Research on Risk Evaluation of Agent Management Projects Based on Clustering Methodology

In risk evaluation of agent management projects, the critical factors which influence rationality of the result are how to estimate risk evaluation index weight and the measurement of risk index. For the risk evaluation index weight account, this thesis proposes risk index weight calculation method which combines AHP method and clustering methodology. This method reduces the anthropic factors of expert authority at a certain extent. As for risk index measuring analysis, the thesis introduces the concept "degree of confidence" and uses Delphi method based on expert trust in order to avoid the subjective influences and make the result of risk measurement having higher confidence level. In order to manage the primary project risk well, it puts forward the risk index rank standard and assesses the risk rank of the every risk index of the project. Take emergency water supply for the section from Beijing to Shijiazhuang in midline of south-to-north water diversion project for example, with the method mentioned above, it makes substantive research on risk evaluation system and evaluation model, and proves the reasonableness, practicability and operability of the evaluation system and evaluation model that the thesis proposed.