A Triple Hybrid Framework Combining Neural ODEs, Thermodynamic Search Algorithm, and Adaptive Step Size Control for Solving Differential Equations

سال انتشار: 1405
نوع سند: مقاله ژورنالی
زبان: انگلیسی
مشاهده: 31

فایل این مقاله در 20 صفحه با فرمت PDF قابل دریافت می باشد

استخراج به نرم افزارهای پژوهشی:

لینک ثابت به این مقاله:

شناسه ملی سند علمی:

JR_GADM-12-1_001

تاریخ نمایه سازی: 22 آبان 1405

چکیده مقاله:

This paper introduces a novel triple hybrid framework for solving ordinary differential equations (ODEs) and differential-algebraic equations (DAEs) by synergistically combining three distinct machine learning and optimization paradigms. The proposed method integrates Neural Ordinary Differential Equations (Neural ODEs) with a Thermodynamic-Inspired Search Algorithm (TSA) and an adaptive machine learning-based step size controller. Neural ODEs provide a continuous-depth representation that naturally captures the differential structure of physical systems, while TSA---inspired by energy minimization, heat exchange, and entropy control---offers robust global exploration capabilities and prevents premature convergence. The third component employs a lightweight neural network that dynamically adjusts step sizes based on local error estimates, enhancing computational efficiency without sacrificing accuracy. This triple hybrid architecture addresses the limitations of existing methods: gradient instability in Physics-Informed Neural Networks, poor exploration in traditional metaheuristics, and the rigidity of fixed step size strategies. The method is evaluated on a diverse suite of benchmark problems including linear and nonlinear boundary value problems, stiff ODEs, systems of ODEs, and index-۱ DAEs. Extensive statistical analysis over ۵۰ independent runs demonstrates superior performance, with average improvements of ۶.۲\% in solution accuracy compared to state-of-the-art hybrid methods, ۲۸\% reduction in standard deviation, and ۳۵\% decrease in computational time through adaptive step sizing. The mesh-free nature of the approach provides differentiable closed-form solutions, making it particularly valuable for applications requiring sensitivity analysis or integration with downstream tasks.

کلیدواژه ها:

Neural Ordinary Differential Equations ، Thermodynamic Search Algorithm ، Adaptive Step Size Control ، Hybrid Metaheuristics ، Differential-Algebraic Equations

نویسندگان

Aliakbar Tajari Siahmarzkooh

Department of Computer Sciences, Golestan University, Gorgan, Iran

مراجع و منابع این مقاله:

لیست زیر مراجع و منابع استفاده شده در این مقاله را نمایش می دهد. این مراجع به صورت کاملا ماشینی و بر اساس هوش مصنوعی استخراج شده اند و لذا ممکن است دارای اشکالاتی باشند که به مرور زمان دقت استخراج این محتوا افزایش می یابد. مراجعی که مقالات مربوط به آنها در سیویلیکا نمایه شده و پیدا شده اند، به خود مقاله لینک شده اند :
  • R. J. LeVeque, Finite difference methods for ordinary and partial ...
  • J. D. Murray, Mathematical biology: I. An introduction, ۳rd ed., ...
  • T. J. Hughes, The finite element method: linear static and ...
  • M. Raissi, P. Perdikaris, and G. E. Karniadakis, Physics-informed neural ...
  • R. T. Q. Chen, Y. Rubanova, J. Bettencourt, and D. ...
  • P. Kidger, On Neural Differential Equations, PhD thesis, University of ...
  • M. Misaghi and M. Yaghoobi, Improved invasive weed optimization algorithm ...
  • M. Kumar and N. Yadav, Solving differential equations with neural ...
  • M. S. Hashemi and A. Haji Badali, A hybrid PSO-ANN ...
  • F. S. Zainuddin and M. A. Mohamed, Genetic algorithm-neural network ...
  • A. K. Qin, V. L. Huang, and P. N. Suganthan, ...
  • M. Shams and S. Effati, Solving optimal control problems using ...
  • A. H. Gandomi and A. H. Alavi, An introduction to ...
  • P. K. Das and H. S. Behera, A hybrid mutated ...
  • S. H. Tabasi, R. Pourgholi, and A. A. Boroujeni, A ...
  • D. Kong, Y. Wang, S. Chen, L. Zhang, and H. ...
  • I. G. Tsoulos, D. Gavrilis, and E. Glavas, Solving differential ...
  • S. Karimkashi and A. A. Kishk, Invasive weed optimization and ...
  • Y. Liu and Z. Zhang, A memory-enhanced particle swarm optimization ...
  • B. J. Zou and L. Tian, Automatic and Structure-Aware Sparsification ...
  • A. R. Mehrabian and C. Lucas, A novel numerical optimization ...
  • P. Pal, T. Kumar, and A. Kumar, Hybrid Weed-Gravitational Evolutionary ...
  • M. H. Taheri, M. Moradi, and M. R. J. Motlagh, ...
  • R. Pourgholi, S. Foadian, and S. H. Hashem Tabasi, EEPM: ...
  • J. C. Butcher, Numerical methods for ordinary differential equations, ۴th ...
  • I. E. Lagaris, A. Likas, and D. I. Fotiadis, Artificial ...
  • H. Dana Mazraeh and K. Parand, GELS-SVR: a combination of ...
  • M. Movahedian Moghaddam, H. Dana Mazraeh, and K. Parand, Symbolic ...
  • H. D. Mazraeh, K. Parand, M. Hosseinzadeh, A. Shams, and ...
  • H. Dana Mazraeh and K. Parand, GEPINN: An innovative hybrid ...
  • نمایش کامل مراجع