融入偏好的公路多目标三维导向线生成模型及算法研究

Alignment zone is the specific route selection zone after the determination of highwaycorridor, which determines the initial location of a route and is the basis of the design of location and geometric parameter. At present, the extensive research on computer route selection was carried out both at home and abroad. But there is almost no model that really applied on a large scale because of the complexity of the problem and the difficulties on dealing with a wide range of terrain data. To reduce the burden of calculation of route selection, this application takes the corridor alignment which represents the space position of route as the researchsubject.Based on conventional HAO model, we are planning to build a 3D corridor model which considers multiple objective factors of alignment zone. RPSO algorithm is an improved particle swarm optimization algorithm proposed by our research group previously, the algorithm has the advantages of rapid convergence and has obvious inflexion,which fit in with the corridor alignment on the characteristic of fast searching. With the combination, the early stage inflection point of convergence is used as termination point of iteration, and the generating efficiency of corridor alignment will be improved.In addition, for that there is no 3D location model which takes the preference into consideration at the moment, this topic is ready to study the Multi-Objective RPSO algorithm with preference which can incorporate preference area of alignment zone and preference objective into model in order to meet the application of 3D corridor model.

路线带是公路走向确定后具体选择路线经过的地带，它所确定的初步位置是进一步线形线位设计的依据。目前，国内外对计算机选线开展了广泛的研究，但是由于选线问题本身的复杂性及处理大范围地形地物数据在现阶段的困难，到目前为止，很难有一种方法或模型能真正大规模地应用。为缩减问题规模，本项目以代表路线带空间位置的导向线为研究对象，基于传统的HAO模型，建立一种考虑路线带多目标因素的三维导向线生成模型。 RPSO算法是课题组先期提出的粒子群改进算法，它具有初期收敛迅速且收敛曲线具有明显拐点的特性，与导向线无需精细化搜索的特点相契合，二者结合，并把早期收敛拐点作为迭代终止点，将提高导向线生成的效率。此外，现有的多目标三维选线模型均未将设计者的偏好引入，本课题拟研究具有偏好的多目标RPSO算法，将路线带偏好区域和偏好目标纳入模型，满足三维走向线模型的应用需要。

公（铁）路是国家基础设施建设的重要组成部分，其投资巨大。其中路线位置直接决定了投资高低，而利用现代智能优化技术理论上可以根据给定的条件计算出线路的位置。本项目提出了公（铁）路智能优化的三维选线模型，自动根据地形的变化，同时确定平面和纵断面的位置。为了提高长距离线路的计算效率，提出了两阶段方法，在所建立的三维选线实验平台上实验，验证了此方法及模型在长距离线路上的提高效率的可行性。在路线生成中，面临各种多目标需求，决策者或设计者的偏好往往影响了路线位置的，本项目提出了考虑偏好的多目标算法，并将创新性地应用于此三维选线模型，验证了算法和应用的可行性。在粒子群算法研究的过程中，发现了粒子群中的个性粒子，实验及理论验证了该粒子的存在提升了粒子群算法及其变种算法的优化性能。