Impact assessment of new mobility services using accelerated activity-based demand modeling

The digital world has gotten closer to us since the first real smartphone came out. Increasingly convenient smartphone apps have greatly simplified our daily lives. As a result, new mobility services such as mobility as a service (MaaS) and mode-sharing services may affect the behaviour of individual travellers significantly. Furthermore, evolving factors concerning the travelling population may also change the mobility landscape. Moreover, economic and land use changes affect the transportation system's performance.
Facing all these societal changes, policymakers want to know how the transportation system will perform and what kind of policies and interventions should be proposed to achieve the desired output. For this purpose, in this thesis, we adopt activity-based travel demand models (ABMs) to estimate future travel demand accurately. ABMs closely replicate actual travelling decisions based on behavioural theories on how people participate in activities in the presence of constraints. In this case, however, the models should also be able to model both traditional and new mobility modes, simulate many scenarios fast and generate stable output.
To address these challenges, this thesis develops a multimodal tour-based mode choice model to handle different transportation modes, including the new mobility services. Next, it shows that a high speedup can be obtained using a computer's graphic process unit. Furthermore, it illustrates that common random numbers could significantly reduce the computation time required. Finally, a large-scale case study is performed, demonstrating the potential of the ABM framework on the impact of different parking policies, MaaS subscriptions and infrastructure interventions.