Numerical calculations accuracy issues in basic programming teaching
Date
2014Author
Brzeziński, Dariusz
Rowińska, Zdzisława
Gocławski, Jarosław
Metadata
Show full item recordAbstract
[...] In the chapter there are discussed basic types of floating-point calculations
errors which results from the application of numerical methods:
• errors resulting from number systems differences (binary vs. decimal),
• errors resulting from differences how the input data is treated (discrete
vs. continuous),
• errors resulting from limitations of binary arithmetic.
There are also served short characteristics of the errors, in many cases supplemented
by examples of their occurrence.
The authors pointed out, that the standard level of accuracy offered by
widely applied developments environments are not sufficient to achieve the
highest precision in case of today's computational problems.
Some methods to achieve this goal are given. There are discussed the effects
of propagation and accumulation of basic types of computational errors
aroused as an effect of arithmetic operations and selected functions transformations.