Lecture 6 Notes

125 archimedes 250 bc zu chongzhi liu hui 480 ad

Info iconThis preview shows page 1. Sign up to view the full content.

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

Unformatted text preview: beJer” En and x* Bounds •  What is “beJer” –  Accuracy, Speed, Precision •  Want x* accurate as quickly as possible –  Least number of steps ~= least number of computa3ons (or fastest execu3on 3me), but it is a start –  How do you analyze approaches? •  Run experiments •  Look at how fast error bound shrinks (rate of convergence) –  Check absolute error for special cases –  Analy3c Methods •  Compare rates of convergence Convergence Rates of Convergence Ayer 5 terms, •  The Gregory–Leibniz series is within 0.2 •  The Nilakantha's series is within 0.002 Some really great series exist (>14 correct decimal digits per term) great En, but difficult En Approxima3ng Func3ons Evolu3on of x* for log(1+y) A func3on at a par3cular point is a number (like pi), so numerical approxima3on material we have studied applies, e.g., x: = ln(1+y) = y – y^2/2 + y^3/3 – y^4/4 + y^5/5 + … x* = y (N=1) = y – y^2/2 (N=2) = y – y^2/2 + y^3/3 (N=3) = y – y^2/2 + y^3/3 – y^4/4 (N=4) = y – y^2/2 + y^3/3 – y^4/4 + y^5/5 (N=5) … En = En(y) (<- bound generally depe...
View Full Document

Ask a homework question - tutors are online