Academic Journals Database
Disseminating quality controlled scientific knowledge

Runge-Kutta Schemes Coefficients Simulation for Comparison and Visual Effects

Author(s): Tajudeen Abiola Ogunniyi Salau | Olusegun Olufemi Ajide

Journal: Engineering
ISSN 1947-3931

Volume: 05;
Issue: 05;
Start page: 530;
Date: 2013;
Original page

Keywords: Runge-Kutta | Visual Effects | Simulation Coefficients | Taylor Series and Butcher Assumptions

Runge-Kutta scheme is one of the versatile numerical tools for the simulation of engineering systems. Despite its wide and acceptable engineering use, there is dearth of relevant literature bordering on visual impression possibility among different schemes coefficients which is the strong motivation for the present investigation of the third and fourth order schemes. The present study capitalise on results of tedious computation involving Taylor series expansion equivalent supplemented with Butcher assumptions and constraint equations of well-known works which captures the essential relationship between the coefficients. The simulation proceeds from random but valid specification of two out of the total coefficients possible per scheme. However the remaining coefficients are evaluated with application of appropriate function relationship. Eight and thirteen unknown coefficients were simulated respectively for third and fourth schemes over a total of five thousand cases each for relevant distribution statistics and scatter plots analysis for the purpose of scheme comparison and visual import. The respective three and four coefficients of the slope estimate for the third and fourth schemes have mix sign for large number of simulated cases. However, none of the two schemes have above three of these coefficients lesser than zero. The percentages of simulation results with two coefficients lesser than zero dominate and are respectively 56.88 and 77.10 for third and fourth schemes. It was observed that both popular third and fourth schemes belong to none of the coefficients being zero classification with respective percentage of 0.72 and 3.28 intotal simulated cases. The comparisons of corresponding scatter plots are visually exciting. The overall difference between corresponding scatter plots and distribution results can be used to justify the accuracy of fourth scheme over its counterpart third scheme.

Tango Jona
Tangokurs Rapperswil-Jona

RPA Switzerland

Robotic Process Automation Switzerland