arithmetic
arithmetic,
shows how arithmetic on a computer can break many mathematical rules.
Location: http://people.sc.fsu.edu/~jburkardt/classes/math1070_2019/arithmetic/arithmetic.html
The notes:
Scripts:
-
cos_taylor.m,
evaluates the Taylor series for cos(x).
-
cos_taylor_test.m,
compares the Taylor series for cos(x) to cos(x) itself.
-
patriot.m,
looks at how a small error in the Patriot missile eventually
becomes a big one.
-
pi_series1.m,
evaluates a series for pi.
-
pi_series2.m,
evaluates a series for pi until better accuracy than 22/7 is achieved.
-
pi_series3.m,
evaluates a series for pi until better accuracy than 355/113 is achieved.
-
plot_poly.m,
plots three versions of the same formula:
p1(x) = x^7-7x^6+21x^5-35x^4+35x^3-21x^2+7x-1;
p2(x) = ((((((x-7)*x+21)*x-35)*x+35)*x-21)*x+7)*x-1;
p3(x) = (x-1)^7.
-
rule_breakers.m,
quickly calculates some numbers that characterize limits
of computational accuracy.
-
tan_difference.m,
tries to estimate the derivative of tan(x) using an ever
decreasing finite difference stepsize.
Text:
Images:
-
cos_taylor_test.png,
a Taylor series for cos(x) fails for large x.
-
plot_poly.png,
three ways to evaluate the same polynomial give different results.
-
tan_difference.png,
how the error in a difference estimate for the derivative
of the tangent function blows up.
-
trident.png,
a Trident missile fails its test.
Last revised on 08 September 2019.