Mechanism Design and Implementation for Lung Exchange
Authors: Suiqian Luo, Pingzhong Tang
IJCAI 2015 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We implement the mechanism based on the algorithms above and test it on simulated data (based on population distributions and realistic problem sizes). Each algorithm has its own advantage and both of them run sufficiently fast on these data. Our experiments show that the algorithms scale up to reasonably large instances and each algorithm has its own advantages. |
| Researcher Affiliation | Academia | Suiqian Luo and Pingzhong Tang Institute for Interdisciplinary Information Sciences, Tsinghua University, Beijing, China luosq13@mails.tsinghua.edu.cn, kenshin@tsinghua.edu.cn |
| Pseudocode | No | The paper describes algorithms (ILP and search) in prose, but does not include structured pseudocode or algorithm blocks. |
| Open Source Code | No | The paper does not provide any explicit statements about making its source code available or include links to code repositories. |
| Open Datasets | No | The paper states it uses "simulated data" based on population statistics and describes how it was generated, but does not provide access information (link, citation to public dataset) for this data. |
| Dataset Splits | No | The paper discusses |
| Hardware Specification | Yes | All our experiments are performed in Linux (open SUSE 13.1), using a PC with four 3.2GHz Intel i5-3470 processors, and 4GB of RAM. |
| Software Dependencies | Yes | We use the CPLEX12.6 software which can take benefit of multiple processors to serve as an ILP solver in our experiment. |
| Experiment Setup | Yes | Our experimental data is carefully simulated based on the statistics of US populations. Live donor lobar transplantation is especially common for these who suffer from cystic fibrosis or pulmonary hypertension [Ergin et al., 2014]. Based on this, we simulate all the patients come from these two classed with ratio 2.66 to 1, according to the statistics. The patient blood type distribution in total is 45.7% O, 40.4% A, 10.5% B, 3.4% AB. The donor blood type distribution is 44% O, 42% A, 10% B, 4% AB. We generate the lung size of each patient uniformly. Compatibility test is based on the blood type compatibility and size compatibility. Besides the donation can be between the same blood-type, blood-type O can donate to any type and blood-type AB can accept any type. The size compatibility is that a patient can only receive from a donor as heavy as himself. After the generation of patients and donors, we randomly group them into n donor-patient triples. The running time is recorded as the time actually passed instead of the total time of which the four processors use. In the search algorithm, we just use one processor to implement and run our algorithm. |