论文部分内容阅读
公路路线设计是一个多目标的问题,传统将多目标转换为单目标优化方法存在目标统一准确性差的缺点,考虑多目标间的冲突性,提出采用多目标遗传算法进行路线优化。主要思想是通过建立多目标模型和路线线形模型,进行平面线形设计,经过纵断面优化后计算各目标函数值。采用快速非支配排序法对目标值进行排序,满足分布性的要求下构造非支配集,反复迭代计算后,输出满足终止条件的非劣解集。利用此算法进行实例验证,结果表明能在兼顾安全、经济、舒适、环保的目标下,给出一序列最优解,为路线方案决策提供直接依据。
The design of highway route is a multi-objective problem. The traditional method of converting multi-objective to single-objective optimization has the disadvantage of poor accuracy of unified target. Considering the conflict between multiple targets, multi-objective genetic algorithm is proposed to optimize the route. The main idea is to establish a planar linear design by establishing a multi-objective model and a linear alignment model. After optimization of the longitudinal sections, the objective function values are calculated. The fast non-dominated sorting method is used to sort the target values to construct the non-dominated set under the condition of satisfying the requirements of distribution. After repeated iterative calculations, the non-dominated set satisfying the termination condition is output. Using this algorithm to verify the example, the result shows that a sequence of optimal solutions can be given under the goal of safety, economy, comfort and environmental protection, providing a direct basis for decision making of the route plan.