Computing the Strategy to Commit to in Polymatrix Games
Authors: Giuseppe De Nittis, Alberto Marchesi, Nicola Gatti
AAAI 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Experimental evaluation. We ran Algorithm 1 on a testbed of OLTPGs, evaluating the running time as a function of the number of players n and the number of actions per player m. Specifically, for each pair pn, mq, times are averaged over 20 game instances, with n P t3, . . . , 10u and m P t4, 6, . . . 10, 15, . . . , 70u. Game instances have been randomly generated, with each payoff uniformly and independently drawn from the interval r0, 100s. All experiments are run on a UNIX machine with a total of 32 cores working at 2.3 GHz, and equipped with 128 GB of RAM. Each game instance is solved on a single core, within a time limit of 7200 seconds. The algorithm is implemented in Python 2.7, while all LP programs are solved with GUROBI 7.0, using the Python interface. Figure 3 contains two plots of the average computing times, as a function of n and m, respectively. |
| Researcher Affiliation | Academia | Giuseppe De Nittis, Alberto Marchesi, Nicola Gatti Politecnico di Milano Piazza Leonardo da Vinci, 32 Milano, Italy {giuseppe.denittis, alberto.marchesi, nicola.gatti}@polimi.it |
| Pseudocode | Yes | Algorithm 1 Exact-P-LFE |
| Open Source Code | No | The paper states the algorithm is implemented in Python 2.7 and uses GUROBI 7.0, but it does not provide an explicit statement about releasing the source code for their methodology or a link to a code repository. |
| Open Datasets | No | Game instances have been randomly generated, with each payoff uniformly and independently drawn from the interval r0, 100s. |
| Dataset Splits | No | The paper describes generating game instances and evaluating algorithm running time, but it does not specify training, validation, or test dataset splits as typical in machine learning experiments. |
| Hardware Specification | Yes | All experiments are run on a UNIX machine with a total of 32 cores working at 2.3 GHz, and equipped with 128 GB of RAM. |
| Software Dependencies | Yes | The algorithm is implemented in Python 2.7, while all LP programs are solved with GUROBI 7.0, using the Python interface. |
| Experiment Setup | Yes | Game instances have been randomly generated, with each payoff uniformly and independently drawn from the interval r0, 100s. Each game instance is solved on a single core, within a time limit of 7200 seconds. |