snl

Sensor Network Localization using a Decentralized Proximal Splitting Method

To run, you will need to clone the development branch of the OARS repository from https://github.com/peterbarkley/oars.git

git clone --branch development https://github.com/peterbarkley/oars.git

Then install OARS by running the following command:

cd oars
pip install -e .

Then you can install the remainder of the project's dependencies by running the following command:

pip install -r requirements.txt

I used the MOSEK Fusion API for reference solutions - you can choose another solver instead by commenting out solve_snl_fusion and uncommenting solve_snl_vec in the python scripts.

You will need to have latex installed in order to write the charts as they are given in the paper. Otherwise you can just comment out the plt.rcParams['text.usetex'] = True line in each of the python scripts.

If you want to run the size comparison, you will need mpi installed, and you will need to run the script on a node with at least 251 cores, or reduce the maximum size in the script.

You can run the project by running the following command:

./paper.sh

Data Files

The 00_alpha_sweep script generates the errors_data.csv file. This can be plotted with 00_plot_alphas.py The 01_mpps_admm_comparison script generates the mpps_admm_comparison.csv file. The 03_matrix_design_time script generates the design_time.csv file. The 05_termination script generates the term_output.csv file. This can be plotted with 05_plot_term.py The 06_size_comparison script generates the experiment_results.h5 file. This can be plotted with 06_plot_size.py

Charts

00_plot_alphas.py generates fig_0_alphas_vs_error_all_alg_log.pdf. 01_mpps_admm_comparison.py generates fig_1_mpps_admm_comparison.pdf and fig_1_mpps_admm_comparison_warm.pdf. 02_matrix_design_and_centrality.py generates fig_2_sk_oars_error.pdf and fig_3_centroid_distance.pdf. 03_matrix_design_time.py generates fig_2_design_time.pdf. 04_point_plot.py generates fig_3_point_plot_dist.pdf. 05_plot_term.py generates fig_4_paired_diffs_mean.pdf and fig_4_early_termination_density_mean.pdf. 06_plot_size.py generates size_comparison_grid_mean.pdf and size_grid_times_mean.pdf.