All-pairs nearly 2-approximate shortest-paths in O ( n 2 polylog n ) time
read more
Citations
Fast approximation algorithms for the diameter and radius of sparse graphs
A simple and linear time randomized algorithm for computing sparse spanners in weighted graphs
Better approximation algorithms for the graph diameter
Approximate distance oracles for unweighted graphs in expected O(n2) time
Faster Algorithms for Approximate Distance Oracles and All-Pairs Small Stretch Paths
References
Gaussian elimination is not optimal
Matrix multiplication via arithmetic progressions
Storing a Sparse Table with 0(1) Worst Case Access Time
Approximate distance oracles
Approximate distance oracles
Related Papers (5)
Frequently Asked Questions (15)
Q2. How can the authors calculate all-pairs in O(n2) time?
Given an undirected unweighted graph G(V, E) on |V | = n vertices, the authors can compute all-pairs 3-approximate distances in O(n2) time.
Q3. What is the stretch factor for all-pairs shortest path?
But the stretch factor may be quite huge for short paths since β depends on ζ as (1/ζ)log 1/ζ , depends inverse exponentially on ρ and inverse polynomially on .
Q4. How can the authors improve the time for a BFS tree?
To improve its preprocessing time to (n2 polylog n), one idea is to perform BFS from R on a spanner (having o(n2) edges) of the original graph.
Q5. How many steps can be used to improve the preprocessing time of a graph?
The preprocessing time of the first two steps in Algorithm The authordescribed above can be bounded by O(n2 log n) with a suitable choice of p.
Q6. What is the new scheme for distance?
For a vertex u and a set R ⊂ V of vertices, Ball(u, R) denotes the set of vertices of the graph, such that the distance from u to these vertices is less than the distance from u to the nearest vertex of the set R.New scheme for approximate distanceLet R ⊂ V be a set of vertices.
Q7. What is the shortest path between u and v?
Case 1: Ball(u, R) = {u} Let v0(= u), v1, · · · , vl(= v) be the shortest path between u and v. Since Ball(u, R) consists of vertex u only, u must be adjacent to nu.
Q8. What is the known algorithm for all-pairs shortest path?
In its most generic version, that is, for directed graph with real edge-weights, the best known algorithm [6] for this problem requires O(mn+n2 log log n) time.
Q9. What is the length of the path Pwu excluding the edge e(u′?
Now the part of the path Pwu excluding the edge e(u′, u) is of length x−1, and can’t be stretched to more than 3(x−1) in the spanner.
Q10. how many times does the first iteration require expected O(m + n/p?
Add a uniform sample of size np from V ′ to R; For every u ∈ V \\R doCompute Ball(u, R); For every v ∈ V \\R doC(v, R)← {u ∈ V | v ∈ Ball(u, R)}; V ′ ← {v ∈ V | |C(v, R)| > 4/p};Return R; }For the first iteration, R is a uniform sample from V , that is, R = Rp. So using Theorem 1, the first iteration requires expected O(m + n/p2) time.
Q11. how many steps are there to report the distance between u and v?
1. Now considering the path from nu to v passing through u, the authors can observe that δ(nu, v) is bounded by 2a + x + b +1. Similarly, analyzing the path from nv to u passing through v, the authors can observe that δ(nv , u) is bounded by 2b + x + a + 1. Therefore, the distance reported by Q(u, v) is bounded as follows.
Q12. How many units of time can be used to improve the running time of a graph?
in the worst case, the authors have been able to improve the running time by a factor of O(n1/6) at the expense of introducing an additive error of just one unit.
Q13. How is the distance between u and v calculated?
In the final and the third case, when the two Balls are non-overlapping, the authors use the global distance information stored at nu and nv.Lemma 1. Given a graph G(V, E) and any two vertices u, v ∈ V , the approximate distance between u and v as reported by the query procedure Q(u, v) is bounded by 2δ(u, v) + 1.Proof.
Q14. What is the known upper bound on the time complexity of this problem?
The best known upper bound on the time complexity of this problem is O(n3 √ log log n/ log n) due to Zwick [10], which is marginally sub-cubic.
Q15. How can the authors use the BFS tree to report distances?
Without any modifications, all their data-structures for reporting approximate distances can also be used to report approximate shortest-paths in optimal time.