efficiency scores and actual project overrun rate (correlation coefficient : -0.832). The intelligent diagnostic system can identify typical efficiency bottlenecks with accuracy of 79.6%. Compared with traditional methods, it saves at least several days of analysis time to several hours, and provides an efficient and reliable decision support for refined management and resource allocation in power distribution network engineering.
Chen et al. (Thu,) studied this question.