Error: Efficient and Robust Measure of Accuracy

Abstract

Most of the numerical methods developed for a large number of calculations are so complex and complicated that their successful implementation depends largely on the use of calculator or a computer which is capable of obtaining results to the desired degree of accuracy. Consequently, for any numerical computation, there occur some errors. This paper focuses on error as a tool to measuring accuracy in finding solution to numerical problems. We considered different types and forms of errors, their sources and existence in our daily routine data and mathematical computation. While some manual computational errors were considered, this paper focuses more on computed errors of solution to some selected differential equations where results were compared using the errors. The findings clearly established that the smaller the error, the better and optimal is the results.