Constrained Multiagent Markov Decision Processes: a Taxonomy of Problems and Algorithms

Authors: Frits de Nijs, Erwin Walraven, Mathijs M. de Weerdt, Matthijs T. J. Spaan

JAIR 2021 | Venue PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Theoretical In this survey we conceptualize these domains and establish a generic problem class based on Markov decision processes. We identify and compare the conditions under which algorithms from the planning literature for problems in this class can be applied... Further we discuss the advantages and disadvantages of these algorithms. We conclude by identifying open problems that are directly related to the conceptualized domains, as well as in adjacent research areas.
Researcher Affiliation Academia Frits de Nijs EMAIL Dept. of Data Science and AI, Faculty of IT, Monash University 20 Exhibition Walk, 3168 Clayton, Australia, Erwin Walraven EMAIL Mathijs M. de Weerdt EMAIL Matthijs T. J. Spaan EMAIL Delft University of Technology Van Mourik Broekmanweg 6, 2628 XE Delft, The Netherlands
Pseudocode Yes Algorithm 1 Column generation for CMDP M (Yost & Washburn, 2000), Algorithm 2 Constrained point-based backup stage (Kim et al., 2011), Algorithm 3 LDD+GAPS for CMDP M (Agrawal et al., 2016)
Open Source Code Yes Furthermore, to aid in the reproducibility of past results and algorithms, we implemented a significant portion of them in an open-source Constrained Planning Toolbox, which may be found on-line.2 The final column of Table 3 indicates whether the algorithm is present in the toolbox at the time of writing the article. ... 2. The toolbox and documentation can be found in the following repository: https://github.com/Alg TUDelft/Constrained Planning Toolbox.
Open Datasets No This is a survey paper and does not present new experimental results or introduce new datasets. The paper discusses various application domains and existing algorithms, but does not provide concrete access information for any specific dataset used in its own analysis.
Dataset Splits No This is a survey paper and does not present new experimental results or use datasets for its own analysis, therefore no dataset split information is provided.
Hardware Specification No This is a survey paper focusing on the conceptualization and taxonomy of problems and algorithms. It does not present new experimental results that would require hardware specifications.
Software Dependencies No This is a survey paper and does not present new experimental results, therefore specific ancillary software dependencies with version numbers for replicating experiments are not provided.
Experiment Setup No This is a survey paper that analyzes existing problems and algorithms, and does not describe a new experimental setup with specific hyperparameters or training configurations.