当前位置: X-MOL 学术SIAM J. Numer. Anal. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Swarm-Based Gradient Descent Meets Simulated Annealing
SIAM Journal on Numerical Analysis ( IF 2.8 ) Pub Date : 2024-12-17 , DOI: 10.1137/24m1657808
Zhiyan Ding, Martin Guerra, Qin Li, Eitan Tadmor

SIAM Journal on Numerical Analysis, Volume 62, Issue 6, Page 2745-2781, December 2024.
Abstract. We introduce a novel method, called swarm-based simulated annealing (SSA), for nonconvex optimization which is at the interface between the swarm-based gradient-descent (SBGD) [J. Lu et al., arXiv:2211.17157; E. Tadmor and A. Zenginoglu, Acta Appl. Math., 190 (2024)] and simulated annealing (SA) [V. Cerny, J. Optim. Theory Appl., 45 (1985), pp. 41–51; S. Kirkpatrick et al., Science, 220 (1983), pp. 671–680; S. Geman and C.-R. Hwang, SIAM J. Control Optim., 24 (1986), pp. 1031–1043]. Similarly to SBGD, we introduce a swarm of agents, each identified with a position, [math] and mass [math], to explore the ambient space. Similarly to SA, the agents proceed in the gradient descent direction, and are subject to Brownian motion. The annealing rate, however, is dictated by a decreasing function of their mass. As a consequence, instead of the SA protocol for time-decreasing temperature, here the swarm decides how to “cool down” agents, depending on their own accumulated mass. The dynamics of masses is coupled with the dynamics of positions: agents at higher ground transfer (part of) their mass to those at lower ground. Consequently, the resulting SSA optimizer is dynamically divided between heavier, cooler agents viewed as “leaders” and lighter, warmer agents viewed as “explorers.” Mean-field convergence analysis and benchmark optimizations demonstrate the effectiveness of the SSA method as a multidimensional global optimizer.


中文翻译:


基于群的梯度下降与模拟退火的结合



SIAM 数值分析杂志,第 62 卷,第 6 期,第 2745-2781 页,2024 年 12 月。

抽象。我们介绍了一种称为基于群的模拟退火 (SSA) 的新方法,用于非凸优化,该方法位于基于群的梯度下降 (SBGD) [J. Lu et al., arXiv:2211.17157;E. Tadmor 和 A. Zenginoglu, Acta Appl. Math., 190 (2024)] 和模拟退火 (SA) [V. Cerny, J. Optim.《理论应用》,第 45 卷(1985 年),第 41-51 页;S. Kirkpatrick 等人,《科学》,第 220 卷(1983 年),第 671-680 页;S. Geman 和 C.-R.Hwang, SIAM J. Control Optim.,第 24 卷(1986 年),第 1031-1043 页]。与 SBGD 类似,我们引入了一组代理,每个代理都由位置 [math] 和质量 [math] 标识,以探索周围空间。与 SA 类似,代理沿梯度下降方向进行,并受布朗运动的影响。然而,退火速率是由它们的质量的递减函数决定的。因此,这里 swarm 不是用于时间递减温度的 SA 协议,而是根据它们自身的累积质量来决定如何“冷却”代理。质量的动态与位置的动态相耦合:位于较高地面的代理将其质量(部分)转移到位于较低地面的代理。因此,生成的 SSA 优化器被动态地划分为被视为“领导者”的较重、较冷的代理和被视为“探索者”的较轻、较温暖的代理。均值场收敛分析和基准优化证明了 SSA 方法作为多维全局优化器的有效性。
更新日期:2024-12-18
down
wechat
bug