No abstract

In recent decades, the issue of interdiction in human-made infrastructures has become increasingly important. The world, after 9/11, recognized the importance of addressing major infrastructures vulnerabilities and protecting them against intentional disruptions. Networks such as transportation networks or electrical grids are one of the most important network infrastructures in which any disruption can cause irreparable costs and customer service declination. Therefore, the phenomenon of interdiction and fortification of network components against possible attacks such as terrorist acts or military operations has been of great importance in the field of research. In this research we develop a tri-level model to obtain optimal edge protect plan for existing network against deliberate attacks on the capacitated vehicle routing problem (CVRP). The vehicle routing problem is one of the most important issues in the distribution and transportation field, where a fleet of vehicles tries to meet demand points in the least possible cost. Protecting edges from deliberate disturbances in this issue has not been addressed in the literature so far. The problem is strictly NP-Hard and we employ the heuristic approaches to solve it. The main idea behind our approach is to iteratively sample the tri-level solution space so that instead of solving the CVRP. This approach called backward sampling framework. We also introduce a Nested Genetic Algorithm (NGA) to accelerate algorithm implementation which numerical experiments prove the effectiveness of this approach in solving the problem.