Solving Transition-Independent Multi-Agent MDPs with Sparse Interactions
Authors: Joris Scharpff, Diederik Roijers, Frans Oliehoek, Matthijs Spaan, Mathijs de Weerdt
AAAI 2016 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | In our experiments we find optimal policies for the maintenance planning problem (MPP, see the introduction) that minimise the (time-dependent) maintenance costs and economic losses due to traffic hindrance. Using this domain we conduct three experiments with Co Re to study 1) the expected value when solving centrally versus decentralised methods, 2) the impact on the number of joint actions evaluated and 3) the scalability in terms of agents. |
| Researcher Affiliation | Academia | Joris Scharpff Delft University of Technology, The Netherlands Diederik M. Roijers University of Amsterdam, The Netherlands Frans A. Oliehoek University of Amsterdam, The Netherlands University of Liverpool, United Kingdom Matthijs T. J. Spaan and Mathijs M. de Weerdt Delft University of Technology, The Netherlands |
| Pseudocode | Yes | Algorithm 1: Co Re(Φ, θN t , h, N) |
| Open Source Code | No | No statement regarding the release of source code or a link to a code repository was found. |
| Open Datasets | No | The paper describes generating its own problem instances and test sets (e.g., 'rand[h]', 'coordint', 'mpp', 'pyra') but does not provide concrete access information (link, DOI, specific repository, or formal citation with authors/year) for a publicly available or open dataset. |
| Dataset Splits | No | The paper describes the generation of problem instances and test sets (e.g., 'mpp', 'rand[h]'), but does not provide specific details on how these were split into training, validation, or testing sets (e.g., percentages, sample counts, or citations to predefined splits). |
| Hardware Specification | No | No specific hardware details (e.g., exact GPU/CPU models, processor types with speeds, or memory amounts) used for running the experiments were provided. |
| Software Dependencies | No | The paper mentions other algorithms like SPUDD and GMAA-ICE*, but does not specify any software dependencies (e.g., library or solver names with version numbers) used for their own implementation. |
| Experiment Setup | No | The paper describes the problem domain and characteristics of the generated test instances (e.g., 'planning horizons 5 to 10', 'random delay probabilities'), but it does not provide specific experimental setup details such as hyperparameter values, optimizer settings, or other concrete training configurations for the Co Re algorithm. |