Tractability of Planning with Loops
Authors: Siddharth Srivastava, Shlomo Zilberstein, Abhishek Gupta, Pieter Abbeel, Stuart Russell
AAAI 2015 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We demonstrate our techniques on several planning problems and show that they apply to challenging real-world tasks such as doing the laundry with a PR2 robot. These results resolve a number of open questions about planning with loops and facilitate the development of new algorithms and applications. |
| Researcher Affiliation | Academia | Siddharth Srivastava1 Shlomo Zilberstein2 Abhishek Gupta1 Pieter Abbeel1 Stuart Russell1 1 Computer Science Division, University of California, Berkeley, CA 94720 2 School of Computer Science, University of Massachusetts, Amherst, MA 01003 |
| Pseudocode | Yes | Algorithm 1: (Progress-Sieve) abstract policy termination test |
| Open Source Code | No | A video of the PR2 robot doing laundry using this approach is available at http://tiny.cc/laundrybot. (This link is for a video, not source code.) |
| Open Datasets | Yes | We experimented with versions of the mining problem from the settlers domain by adding various items to the production chain (e.g., tools can be produced when there is a sufficient amount of iron), as well as the problems considered by (Srivastava et al. 2011b) and those by (Bonet, Palacios, and Geffner 2009) that have memoryless solutions. |
| Dataset Splits | No | No specific dataset split information (percentages, sample counts, or detailed splitting methodology) is provided. |
| Hardware Specification | Yes | All the problems were solved in at most 1s (on a 1.7GHz Intel Core i5 Mac) |
| Software Dependencies | No | The implementation is in Java and uses FF (Hoffmann and Nebel 2001; Hoffmann 2003) as the classical planner. (Specific version numbers for Java or FF are not provided.) |
| Experiment Setup | Yes | We used a Kinect sensor to obtain RGBD data and Open CV s contour generation based on color masks to detect clothes as objects with a specific range of colors (close to red). |