Title :
Shortest Path Tree Computation in Dynamic Graphs
Author :
Chan, Edward P F ; Yang, Yaya
Author_Institution :
Sch. of Comput. Sci., Univ. of Waterloo, Waterloo, ON
fDate :
4/1/2009 12:00:00 AM
Abstract :
Let G = (V, E, omega) be a simple digraph, in which all edge weights are nonnegative real numbers. Let G´ be obtained from G by an application of a set of edge weight updates to G. Let sisinV and let Ts and Ts´ be Shortest Path Trees (SPTs) rooted at s in G and G´, respectively. The Dynamic Shortest Path (DSP) problem is to compute Ts´ from Ts. Existing work on this problem focuses on either a single edge weight change or multiple edge weight changes in which some of them are incorrect or are not optimized. We correct and extend a few state-of-the-art dynamic SPT algorithms to handle multiple edge weight updates. We prove that these algorithms are correct. Dynamic algorithms may not outperform static algorithms all the time. To evaluate the proposed dynamic algorithms, we compare them with the well-known static Dijkstra algorithm. Extensive experiments are conducted with both real-life and artificial data sets. The experimental results suggest the most appropriate algorithms to be used under different circumstances.
Keywords :
directed graphs; number theory; set theory; tree data structures; trees (mathematics); digraph; dynamic graph; multiple edge weight update; nonnegative real number; set theory; shortest path tree computation; Data mining; Digital signal processing; Distance measurement; Graph theory; Heuristic algorithms; Optimization; Proposals; Dynamic; Dynamic Algorithms; Dynamic shortest path; Fully- and Semi-Dynamic Algorithms.; Graphs; Shortest Path; Shortest Path Trees; Shortest Paths; dynamic algorithms; dynamic graphs; graph algorithms; routing protocol.; shortest path trees;
Journal_Title :
Computers, IEEE Transactions on
DOI :
10.1109/TC.2008.198