不确定层次分析下的桥梁评估最优指标权重确定

针对不确定层次分析法在桥梁评估中的不足,通过对不确定性判断原理进行分析,利用元素相离度的概念提出了一种求解指标权重的新方法,即根据不确定判断矩阵中的未知权重构造了一种目标优化模型,并利用计算智能中的实数遗传算法将优化模型转化为无约束的目标优化函数,通过复制、交叉、变异等遗传算子进行全局优化搜索,从而获取指标的最优权重。通过对2个算例的指标进行不确定判断,并对不同优化算法获取的优化结果进行分析比较,分析结果表明,新方法不仅能搜索到更优解,而且能获得更优权重,因此,验证了采用新方法的可行性。

Determination of the Optimized Index Weights for Bridge Evaluation under Uncertain Hierarchy Analysis

The approach of uncertain hierarchy analysis is applied more and more into the bridge evaluation in recent years.But the bridge index weights obtained from general method can't judge the importance of bridge indexes.By analyzing the principle of uncertain judgment,a new method of calculating the index weights is proposed based on the principle of near distance degree,in which a model of objective optimization for the unknown index weights among the judgment matrix of uncertain hierarchy analysis is set up.Then the optimal model is transformed to the nonrestrictive optimal function by using the approach of real genetic algorithm(RGA).The global optimal searching is also carried out according to the operation of copy,chiasm and aberrance to obtain optimal index weights.At last,two examples are established to test the validity,some different methods are also executed,and the results are analyzed and comprised.The results show that the better solution can be obtained form the new method and it verified the feasibility.