Monte-Carlo Method
(Difference between revisions)
m |
(dicking with better code.) |
||
Line 8: | Line 8: | ||
Code: | Code: | ||
− | for i = 1: | + | % Monte-carlo simulator for solving area of a quarter-circle of radius 1, |
− | x = rand; | + | % dictated by bounds of x^2+y^2 = 1 |
+ | |||
+ | % Set up arrays and other constants | ||
+ | x = 0; % x coordinate | ||
+ | y = 0; % y coordinate | ||
+ | total = 10000; % amount of points to generate | ||
+ | count = 0; % number of points inside area | ||
+ | |||
+ | % monte-carlo solver | ||
+ | for i = 1:total | ||
+ | % generate a random point in 0<x<1 and 0<y<1 | ||
+ | x = rand; % random number between 0 and 1 | ||
y = rand; | y = rand; | ||
− | if x^2+y^2 <=1 | + | % monte-carlo check if inside bounds |
− | + | if (x^2+y^2 <= 1) | |
+ | count = count + 1; | ||
end | end | ||
− | |||
end | end | ||
− | area = | + | |
+ | % calculate 'true' area and monte-carlo solved area | ||
+ | circlearea = pi*1^2; % 'true' area | ||
+ | montearea = count/total*4; % monte-carlo area | ||
This, when run on my computer, produced the result of 0.783. The area of a quarter circle works out to be π/4 if the radius is 1. π/4 = 0.7854, so you can see that this is quite close. More iterations than 1000 would produce a closer answer. | This, when run on my computer, produced the result of 0.783. The area of a quarter circle works out to be π/4 if the radius is 1. π/4 = 0.7854, so you can see that this is quite close. More iterations than 1000 would produce a closer answer. |
Revision as of 22:00, 21 August 2008
This is a small, simple example of utilising the Monte-Carlo method to calculate the value of pi. It's easy because it is only a 2-dimensional problem and thus is easy to visualise, and is trivial in its code implementation.
Picture the graph shown in the parent article - that of a quarter circle plotted on a unit x-y graph:
Below is a small example piece of Matlab code that does the above simulation:
Code:
% Monte-carlo simulator for solving area of a quarter-circle of radius 1, % dictated by bounds of x^2+y^2 = 1
% Set up arrays and other constants x = 0; % x coordinate y = 0; % y coordinate total = 10000; % amount of points to generate count = 0; % number of points inside area
% monte-carlo solver for i = 1:total % generate a random point in 0<x<1 and 0<y<1 x = rand; % random number between 0 and 1 y = rand; % monte-carlo check if inside bounds if (x^2+y^2 <= 1) count = count + 1; end end
% calculate 'true' area and monte-carlo solved area circlearea = pi*1^2; % 'true' area montearea = count/total*4; % monte-carlo area
This, when run on my computer, produced the result of 0.783. The area of a quarter circle works out to be π/4 if the radius is 1. π/4 = 0.7854, so you can see that this is quite close. More iterations than 1000 would produce a closer answer.