Trails Forward (Agent Based Modeling)

The Trails Forward project is a multidisciplinary collaboration focused on developing an educational video game intended to increase awareness regarding the environmental, economic, and institutional factors around land-use conflicts. At the same time, this gaming platform is designed to be used as a unique simulation platform that can be used to investigate the consequences of manipulating these factors in different ways. Game dynamics generally emerge as an interaction between policy specification, player input, and the resultant behavior of autonomous computer-based agents. These agents represent autonomous individuals within the world, and represent a core feature of the Trails Forward architecture. The development of these agents is extremely flexible, and can be tailored to represent a wide range of potential ‘roles’, ranging from home buyers to wildlife species. By acting in an autonomous and self-interested fashion, the cumulative behavior of the set of agents can be used to describe the entire system. This allows for a nearly endless variety of research questions to be addressed, from exploring the habitat utilization of various species in developing housing markets, to examining home buyer behavior in response to a range of potential market constraints. Indirect effects can be examined as well, such as the effect of rising timber prices on wildlife habitat. One of the greatest benefits is that this environment allows us to examine complex relationships and dynamics that are either impracticable or impossible to experimentally manipulate in the real world. The behaviors of these agents are defined pragmatically, allowing researchers to easily tailor agent behavior to suit their specific requirements or interests. The system is designed so that the simulated world’s semantics are similar to NetLogo, which allows models developed using NetLogo to easily be ported over to integrate with the Trails Forward game architecture.

Through this system design, the potential exists to specify different algorithms to enrich the landscape dynamics within the game in order to tailor the simulated environment to better address specific questions posed by the researcher. For example, the current iteration of the game will utilize growth and yield models specific to different forest types in order to simulate the production of various wood products within the game. This allows agents (both human and computer-based) the potential to make more detailed and informed decisions regarding land-use decisions. They can then evaluate potential opportunity costs associated with specific decisions, such as the financial implications of altering a timber harvesting plan to provide a target amount of habitat for a particular species. This feature could be extended to incorporate algorithms describing the growth of agricultural crops based on underlying data such as soil type, elevation, aspect, etc.

more info…