Finding Optimal Solutions in HTN Planning - A SAT-based Approach

Authors: Gregor Behnke, Daniel Höller, Susanne Biundo

IJCAI 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We implemented the given techniques within the PANDA planning framework using its implementation of a propositional encoding for HTN planning [Behnke et al., 2019a; Behnke et al., 2018b; Behnke et al., 2018a]. The code can be downloaded at https://www.uni-ulm.de/en/in/ki/panda/. We use the 144 instances from the most recent evaluations of satisficing HTN planners [Behnke et al., 2019a; H oller et al., 2018]. HTN2STRIPS max PB was only run on the 109 tail-recursive instances, as it can only find optimal solution on them. Each planner was given 10 minutes of runtime and 4 GB of RAM on an Intel E5-2660. For the propositional encoding, we used three SAT solvers: cryptominisat (cms) 5.5 [Soos, 2018], exp MV [Chowdhury et al., 2018], and Maple LCM [Ryvchin and Nadel, 2018], some of the best performing SAT solvers in the 2018 SAT race. Results The number of optimally solved planning problems for each planner is shown in Tab. 1. The runtime behaviour is shown in Fig. 2.
Researcher Affiliation Academia Gregor Behnke , Daniel H oller and Susanne Biundo Institute of Artificial Intelligence, Ulm University, Ulm, Germany
Pseudocode Yes Algorithm 1 Calculate K4(ℓ) base algorithm; Algorithm 2 Calculate K4(ℓ)
Open Source Code Yes The code can be downloaded at https://www.uni-ulm.de/en/in/ki/panda/.
Open Datasets Yes We use the 144 instances from the most recent evaluations of satisficing HTN planners [Behnke et al., 2019a; H oller et al., 2018].
Dataset Splits No The paper mentions using '144 instances' from benchmark evaluations but does not provide specific details on how these instances are split into training, validation, or test sets for reproduction, nor does it specify cross-validation.
Hardware Specification Yes Each planner was given 10 minutes of runtime and 4 GB of RAM on an Intel E5-2660.
Software Dependencies Yes For the propositional encoding, we used three SAT solvers: cryptominisat (cms) 5.5 [Soos, 2018], exp MV [Chowdhury et al., 2018], and Maple LCM [Ryvchin and Nadel, 2018], some of the best performing SAT solvers in the 2018 SAT race.
Experiment Setup Yes Each planner was given 10 minutes of runtime and 4 GB of RAM on an Intel E5-2660.