A Model-Free Approach for Solving Choice-Based Competitive Facility Location Problems Using Simulation and Submodularity

This archive is distributed in association with the INFORMS Journal on Computing under the MIT License.

The software and data in this repository are a snapshot of the software and data that were used in the research reported on in the paper A Model-Free Approach for Solving Choice-Based Competitive Facility Location Problems Using Simulation and Submodularity by Robin Legault and Emma Frejinger.

Cite

To cite the contents of this repository, please cite both the paper and this repo, using their respective DOIs.

https://doi.org/10.1287/ijoc.2023.0280

https://doi.org/10.1287/ijoc.2023.0280.cd

Below is the BibTex for citing this snapshot of the repository.

@misc{legault2024model,
  author =        {Legault, Robin and Frejinger, Emma},
  publisher =     {INFORMS Journal on Computing},
  title =         {A Model-Free Approach for Solving Choice-Based Competitive Facility Location Problems Using Simulation and Submodularity},
  year =          {2024},
  doi =           {10.1287/ijoc.2023.0280.cd},
  note =          {Available for download at https://github.com/INFORMSJoC/2023.0280},
}  

Description

This software implements a partial Benders decomposition method for solving choice-based competitive facility location problems.

This repository is made of the following directories and subdirectories

1. src: implementation of the algorithm and other necessary source code
2. data: instances used to generate results in the paper
3. scripts: scripts used to replicate the experiments in the paper
4. results: raw results and figures reported in the paper

Building

Requirements

Please see Project.toml for the packages versions and Manifest.toml for a complete snapshot of the Julia 1.8.5 environment used by the authors. The commercial solver CPLEX is required.

Results

The detailed results of our experiments are provided in folder results/results_paper. Text files reproducing the content of the tables of Section 6 and Appendix E based on these results can be generated and stored in folder results by running the following scripts.

scripts/tables_MNL.jl
scripts/tables_MMNL.jl

Replicating

To replicate the experiments of Section 6.1 (Conditional MNL Instances), run the following script:

scripts/run_experiments_MNL.jl

To replicate the experiments of Section 6.2 (Generative MMNL Instances), first generate the instances by running the following script:

scripts/generate_instances_MMNL.jl

then run the following script:

scripts/run_experiments_MMNL.jl

The detailed results will be stored in folder results.

Support

For support in using this software, submit an issue.