There are many issues in the fields of physics, chemistry, biology, and astronomy
which can be solved through differential equation formulations. In general, the
completion of differential equations can be done analytically or by numerical
methods. If the completion is done analytically, usually it is done through calculus
theories, and may require a longer time to solve. Anticipating the difficulties
posed by differential equations analysis, numerical method is being used instead.
This numerical completion provides solution in the form of approach and being
carried out by visiting the initial value which then needs to be advanced gradually,
step by step. Utilizing computers in solving differential equations would also help
develop the application of numerical methods. Therefore, this study is expected to
be able to improve the existing methods. This research will compare the accuracy
of different methods, the Runge-Kutta Fehlberg and Adams-Moulton methods, in
completing differential equations, which is limited to ordinary differential
equations of first order and second order. It is found that there is general
difference between the two method with Runge-Kutta Fehlberg method being the
one-step method with an uncertain step size, while the Adams-Moulton method
being the double steps method. Comparison of accuracy is obtained
5116 Gerardus Polla
through comparing the value of differential equations results numerically with
differential equations result obtained from MatLab version 5.3. A number of
experiments on the completion of linear ordinary differential equations of first
order and second order are done through computerization to compare the accuracy
between the Runge-Kutta Fehlberg and Adams-Moulton methods. In addition,
accuracy is being pointed out through relative error. From the research results
with the non-parametric statistical test it can be seen that the Jackknife for linear
ordinary differential equations of first order and second order, the Runge-Kutta
Fehlberg method has an average relative error greater than the Adams-Moulton
method. It can be concluded that the Adams-Moulton method has more rigorous
accuracy than the Range-Kutta Fehlberg method in solving linear ordinar