Numerical Analysis - Errors
Uploaded by
Miguel YorroNumerical Analysis - Errors
Uploaded by
Miguel YorroModel: Involves simplification or omission in the
model equation
Data: Errors in measurement or estimation of raw data Numerical Method: Errors based on some
approximation
Representation of Numbers: for example, p
cannot be represented exactly by a finite number of digits
Arithmetic: Mistakes in carrying out operations such
as addition or multiplication
Number Representation: Numerical calculation can
involve numbers that cannot be represented exactly by a finite number of digits. For instance 2/3 (base 10), or irrational numbers (p)
Round-off Error: Reducing the number to a
significant digit
Truncation Error: Reducing an infinite series or
iterative process
Absolute Error: Absolute difference between the exact
number x and the approximate number X
Relative Error: Ratio of the absolute error and the
absolute exact number
Implicit Error: Errors of the original numbers or factors Accumulated Error: Total cumulative error at any given
step or iteration
Consider two numbers Under Addition/Subtraction: The magnitude of
the propagated error is not more than the sum of the initial absolute errors
Under Multiplication: The maximum relative error
propagated is approximately the sum of the initial relative errors
Addition of a small but non-zero number may have no effect
5.18x102 + 4.37x10-1 = 5.18x102 + 0.00437x102 = 5.18437x102 ~ 5.18x102
The multiplicative inverse may not exist; a*(1/a) 1
a = 3.00x100 1/a is 3.33x10-1 a* (1/a) is 9.99 * 10-1
The associative property may not apply; (a+b)+c a+(b+c)
a = 6.31x101, b = 4.24x100, c = 2.47x101, then (a+b)+c = (6.31x101 + 0.424x101) + 2.47x101 6.73x101 + .0247x101 6.75x101, whereas a+(b+c) = 6.31x101 + (4.24x100 + 2.47x100) 6.31x101 + 4.49x100 6.31x101+4.49x100 6.31x101+ 0.449x101 6.76x101
Subtracting a number from another nearly equal number may result in loss of significance or cancellation error.
1 - cos(0.05) = 1-0.99875 1.00x100 - 0.999x100 1.00x10-3
Representing a function as an infinite series of terms involving simpler functions Taylor series: where z is between a and x +1
= + 1! + 2!
2
+ +
( ) +
+ 1 !
( )+1
() = + ()
()
For the number 5/3 = 1.66666 determine the magnitude of round-off error when it is represented by a number obtained from the decimal form by: 1. Chopping to 3S 2. Chopping to 3D 3. Rounding to 3S 4. Rounding to 3D Evaluate the following operations as accurately as possible, including all errors 1. 8.24 + 5.33 2. 124.53 124.52 3. 4.27 x 3.13 4. 9.48 x 0.513 6.72 5. 0.25 x 2.84 0.64 6. 1.73 2.16 +0.08 + 1.00 -2.23 -0.97 +3.02
Modify the previous assignment using a Taylor series approximation about a = 0 for cos and sin. Choose where to truncate the series and compare the results with those from your previous program using the intrinsic functions. How big are the discrepancies?
