Commit b1414ded authored by Taddeüs Kroes's avatar Taddeüs Kroes

Merge branch 'master' of ssh://vo20.nl/home/git/repos/uva

parents 50bdf230 5c46eed8
CC=clang
CFLAGS=-Wall -Wextra -pedantic -std=c99 -D_GNU_SOURCE -g
CC=gcc
CFLAGS=-Wall -Wextra -pedantic -std=c99 -D_GNU_SOURCE -g -O0
LFLAGS=-lm
all: q1 q2 q3 q4 q5 q6 q7 report.pdf
......
......@@ -20,12 +20,12 @@ double gauss(func_ptr f, double a, double b) {
return half_dx * (f(x1) + f(x2));
}
double integral(func_ptr f, method_ptr method, double a, double b, double dx) {
int steps = (int)((b - a) / dx), i;
double x, surface = 0;
double integral(func_ptr f, method_ptr method, double a, double b, int steps) {
int i;
double surface = 0, dx = (b - a) / steps;
for( i = 0; i < steps; i++)
surface += method(f, x = a + i * dx, x + dx);
surface += method(f, a + i * dx, a + (i + 1) * dx);
return surface;
}
......@@ -10,6 +10,6 @@ double rectangle(func_ptr f, double a, double b);
double trapezoidal(func_ptr f, double a, double b);
double simpson(func_ptr f, double a, double b);
double gauss(func_ptr f, double a, double b);
double integral(func_ptr f, method_ptr method, double a, double b, double dx);
double integral(func_ptr f, method_ptr method, double a, double b, int steps);
#endif
......@@ -4,8 +4,6 @@
#include "func_ptr.h"
#include "integral.h"
#define DX 1e-5
double f1(double x) {
return pow(M_E, -x);
}
......@@ -14,34 +12,35 @@ double f2(double x) {
return x * f1(x);
}
#define PRINT_INTEGRAL(func, method, a, b) (printf(#func " from " #a " to " \
#b " using %-19s %.11e\n", #method " method:", integral(&func, &method, a, b, DX)))
#define PRINT_INTEGRALS(func, a, b, real) { \
double _i; \
PRINT_INTEGRAL(func, a, b, rectangle, (real)); \
PRINT_INTEGRAL(func, a, b, trapezoidal, (real)); \
PRINT_INTEGRAL(func, a, b, simpson, (real)); \
PRINT_INTEGRAL(func, a, b, gauss, (real)); \
}
#define PRINT_INTEGRAL(func, a, b, method, real) { \
_i = integral(&func, &method, a, b, atoi(argv[1])); \
printf(#func " from " #a " to " #b " using %-19s %.80e (%.2f%%)\n", \
#method " method:", _i, fabs((real - _i) / real * 100)); \
}
int main(int argc, char **argv) {
if( argc != 2 ) {
printf("Usage: %s STEPS", argv[0]);
return -1;
}
int main(void) {
PRINT_INTEGRAL(f1, rectangle, 0, 1);
PRINT_INTEGRAL(f1, trapezoidal, 0, 1);
PRINT_INTEGRAL(f1, simpson, 0, 1);
PRINT_INTEGRAL(f1, gauss, 0, 1);
PRINT_INTEGRALS(f1, 0, 1, (M_E - 1) / M_E);
puts("");
PRINT_INTEGRAL(f2, rectangle, 0, 2);
PRINT_INTEGRAL(f2, trapezoidal, 0, 2);
PRINT_INTEGRAL(f2, simpson, 0, 2);
PRINT_INTEGRAL(f2, gauss, 0, 2);
PRINT_INTEGRALS(f2, 0, 2, 1 - 3/pow(M_E, 2));
puts("");
PRINT_INTEGRAL(f2, rectangle, 0, 20);
PRINT_INTEGRAL(f2, trapezoidal, 0, 20);
PRINT_INTEGRAL(f2, simpson, 0, 20);
PRINT_INTEGRAL(f2, gauss, 0, 20);
PRINT_INTEGRALS(f2, 0, 20, 1 - 21/pow(M_E, 20));
puts("");
PRINT_INTEGRAL(f2, rectangle, 0, 200);
PRINT_INTEGRAL(f2, trapezoidal, 0, 200);
PRINT_INTEGRAL(f2, simpson, 0, 200);
PRINT_INTEGRAL(f2, gauss, 0, 200);
PRINT_INTEGRALS(f2, 0, 200, 1 - 201/pow(M_E, 200));
puts("");
PRINT_INTEGRAL(sin, rectangle, 0, 8 * M_PI);
PRINT_INTEGRAL(sin, trapezoidal, 0, 8 * M_PI);
PRINT_INTEGRAL(sin, simpson, 0, 8 * M_PI);
PRINT_INTEGRAL(sin, gauss, 0, 8 * M_PI);
PRINT_INTEGRALS(sin, 0, 8 * M_PI, 0.0);
return 0;
}
......@@ -36,7 +36,7 @@ integrals = [(lambda x: exp(-x), 0, 1),
(lambda x: x * exp(-x), 0, 200),
(lambda x: sin(x), 0, 8 * pi)]
methods = [rectangle, trapezoidal, simpson]
methods = [rectangle, trapezoidal, simpson, gauss]
DX = 1e-3
......@@ -44,7 +44,5 @@ for i in integrals:
for method in methods:
print 'integrate from %d to %d using %-11s: %.16f' \
% (i[1], i[2], method.func_name, integrate(method, DX, *i))
print 'integrate from %d to %d using %-11s: %.16f' \
% (i[1], i[2], 'gauss', gauss(*i))
print '-------------'
......@@ -86,11 +86,94 @@ beoogde nauwkeurigheid.
