Multi-Agent Based Simulation of Commuting in Urban Areas
Transportation policies and infrastructure investments can have substantial consequences on the society and travelers’ behavior. Thus it is very important to assess the impacts of such changes before implementation. One way for doing such impact assessment is to perform experimental studies in real world to get high realistic results of the evaluation. However, there are some constraints associated to the real-world experimental studies, e.g. they are very expensive and time-consuming and sometimes infeasible. In this project we try to address the challenges stated above by developing a simulation tool in order to estimate the effects of applying different transportation policies or investments. The tool can act as a decision support system for policy makers in order to investigate several what-if scenarios. Some sample policies that can be investigated by the tool can be: Changed schedules (e.g., bus and train), Pricing schemes (e.g., ticket prices, etc.), Taxes and fees (e.g., congestion fees, fuel taxes, etc.)
All “commuters” of the urban area will be modelled on an individual level. In order to model the commuters in such a detailed level, we include the information regarding commuters’ life-style and travel behavior. Examples of such information are: socio-demographic information, mobility tool ownership data (car ownership and public transport season ticket), etc. Furthermore, the model includes all possible modes of transportation for a commuter to produce more realistic results. This decision is made to support how commuters choose their transport mode in reality, where they can decide to switch to another mode of transportation is special circumstances. The output of the model is choice of transportation (which might be a chain of switching between different transport modes), start and end time of the travel, and cost of the travel for each agent. The output data will be calculated for every single agent, but the results can be aggregated to show a general overview of the modal share, travel time and cost for commuters.
The results from the simulation will be sent to a visualization module which is developed in this project in order to make the results more understandable for the users (policy makers or transport planners).
Advisors: Professor Paul Davidsson, Docent Jan Persson, MaH/BTH.
University: : Malmö University (MaH)