RAO*: An Algorithm for Chance-Constrained POMDP’s
Authors: Pedro Rodrigues Quemel e Assis Santana, Sylvie Thiébaux, Brian Williams
AAAI 2016 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We demonstrate the usefulness of RAO in two challenging domains of practical interest: power supply restoration and autonomous science agents. ... This section provides empirical evidence of the usefulness and general applicability of CC-POMDP s as modeling tool for risk-sensitive applications, and shows how RAO performs when computing risk-bounded policies in two challenging domains of practical interest: automated planning for science agents (SA) ... and power supply restoration (PSR) ... All models and RAO were implemented in Python and ran on an Intel Core i7-2630QM CPU with 8GB of RAM. Our SA domain... The PSR domain... We evaluated the performance of RAO in both domains under various conditions, and the results are summarized in Tables 1 (higher utility is better) and 2 (lower cost is better). |
| Researcher Affiliation | Collaboration | Massachusetts Institute of Technology, CSAIL +The Australian National University & NICTA 32 Vassar St., Room 32-224, Cambridge, MA 02139 Canberra ACT 0200, Australia {psantana,williams}@mit.edu Sylvie.Thiebaux@anu.edu.au |
| Pseudocode | Yes | Algorithm 1 RAO, Algorithm 2 expand-policy, Algorithm 3 update-policy |
| Open Source Code | No | The paper does not provide any explicit statements about making its source code available or links to a code repository. |
| Open Datasets | Yes | Our SA domain is based on the planetary rover scenario described in (Benazera et al. 2005). In the PSR domain (Thi ebaux and Cordier 2001), the objective is to reconfigure a faulty power network by switching lines on or off so as to resupply as many customers as possible. |
| Dataset Splits | No | The paper discusses the problem formulation and policy execution horizon but does not provide specific details on training, validation, or test dataset splits (e.g., percentages or sample counts). |
| Hardware Specification | Yes | All models and RAO were implemented in Python and ran on an Intel Core i7-2630QM CPU with 8GB of RAM. |
| Software Dependencies | No | The paper states that the models and RAO were "implemented in Python" but does not specify the Python version or any other software dependencies with version numbers. |
| Experiment Setup | No | The paper mentions experimental conditions like "various time windows and risk levels" for SA and "various numbers of faults and risk levels" for PSR. However, it does not provide specific experimental setup details such as hyperparameters (e.g., learning rates, batch sizes), model initialization, or specific training configurations typical for machine learning experiments. |