Much has been written on shortest path problems with weight, or resource, constraints. However, relatively little of it has provided systematic computational comparisons for a representative selection of algorithms. Furthermore, there has been almost no work showing numerical performance of scaling algorithms, although worst-case complexity guarantees for these are well known, nor has the effectiveness of simple preprocessing techniques been fully demonstrated. Here, we provide a computational comparison of three scaling techniques and a standard label-setting method. We also describe preprocessing techniques which take full advantage of cost and upper-bound information that can be obtained from simple shortest path information. We show that integrating information obtained in preprocessing within the label-setting method can lead to very substantial improvements in both memory required and run time, in some cases, by orders of magnitude. Finally, we show how the performance of the label-setting method can be further improved by making use of all Lagrange multiplier information collected in a Lagrangean relaxation first step.
A label setting algorithm for solving the Elementary Resource Constrained Shortest Path Problem, using node resources to forbid repetition of nodes on the path, is implemented. A state-space augmenting approach for accelerating run times is considered. Several augmentation strategies are suggested and compared numerically.
Abstract. In this paper we present a study of several Ant Colony Optimization (ACO) algorithms for the Set Covering Problem. In our computational study we emphasize the influence of different ways of defining the heuristic information on the performance of the ACO algorithms. Finally, we show that the best performing ACO algorithms we implemented, when combined with a fine-tuned local search procedure, reach excellent performance on a set of well known benchmark instances.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.