DEVELOPMENT OF GRAPH-BASED NUMERICAL METHODS FOR SOLVING PARTIAL DIFFERENTIAL EQUATIONS

Authors

  • Adly Farius Disayangi Laia Universitas Sumatera Utara
  • Mardiningsih -
  • Saib Suwilo

DOI:

https://doi.org/10.32734/jormtt.v8i2.25971

Keywords:

partial differential equations, heat equation, numerical methods, Graph Theory, Laplacian Matrix

Abstract

Partial differential equations (PDEs) are widely used to model physical phenomena such as heat transfer and diffusion, but analytical solutions are often difficult to obtain, requiring numerical approaches. This study develops a graph-based numerical method for solving one-dimensional and two-dimensional heat equations on regular and irregular domains. The discrete domain is represented as a graph, where vertices denote discrete points and edges describe relationships among them. A graph Laplacian matrix is then constructed to compute the numerical solution. The results show that the proposed method produces solutions that closely approximate the exact solution and those obtained using the conventional finite difference method. For the one-dimensional heat equation, the solution at a point changed from 0.125 to 0.127488 at the fifth iteration, with relative errors ranging from 6.94% to 27.15%. For the two-dimensional case, the solution at a point changed from 0.5 to 0.535785, with relative errors ranging from 5.56% to 23.46%. In the irregular L-shaped domain, the solution at a point changed from 1.23 to 1.154866, with relative errors ranging from 4.65% to 20.35%. These findings indicate that the graph-based numerical method is flexible and effective for solving PDEs on complex domains without the need for complicated grid reconstruction.

Downloads

Download data is not yet available.

Downloads

Published

2026-11-18

How to Cite

[1]
A. F. D. Laia, M. -, and S. Suwilo, “DEVELOPMENT OF GRAPH-BASED NUMERICAL METHODS FOR SOLVING PARTIAL DIFFERENTIAL EQUATIONS”, J. of Research in Math. Trends and Tech., vol. 8, no. 2, pp. 95–110, Nov. 2026.