Barrier Algorithms for Constrained Non-Convex Optimization

Authors: Pavel Dvurechensky, Mathias Staudigl

ICML 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Theoretical Our main goal is theoretical and we show that barrier algorithms possess favorable complexity beyond convexity.
Researcher Affiliation Academia 1Weierstrass Institute for Applied Analysis and Stochastics, Berlin, Germany 2University of Mannheim, Mannheim, Germany.
Pseudocode Yes Algorithm 1: First-Order Adaptive Barrier Method FOABM(µ, ε, L0, x0); Algorithm 2: Second-Order Adaptive Barrier Method SOABM(µ, ε, M0, x0)
Open Source Code No The paper does not provide any statement or link indicating that open-source code for the described methodology is available.
Open Datasets No The paper explicitly states, "We do not perform numerical experiments for two reasons. 1) Our main goal is theoretical and we show that barrier algorithms possess favorable complexity beyond convexity. 2) We are not aware of any baseline algorithms with similar complexity that can accommodate with general set constraints, and produce feasible iterates without involving a projection step." As such, no datasets were used for training.
Dataset Splits No The paper explicitly states, "We do not perform numerical experiments for two reasons. 1) Our main goal is theoretical and we show that barrier algorithms possess favorable complexity beyond convexity. 2) We are not aware of any baseline algorithms with similar complexity that can accommodate with general set constraints, and produce feasible iterates without involving a projection step." As such, no datasets were used for validation.
Hardware Specification No The paper explicitly states, "We do not perform numerical experiments for two reasons. 1) Our main goal is theoretical and we show that barrier algorithms possess favorable complexity beyond convexity. 2) We are not aware of any baseline algorithms with similar complexity that can accommodate with general set constraints, and produce feasible iterates without involving a projection step." Therefore, no hardware specifications are mentioned for experiments.
Software Dependencies No The paper explicitly states, "We do not perform numerical experiments for two reasons. 1) Our main goal is theoretical and we show that barrier algorithms possess favorable complexity beyond convexity. 2) We are not aware of any baseline algorithms with similar complexity that can accommodate with general set constraints, and produce feasible iterates without involving a projection step." Therefore, no software dependencies with version numbers are mentioned for experiments.
Experiment Setup No The paper explicitly states, "We do not perform numerical experiments for two reasons. 1) Our main goal is theoretical and we show that barrier algorithms possess favorable complexity beyond convexity. 2) We are not aware of any baseline algorithms with similar complexity that can accommodate with general set constraints, and produce feasible iterates without involving a projection step." Therefore, no experimental setup details are provided.