| The highway alignment optimization algorithms available at present use the total cost as the objective function. The travel time cost, vehicle operating cost, accident cost, earthwork cost, land acquisition cost, and pavement and construction costs are the basic components of the total cost. This is a single-objective optimization approach and it has limited capability in considering cost component with different units. This approach uses the information of environmental impact, socioeconomic impact, impacts on historic sites and other sensitive areas in the total cost by transforming them to monetary values. It has limitation in yielding a set of alternatives with different level of trade-off among the objective values. Moreover, this process cannot yield a set of alternative solutions from a single execution of the method. The dissertation presents a multi-objective approach to overcome these shortcomings.;Analyzing different cost components and objectives for highway alignment optimization, it is observed that depending on the study area, some of the obtained alignments may be conflicting in nature, i.e., minimizing one cost component or objective value may yield an alignment with higher cost component or objective values. This leads to the study of the alternatives that might be obtained by trade-off among different cost components and other objective values. The multi-objective optimization has the potential to yield a set of alternatives with trade-off option. Generally, a Pareto-optimal front is designed to examine the trade-offs among two different objectives.;In the available highway alignment optimization approach, Points of Intersections (PIs) are considered as the decision variables. The highway alignment objectives depend on factors such as traffic growth, study area, and economic indices, which makes it difficult to represent them as an explicit function of the desired decision variables. To add to the complexity, the cost components and objectives are not continuous within the search space. Available multi-objective optimization algorithms do not address this issue. Therefore, there is a need to develop a multi-objective optimization algorithm that can efficiently and effectively optimize problems such as highway alignment optimization.;In this research, Genetic Algorithm (GA) based multi-objective optimization algorithms are specially developed for highway alignment design and optimization. This will have the capability to yield a set of trade-off alternatives. Mathematical formulations are developed to estimate the non-monetary objective values. The efficiency of generation of alternatives depends on the genetic algorithm reproduction operators. A special set of reproduction operators are also developed for this research work. All these formulation and estimation processes are computer coded in C and Avenue to obtain the alternatives. Effectiveness of the developed methodology is ensured by application of the multi-objective highway alignment optimization model to problems with discrete search domain and objective functions with indirect decision variables. Also, the developed model has the ability to converge to a Pareto-optimal front for two objectives. The trade-offs among individual objectives and cost components are analyzed through extensive sensitivity analysis. The Pareto-optimal front is a tool to graphically represent trade-offs for two objectives, but fails for more than two objectives. Therefore, effective representation of multi-objectives will also be pursued in this research work.;Keywords. Multi-objective optimization, highway alignment optimization, genetic algorithms, environmental impact, economic impact, social impact, geographical information system (GIS) database. |
