cutLPK

A scalable linear programming based algorithm for K-Means clustering

Dependencies

The project relies on the following libraries:

• cuPDLP: A high-performance CUDA-based LP solver. cuPDLP GitHub Repository
• Eigen: A C++ template library for linear algebra.
• OpenMP: An API for parallel programming in C, C++, and Fortran.

Building the Project

The project uses CMake for configuration and building. Below are the steps to build the project:

1. Navigate to the project directory:

   cd path/to/cutLPK

2. Open the config.cmake file in a text editor and specify the dependency paths according to your system setup.

3. Create and enter a build directory:

   mkdir build
   cd build

4. Configure the project with CMake:

   cmake ..

5. Build the project:

   cmake --build

Example Usage

./cutLPK iris.csv 3

Parameters

The following parameters can be set for the algorithm. To change a parameter value, use the format -param_name=param_value.

Parameter	Default Value	Description
solver	"cupdlp"	The LP solver to use
output_file	"logFile.txt"	Output file for logs
output_level	1	Level of output detail
random_seed	1	Seed for random number generation
max_cuts_init	1.5e7	Maximum number of initial cuts
max_cuts_added_iter	3e7	Maximum number of violated cuts per iteration
max_separation_size	1.5e7	Maximum separation size
warm_start	2	Warm start option(1 is deterministic choose the initial cuts and 2 is randomly sampling)
t_upper_bound	K	Upper bound for t (K is the number of clusters)
time_limit_lp	180	Time limit for LP in each cutting plane iteration (seconds)
time_limit_all	14400	Overall time limit (seconds)
solver_tolerance_per_iter	1e-4	Solver tolerance per iteration
cuts_vio_tol	1e-4	Cut violation tolerance
cuts_act_tol	1e-4	Cut activation tolerance
opt_gap	1e-4	Optimality gap

Acknowledgments

This project includes a copy of the cuPDLP-C solver. We have made slight modifications to the termination check to accommodate early termination when solving linear programming problems within our algorithm framework. We recommend using the version of cuPDLP-C included in this project for optimal performance of the algorithm.