|Ph.D Student||Levi Yedidiya|
|Subject||Optimization Model for Urban Arrays on Very Large Floating|
|Department||Department of Civil and Environmental Engineering||Supervisors||Professor Yehiel Rosenfeld|
|Professor Shlomo Bekhor|
|Full Thesis text - in Hebrew|
This study deals with developing a method and tools for optimal urban planning of new urban arrays on regions without topographical and other constraints. Such regions can be found when addressing the issue of urban planning on artificial islands, which was the trigger of this research. The model developed in this study was demonstrated for this case.
The optimization problems related to urban planning are complex, since they need to consider the customers’ behavior. Only a limited number of studies were published about the urban land-use design problem and the combined network and land-use problem. Some limitations of those models were noticed in the literature review. The limitations included, among others : lack of reference to the limited area resource, which calls for maximizing the utilization of the area; using a limited transportation model, taking into account only private cars as the transportation mode; and demonstration of the models only on small networks. Most of the models used heuristics to solve the problem, mostly Genetic Algorithm.
The aim of this study is to develop a mathematical optimization model for urban planning while addressing three components: (1) the geometric shape of the new urban area, (2) the land use composition and location, and (3) the transportation system, which address several means of transportation. In addition to the problem formulation, the thesis develops an effective solution method to the problem, suitable to its unique properties.
The problem is formulated as a multi-objective bi-level optimization problem, looking for a set of non-dominated solutions. Discrete decision variables were defined to describe the land uses and transportation network components. Three objective functions were chosen as criteria to rank the different scenarios: (1) minimum costs, (2) minimum system travel time, and (3) maximum utilization of the island area.
An iterative innovative three-step algorithm is developed to solve the problem, based on the genetic algorithm. The proposed model and algorithm are tested and demonstrated using two case studies. The results showed a good convergence rate.
A “cluster analysis” was performed on the Pareto front results, using an improved K-means clustering technique, and a similarity was found between the obtained clusters and the land use patterns. The idea of dividing to clusters or patterns may help the planners while using the optimization tool with the understanding that small differences between scenarios are almost irrelevant. A sensitivity analysis was performed, with regard to the input variables and the parameters of the genetic algorithm. The results show that the model is quite robust.
The contribution of this research is threefold: A theoretical contribution relates to the mathematical formulation of the problem and to the expansion of the lower level problem to account for the mode split between public and private transportation. A methodological contribution relates to the development of the algorithm and the way that the results are analyzed using clustering technique. An applied contribution relates to the development of an optimization tool, which may assist city planners and decision makers while planning new urban areas.