Bi-objective evolutionary approach to the design of patrolling schemes for improved border security

Patrolling is vital to law enforcement missions and public safety, as patrollers are the most visible entities to perpetrators and first responders whenever there is a breach in security. Factors such as the size and geographical diversity of patrolling zones increase the complexity of protecting national borders, creating a need for advanced techniques to design adequate patrolling schemes that help patrol units prevent and deter potential suspicious activities. This study focuses on optimizing the deployment of personnel to patrol designated areas known for illicit cross-border activities. A multi-objective heuristic approach is proposed to design patrolling schemes with the intention to minimize vulnerability and cost. An evolutionary algorithm is used to find solutions, and the second-order stochastic dominance (SSD) approach is used to rank those solutions in order to obtain the approximate Pareto set of "pseudo-optimal" solutions, which are characterized by the routes followed by patrol units. Illustrative examples are provided to discuss how the model is applied. (C) 2017 Elsevier Ltd. All rights reserved.