| Models of complex, adaptive, and dynamic geographic systems can offer significant insight into the underlying mechanisms that drive system dynamics and assist decision-makers as they attempt to manage these systems in a sustainable manner. Geographically-aware, mobile, and adaptive entities often play a key role in guiding the dynamics of these geographic systems. Agent-based models have been recognized as an appropriate computational approach for modeling these systems. The purpose of this dissertation research is to address and solve key research issues associated with agent-based models of complex adaptive geographic systems. These issues revolve around the explicit representation of geographic context and the modeling of spatial cognition and learned behavior in agents. A modeling framework for the representation of geographically-aware intelligent agents is proposed and implemented as part of this research. This framework integrates functionality from agent-based models, machine learning, and geographic information systems into a single platform that provides support for modeling adaptive behavior in individuals within spatially-explicit environments. Agents derived from this framework are contextually-aware, knowledge-driven, and adaptive and, thus, can potentially solve complex geographic problems. A model of the adaptive movement behavior of elk represented as geographically-aware intelligent agents was used to examine the utility of this framework. Experimental results suggest that this framework is well-suited to the resolution of complex adaptive geographic problems. |
