1. https://arxiv.org/abs/1906.01227
2. https://github.com/chaitjo/graph-convnet-tsp

Commands to run

1. Activate Environment or create from requirements.txt

source {environment name}/bin/activate

2. Various parameters are listed in makefile, run commmand

make parameter_1={value_1} parameter_2={value_2{} .. 

Reults will be formed in the result_phase2_rl folder

Sample Dataset

A sample SBM graph of 50 nodes is given in the data folder. The structure of input graph is as follows:

<Graph_name> //input to the model
    <test_set> 
        <1>...<n>
    <val_set> 
        <1>...<n>
    <train_set> 
        <1>...<n>
// Each folder in train/val/test set should have a graph.txt and graph_stats.txt

Running Baselines

Code in Baselines folder

HMETIS/Spectral Clustering

1. Run a convert.py which takes a graph folder which contains graph.txt ans graph_stats.txt

python3 convert.py data/sample_graph/test_set/1/

2. Now,you can use visualise_cuts scripts to get values of various metrics using the cut formed

GAP

1. In Baseline run make gap