koallen / edge-disjoint-partitioning

Reproducing the SIGMOD paper "Graph Indexing for Shortest-Path Finding over Dynamic Sub-Graphs"

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Edge Disjoint Partitioning (EDP)

Reproducing the SIGMOD paper "Graph Indexing for Shortest-Path Finding over Dynamic Sub-Graphs"

Compilation

This code is written in C++, make sure you have a C++ compiler (e.g. g++, clang++, etc.).

To compile the program, just go into src directory and do

$ make

It will generate 2 executables, edp_main is used for the experiments while edp_test is for debugging.

Running experiments

After compilation, just do (inside src)

$ ./edp_main /path/to/graph <number-of-labels> <number-of-vertices> > output.txt

This writes the index size after each query to the file output.txt. For example,

$ ./edp_main ../datasets/youtube_subset.txt 5 2000 > output.txt

Datasets

Below are some details for the dataset we used.

Subset of Youtube

  • Directed graph
  • |V|: 1940
  • |E|: 290234
  • Number of labels: 5

Subset of Biogrid

  • Undirected graph
  • |V| = 8783
  • |E| = 199009
  • Number of labels = 7

String

  • Unidrected graph
  • |V| = 4458
  • |E| = 161367
  • Number of labels = 6

About

Reproducing the SIGMOD paper "Graph Indexing for Shortest-Path Finding over Dynamic Sub-Graphs"


Languages

Language:C++ 88.7%Language:Gnuplot 8.5%Language:Makefile 1.8%Language:Shell 1.0%