Three New Algorithms to Solve N-POMDPs

Authors: Yann Dujardin, Tom Dietterich, Iadine Chads

AAAI 2017 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Complexity results are provided for each of the algorithms, and they are tested on well-known benchmarks.
Researcher Affiliation Collaboration Yann Dujardin CSIRO yann.dujardin@csiro.au Tom Dietterich School of EECS Oregon State University tgd@oregonstate.edu Iadine Chad es CSIRO iadine.chades@csiro.au
Pseudocode Yes Algorithm 1 α-min-2-fast
Open Source Code No The paper does not provide an explicit statement or link for the open-source code of the described methodology.
Open Datasets Yes We conducted experiments on three benchmark problems1 given an initial policy Γ provided by SARSOP and the corresponding lower bound V (b0), where b0 the initial belief (at t = 0). ... 1http://www.pomdp.org/examples/
Dataset Splits No The paper uses benchmark problems but does not explicitly provide specific dataset split information (e.g., percentages, sample counts, or citations to predefined splits) for reproducing data partitioning.
Hardware Specification No The paper does not provide specific hardware details (e.g., exact GPU/CPU models, processor types, or memory amounts) used for running its experiments.
Software Dependencies No The paper mentions 'SARSOP' as an existing POMDP solver and refers to 'mixed integer linear programs' but does not specify software names with version numbers for replication.
Experiment Setup Yes The time-horizon chosen for the experiments was T = 10. The number of β-points for αmin-2-solve was 100.