Multi-Layer Multi-Index Comprehensive Evaluation for Dike Safety

Abstract

Dike safety is influenced by many factors. The contribution of every factor can be reflected with multiple indexes. To implement the reasonable evaluation for dike safety, a multiple indexes-based layer-by-layer evaluation approach is proposed in this paper. Extension theory, genetic algorithm and analytical hierarchy process are combined to identify the effects of all indexes on dike safety. First, a multi-layer system is built to describe the relationship between dike safety and its main influencing factors, evaluation indexes. Safety level and corresponding quantitative criteria of evaluation indexes are given. Second, the matter element analysis method in extension theory is introduced to establish the multi-index evaluation model. The relative importance of evaluation indexes can be determined and dike safety can be evaluated dynamically with the model. Third, to decrease artificial participation during the evaluating process and make the assessment result more objective, an accelerating genetic algorithm is integrated into analytical hierarchy process to calculate the weights of lower layer indexes in multi-layer evaluation system of dike safety. Finally, the proposed approach is applied to an actual dike case. Results show that the methodology is capable of more objectively judging whole dike safety, can help with identifying potential risk and hazard level.