Impression : Jurnal Teknologi dan Informasi
Vol. 5 No. 2 (2026): Juli 2026

Analisis Kinerja Implementasi Algoritma Broyden pada Skema Implisit Persamaan Diferensial Biasa Sistem Stiff

Hery Andi Sitompul (Universitas HKBP Nommensen Medan)
Asmina Herawaty Sinaga (Universitas HKBP Nommensen Medan)
Dewi Sholeha (Universitas Prima Indonesia Medan)
Alimin Purba (Universitas Deli Sumatera Medan)



Article Info

Publish Date
21 Jul 2026

Abstract

Penyelesaian sistem persamaan diferensial biasa (ODE) yang kaku umumnya dilakukan secara numerik menggunakan skema implisit. Namun, pendekatan ini membutuhkan komputasi yang kompleks dan memakan waktu karena perlunya penyelesaian sistem persamaan non-linear pada setiap diskretisasi domain solusi. Metode Newton-Raphson adalah pilihan yang paling populer karena konvergensi dan stabilitasnya; meskipun demikian, metode ini membutuhkan komputasi matriks Jacobian dan inversnya pada setiap langkah waktu, yang menimbulkan hambatan signifikan terhadap efisiensi waktu. Studi ini bertujuan untuk menganalisis kinerja implementasi algoritma Broyden, yang secara efisien memperbarui matriks Jacobian dalam skema implisit untuk sistem ODE yang kaku. Kinerja algoritma dievaluasi berdasarkan tiga indikator utama: akurasi konvergensi, kecepatan komputasi (waktu CPU), dan jumlah iterasi. Pengujian dilakukan menggunakan beberapa kasus uji sistem kaku standar dengan variasi ukuran langkah (h) dan rasio kekakuan. Hasil simulasi menunjukkan bahwa algoritma Broyden berhasil meningkatkan kinerja skema implisit dengan stabilitas dan akurasi yang sangat baik. Oleh karena itu, implementasi algoritma Broyden dalam skema implisit terbukti menjadi alternatif yang efektif dan efisien untuk menyelesaikan sistem ODE kaku berskala besar. Solving stiff systems of ordinary differential equations (ODEs) is generally performed numerically using implicit schemes. However, this approach requires complex and time-consuming computation due to the necessity of solving a system of non-linear equations at each discretization of the solution domain. The Newton-Raphson method is the most popular choice owing to its convergence and stability; nonetheless, it requires computing the Jacobian matrix and its inverse at every time step, which poses a significant bottleneck to time efficiency. This study aims to analyze the performance of the Broyden algorithm implementation, which efficiently updates the Jacobian matrix within the implicit scheme for stiff ODE systems. The algorithm's performance was evaluated based on three key indicators: convergence accuracy, computational speed (CPU time), and the number of iterations. Testing was conducted using several standard stiff system test cases with variations in step size () and stiffness ratios. The simulation results demonstrate that the Broyden algorithm successfully enhances the performance of the implicit scheme with excellent stability and accuracy. Consequently, the implementation of the Broyden algorithm in implicit schemes proves to be an effective and efficient alternative for solving large-scale stiff ODE systems.

Copyrights © 2026






Journal Info

Abbrev

jti

Publisher

Subject

Automotive Engineering Chemical Engineering, Chemistry & Bioengineering Civil Engineering, Building, Construction & Architecture Computer Science & IT Electrical & Electronics Engineering

Description

Impression accepts articles in the fields of Electrical Engineering, Mechanical Engineering, Civil Engineering, Marine Technology Industrial Engineering, Marine Fisheries Technology, Agricultural Technology, Informatics Engineering, Information Systems, Computer, Expert systems, Decision Support ...