Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in [1].
Optimal Policies for Observing Time Series and Related Restless Bandit Problems
Authors: Christopher R. Dance, Tomi Silander
JMLR 2019 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We discuss computation of that index, give closed-form formulae for it, and compare the performance of the associated index policy with heuristic policies. ... 4. Numerical Experiments We discuss algorithms for computing the Whittle index given in Theorem 1, we present closed-form expressions for that index and compare the performance of the Whittle index policy with two previously-proposed heuristics. ... Figure 8 compares the costs incurred by these heuristics. |
| Researcher Affiliation | Industry | Christopher R. Dance EMAIL Tomi Silander EMAIL NAVER LABS Europe 6 chemin de Maupertuis Meylan, 38240, France |
| Pseudocode | No | The paper does not contain any explicit pseudocode blocks or algorithms. It primarily presents mathematical derivations, theorems, and proofs. |
| Open Source Code | No | The paper does not provide a direct link to source code or an explicit statement indicating that the code for the described methodology is publicly available. While it mentions previous work and algorithms, it does not offer access to the implementation used in this paper. |
| Open Datasets | No | The numerical experiments described in Section 4.4 utilize a 'simple scenario' with parameters like 'n = 10 projects' and 'initial variance is xi,0 = 4', indicating a simulated setup rather than a publicly available dataset. No specific link, DOI, or formal citation for a public dataset used in their experiments is provided. |
| Dataset Splits | No | The paper describes a simulated scenario for its numerical experiments in Section 4.4, but it does not mention any training, testing, or validation dataset splits. Such splits are typically not relevant for the type of theoretical and simulation-based analysis presented. |
| Hardware Specification | No | The paper does not provide any specific details about the hardware used to conduct the numerical experiments, such as CPU or GPU models, or memory specifications. |
| Software Dependencies | No | The paper does not list specific software dependencies with version numbers (e.g., programming languages, libraries, or scientific computing packages) that were used for implementation or analysis. |
| Experiment Setup | Yes | In detail, there are n = 10 projects, the number of observations per round is m = 1, the cost at time t is 40x1,t + Pn i=2 xi,t, and the initial variance is xi,0 = 4 for all projects. The discount factor is β = 0.99 and the total cost for a given method is computed as P199 t=0 βt(40x1,t + Pn i=2 xi,t). The projects have the same map-with-a-gap given by φ0(x) = x + 1 and φ1(x) = 1/(0.1 + 1/(x + 1)). |