Iterative Linearized Control: Stable Algorithms and Complexity Guarantees

Authors: Vincent Roulet, Siddhartha Srinivasa, Dmitriy Drusvyatskiy, Zaid Harchaoui

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

Reproducibility Variable Result LLM Response
Research Type Experimental 5. Experiments We illustrate the performance of the algorithms considered including the proposed accelerated regularized Gauss Newton algorithm on two classical problems drawn from Li & Todorov (2004): swing-up a pendulum, and move a twolinks robot arm.
Researcher Affiliation Academia 1Department of Statistics, University of Washington, Seattle, USA 2Paul G. Allen School of Computer Science & Engineering, University of Washington, Seattle, USA 3Department of Mathematics, University of Washington, Seattle, USA.
Pseudocode Yes Algorithm 1 Accelerated Regularized Gauss-Newton
Open Source Code Yes The code for this project is available at https://github.com/vroulet/ilqc.
Open Datasets No The paper refers to 'inverted pendulum' and 'twolinks arm' as classical problems drawn from Li & Todorov (2004), but it does not provide concrete access information (link, DOI, repository, or specific citation for a dataset file) for any publicly available or open dataset used in the experiments.
Dataset Splits No The paper does not provide specific dataset split information (exact percentages, sample counts, citations to predefined splits, or detailed splitting methodology) for training, validation, or testing.
Hardware Specification No The paper does not provide specific hardware details (exact GPU/CPU models, processor types, or memory amounts) used for running its experiments.
Software Dependencies No The paper mentions Py Torch and Tensor Flow but does not specify their version numbers or any other key software components with specific versions needed for replication.
Experiment Setup Yes For ILQR, we use an Armijo line-search to compute the next step. For both the regularized ILQR and the accelerated regularized ILQR, we use a constant step-size sequence tuned after a burn-in phase of 5 iterations.