Cutting Plane Selection with Analytic Centers and Multiregression - Density Filtering Plots

<p>This repository contains the detailed plots for the density filtering experiments of the paper &quot;Cutting Plane Selection with Analytic Centers and Multiregression&quot;.<br> The data were generated by the experiments available on the associated [repository](https://github.com/Opt-Mucca/Analytic-Center-Cut-Selection).</p> <p>In the following naming conventions, `X` stands for density filtering at X%.<br> The plots `root_direct_filter_X.pdf` show the relative primal-dual difference at the end of the separation at the root node for each instance-seed pair.<br> The plots `root_avg_filter_X.pdf` show the relative primal-dual difference at the end of the separation at the root node for each instance with all seeds averaged with a geometric mean.<br> The plots `root_mindev_filter_X.pdf` show the relative primal-dual difference at the end of the separation at the root node for each instance with the minimum deviation between the baseline and filtered method.</p> <p>In the following naming conventions, `Y` stands for the quantity of interest including:<br> - `lp_iterations`: number of LP iterations to solve the problem (only when solved to optimality)<br> - `node`: number of nodes taken to solve the problem (only when solved to optimality)<br> - `time`: time taken to solve the problem<br> - `node_throughput`: number of nodes processed per second<br> - `lpiter_throughput`: number of LP iterations performed per second</p> <p>The plots `tree_direct_filter_Y_rel_X.pdf` show the relative primal-dual difference at the end of the separation at the root node for each instance-seed pair.<br> The plots `tree_agg_filter_Y_rel_X.pdf` show the relative primal-dual difference at the end of the separation at the root node for each instance with seeds aggregated with a geometric mean.</p> <p>&nbsp;</p>