Ivan Contreras

Stochastic Uncapacitated Hub Location

We study stochastic uncapacitated hub location problems in which uncertainty is associated to demands and transportation costs. We show that the stochastic problems with uncertain demands or dependent transportation costs are equivalent to their associated deterministic expected value problem (EVP), in which random variables are replaced by their expectations. In the case of uncertain independent transportation costs, the corresponding stochastic problem is not equivalent to its EVP and specific solution methods need to be developed. We describe a Monte-Carlo simulation-based algorithm that integrates a sample average approximation scheme with a Benders decomposition algorithm to solve problems having stochastic independent transportation costs. Numerical results on a set of instances with up to 50 nodes are reported.

The Tree of Hubs Location Problem

This paper presents the Tree of Hubs Location Problem. It is a network hub location problem with single assignment where a fixed number of hubs have to be located, with the particularity that it is required that the hubs are connected by means of a tree. The problem combines several aspects of location, network design and routing problems. Potential applications appear in telecommunications and transportation systems, when set-up costs for links between hubs are so high that full interconnection between hub nodes is prohibitive. We propose an integer programming formulation for the problem. Furthermore, we present some families of valid inequalities that reinforce the formulation and we give an exact separation procedure for them. Finally, we present computational results using the well-known AP and CAB data sets.

Benders Decomposition for Large-Scale Uncapacitated Hub Location

This paper describes an exact algorithm capable of solving large-scale instances of the well-known uncapacitated hub location problem with multiple assignments. The algorithm applies Benders decomposition to a strong path-based formulation of the problem. The standard decomposition algorithm is enhanced through the inclusion of several features such as the use of a multicut reformulation, the generation of strong optimality cuts, the integration of reduction tests, and the execution of a heuristic procedure. Extensive computational experiments were performed to evaluate the efficiency and robustness of the algorithm. Computational results obtained on classical benchmark instances (with up to 200 nodes) and on a new and more difficult set of instances (with up to 500 nodes) confirm the efficiency of the algorithm.

General network design: A unified view of combined location and network design problems

This paper presents a unified framework for the general network design problem which encompasses several classical problems involving combined location and network design decisions. In some of these problems the service demand relates users and facilities, whereas in other cases the service demand relates pairs of users between them, and facilities are used to consolidate and re-route flows between users. Problems of this type arise in the design of transportation and telecommunication systems and include well-known problems such as location-network design problems, hub location problems, extensive facility location problems, tree-star location problems and cycle-star location problems, among others. Relevant modeling aspects, alternative formulations and possible algorithmic strategies are presented and analyzed.

The Dynamic Uncapacitated Hub Location Problem

This paper presents a dynamic (or multi-period) hub location problem. It proposes a branch-and-bound algorithm that uses a Lagrangian relaxation to obtain lower and upper bounds at the nodes of the tree. The Lagrangian function exploits the structure of the problem and can be decomposed into smaller subproblems that can be solved efficiently. In addition, some reduction procedures based on the Lagrangian bounds are implemented. These yield a considerable reduction of the size of the problem and thus help reduce the computational burden. Numerical results on a set of instances with up to 100 nodes and 10 time periods are reported.

Tight bounds from a path based formulation for the tree of hub location problem

This paper considers the tree of hub location problem. We propose a four index formulation which yields much tighter LP bounds than previously proposed formulations, although at a considerable increase of the computational burden when obtained with a commercial solver. For this reason we propose a Lagrangean relaxation, based on the four index formulation, that exploits the structure of the problem by decomposing it into independent subproblems which can be solved quite efficiently. We also obtain upper bounds by means of a simple heuristic that is applied at the inner iterations of the method that solves the Lagrangean dual. As a consequence, the proposed Lagrangean relaxation produces tight upper and lower bounds and enable us to address instances up to 100 nodes, which are notably larger than the ones previously considered in the literature, with sizes up to 20 nodes. Computational experiments have been performed with benchmark instances from the literature. The obtained results are remarkable. For most of the tested instances we obtain or improve the best known solution and for all tested instances the deviation between our upper and lower bounds, never exceeds 10%.

Lagrangean relaxation for the capacitated hub location problem with single assignment

This article considers the capacitated hub location problem with single assignment. We propose a Lagrangean relaxation to obtain tight upper and lower bounds. The Lagrangean function that we formulate exploits the structure of the problem and can be decomposed into smaller subproblems that can be solved efficiently. In addition, we present some simple reduction tests, based on the Lagrangean relaxation bounds that allows us to reduce considerably the size of the formulation and thus, to reduce the computational effort. Computational experiments have been performed with both benchmark instances from literature and with some new larger instances. The obtained results are impressive. For all tested instances (ranging from 10 to 200 nodes), we obtain or improve the best known solution and the obtained duality gaps, between our upper and lower bounds, never exceed 3.4%.

Branch and Price for Large-Scale Capacitated Hub Location Problems with Single Assignment

This paper presents a branch-and-price algorithm for the capacitated hub location problem with single assignment, in which Lagrangean relaxation is used to obtain tight lower bounds of the restricted master problem. A lower bound that is valid at any stage of the column generation algorithm is proposed. The process to obtain this valid lower bound is combined with a constrained stabilization method that results in a considerable improvement on the overall efficiency of the solution algorithm. Numerical results on a battery of benchmark instances of up to 200 nodes are reported. These seem to be the largest instances that have been solved to optimality for this problem.

Exact Solution of Large-Scale Hub Location Problems with Multiple Capacity Levels

This paper presents an extension of the classical capacitated hub location problem with multiple assignments in which the amount of capacity installed at the hubs is part of the decision process. An exact algorithm based on a Benders decomposition of a strong path-based formulation is proposed to solve large-scale instances of two variants of the problem: the splittable and nonsplittable commodities cases. The standard decomposition algorithm is enhanced through the inclusion of features such as the generation of strong optimality cuts and the integration of reduction tests. Given that in the nonsplittable case the resulting subproblem is an integer program, we develop an efficient enumeration algorithm. Extensive computational experiments are performed to evaluate the efficiency and robustness of the proposed algorithms. Computational results obtained on benchmark instances with up to 300 nodes and five capacity levels confirm their efficiency.

Hub Location Problems

Hub Location Problems (HLPs) lie at the heart of network design planning in transportation and telecommunication systems. They are a challenging class of optimization problems that focus on the location of hub facilities and on the design of hub networks. This chapter overviews the key distinguishing features, assumptions and properties commonly considered in HLPs. We highlight the role location and network design decisions play in the formulation and solution of HLPs. We also provide a concise overview of the main developments and most recent trends in hub location research. We cover various topics such as hub network topologies, flow dependent discounted costs, capacitated models, uncertainty, dynamic and multi-modal models, and competition and collaboration. We also include a summary of the most successful integer programming formulations and efficient algorithms that have been recently developed for the solution of HLPs.