Efficient Computation of Emergent Equilibrium in Agent-Based Simulation
Authors: Zehong Hu, Meng Sha, Moath Jarrah, Jie Zhang, Hui Xi
AAAI 2016 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | The experimental results show that our framework outperformed Monte Carlo experiments in terms of computation efficiency while maintaining the accuracy. |
| Researcher Affiliation | Collaboration | 1Rolls-Royce@NTU Corporate Lab, Nanyang Technological University, Singapore 2School of Computer Engineering, Nanyang Technological University, Singapore 3Rolls-Royce Singapore Pte Ltd, Singapore |
| Pseudocode | Yes | Algorithm 1: Macro-Micro Bridge; Algorithm 2: Pseudo-Arclength Equilibrium Solver |
| Open Source Code | No | The paper does not provide an explicit statement about releasing source code or a link to a code repository. |
| Open Datasets | No | The paper evaluates the framework using two agent-based simulation models (opinion dynamics and labour market models) which are simulated rather than relying on external publicly available datasets with access information. Therefore, no public dataset access is provided. |
| Dataset Splits | No | The paper uses Monte Carlo experiments for comparison and specifies parameters for its simulations, but it does not describe specific training, validation, or test dataset splits in a traditional machine learning sense, as it relies on simulations rather than pre-existing datasets. |
| Hardware Specification | Yes | In executing our experiment, we use MATLAB 2013a running on Intel Xeon CPU E5-1650, 16 Gi B RAM and Ubuntu 14.04LTS operating system. |
| Software Dependencies | No | The paper mentions 'MATLAB 2013a' and 'Ubuntu 14.04LTS operating system' as part of the experimental setup, and 'Sparse Grid Interpolation Toolbox' is mentioned as used, but no specific version numbers are provided for these or other libraries/packages. |
| Experiment Setup | Yes | The total number of agents is set to 5000. The macroscopic step δt is chosen as 6000 microscopic simulation steps, and the number of microscopic realizations is set to 10. The confidence bound ϵ is chosen as 0.5. For the labour market model: The total number of firms and works is set to be 200 and 1000, respectively. The macroscopic time step δt is selected as 50 microscopic simulation steps, and the number of microscopic realizations is set to 50. ρ = 0.80 and ρ = 0.81 are used as the two starting points. δs is set to be 0.1. |