Multi-objective border patrolling optimization using game theory and evolutionary algorithms
Border security has evolved significantly since the days when no more than 75 Mounted Guards patrolled the U.S. border in the early 1900's. The border region includes thousands of miles of both land and maritime borders that must be controlled, as well as commercial transportation networks. One of the primary components of the Department of Homeland Security (DHS) is the U.S. Customs and Border Protection (CBP) agency. One of the most typical activities performed by CBP is patrolling the expansive open areas in between official points of entry (POEs) to prohibit illegal entry attempts. Roving patrols are routinely conducted by officers using many different modes of transportation depending on the terrain: foot, trucks, ATVs, boats, snowmobiles, etc. In addition to directly observing and banning illegal activity, CBP officers may conduct "sign cutting" operations which seek to identify recent signs of illegal entry (e.g., footprints or vehicle tracks). There are many factors that go into the patrolling policy for a particular area, including the type of terrain, the proximity to cities or major transportation arteries, the types of resources and infrastructure available in the area, the historical traffic pattern and apprehension rates, and so forth. The diversity and volume of illegal activity that must be controlled, and the variety of resources that can be deployed to secure the border make border security a complex mission. A key operational problem encountered by those charged with the task of border security is the scheduling and deployment of patrols. Given the complexity of the CBP patrolling problem (and other similar security/defense problems), it is clear that computational decision support tools have the potential to greatly improve resource allocation policies, and to reduce the burden on human schedulers. Therefore, this dissertation presents new methods for generating solutions to large-scale patrolling optimization problems. Three different methods are presented which can be used for different scenarios. The first method presents a new framework for finding Pareto-optimal solutions to multi-objective patrolling problems. The objectives considered in the model are the minimization of worst idleness (amount of time that each location is without protection since last visit), the minimization of the infiltration ratio (success of the attacker to cross the border), and the minimization of the total patrolling cost. The second contribution is an extension to the first multiple objective patrolling problem. The second approach presents a hybrid model that combines a multiple objective evolutionary algorithm with game theory techniques to find an optimal patrolling strategy that provides the best objective values for the three objectives considered. For these two initial approaches, the solution methods are based on evolutionary optimization techniques in the interests of scalability and therefore, a variety of Pareto-optimal patrolling strategies are presented as the output to these problems. These Pareto-optimal solutions can then be presented to human analysts for decision making, visualization, or further post-Pareto analysis. Currently, the U.S. Border Patrol operates in three operational border enforcement environments: urban, rural, and remote. Models developed to protect urban areas do not necessarily perform well when applied to the protection of remote areas. Securing remote areas presents different challenges due to a wide range of terrain conditions, large areas to be covered, limited resources available, etc. It is in this environment that the U.S. Border Patrol does not have a persistent law enforcement presence and therefore seeks to increase its situational awareness without applying the same density of resources than resources allocated to urban and rural areas. The third method presented in this dissertation describes the development of a decision support model to assists the U.S. Border Patrol with the deployment of the most appropriate CBP change detection resources available to effectively monitor the "remote" environment and to better respond to cross-border intrusions.
Aguirre, Oswaldo, "Multi-objective border patrolling optimization using game theory and evolutionary algorithms" (2014). ETD Collection for University of Texas, El Paso. AAI3682446.