Much research work in recreation behaviour has been done over the last decade to more adequately understand societal and individual attitudes toward wilderness, reasons for participating in wilderness areas, and factors affecting wilderness behaviour, including the influences of management, crowding and conflict. The focus of this research is on empirically assessing recreator behaviour as a means of providing guidelines for improving and managing wilderness use. This work, while providing empirical data for understanding behavioural characteristics of a variety of recreation groups, has not been effectively incorporated into spatial simulation systems to assess the impact of alternative management decisions upon wilderness landscapes over time. While some have employed Geographic Information Systems (GIS) to provide a precise, geo-referenced database of real world data as a foundation for some simulations, none have employed the use of artificial agents in simulating complex social interactions and decision-making in the context of wilderness landscapes. Since it is likely that different decision-making contexts may result in different actions, the design of a human-like artificial agent must be bounded or grounded with real data. This form of agent design has been referred to as "calibrated agents", where the agents are calibrated and grounded empirically to actual rather than idealised human behaviour. One problem with this approach is in the design and calibration of these agents. This paper describes some of the issues confronted in designing human-like agents calibrated against recreators who use the Broken Arrow Canyon experimental area in the Coconino National Forest of the Sedona Ranger District. Issues related to calibrating the agent's visual, mobility and goal-oriented systems against real world data is discussed. Results of this work suggest that agents calibrated against experimental data may be usable as a basis for agent behaviour in larger theoretical models to assist the forest manager in answering questions about recreation management in a more general setting.
|Original language||English (US)|
|State||Published - Apr 1 1996|
ASJC Scopus subject areas
- Computer Networks and Communications
- Artificial Intelligence