A Simple Yet Efficient Rank One Update for Covariance Matrix Adaptation

2017/10/11 by Zhenhua Li, Qingfu Zhang, Li, Zhenhua +1
Computer Science · #FOS: Computer and information sciences #Face and Expression Recognition #Image and Signal Denoising Methods #Neural Networks and Applications #Neural and Evolutionary Computing (cs.NE)

paper · pdf · doi:10.48550/arxiv.1710.03996

openalex publication_date 2017/10/11 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

In this paper, we propose an efficient approximated rank one update for covariance matrix adaptation evolution strategy (CMA-ES). It makes use of two evolution paths as simple as that of CMA-ES, while avoiding the computational matrix decomposition. We analyze the algorithms' properties and behaviors. We experimentally study the proposed algorithm's performances. It generally outperforms or performs competitively to the Cholesky CMA-ES.

Related

• Natural Gradient Interpretation of Rank-One Update in CMA-ES
• Covariance Matrix Adaptation Evolution Strategy Assisted by Principal Component Analysis
• Diagonal Acceleration for Covariance Matrix Adaptation Evolution Strategies
• Covariance Matrix Adaptation Evolution Strategy for Low Effective Dimensionality
• CMA-ES with Two-Point Step-Size Adaptation