Font Size: a A A

Three-Dimensional Interactive Highway Alignment Optimization Based On GPU Parallel Genetic Algorithm

Posted on:2014-06-26Degree:MasterType:Thesis
Country:ChinaCandidate:Y LiuFull Text:PDF
GTID:2272330452462707Subject:Computer Science and Technology
Abstract/Summary:PDF Full Text Request
Comunication and Transportation is very important to social economy, and highway isone of the most historic and most widely-used transportation modes. Highway has a greatsignificance to development of a country and progress of society. Highway AlignmentOptimization (HAO) is one of the most important aspects of highway design, for theinefficiency of traditional HAO methods, nowdays Genetic Algorithm based HAO methodbecomes a hot spot in this area. In this thesis, with the background of Gaoping–Xinxianghighway design visualization platform, a3D HAO method using parallel genetic algorithm onGPU is proposed. The main research content in this thesis is as follows:(1) A non-feasible solution pre-screening method in fitness evaluation.In order to reduce the number of population evaluation in genetic algorithm, non-feasiblesolutions which violate design constrains are pre-screened. By removing the non-feasiblesolutions, the population evaluation time reduces in a certain degree so that HAO efficiency isimproved.(2)3D HAO based on GPU parallel genetic algorithm.For the inefficiency of traditional HAO methods, highway alignment optimizationprocess is deeply studied and is resolved into several subprocedures. For example, fitnessfunction is partitioned into horizontal alignment fitting, vertical alignment fitting, non-feasiblesolution pre-screening, pile structures computation and earthwork computation, etc. For everythese procedures, its parallel model is proposed.(3) A3D visualization system for HAO.For the convenience of visual highway design in multiple dimensions, a3D visualizationsystem for HAO is implementd. This visual system has flat, longitudinal and3D views whichrealize “WYSWYG” for highway design. The flat view show highway horizontal alignment,longitudinal view gives a vertical section of highway alignment, and3D view displays thewhole geographical environment including terrain and highway in three-dimensional. In thisvisual system, a simple3D model merging method is implemented.
Keywords/Search Tags:Highway alignment optimization, parallel genetic algorithm, 3D, GPU, CUDA
PDF Full Text Request
Related items