Predictive modeling of flexible EHD pumps using Kolmogorov–Arnold Networks

We present a novel approach to predicting the pressure and flow rate of flexible electrohydrodynamic pumps using the Kolmogorov–Arnold Network. Inspired by the Kolmogorov–Arnold representation theorem, KAN replaces fixed activation functions with learnable spline-based activation functions, enabling...

Full description

Saved in:
Bibliographic Details
Main Authors: Yanhong Peng, Yuxin Wang, Fangchao Hu, Miao He, Zebing Mao, Xia Huang, Jun Ding
Format: Article
Language:English
Published: Elsevier 2024-12-01
Series:Biomimetic Intelligence and Robotics
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S2667379724000421
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850101069917454336
author Yanhong Peng
Yuxin Wang
Fangchao Hu
Miao He
Zebing Mao
Xia Huang
Jun Ding
author_facet Yanhong Peng
Yuxin Wang
Fangchao Hu
Miao He
Zebing Mao
Xia Huang
Jun Ding
author_sort Yanhong Peng
collection DOAJ
description We present a novel approach to predicting the pressure and flow rate of flexible electrohydrodynamic pumps using the Kolmogorov–Arnold Network. Inspired by the Kolmogorov–Arnold representation theorem, KAN replaces fixed activation functions with learnable spline-based activation functions, enabling it to approximate complex nonlinear functions more effectively than traditional models like Multi-Layer Perceptron and Random Forest. We evaluated KAN on a dataset of flexible EHD pump parameters and compared its performance against RF, and MLP models. KAN achieved superior predictive accuracy, with Mean Squared Errors of 12.186 and 0.012 for pressure and flow rate predictions, respectively. The symbolic formulas extracted from KAN provided insights into the nonlinear relationships between input parameters and pump performance. These findings demonstrate that KAN offers exceptional accuracy and interpretability, making it a promising alternative for predictive modeling in electrohydrodynamic pumping.
format Article
id doaj-art-acf3cd2ba81040e19e378a942a266d14
institution DOAJ
issn 2667-3797
language English
publishDate 2024-12-01
publisher Elsevier
record_format Article
series Biomimetic Intelligence and Robotics
spelling doaj-art-acf3cd2ba81040e19e378a942a266d142025-08-20T02:40:08ZengElsevierBiomimetic Intelligence and Robotics2667-37972024-12-014410018410.1016/j.birob.2024.100184Predictive modeling of flexible EHD pumps using Kolmogorov–Arnold NetworksYanhong Peng0Yuxin Wang1Fangchao Hu2Miao He3Zebing Mao4Xia Huang5Jun Ding6College of Mechanical Engineering, Chongqing University of Technology, Chongqing 400054, China; Corresponding author.Department of Mechanical Systems Engineering, Nagoya University, Tokai National Higher Education and Research, Nagoya 464-8603, Japan; School of Energy and Power, Jiangsu University of Science and Technology, Zhenjiang 212100, ChinaCollege of Mechanical Engineering, Chongqing University of Technology, Chongqing 400054, ChinaCollege of Mechanical Engineering, Chongqing University of Technology, Chongqing 400054, ChinaFaculty of Engineering, Yamaguchi University, Yamaguchi 755-8611, JapanCollege of Mechanical Engineering, Chongqing University of Technology, Chongqing 400054, ChinaCollege of Mechanical Engineering, Chongqing University of Technology, Chongqing 400054, ChinaWe present a novel approach to predicting the pressure and flow rate of flexible electrohydrodynamic pumps using the Kolmogorov–Arnold Network. Inspired by the Kolmogorov–Arnold representation theorem, KAN replaces fixed activation functions with learnable spline-based activation functions, enabling it to approximate complex nonlinear functions more effectively than traditional models like Multi-Layer Perceptron and Random Forest. We evaluated KAN on a dataset of flexible EHD pump parameters and compared its performance against RF, and MLP models. KAN achieved superior predictive accuracy, with Mean Squared Errors of 12.186 and 0.012 for pressure and flow rate predictions, respectively. The symbolic formulas extracted from KAN provided insights into the nonlinear relationships between input parameters and pump performance. These findings demonstrate that KAN offers exceptional accuracy and interpretability, making it a promising alternative for predictive modeling in electrohydrodynamic pumping.http://www.sciencedirect.com/science/article/pii/S2667379724000421Kolmogorov–Arnold NetworksElectrohydrodynamic pumpsNeural network
spellingShingle Yanhong Peng
Yuxin Wang
Fangchao Hu
Miao He
Zebing Mao
Xia Huang
Jun Ding
Predictive modeling of flexible EHD pumps using Kolmogorov–Arnold Networks
Biomimetic Intelligence and Robotics
Kolmogorov–Arnold Networks
Electrohydrodynamic pumps
Neural network
title Predictive modeling of flexible EHD pumps using Kolmogorov–Arnold Networks
title_full Predictive modeling of flexible EHD pumps using Kolmogorov–Arnold Networks
title_fullStr Predictive modeling of flexible EHD pumps using Kolmogorov–Arnold Networks
title_full_unstemmed Predictive modeling of flexible EHD pumps using Kolmogorov–Arnold Networks
title_short Predictive modeling of flexible EHD pumps using Kolmogorov–Arnold Networks
title_sort predictive modeling of flexible ehd pumps using kolmogorov arnold networks
topic Kolmogorov–Arnold Networks
Electrohydrodynamic pumps
Neural network
url http://www.sciencedirect.com/science/article/pii/S2667379724000421
work_keys_str_mv AT yanhongpeng predictivemodelingofflexibleehdpumpsusingkolmogorovarnoldnetworks
AT yuxinwang predictivemodelingofflexibleehdpumpsusingkolmogorovarnoldnetworks
AT fangchaohu predictivemodelingofflexibleehdpumpsusingkolmogorovarnoldnetworks
AT miaohe predictivemodelingofflexibleehdpumpsusingkolmogorovarnoldnetworks
AT zebingmao predictivemodelingofflexibleehdpumpsusingkolmogorovarnoldnetworks
AT xiahuang predictivemodelingofflexibleehdpumpsusingkolmogorovarnoldnetworks
AT junding predictivemodelingofflexibleehdpumpsusingkolmogorovarnoldnetworks