A visualization and simulation tool that will generate effective patrolling strategies to protect the U.S. borders from illegal intrusion using game theoretic methods and models.

摘要

In the recent decade, the United States Border Patrol has increased the presence of border security officers at points of entry along the U.S.- Mexico border. In response to this increased presence of border security officials at ports of entries, illegal intruders have rerouted their intrusions in between ports of entry, in harsher terrain areas of the U.S. (e.g., the Arizona desert). The U.S. Border Patrol could benefit from a tool that plans effective border patrolling routes, and helps border security officials to make risk-based decisions for resource allocation. We are currently developing a tool called Genetic Algorithm for a Map-based Multi-Agent System, or GAMMASys, which contains tools, that allows for the creation of effective border patrolling strategies using visualization, simulation, randomization, and automation. These features allow border security officials to observe the relationship between adversaries (i.e., patrollers and intruders) using a visualized terrain tool. This allows for multiple patrolling strategies to be simulated through games using a defined adversarial game theoretic model; allows for intelligent and unpredictable decisions to be made by adversaries using randomization; and allows for effective patrolling strategies to be generated automatically using crossover and mutation techniques with the genetic algorithm. The contribution of this research is to provide a sophisticated tool that can apply all these aspects to generate effective patrolling strategies for the real-world border zones, and answer questions about the cause and effect relationship of intelligent decisions made by adversaries in the border security domain. These questions regard: altering patroller/intruder intelligence levels in the game theoretic model, efficient parameter combinations for high solution quality in lower amounts of time, and resource allocation. The ultimate intention is to construct scenarios to evaluate the relationships between intelligent adversaries and the effects of their decisions.