Why does my code get stuck when trying to convert a double integral into a double?

1 次查看（过去 30 天）

Guillermo 2023-5-4

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1957814-why-does-my-code-get-stuck-when-trying-to-convert-a-double-integral-into-a-double

评论： Walter Roberson 2023-5-5

syms x y z
f(x,y) = sqrt(1+((4*y)/(121*sqrt((4*x^2)/121 + (4*y^2)/121))^2)+((4*x)/(121*sqrt((4*x^2)/121 + (4*y^2)/121))^2))
minX= -sqrt(220.523)
maxX= sqrt(220.523)   %input("Limite superior de X:")
intX = int(f,x,minX,maxX)
minY= -sqrt(220.523-x^2)  %input("Limite inferior de Y:")
maxY= sqrt(220.523-x^2)   %input("Limite superior de Y:")
intY = int(f,y,minY,maxY)
integralDobleTipoUno = int(intY,x,minX,maxX)
integralDobleTipoDos= int(intX,y,minY,maxY)
var = double(integralDobleTipoUno)

When reaching "var", it loads forever

采纳的回答

Walter Roberson 2023-5-4

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1957814-why-does-my-code-get-stuck-when-trying-to-convert-a-double-integral-into-a-double#answer_1228924

if you are after the numeric solution then switch to vpaintegral()

8 个评论
显示 6更早的评论隐藏 6更早的评论

Walter Roberson 2023-5-4

For one thing, remember GIGO, "garbage in, garbage out".

When you write 220.523 you are going to get some number between approximately 220523/1000 ± 10^-12. Not necessarily exactly the same number for Windows as for Mac Intel or Mac ARM or Linux Intel. I think it should be consistent between Windows Intel and Windows AMD but I would not want to guarantee.

Anyhow you get some particular binary double precision floating point number. You do not get 220523/1000.

And then you sqrt() that as a double precision floating point number. The result can be different for Intel vs AMD vs ARM, and in theory could be different for different models within those architectures, or could in theory depend on whether it was stand-alone or part of a vectorized operation.

Then you apply those increasingly uncertain numbers as the boundaries of integration of an integral that has a division, and you wonder why different software systems give you different answers.

int() is a request to produce indefinitely precise results, exact values, algebraic numbers if possible, but you are using inexact boundaries. You were never going to get back meaningful results.