RBSim3: Agent-based simulations of human behaviour in GIS environments using hierarchical spatial reasoning

This paper describes recent advancements in the development of an agent based simulation model called Recreation Behaviour Simulator (RBSim). RBSim allows recreation managers to investigate alternative management scenarios for linear recreation systems such as roads, trails, and rivers. The simulation program integrates Geographic Information Systems to represent natural environments with network algorithms and agent based simulations to model human behaviour. As a management tool, RBSim allows managers to alter characteristics of the environment through standard GIS techniques, as well as altering arrival schedules of visitors, global events such as weather, and the size and location of facilities such as parking areas for vehicles, campgrounds, and viewing platforms. Statistical output from the simulator may then be analyzed to reveal bottlenecks, crowding conditions, excessive queuing times for facilities, and the number of visual encounters. The architecture of this model has been described in detail in other papers (Itami et al 2002, Itami and Gimblett, 2001). This paper will focus on the spatial hierarchy used to simulate trips, and the use of hierarchical probabilistic rules to define complex spatial reasoning and context specific decision making in recreation agents. These concepts are demonstrated in a day use simulation for Port Campbell National Park in Victoria, Australia and compared to visitor use data collected for the study area.