基于交互式多目标遗传算法的混凝土桥面板维修优化

Maintenance Optimization of Concrete Bridge Deck Based on Interactive Multiobjective Genetic Algorithm

为了保证桥梁结构安全,并使维修效果最好,研究了如何确定桥面板最佳维修策略的问题。建立了桥面板维修的多目标优化模型。提出一种基于交互式遗传算法的维修策略优化方法,通过非精确偏好信息引导遗传搜索,使种群朝着多个目标的最佳综合性能进化。引入适应值共享方法增加种群的多样性,最终获得符合决策者偏好的解集。并通过算例进行了分析验证。结果表明,基于交互式多目标遗传算法的混凝土桥面板维修优化方法可以在有限的桥梁维修资源和良好的维修效果之间进行折衷,不仅获得一组比较理想的维修方法组合,还可以考虑决策者的个人倾向,降低了维修策略的选择难度,为桥梁管理者提供了有力的决策支持。

In order to assure bridge structure safety and obtain most satisfactory result from maintenance action, the problem of selecting the optimal maintenance strategy of concrete bridge deck was investigated. A maintenance strategy optimization method based on interactive genetic algorithm was presented by way of inputting inaccurate preference information and leading population into optimal comprehensive performance of multiple objects. Fitness sharing technique was introduced to improve the diversity of population and obtain the set of solutions satisfying decision-maker＇s preference ultimately. Analysis of a realistic example indicates that the maintenance optimization method of concrete bridge deck based on interactive genetic algorithm can balance limited bridge maintenance resources with good maintenance effect and supply a set of tradeoff solutions as optimization result. Therefore, the difficulty of selecting optimal maintenance strategy can be reduced to a large extent and the optimization method proposed herein is able to present powerful decision-making sustainment to bridge managers.