Probabilistic Planning with Reduced Models
Authors: Luis Pineda
AAAI 2014 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We evaluated our continual planning paradigm on the well-known racetrack domain and showed that M1 l reductions can be used to quickly compute near-optimal plans. The goal was to minimize the combined cost of planning and execution time, accounting for 1 second of execution time per action. Figure 1 (left) shows the relative increase in expected combined cost with respect to a theoretical lower bound optimal cost ignoring planning time for 6 planning methods. The best results were obtained using M1 1- and M1 2-reductions (M11 and M12, respectively). Additionally, using an initial version of a PPDDL compiler, we applied our approach to several IPPC 08 domains using the IPC-style of evaluation: giving the planner a fixed amount of time to solve several rounds of the same problem. Figure 1 (right) shows that a planner using M1 1-reductions successfully solves many problem instances, results that are on par with those reported for state-of-the-art planners in these domains (Trevizan and Veloso 2012). |
| Researcher Affiliation | Academia | Luis Pineda School of Computer Science University of Massachusetts Amherst, MA 01003, USA lpineda@cs.umass.edu |
| Pseudocode | No | The paper describes algorithms and methods verbally but does not include any structured pseudocode or algorithm blocks. |
| Open Source Code | No | The paper mentions 'using an initial version of a PPDDL compiler' and states 'my plan is to expand the support of the currently limited PPDDL compiler' but does not provide any concrete access (link or explicit statement of public release) to the source code for the methodology described. |
| Open Datasets | Yes | We evaluated our continual planning paradigm on the well-known racetrack domain... Additionally, using an initial version of a PPDDL compiler, we applied our approach to several IPPC 08 domains using the IPC-style of evaluation... |
| Dataset Splits | No | The paper refers to evaluations on specific domains (racetrack, IPPC 08) but does not provide specific details on how the datasets were split into training, validation, or test sets, nor does it mention cross-validation. |
| Hardware Specification | No | The paper does not provide any specific details about the hardware (e.g., CPU, GPU models, memory, or cloud instances) used for running the experiments. |
| Software Dependencies | No | The paper mentions a 'PPDDL compiler' and refers to algorithms like 'LAO*' and 'LRTDP' and solvers like 'FF-Replan' but does not specify any version numbers for these software components or libraries. |
| Experiment Setup | Yes | The goal was to minimize the combined cost of planning and execution time, accounting for 1 second of execution time per action... giving the planner a fixed amount of time to solve several rounds of the same problem. |