Collaborative Research: Dynamic Resource Allocation Models for Law Enforcement Operations Against Illegal Drug Trafficking

这项合作研究拨款研究了新的优化模型，用于确定执法部门为阻止非法毒品网络而做出的资源分配决策。这些决定的重点是安排执法活动，以便成功地监控、打击和逮捕非法毒品网络中的犯罪分子。这些决策将使用一个新的动态拦截框架和一个马尔可夫决策过程框架来建模。拟议的模型特别纳入了平衡情报行动和物理封锁的需要，前者有助于识别新的罪犯，并针对已知的高级罪犯立案，后者将罪犯从网络中清除。基于约束规划的优化算法将被提出并集成到传统的整数规划技术中，以利用整个网络中与识别、瞄准和逮捕罪犯相关的调度决策的结构来求解所提出的模型。如果成功，本研究的结果将提供工具，帮助更好地理解如何利用稀缺的执法资源来减少非法毒品贩运。这包括更好地了解：(1)如何利用各个城市的国家机构(如禁毒署)的资源与当地执法部门合作；(2)如何最好地使目前受到执法部门监视的罪犯组合多样化；(3)罪犯之间的信息流动如何影响拦截工作。这一政策驱动型分析的结果将通过与当地执法机构的合作与从业者分享。该项目将使一年级的工程学学生参与尖端研究，并通过在能够利用基于分解的解决方案方法的超级计算资源上实施专门的算法来发展他们的高性能计算技能。

This collaborative research grant investigates new optimization models for determining resource allocation decisions of law enforcement to interdict illegal drug networks. These decisions focus on scheduling the activities of law enforcement in order to successfully monitor, target, and arrest criminals in an illegal drug network. These decisions will be modeled using a novel dynamic interdiction framework and a Markov Decision Process framework. The proposed models specifically incorporate the need to balance intelligence operations, which help to identify new criminals and build cases against known high-ranking criminals, and physical interdictions, which remove criminals from the network. Specialized constraint programming-based optimization algorithms will be proposed and integrated into traditional integer programming techniques in order to exploit the structure of the scheduling decisions associated with identifying, targeting, and arresting criminals throughout the network to solve the proposed models.If successful, the results of this research will provide tools that help to better understand how to utilize scarce law enforcement resources to mitigate illegal drug trafficking. This includes a better understanding of: (i) how to utilize resources of national agencies (e.g., the Drug Enforcement Agency) across various cities in partnering with local law enforcement, (ii) how to best diversify the portfolio of criminals that are currently under surveillance by law enforcement, and (iii) how the flow of information between criminals impacts interdiction efforts. The results of this policy-driven analysis will be shared with practitioners through collaborations with local law enforcement agencies. The project will engage freshmen engineering students in cutting-edge research and develop their high-performance computing skills by implementing the specialized algorithms on supercomputing resources capable of exploiting decomposition-based solution approaches.