Direct-search methods for decentralized blackbox optimization

Derivative-free optimization algorithms are particularly useful for tackling blackbox optimization problems where the objective function arises from complex and expensive procedures that preclude the use of classical gradient-based methods. In contemporary decentralized environments, such functions are defined locally on different computational nodes due to technical or privacy constraints, introducing additional challenges within the optimization process. In this paper, we adapt direct-search methods, a classical technique in derivative-free optimization, to the decentralized setting. In contrast with zeroth-order algorithms, our algorithms rely on positive spanning sets to define suitable search directions, while still possessing global convergence guarantees thanks to carefully chosen stepsizes. Numerical experiments highlight the advantages of direct-search techniques over gradient-approximation-based strategies.

• Publication year

  2025

• Venue

  Unknown venue

• Publication date

  2025-04-05

• Fields of study

  Mathematics, Computer Science

• Identifiers
  arXiv 2504.04269
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• Adaptive and Communication-Efficient Zeroth-Order Optimization for Distributed Internet of Things

  2024cited by this paper
• Direct-search methods in the year 2025: Theoretical guarantees and algorithmic paradigms

  2024cited by this paper
• When Decentralized Optimization Meets Federated Learning

  2023cited by this paper
• Adaptive Consensus: A Network Pruning Approach for Decentralized Optimization

  2023cited by this paper
• Decentralized gradient tracking with local steps

  2023cited by this paper
• Distributed Zero-Order Algorithms for Nonconvex Multiagent Optimization

  2021cited by this paper
• Communication-Efficient Decentralized Zeroth-order Method on Heterogeneous Data

  2021cited by this paper
• Decentralized Zeroth-Order Constrained Stochastic Optimization Algorithms: Frank–Wolfe and Variants With Applications to Black-Box Adversarial Attacks

  2020cited by this paper
• Distributed Zeroth Order Optimization Over Random Networks: A Kiefer-Wolfowitz Stochastic Approximation Approach

  2018cited by this paper
• ZONE: Zeroth-Order Nonconvex Multiagent Optimization Over Networks

  2017influential reference
• Derivative-Free and Blackbox Optimization

  2017cited by this paper
• Distributed nonconvex multiagent optimization over time-varying networks

  2016cited by this paper
• On Nonconvex Decentralized Gradient Descent

  2016influential reference
• NEXT: In-Network Nonconvex Optimization

  2016cited by this paper
• Direct Search Based on Probabilistic Descent

  2015cited by this paper
• On the Convergence of Decentralized Gradient Descent

  2013influential reference
• Worst case complexity of direct search

  2013cited by this paper
• Stochastic First- and Zeroth-Order Methods for Nonconvex Stochastic Programming

  2013cited by this paper
• Derivative free optimization methods

  2012cited by this paper
• Introduction to derivative-free optimization

  2010cited by this paper
• Distributed Subgradient Methods for Multi-Agent Optimization

  2009cited by this paper
• Benchmarking Derivative-Free Optimization Algorithms

  2009influential reference
• Asynchronous Parallel Generating Set Search for Linearly Constrained Optimization

  2008cited by this paper
• Revisiting Asynchronous Parallel Pattern Search for Nonlinear Optimization

  2005cited by this paper
• Optimization by Direct Search: New Perspectives on Some Classical and Modern Methods

  2003influential reference
• Asynchronous Parallel Pattern Search for Nonlinear Optimization

  2001cited by this paper

• No citing papers are available for this paper.