Computing k-shortest path for nuclear spent fuel highway transportation

Abstract

Four k-shortest path algorithms were implemented to select a set of optimal routes for transporting nuclear spent fuel on the U.S. Interstate highway system. Origin-destination pairs drawn from a selection of six nuclear reactors and two potential repository sites across the U.S. were used for evaluation of the computational performance of each algorithm. The first two algorithms maintained a priority queue under two different management systems, a linear search and a binary heap. The third was an adaptive algorithm which exploited the knowledge of the Euclidean distances between nodes to reduce search effort. The relative effectiveness of these algorithms in determining the k-shortest path was examined in terms of computation time. Finally, a simple (or loopless) path algorithm, which required only a minor modification of the original k-shortest path algorithm, was introduced for situations in which a large set of simple paths was needed.