用于线性约束下优化的协方差矩阵自适应进化策略
神经与进化计算
2018-09-24 v2 最优化与控制
摘要
本文针对带线性约束的优化问题，提出了一种协方差矩阵自适应进化策略（CMSA-ES）。所提算法称为线性约束CMSA-ES（lcCMSA-ES）。它使用一种特殊构造的变异算子并结合基于投影的修复来满足约束。lcCMSA-ES在由约束定义的线性流形上自行演化。目标函数仅在可行搜索点处求值（内点法）。这一性质在仿真优化与有限元方法等应用领域常常被要求。该算法在多种不同测试问题上进行了测试，展现出可观的结果。
引用
@article{arxiv.1806.05845,
title = {A Covariance Matrix Self-Adaptation Evolution Strategy for Optimization under Linear Constraints},
author = {Patrick Spettel and Hans-Georg Beyer and Michael Hellwig},
journal= {arXiv preprint arXiv:1806.05845},
year = {2018}
}
备注
This is a PREPRINT of an article accepted by IEEE TRANSACTIONS ON EVOLUTIONARY COMPUTATION. >>> COPYRIGHT 2018 IEEE <<< Manuscript received Jan, 2018; revised Jun, 2018; accepted Sep, 2018. Due to size limitations, this manuscript comprises figures with reduced resolution. The work was supported by the Austrian Science Fund FWF under grant P29651-N32. Content: 10 pages + supplementary material