Incorporating spatio-temporal knowledge in an Intelligent Agent Model for natural resource management

Space and time are intrinsic components of the decision-making process in natural resource management. Decisions to extract resources from a specific location have consequences for all future decisions as they may lead to profitable opportunities or, conversely, towards unfavorable outcomes. As such, the spatio-temporal nature of decision-making should be acknowledged and incorporated into models developed to assist the management of natural resources. The objective of this research is to develop an Intelligent Agent Model that is able to learn through repetitive simulation how to make decisions regarding natural resource extraction. Specifically, an agent is guided by heuristic algorithms to search a natural landscape and learn which locations hold the highest profits and when it is best to extract the resource in order to improve the potential of future opportunities. The model is implemented using hypothetical and real data sets to emulate the process of harvesting trees in natural forests in order to maximize profits while respecting spatial constraints that are imposed in order to conserve various aspects of the forest. Simulation results reveal the ability of the Intelligent Agent Model to utilize spatio-temporal knowledge in order learn how to devise optimal solutions in a variety of scenarios. Furthermore, the model demonstrates how the timing of decisions is linked to the spatial constraints imposed on the operation. The findings from this research can be used to inform natural resource management about the importance of the relationship between the location and timing of resource-based activities.