On the approximation of the minimum disturbance p-facility location problem
Discrete Applied Mathematics - Special issue: Third ALIO-EURO meeting on applied combinatorial optimization
Applying Bundle Methods to the Optimization of Polyhedral Functions: An Applications-Oriented Development
A Local-Search-Based Heuristic for the Demand-Constrained Multidimensional Knapsack Problem
INFORMS Journal on Computing
A Local-Search-Based Heuristic for the Demand-Constrained Multidimensional Knapsack Problem
INFORMS Journal on Computing
On the remoteness function in median graphs
Discrete Applied Mathematics
The multi-depot capacitated location-routing problem with fuzzy travel times
Expert Systems with Applications: An International Journal
Capacitated location-routing problem with time windows under uncertainty
Knowledge-Based Systems
Hi-index | 0.00 |
The problem of simultaneously locating obnoxious facilities and routing obnoxious materials between a set of built-up areas and the facilities is addressed.Obnoxious facilities are those facilities which cause exposure to people as well as to the environment i.e. dump sites, chemical industrial plants, electric power supplier networks, nuclear reactors and so on. A discrete combined location-routing model, which we refer to as Obnoxious Facility Location and Routing model (OFLR), is defined. OFLR is a NP-hard problem for which a Lagrangean heuristic approach is presented. The Lagrangean relaxation proposed allows to decompose OFLR into a Location subproblem and a Routing subproblem; such subproblems are then strengthened by adding suitable inequalities. Based on this Lagrangean relaxation two simple Lagrangean heuristics are provided. An effective Branch and Bound algorithm is then presented, which aims at reducing the gap between the above mentioned lower and upper bounds. Our Branch and Bound exploits the information gathered while going down in the enumeration tree in order to solve efficiently the subproblems related to other nodes. This is accomplished by using a bundle method to solve at each node the Lagrangean dual. Some variants of the proposed Branch and Bound method are defined in order to identify the best strategy for different classes of instances. A comparison of computational results relative to these variants is presented.