Multiobjective risk analysis in a key water control project management

In accordance with a management problem of a key water control project, based on the discussion of major risk factors and risk evaluation, a framework of multiobjective risk hierarchical analysis is proposed by use of large-scale system theory. The overall model and its submodels (flood control, power generation, and irrigation, etc.) are presented for risk management, in which the performance indices, such as resilience, vulnerability, etc., are introduced to measure the risk-loss levels (duration, intensity, and frequency). These submodels are solved by dynamic programming and multiobjective stochastic dynamic programming with invertibility respectively. Thus the noninferior solution set of the overall model and the relationship among risk rates and system losses can be obtained. In the meantime, the best compromise solution of reconciling these conflicting objectives is determined by use of the mean-rate method of multiobjective decision making. The proposed framework and the models are applied to a multiobjective risk management problem of a key water control project in China.