Decentralized Mean Field Games
Authors: Sriram Ganapathi Subramanian, Matthew E. Taylor, Mark Crowley, Pascal Poupart9439-9447
AAAI 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | In this section we study the performance of our algorithms. The code for experiments has been open-sourced (Subramanian 2021). We provide the important elements of our domains here, while the complete details are in Appendix I. The first five domains belong to the MAgent environment (Zheng et al. 2018). |
| Researcher Affiliation | Academia | 1 University of Waterloo, Waterloo, Ontario, Canada 2 University of Alberta, Edmonton, Alberta, Canada 3 Vector Institute, Toronto, Ontario, Canada 4 Alberta Machine Intelligence Institute (Amii), Edmonton, Alberta, Canada |
| Pseudocode | Yes | Algorithm 1: Q-learning for DMFG |
| Open Source Code | Yes | The code for experiments has been open-sourced (Subramanian 2021). ... Subramanian, S. G. 2021. Decentralized Mean Field Games. https://github.com/Sriram94/DMFG. |
| Open Datasets | Yes | In our environment, ride requests come from the open source New York Yellow Taxi dataset (NYYellow Taxi 2016). The road network (represented as a grid with a finite set of nodes or road intersections) contains a simulated set of vehicles (agents) that aim to serve the user requests. Further details about this domain are in Appendix I. ... NYYellow Taxi. 2016. New York yellow taxi dataset. http:// www.nyc.gov/html/tlc/html/about/triprecorddata.shtml. |
| Dataset Splits | No | The paper does not explicitly provide information on validation dataset splits. It only mentions 'training data' and a 'test set'. |
| Hardware Specification | No | The paper mentions support from 'Compute Canada' and various institutes, implying computational resources were used, but it does not specify any exact hardware details such as CPU/GPU models or memory. |
| Software Dependencies | No | The paper mentions the use of neural networks, PPO, DDPG, and DQN algorithms but does not provide specific version numbers for any software libraries, frameworks, or dependencies. |
| Experiment Setup | Yes | Detailed description of the algorithms are in Appendix H (see Algs. 2 and 3). A complexity analysis is in Appendix K, and hyperparameter details are in Appendix J. |