The Design of Intelligent Agents: A Layered Approach
The Design of Intelligent Agents: A Layered Approach
Proceedings of the 38th conference on Winter simulation
DS-RT '09 Proceedings of the 2009 13th IEEE/ACM International Symposium on Distributed Simulation and Real Time Applications
Winter Simulation Conference
SessionSim: activity-based session generation for network simulation
Winter Simulation Conference
Cybersim: geographic, temporal, and organizational dynamics of malware propagation
Proceedings of the Winter Simulation Conference
| Hi-index | 0.00 |
We introduce ActivitySim, a simulator that generates daily activities for a population of millions of individual agents, each characterized by a set of demographic attributes that is based on US census data. The daily schedules for each agent that consist of a sequence of typical activities, such as sleeping, shopping, working etc., with their respecting starting and ending times. Further, each of the scheduled activities is geo-located at an appropriate location for the type of activity and for the personal situation of the agent, like a business or private residence. ActivitySim has been developed as part of a larger effort to understand the interdependencies among national infrastructure networks and their demand profiles that emerge from the different activities of individuals in both baseline as well as emergency scenarios, such as hurricane evacuations. We present the scalable software engineering principles underlying ActivitySim, the socio-technical modeling paradigms that drive the activity generation, and proof-of-principle results for a 2.6 M agent scenario in the Twin Cities, MN, area.