Using Sensors to Interdict Nuclear Material Smuggling

摘要

We describe a stochastic network interdiction model for locating sensors that detect nuclear material. Arnnuclear material smuggler selects a path through a transportation network that maximizes the probabilityrnof avoiding detection. An interdictor installs sensors to minimize that maximum probability. We formulaternthis problem as a bi-level stochastic mixed-integer program, and then focus on a special case in which thernunderlying network is bipartite. We show that a class of valid inequalities, called step inequalities, canrnsignificantly reduce computational effort.