基于多目标的路面养护决策优化模型的应用研究

Multiobjective optimization in pavement maintenance and rehabilitation programming

传统路面养护决策中需要将道路使用者不同的效益换算为费用进行寿命周期费用分析,这样的换算过程不仅不合理,而且还存在很大的任意性和不确定性。通过建立路网中某一个路段在分析期内所需投入养护资金最小和所获取效益最大的多个目标的优化模型,对这个项目在整个分析期内养护资金需求的问题进行研究。模型求解时利用多目标优化中非支配解(Pareto解)的概念,并利用遗传算法求得优化模型的非支配解。通过实例计算,对结果进行分析,验证模型的有效性。

Traditionally, the fond for maintenance and rehabilitation are acquired on the basis of the Life-Cycle Cost Analysis, which needs to convert the user benefit into a single monetary value. But it is unreasonable and uncorrectable to convert some of the user benefit into monetary. A multiple objectives optimal model which opti- mize or prioritize on the combinations of cost and benefit （effectiveness） is set up in this paper. For solving multi-objective optimization problems, the concepts of Pareto optimal solution are adopted. This paper develops a genetic-algorithm-based procedure for solving multi-objective project level pavement maintenance programming problems. The solutions we obtained in this dissertation should put into practice in order to testify their rationality and feasibility.