A Risk-averse Inventory-based Supply Chain Protection Problem with Adapted Stochastic Measures under Intentional Facility Disruptions: Decomposition and Hybrid Algorithms

Abstract

Owing to rising intentional events, supply chain disruptions have been considered by setting up a game between two players, namely, a designer and an interdictor contesting on minimizing and maximizing total cost, respectively. The previous studies have found the equilibrium solution by taking transportation, penalty and restoration cost into account. To contribute further, we examine how incorporation of inventory cost influences the players' strategies. Assuming risk-averse feature of the designer and fully optimizing property of the interdictor with limited budget, the conditional-value-at-risk is employed to be involved in total cost. Using special order sets of type two and duality role, the linearized tri-level problem is solved by column-and-constraint generation and benders decomposition algorithms in terms of small-sized instances. In terms of larger-sized instances, we also contribute to prior studies by hybridizing corresponding algorithms with bio-geography based optimization method. Another non-trivial extension of our work is to define adapted stochastic measures based on the proposed mean-risk tri-level formulation. Borrowing instances from prior papers, the computational results indicate the managerial insights on players' decisions, the model's efficiency and performance of the algorithms.