Decentralized Noncooperative Games with Coupled Decision-Dependent Distributions
Authors: Wenjing YAN, Xuanyu Cao
NeurIPS 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Numerical experiments further confirm the effectiveness of our algorithm and theoretical results. |
| Researcher Affiliation | Academia | Wenjing Yan Xuanyu Cao Department of Electronic and Computer Engineering The Hong Kong University of Science and Technology wj.yan@connect.ust.hk, eexcao@ust.hk |
| Pseudocode | Yes | Algorithm 1 Decentralized Stochastic Primal-Dual Algorithm: The Procedures at Player i, i [n]: |
| Open Source Code | No | Does the paper provide open access to the data and code, with sufficient instructions to faithfully reproduce the main experimental results, as described in supplemental material? Answer: [No] |
| Open Datasets | Yes | The simulation setup is based on dataset from a prior Kaggle competition.2 Our study focuses on three ride-share platforms (Uber, Lyft, and Via) and eight competing areas within New York.2The data is publicly available at https://www.kaggle.com/brllrb/uber-and-lyft-dataset-boston-ma |
| Dataset Splits | No | The paper describes simulation parameters and settings for the Cournot game and ride-share market experiments, including initial conditions and distributions for various parameters, but it does not specify explicit train/validation/test dataset splits in the conventional machine learning sense for data partitioning. |
| Hardware Specification | No | The paper describes numerical experiments conducted for a networked Cournot game and a ride-share market but does not specify any particular hardware details such as GPU models, CPU types, or memory used for these experiments. |
| Software Dependencies | No | The paper describes its algorithms and numerical experiments but does not list any specific software dependencies or libraries with version numbers required to reproduce the experiments. |
| Experiment Setup | Yes | In the simulation, we set n = 5 and m = 3. The production capacity Qi is randomly and uniformly drawn from [10, 12] for all i [5], and the market s capacity Bj is randomly and uniformly drawn from [10, 15] for all j [m]. All entries in di, i [n] are randomly and uniformly drawn from [1, 1.5]. The distribution of ξo j is set to min(max(N(2.5, 1), 2.5), 7.5). The performative power αj is randomly and uniformly drawn from ( 1, 0], for all j [3]. Other settings are: Λj = 10, cj = 10, dj = 5 and τj = 2, j [3]. The simulation setup is based on dataset from a prior Kaggle competition. Our study focuses on three ride-share platforms (Uber, Lyft, and Via) and eight competing areas within New York. We randomly and uniformly assign the total number of rides, Qi, from the range [200, 400] for each platform i [3]. Similarly, the accommodated capacity, Bj, is randomly and uniformly drawn from [50, 150] for all j [8]. All entries in di, i [n] are randomly and uniformly drawn from [0.2, 2.2]. The distribution of ξo j is set as min(max(N(1, 1), 1), 5). Additionally, we set the following values for all areas j [8]: Λj = 5, cj = 5, dj = 5, and τj = 2. |