A. Maji, M.K. Jha
Pages: 51-72
Abstract
A Multi-Objective Evolutionary Algorithm (MOEA) framework for highway route planning is proposed in this paper to simultaneously optimize the environmental, economic and social impacts along with the route cost. The proposed model is capable of generating multiple alternatives with varying trade-offs between four objectives (i.e., constrained environmental impact and unconstrained route cost, economic and social impacts). At present, transportation planners rely on engineering judgment to develop these alternatives; however, they are unable to fully explore the spatial information of land parcels. Thus, true optimum alternatives for detail design is not guaranteed. On the other hand, the available highway route optimization models combine all objectives into one and solve it as a single objective highway route planning problem. Essentially these models yield only one route alignment and are unable to simultaneously optimize all highway route planning objectives. The proposed MOEA framework for highway route planning uses genetic algorithm for the optimization process and geographical information system to estimate environmental, economic and social impacts. A total of eight different types of mutation and crossover operators are specifically developed to generate offspring candidate solutions from the non-dominated and dominated parent solution pools. These offspring candidate solutions help in evolving to optimum nondominated solution set or route alternatives. A study area with real world geo-spatial information on environmental, economic and social parameters is considered to demonstrate the performance of the proposed model. The proposed model yielded 200 mutually exclusive non-dominated alternatives at the end of 48 generations. Overall, it has the capability to provide transportation planners a set of true optimal route alternatives for more than one route planning objectives.
Keywords: multi-objective optimization; highway route planning; transportation planning; evolutionary algorithms; environmental impact; geographical information system (GIS)