Research On Models And Methods Of Project Risk Assessment Based On DEA

Project risk assessment technology has become an important topic of system reliability engineering, it is used in the practice of enterprise project management widely. As the project's risks have some characteristics such as non-linear, complex, dynamic, uncertain and so on, research and application of project risk assessment technology attract the attention of entrepreneurs and academics increasingly.In this thesis, considering both positive and negative factors that affect the risk of project risk assessment synthetically, introducing data envelopment analysis, the coordinate preference project risk assessment model based on DEA is established by analysing the equivalence of DEA efficient solutions and Pareto solutions of multiobjective programming. Conception of risk envelope is provided. By introducing fuzzy comprehensive evaluation method based on entropy weight and selecting two moving factors to reduce the risk envelope layer by layer, the method of sorting risk by moving risk surface is improved. The numerical example of software project is given to verify feasible and valid of the model and method that have the capacity of optimizing the opposite evaluation of the assessment and ranking the risk of decision making programs absolutely.As both positive and negative factors'data are easy to appear errors and changes, two models of DEA sensitivity analysis are improved. Then the sensitivity analysis of coordinate preference risk assessment model based on DEA is given. It presents some sufficiency conditions which can hold the efficiency and low risk of decision making units, while positive and negative factors of decision making units are absolute changed and relative changed simultaneously. The numerical example indicates that this method can provide effective decision making for avoiding and controlling the risk of the project.This thesis also studies the risk assessment problem that has subjective preference and opposite evaluation, and analyzes the flaws of traditional cross-efficiency evaluation method in DEA. On the basis of the cone ratio DEA model and the coordinate preference project risk assessment model based on DEA, DEA risk assessment model based on preference cone is established, subjective evaluation and objective evaluation are combined coordinately. It presents a cross-efficiency evaluation method based on DEA model with preference cone using in project risk assessment, reflects cross-evaluation mechanism between self-evaluation and peer-evaluation. The numerical example indicates that this method is quicker in calculative process and all units can be ranked in full order. Because of adding subjective preference informations, the result of this method has access to the actual conditions.