Sharing information in all pairs shortest path algorithms

Authors:
Tadao Takaoka;Mashitoh Hashim
Affiliations:
University of Canterbury, Christchurch, New Zealand;University of Canterbury, Christchurch, New Zealand
Venue:
CATS 2011 Proceedings of the Seventeenth Computing on The Australasian Theory Symposium - Volume 119
Year:
2011

Citing 10
Cited 0

Fibonacci heaps and their uses in improved network optimization algorithms

Journal of the ACM (JACM)
Faster algorithms for the shortest path problem

Journal of the ACM (JACM)
Finding the hidden path: time bounds for all-pairs shortest paths

SIAM Journal on Computing
Shortest path algorithms for nearly acyclic directed graphs

Theoretical Computer Science - Special issue: graph theoretic concepts in computer science
A data structure for manipulating priority queues

Communications of the ACM
Improved shortest path algorithms for nearly acyclic graphs

Theoretical Computer Science
Integer priority queues with decrease key in constant time and the single source shortest paths problem

Proceedings of the thirty-fifth annual ACM symposium on Theory of computing
A new approach to all-pairs shortest paths on real-weighted graphs

Theoretical Computer Science - Special issue on automata, languages and programming
All-pairs shortest paths for unweighted undirected graphs in o(mn) time

SODA '06 Proceedings of the seventeenth annual ACM-SIAM symposium on Discrete algorithm
Solving shortest paths efficiently on nearly acyclic directed graphs

Theoretical Computer Science

Quantified Score

Hi-index	0.00

Visualization

Abstract

We show two improvements on time complexities of the all pairs shortest path (APSP) problem for directed graphs that satisfy certain properties. The idea for speed-up is information sharing by n single source shortest path (SSSP) problems that are solved for APSP. We consider two parameters, in addition to the numbers of vertices, n, and edges, m. First we improve the time complexity of O(mn + n2 √log c) to O(mn + nc) for the APSP problem with the integer edge costs bounded by c. When c ≤ O(n√log n), this complexity is better than the previous one. Next we consider a nearly acyclic graph. We measure the degree of acyclicity by the size, r, of a given set of feedback vertices. If r is small, the given graph can be considered to be nearly acyclic. We improve the existing time complexity of O(mn + r3) for the all pairs shortest path problem to O(mn + rn log n) by some kind of information sharing. This complexity is better than the previous one for all values of r under a reasonable assumption of m ≥ n.