Fast Optimal Clearing of Capped-Chain Barter Exchanges

Authors: Benjamin Plaut, John Dickerson, Tuomas Sandholm

AAAI 2016 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental On that real data and demographically-accurate generated data, our new solver scales significantly better than the prior leading approaches.
Researcher Affiliation Academia Benjamin Plaut Carnegie Mellon University bplaut@andrew.cmu.edu John P. Dickerson Carnegie Mellon University dickerson@cs.cmu.edu Tuomas Sandholm Carnegie Mellon University sandholm@cs.cmu.edu
Pseudocode No The paper describes algorithmic concepts like Bellman-Ford and branch-and-price methods in prose, but it does not include any structured pseudocode blocks or algorithm listings.
Open Source Code No The paper does not provide any specific links to source code repositories or explicit statements about the release of their implementation's code.
Open Datasets Yes We first test on real data from the United Network for Organ Sharing (UNOS) nationwide kidney exchange, which now contains 143 transplant centers... We also used demographically-accurate generated data, sampled from the set of all pairs and altruists who had entered the UNOS exchange by Nov. 2014.
Dataset Splits No The paper describes experiments on 'real UNOS match runs' and 'Generated UNOS data' but does not specify any training, validation, or test dataset splits, percentages, or cross-validation methodologies.
Hardware Specification No On each problem instance, each solver was given access to 28GB of RAM, 4 cores, and 60 minutes of wall time.
Software Dependencies No The paper mentions types of solvers (e.g., 'integer program solver', 'branch and price') but does not specify any particular software names with version numbers for dependencies or the solvers used in their experiments.
Experiment Setup Yes The cycle cap was set to 3, as is almost ubiquitous in practice (also at UNOS). We varied the chain cap. On each problem instance, each solver was given access to 28GB of RAM, 4 cores, and 60 minutes of wall time.