Robust-Adaptive Control of Linear Systems: beyond Quadratic Costs

Authors: Edouard Leurent, Odalric-Ambrym Maillard, Denis Efimov

NeurIPS 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Finally, in Section 6 we demonstrate the applicability of Algorithm 1 in two numerical experiments: a simple illustrative example and a more challenging simulation for safe autonomous driving. We show in Table 1(a) the results of 100 simulations of a single episode: the robust agent performs worse than the nominal agent on average, but manages to ensure safety while the nominal agent collides with obstacles in 4% of simulations.
Researcher Affiliation Collaboration Edouard Leurent Univ. Lille, Inria, CNRS, Centrale Lille, UMR 9189 CRISt AL, Renault F-59000 Lille, France edouard.leurent@inria.fr
Pseudocode Yes Algorithm 1 Robust Estimation, Prediction and Control
Open Source Code Yes Code and videos available at https://eleurent.github.io/robust-beyond-quadratic/.
Open Datasets Yes We consider the highway-env environment [25] for simulated driving decision problems. [25] Leurent, E. An environment for autonomous driving decision-making. https://github.com/ eleurent/highway-env, 2018.
Dataset Splits No The paper does not provide specific percentages, counts, or explicit methodology for training/validation/test dataset splits.
Hardware Specification No The paper does not provide specific hardware details (e.g., exact GPU/CPU models, memory amounts) used for running its experiments.
Software Dependencies No The paper does not provide specific ancillary software details (e.g., library or solver names with version numbers) needed to replicate the experiment.
Experiment Setup No The paper describes general experimental settings like "K planning iterations" and "3000 episodes" for DQN, but does not provide specific hyperparameter values or detailed training configurations for the models.