Abstract

ABSTRACT OF THE THESISGeneration of realistic hypothetical urban network configurations for simulated studies of future mobility optionsByNegin ShariatMaster of Science in Civil and Environmental EngineeringUniversity of California, Irvine, 2021Professor R(Jay) Jayakrishnan, ChairTraditional network generators are typically used for pure network studies within idealized optimization or modeling contexts in fields such as Operations Research and Computer Science. But for rare exceptions, such schemes for generating hypothetical networks have not been used in transportation studies and research. This is despite the recognition in many studies that the results in each study were not generalizable or transferable due the study being on limited networks that are often idealized forms of real networks. Newer mobility paradigms envisaged in the near future also make it important to now develop generators of data on hypothetical future network geometries and layouts, as well as the associated supply and demand. The reason for not using network generators in transportation studies is the myriad complexities that need to be addressed in the urban network context. This study describes a network generation framework that focuses on several such complexities (i.e. parameters and their interrelationships) with respect to network form, node density, link connections, freeway link generation, ramp generation, etc. The proposed framework generates not only different sizes and topologies but also more detailed data usable in agent-based models like activity locations. This framework also highlights the demand-side information needed in these models and generates randomly distributed agent trips with defined activities so that they are also usable in activity-based models. Illustrative results are provided for a few candidate networks generated by the proposed network generation methodology and the study discusses several associated details of conceptual and practical significance. Finally, a simple method is suggested for generating transit network (nodes and links) within the base network, and a method for improving it is also discussed.