基于熵和证据理论的NPD项目复杂性模糊评价

Complexity Fuzzy Evaluation of NPD Projects Based on Entropy and Evidence Theory

基于文献综述和专家调查,构建了NPD项目复杂性评价指标体系,并运用"结构熵权法"确定了各指标权重。基于证据"可信度"和"确定性",定义了证据"有效性",结合证据冲突全局分配的思想,改进了证据合成方法;运用改进的证据合成方法综合专家评价信息,构造了"模糊评价矩阵"。依据模糊评价原理对NPD项目复杂性进行评价,并结合算例分析说明了该方法的可行性和实用性。

Implementation of NPD （ New Product Development） projects has become an important measure for enterprises to constantly gain new competitive advantages. However, some studies discover that the degree of complexity in NPD project is a direct source of NPD project risks. Moreover, traditional project management seems to be not an effective measure to deal with the growing complexity of NPD projects. Hence, identifying existing NPD project complexity sources and assessing the complexity of NPD project have become crucial issues for monitoring variations of complexity, quantifying the relationship between complexity and other variables （ such as performances, risks, ere）, and improving existing NPD project management theories and methods. An effective evaluation of NPD project complexity depends on establishing a perfect evaluation index system as well as dealing with fuzzy or uncertain information in the evaluation process. The first part of this study constructs a complexity evaluation index system for NPD projects. We identifysourees of NPD project complexity and construct an initial index system according to a system analysis of NPD projects and a comprehensive literature review. The initial index system is revised and improved via the Delphi method. The second part of this study proposes a fuzzy assessment methodbased on structural entropy weight and D-S evidence theory. We analyze the validity and advantage of applying the structural entropy weight method to ascertaining the complexity index weight of NPD projects. Based on the ＂reliability ＂and ＂certainty＂ of evidence, our research defines the ＂effectiveness＂ of evidence, and proposes a new approach of distinguishing the importance among different evidence. Moreover, we integrate D-S evidence with the ＂Conflict global distribution＂ idea. Moreover, this paper presents a method of using the improved D-S evidence theory to construct a ＂fuzzy evaluation matrix＂. The final part of this studyevaluates the complexity of one NPD project by demonstrating the application process of our proposed method. The calculation results reflect the effectiveness and feasibility of the proposed method. In conclusion, this paper establishes an index system for NPD project complexity evaluation with good explanation ability mainly through qualitative analysis. It is still necessary to further make quantitative evaluation in order to verify the rationality and validity of the index system in the near future. The proposed evaluation method can effectively calculate the index weight and process uncertain information. The proposed model can not only assess NPD project complexity, but also solve problems related to multi-attribute group evaluation and multi-attribute group decision making.