Scribd
Upload
38 views1 page

1.1 Round Off Errors

1. The document discusses types of errors that occur when numerical calculations are rounded off or approximated, including: - Round off errors that occur when values are rounded to a specified number of decimal places or significant figures - Propagated errors that occur when substituting an approximated value into a function 2. Examples are given of calculating the absolute and relative errors from rounding common values like π and e to different decimal places. 3. The maximum predicted error and actual error are calculated for substituting rounded values into simple functions like sine, squares, and exponents based on bounds for the derivative of the function.

Uploaded by

hagls
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
38 views1 page

1.1 Round Off Errors

1. The document discusses types of errors that occur when numerical calculations are rounded off or approximated, including: - Round off errors that occur when values are rounded to a specified number of decimal places or significant figures - Propagated errors that occur when substituting an approximated value into a function 2. Examples are given of calculating the absolute and relative errors from rounding common values like π and e to different decimal places. 3. The maximum predicted error and actual error are calculated for substituting rounded values into simple functions like sine, squares, and exponents based on bounds for the derivative of the function.

Uploaded by

hagls
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 1

1 Errors:

1.1 Round off Errors:

1.1.1 Exercises: Write down the absolute and relative errors that occur when numbers, 31/13, 
√1001, √999, e-5 and sin(1) are rounded off to specified number of decimal places or significant
digits.

Rounded Absolute Relative Rounded Absolute Relative


x Exact Value
to 2 d.p. error error to 3 d.p. error error
31/13 2.384615385 2.38 0.00461538 0.001939237 2.385 0.000384615 0.000161264
 3.141592654
√1001 31.63858404
e-5 0.006737947
sin(1) 0.841470985

Rounded Absolute Relative Rounded Absolute Relative


x Exact Value
to 4 s.d. error error to 4 s.d. error error
31/13 2.384615385 2.385 0.00038462 0.000161264 2.3846 1.53846E-05 6.45165E-06

 3.141592654 3.142 0.00040735 0.000129646 3.1416 7.34641E-06 2.33843E-06


√1001 31.63858404
e-5 0.006737947
sin(1) 0.841470985

1.2 Propagated errors: Let y  f ( x ) be a real valued function continuously differentiable near a
point x  x 0 and let f ( x 0 )  y 0 . Suppose, we want to evaluate an approximate value for y 0 , by

substituting an approximate value x1  x 0   x for x . The error y  f ( x 0   x )  f ( x ) is

called the propagated error. By Taylor’s theorem, there exists a constant  such that

y  f () x . Therefore, if f ( x )  M ,  x , then an error bound is given by y  M x .

1.2.1 Exercises: Determine the maximum error predicted and the actual error that occur when
an expression y  f ( x 0 ) is evaluated by substituting rounded off values for x 0 .

Max. error actual error


f(x) x0 x1 δx y0= f(x0) y1= f(x1)
predicted (M δx) |δy| = |y1-y0|
cos(ξ) δx < δx
sin(x) 31/13 2.385 0.000385 0.686727 0.686448 0.000279
<0.000385
2
 2x δx < 7 δx
x 3.142 0.000407 9.869604 9.872164 < 0.002849
0.002560

1/(1+x) 1001 31.639

ex -4.9999 -5

You might also like