% }}}
\section{Benadering van $\sqrt{2}$} % {{{
\label{sec:benadering van sqrt 2}
% }}}
\section{Newton-Raphson} % {{{
\label{sec:Newton-Raphson}
% }}}
\section{Vier numerieke integraal routines} % {{{
\label{sec:vier numerieke integraal routines}
We hebben de volgende vier methodes ge\"implementeerd om een numerieke integraal
uit te rekenen:
\begin{description}
\item[Rectangle rule]
$$rectangle(f, a, b) = (b - a) \cdot f(\frac{a + b}{2})$$
\item[Trapezoidal rule]
$$trapezoidal(f, a, b) = \frac{(b - a) \cdot (f(a) + f(b))}{2}$$
\item[Simpson's rule]
Simpson's rule kan worden berekend door middel van de bovenstaande methodes
``rectangle'' en ``trapezoidal'':
$$simpson(f, a, b) = \frac{2 \cdot rectangle(f, a, b) + trapezoidal(f, a, b)}{3}$$
\item[Two-point Gauss integratie]
$$gauss(f, a, b) = (b - a) \frac{f(a + \frac{b - a}{2} \cdot (1 -
\frac{1}{\sqrt{3}})) + f(a + \frac{b - a}{2} \cdot (1 +
\frac{1}{\sqrt{3}}))}{2}$$
\end{description}
Naast de bovengenoemde methodes hebben we ook een integraal functie gemaakt die
de methodes aanroept:
\begin{verbatim}
double integral(f, method, a, b, dx):
steps = (int) ((b - a) / dx)
surface = 0
for(i = 0; i < steps; i++):
surface += method(f, a + i * dx, a + (i + 1) * dx)
return surface
\end{verbatim}
\noindent Daarnaast hebben we drie functies gemaakt (waarbij \texttt{sin(x)}
onderdeel is van glibc) die de drie verschillende integranten voorstellen:
\begin{verbatim}
double f1(x):
return pow(M_E, -x)
double f2(x):
return x * f1(x)
double sin(x);
\end{verbatim}
\noindent Dit heeft het volgende resultaat gegeven als we het programma met
1.000.000 steps uitvoeren:
\begin{verbatim}
f1 from 0 to 1 using rectangle method: 6.32120558829e-01
f1 from 0 to 1 using trapezoidal method: 6.32120558829e-01
f1 from 0 to 1 using simpson method: 6.32120558829e-01
f1 from 0 to 1 using gauss method: 6.32120558829e-01
f2 from 0 to 2 using rectangle method: 5.93994150290e-01
f2 from 0 to 2 using trapezoidal method: 5.93994150290e-01
f2 from 0 to 2 using simpson method: 5.93994150290e-01
f2 from 0 to 2 using gauss method: 5.93994150290e-01
f2 from 0 to 20 using rectangle method: 9.99999956716e-01
f2 from 0 to 20 using trapezoidal method: 9.99999956715e-01
f2 from 0 to 20 using simpson method: 9.99999956716e-01
f2 from 0 to 20 using gauss method: 9.99999956716e-01
f2 from 0 to 200 using rectangle method: 1.00000000001e+00
f2 from 0 to 200 using trapezoidal method: 9.99999999965e-01
f2 from 0 to 200 using simpson method: 9.99999999998e-01
f2 from 0 to 200 using gauss method: 9.99999999998e-01
sin from 0 to 8 * M_PI using rectangle method: -1.04730878509e-13
sin from 0 to 8 * M_PI using trapezoidal method: -1.08168370335e-13
sin from 0 to 8 * M_PI using simpson method: -8.63149135973e-14
sin from 0 to 8 * M_PI using gauss method: -8.55165510887e-14
\end{verbatim}
% }}}
\end{document}
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment