基于分步编码改进遗传算法的铁路智能选线

Intelligent Railway Aignment Optimization Based on Stepwise Encoding Genetic Algorithm

为实现铁路三维空间线形的智能优化,建立了综合考虑建设和运营费用、环境影响代价及线路约束的线形优化模型;针对平面、纵面控制点分别给出了分布方法,并构建了由平面交点偏移距、平曲线半径、变坡点标高3个基因片段组成的基因序列.在此基础上,提出了平面-纵面-完整基因的分布编码方法,设计了相应的多种交叉、变异算子实现线路方案的进化.实例分析表明:本方法克服了现有遗传算法产生的线路方案平、竖曲线重叠的缺陷;可生成满足线路约束条件且综合费用较省的线路方案群,最优的智能选线方案比人工定线方案综合费用节省了6.5%.

In order to achieve the optimization of railway 3D alignments,an optimization model was built,in which comprehensive factors including construction,operation,environment,and constraints were embedded.Different distributions for horizontal and vertical control points were presented.A genetic series that consists of offsets of intersection points,the radii of the circular curve,and the elevations of grade change points was designed.Then a stepwise horizontal-vertical-integral genetic encoding method was put forward,and the genetic operators for crossover and mutation were also designed to achieve the optimization of railway alignments.The results of application examples indicate that this method can overcome the shortcoming of the overlapping of horizontal curves and vertical curves,and yield a group of alignments with low comprehensive costs and meanwhile conforming to railway constraints.The optimal alignment obtained by this optimization method can reduce the overall comprehensive cost by 6.5％ than the alignment obtained by human work.