Scheduling weekly flexible activities in a large-scale multi-agent mobility simulator
Ordoñez Medina, Sergio A.
- Conference Paper
Rights / licenseIn Copyright - Non-Commercial Use Permitted
Multi-agent transport simulation models are tools to evaluate and compare future scenarios to support decision making. It is now possible to simulate the mobility of an urban population in the order of 10^7 agents. In activity-based models, the mobility demand of a population can be estimated by specifying the time and location of individual human activities. However, most of the existing activity-based models are designed for a period of one day. This work introduces a methodology for scheduling flexible activities for a week time horizon using independent daily observation from a travel survey. Purpose-specific start time and duration joint distributions are extracted for each day of the week. A utility-maximization algorithm based on spatiotemporal graphs uses these distributions to schedule flexible activities during specific free time windows. The method is used by agents from a synthetic population of Singapore to schedule on-the-fly six flexible activity types during a large- scale mobility simulation. Results show accurate start time and duration distributions by activity type and day of the week comparing with the Household Interview Travel Survey observations of Singapore.. Show more
External linksSearch via SFX
SubjectActivity scheduling; Week time horizon; Flexible activities; Mobility simulation; Multi-day; MATSim
Organisational unit08058 - Singapore ETH Centre (SEC) / Singapore ETH Centre (SEC)
03521 - Axhausen, Kay W.
02226 - NSL - Netzwerk Stadt und Landschaft / NSL - Network City and Landscape
MoreShow all metadata