Planning with Hidden Parameter Polynomial MDPs

Authors: Clarissa Costen, Marc Rigter, Bruno Lacerda, Nick Hawes

AAAI 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We evaluate two domains and empirically show that the polynomial, closed-form, belief representation results in better plans than a sampling-based belief representation.
Researcher Affiliation Academia Clarissa Costen, Marc Rigter, Bruno Lacerda, Nick Hawes Oxford Robotics Institute, University of Oxford {clarissa, mrigter, bruno, nickh}@robots.ox.ac.uk
Pseudocode Yes Algorithm 1: Bayes-Adaptive Monte Carlo Planning
Open Source Code No The paper does not provide an explicit statement or link to the open-source code for the methodology described.
Open Datasets Yes The forecast for the longitudinal and latitudinal velocity of the water is known ahead of time, using data from Copernicus (2022).
Dataset Splits No The paper describes experimental simulations and sampling of true parameters, but it does not specify explicit training, validation, or test dataset splits for a fixed dataset, as the data is generated through interaction with the simulated environments.
Hardware Specification Yes We ran the simulations on Linux OS with an Intel Core i9-11900H processor and 16GB of RAM. We ran the trials on Cent OS with an Intel Xeon Platinum 8268 CPU at 2.90 GHz Processor, with 16GB of RAM.
Software Dependencies No The paper does not provide specific software names with version numbers (e.g., programming languages, libraries, or solvers with their versions) that would be needed to replicate the experiment.
Experiment Setup Yes We set T to 20 seconds. We also varied the size of the set of samples used to represent the posterior distribution, M, for BAMCP-G and -PF. applied dimension reduction to HP2-MDP, where the highest degree of the polynomial was fixed to 12.