15-Errors-in-Computation-4UP

# 15-Errors-in-Computation-4UP - Errors Experiments Errors in...

This preview shows pages 1–3. Sign up to view the full content.

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: Errors Experiments Errors in Computation Dhavide Aruliah UOIT MATH 2070U c D. Aruliah (UOIT) Errors in Computation MATH 2070U 1 / 22 Errors Experiments Errors in Computation 1 Quantification of errors 2 Computational experiments Polynomial evaluation Summation Cancellation Overflow/underflow c D. Aruliah (UOIT) Errors in Computation MATH 2070U 2 / 22 Errors Experiments Absolute and relative errors Definition Let a ∈ R be an exact value and let e a ∈ R be an approximation of a Absolute error in e a 4 a : = | a- e a | Relative error in e a 4 a a , ( a 6 = ) Example ( a = √ 2, e a = 1.414) 4 a = √ 2- 1.414 = 0.0002135 . . . 4 a a = 0.0002135 . . . √ 2 = 0.0001510 . . . c D. Aruliah (UOIT) Errors in Computation MATH 2070U 4 / 22 Errors Experiments Example ( a = √ 2, e a = 1.414) 4 a = 0.0002135 . . . , 4 a a = 0.0001510 . . . |4 a | ≤ 0.00022 4 a a ≤ 0.00016 |4 a | ≤ 0.0003 4 a a ≤ 0.0002 The following are equivalent: a = √ 2, e a = 1.414, |4 a | ≤ 0.22 × 10- 3 a = 1.414 ± 0.22 × 10- 3 1.41378 ≤ a ≤ 1.41422 c D. Aruliah (UOIT) Errors in Computation MATH 2070U 5 / 22 Errors Experiments Rounding versus chopping Rounding to three decimals 0.123767 is rounded to 0.12 4 0.774501 is rounded to 0.77 5 0.6325 is rounded to 0.63 2 0.6335 is rounded to 0.63 4 Chopping to three decimals 0.123767 is chopped to 0.12 3 0.774501 is chopped to 0.77 4 0.6325 is chopped to 0.63 2 0.6335 is chopped to 0.63 3 0.69999 is chopped to 0.69 9 c D. Aruliah (UOIT) Errors in Computation MATH 2070U 6 / 22 Errors Experiments Significant digits Significant digits: leading nonzero digits in a number, e.g., 1.7320 has 5 significant digits, 0.0491 has 3 significant digits “Correct” significant digits problematic concept, e.g., a = 1.00000, e a = 1.00499, 4 a a = 4.99 × 10- 3 b = 9.00000, e b = 8.99899, 4 b b = 1.12 × 10- 4 e a agrees with a to 3 or 4 significant digits (by most definitions)...
View Full Document

## This note was uploaded on 02/23/2010 for the course MATH 2070 taught by Professor Aruliahdhavidhe during the Spring '10 term at UOIT.

### Page1 / 5

15-Errors-in-Computation-4UP - Errors Experiments Errors in...

This preview shows document pages 1 - 3. Sign up to view the full document.

View Full Document
Ask a homework question - tutors are online