Maxim Janzen

Decision Making in an Agent-Based Simulation of Long-Distance Travel Demand

Abstract Analysis of long-distance travel demand has become more relevant in recent times due to the growing share of traffic induced by journeys related to remote activities. Consequently, there is a need of reliable long-distance travel forecasting tools like agent-based simulations. This paper presents a target-based simulation that simulates long-distance travel behavior for a long period of time. It is shown how decisions are modelled in this simulation. Activity type, duration, destination and mode are chosen simultaneously with respect to time and monetary budgets. The presented approach uses a heuristic to reduce the choice set followed by optimizing a discomfort function.

Measuring long-distance travel

Analysis of long-distance travel demand has become more relevant in recent times. The reason is the growing share of traffic induced by journeys related to remote activities, which are not part of daily life. In today’s mobile world, these journeys are responsible for almost 50 percent of the overall traffic. Traditionally, surveys have been used to gather data needed for the analysis of travel demand. Due to the high response burden and memory issues, respondents are known to underreport the number of journeys. The question of the real number of long-distance journeys remains unanswered without additional data sources. This paper uses an alternative data source, mobile phone data, and compares its results for long-distance travel demand with the national household travel survey (Mikrozensus Verkehr 2015). We take a sample of mobile phone data covering 12 months, and identify the number of long-distance trips. The results indicate that the number of long-distance trips is close in these two data sources.

Destination and mode choice in an agent-based simulation of long-distance travel demand

Analysis of long-distance travel demand has become more relevant in recent times. The reason is the growing share of traffic induced by journeys related to remote activities, which are not part of daily life. In today’s mobile world, these journeys are responsible for almost 50 percent of the overall traffic. Consequently, there is a need of reliable long-distance travel forecasting tools. A potential tool is agent-based simulation. Due to the complex task of destination choice modelling, there are just few agent-based simulations available. This paper presents a continuous target-based simulation that simulates long-distance travel behavior for a long period of time. It is shown how destination choice and mode choice is modelled in this agent-based simulation. Destination and mode are chosen simultaneously along with activity type and activity duration. The presented approach uses a heuristic to reduce the choice set since the underlying multi-dimensional optimization problem is too hard to be solved directly with acceptable computational effort. Afterwards the best combination of destination, mode and activity is determined based on the agents’ projected discomfort.

Purpose imputation for long-distance tours without personal information

Analysis of long-distance travel demand has become more relevant in recent times. The reason is the growing share of traffic produced by journeys to remote activities, which are not part of daily life. In today?s mobile world, these journeys are responsible for almost 50 percent of overall traffic. Traditionally, surveys have been used to gather data needed for the analysis of travel demand. Due to the high response burden and memory issues, respondents are known to underreport the number of journeys. Thus, alternative data sources are becoming more important. These sources collect the data passively, e.g. using GPS or GSM networks. The limitation of passively collected data is the lack of semantic information, especially trip purposes. Additionally, socio-demographic information is also missing making it difficult to impute the purpose. This paper shows how one can predict the tour purpose without the need of socio-demographic information. The solution extends the well known random forest approach. Attributes of the tours are used in order to overcome the lack of personal information. The training set for the algorithm was taken from a national travel survey.

Report for the Collaboration between Orange Labs France and IVT, ETH Zurich

To date, travel demand generation for microscopic traffic flow simulation and travel demand model focuses on reproducing and predicting daily behavior. This stands in contrast to the significant part of traffic volume caused by journeys related to activities not usually undertaken during daily life. In order to investigate the long-distance travel behavior a collaboration of IVT, ETH Zurich and the Orange Labs, France is initiated. The idea of this collaboration is to analyze available mobile phone positioning data with a focus on reconstruction of long-distance trips.