Privacy-Preserving Policy Iteration for Decentralized POMDPs
Authors: Feng Wu, Shlomo Zilberstein, Xiaoping Chen
AAAI 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Our experimental results on several common Dec-POMDP benchmark problems confirm the effectiveness of our approach. We implemented our algorithm and tested it on 6 common benchmark problems2 for Dec-POMDPs (i.e., Dec-Tiger, Broadcast Channel, Meeting in a 3 3 Grid, Box Pushing, Recycling Robots, and Mars Rovers). For each problem instance, we first ran our algorithm to generate policies and then evaluated the policies by simulation. |
| Researcher Affiliation | Academia | Feng Wu, Shlomo Zilberstein, Xiaoping Chen School of Computer Science and Technology, University of Science and Technology of China, CHN College of Information and Computer Sciences, University of Massachusetts Amherst, USA wufeng02@ustc.edu.cn, shlomo@cs.umass.edu, xpchen@ustc.edu.cn |
| Pseudocode | Yes | Algorithm 1: Secure Value Estimation. Algorithm 2: Secure Policy Improvement. |
| Open Source Code | No | The paper does not provide any explicit statement or link indicating that the source code for the described methodology is publicly available. |
| Open Datasets | Yes | We implemented our algorithm and tested it on 6 common benchmark problems2 for Dec-POMDPs (i.e., Dec-Tiger, Broadcast Channel, Meeting in a 3 3 Grid, Box Pushing, Recycling Robots, and Mars Rovers). ... 2http://masplan.org/problem domains |
| Dataset Splits | No | The paper mentions running 'trials' and selecting 'Nb samples' but does not specify explicit training/validation/test dataset splits, percentages, or sample counts. |
| Hardware Specification | No | The paper mentions 'a machine with 40 cores' but does not provide specific hardware details such as CPU models, GPU types, or memory specifications. |
| Software Dependencies | No | The paper mentions 'the common Java implementation of the Paillier cryptosystem' but does not provide specific version numbers for Java or the cryptosystem library. |
| Experiment Setup | Yes | In the algorithm, we set the number of trials N = 1000 and the size of best trials Nb = 10. |