Models and algorithms for a rail transit line alignment using GIS and genetic algorithm

The increase in commuting populations and transit ridership in urban areas has given rise to the need for building new transit lines or extending existing ones. The objective of this dissertation research is to perform a microscopic analysis and develop models and algorithms for a rail transit alignment using Geographical Information System (GIS) and Genetic Algorithm (GA) in a given study corridor. Previous research suggests that several analytical methods have been developed to design various components of a transit system, such as optimization of station spacings, route spacing, and route length for rail transit alignments. While all this research has contributed substantially towards the development of a transit system, the applicability of these theoretical models remains limited for real-world problems.;The present research aims to perform a microscopic analysis in a given study corridor to obtain the locations and sequence of stations considering many-to-one travel pattern, variable demand for transit, identification of feasible locations for stations and optimization of various objective functions for station locations for an optimal rail transit alignment through GIS-GA integration.;An optimization algorithm is developed for optimizing station locations for three types of rail transit systems, namely (1) an on-the-ground rail transit system; (2) an underground cut and cover rail transit system, and (3) an underground deep tunnel rail transit system. The algorithm is developed in two stages. In the first stage, the search space is defined by identifying the potential station locations using a GIS and in the second stage a GA is applied to perform the optimal search.;The optimization model for station location is developed using several objective functions of demand and cost as both influence the optimal rail transit alignment. The first objective is to minimize the total system cost per person, which is a function of user cost, operator cost, location cost and variable demand. A bilevel optimization model is introduced for the station location optimization problem. In this model a number of clusters of riders are determined from local population data at the lower level in order to estimate demand and optimize the objective functions at the upper level. The second objective is to maximize the ridership or the service coverage. The user cost per person is minimized separately as the third objective because the user cost is one of the most important decision-making factors for using the transit system. A decision can be made based on the preferred parameters by a transit-planner based on the results obtained using different objective functions. A variable demand case is considered in the research which replicates a realistic scenario that can be expected while planning a rail transit line Optimal solutions are obtained by running an iterative process of re-estimating the variable transit demand by varying the locations and sequence of stations, which results in change in travel time and potential transit demand.;Once the optimal station locations are obtained, the stations are interconnected using a suitable rail line alignment. This is done by connecting each pair of stations using a GIS and GA based algorithm (customized Highway Alignment Optimization model).;The proposed model is applied on an artificially generated study area of size 20 miles x 20 miles around Washington DC in USA. The efficiency of the proposed algorithm is verified on small scale examples first. Then the model is applied on radial corridors of 12 miles x 3 miles and a diagonal corridor of area 22 mile x 3 mile within the study area to obtain optimal rail transit alignments. The results show that a new rail line alignment can be established in a comprehensive, consecutive and automated process using the proposed model. Better results are obtained for radial corridors than for the diagonal corridor as many-to-one travel pattern is considered for the study. The iterative approach in the algorithm increases computational complexities but produces near optimal solution for a rail transit alignment. The results give a planner an initial idea about which objective functions to use for a specific type of study corridor under consideration.