| The build-operate-transfer (BOT) approach is one of the privatization mechanisms for promoting transportation infrastructure development by using private funds to undertake new infrastructure facility. There are many factors affecting the achievement of the BOT project. Two of these factors are the selection of toll charge and roadway capacity of the BOT project and the evaluation of the relevant benefits to the private sectors, government, and road users. In this dissertation, the optimal selection of both toll level and road capacity is formulated as a stochastic bi-level mathematical program with uncertain demand. The upper level is to maximize the objective for different parties' viewpoints, and the lower level describes route choice behavior of road users. A simulation-based genetic algorithm (SGA) procedure is used to solve the stochastic bi-level BOT network design model. Since the stochastic simulation is embedded in the optimization process, probabilistic risk analysis of the project performances of the different perspectives can be assessed. In addition, the BOT model is also extended to consider multi-class users in terms of the income levels with the spatial equity issue.; Modeling a multi-objective BOT network design problem is another task of this dissertation. Several multi-objective models are investigated: mean-variance BOT models, a stochastic profit-social welfare tradeoff model, and a three-objective optimization model. The simulation-based multi-objective genetic algorithm (SMOGA) procedure is used to generate a set of Pareto solutions that allow the decision makers to assess the tradeoff among different objectives of the design scheme. With a clear output in the form of the real tradeoffs, the decision makers can communicate easily with the involved parties on the issue of the conflict among the different requirements. Finally, the developed BOT network design model is applied to the real case study. |
