Collaborative Planning with Encoding of Users’ High-Level Strategies
Authors: Joseph Kim, Christopher Banks, Julie Shah
AAAI 2017 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Through human subject experimentation, we empirically demonstrate that this approach results in statistically significant improvements to plan quality, without substantially increasing computation time. |
| Researcher Affiliation | Academia | Joseph Kim Massachusetts Institute of Technology 77 Massachusetts Avenue Cambridge, MA 02139 joseph kim@csail.mit.edu Christopher J. Banks Norfolk State University 700 Park Ave Norfolk, VA 23504 cjbanks@mit.edu Julie A. Shah Massachusetts Institute of Technology 77 Massachusetts Avenue Cambridge, MA 02139 julie a shah@csail.mit.edu |
| Pseudocode | No | The paper does not contain any pseudocode or algorithm blocks. |
| Open Source Code | No | The paper mentions using existing tools like VAL4, LPRPG-P, and OPTIC, but does not provide any statement or link for the open-source code of its own described methodology. |
| Open Datasets | Yes | We selected problems from two domains: Zenotravel and Satellite, both presented during the third IPC. We used the following benchmark problems for Zenotravel: pfile15 (propositional), pfile13 (numerical) and pfile14 (temporal). For Satellite, we used pfile12 (p), pfile5 (n), and pfile9 (t). |
| Dataset Splits | No | The paper describes using benchmark problems from the International Planning Competitions and a human planning dataset, but it does not specify explicit training, validation, or test dataset splits for these problems or data. |
| Hardware Specification | Yes | All tests were performed using an Intel Xeon Processor (2.27 GHz, 12MB Cache, 16 cores) with 16GB of RAM. |
| Software Dependencies | No | The paper mentions specific software tools such as LPRPG-P, OPTIC, and VAL4, but does not provide their version numbers. |
| Experiment Setup | Yes | We ran the automated planners (anytime planners) and report the best-quality plans generated within 30 min of CPUtime. [...] Penalty weights (λ) on encoded preferences were set to equal 20% of the estimated problem cost. |