Bi-objective optimization models for network interdiction

Abstract

This paper designs models for the network interdiction problem. The interdiction problem under study has two contradicting goals: disrupting the network to minimize the profit of one set of agents, while as much as possible preserve the profit of another set of agents. Three bi-objective optimization methods are employed to form the optimal objectives. Also, we develop two formulations (MILP and multi-stage LP) used to deal with congestion cost which is a piecewise cost function. A numerical instance is also presented to better illustrate those models.