Shortest paths algorithms: theory and experimental evaluation
Mathematical Programming: Series A and B
Communications of the ACM
Using Multi-level Graphs for Timetable Information in Railway Systems
ALENEX '02 Revised Papers from the 4th International Workshop on Algorithm Engineering and Experiments
Shortest Path Algorithms: An Evaluation Using Real Road Networks
Transportation Science
Routing Optimization for Waste Management
Interfaces
Computing the shortest path: A search meets graph theory
SODA '05 Proceedings of the sixteenth annual ACM-SIAM symposium on Discrete algorithms
Waste collection vehicle routing problem with time windows
Computers and Operations Research
A Multiple Pairs Shortest Path Algorithm
Transportation Science
Highway hierarchies hasten exact shortest path queries
ESA'05 Proceedings of the 13th annual European conference on Algorithms
| Hi-index | 0.00 |
This paper considers the generation of the origin-destination (OD) matrix, basic data in any vehicle routing or traveling salesman problem. An OD matrix must be generated by calculating the shortest paths between some nodes. Candidate methods for this are repetitive use of one-to-all shortest path algorithms such as Dijkstra's algorithm, use of all-to-all shortest path algorithms such as the Floyd-Warshall algorithm, and use of specifically designed some-to-some shortest path algorithms. This paper compares the performance of several shortest path algorithms in computing OD matrices on real road networks. Dijkstra's algorithm with approximate bucket data structure performed the best for most of the networks tested. This paper also proposes a grouping-based algorithm for OD matrix generation. Although it is an approximation approach, it has several good characteristics: it can find the exact shortest distances for most OD pairs; it guarantees that all the calculated shortest path distance values have corresponding paths; the deviation of any distance from the corresponding true shortest distance is small; and its computation time is short.