how to remove difference error of .1e^-16

5 次查看（过去 30 天）

bhan stokes 2017-8-22

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/353598-how-to-remove-difference-error-of-1e-16

编辑： Jan 2017-8-22

I am working on a program where I am concatenating a number(say 0.3654) after converting it to binary with other bits and then retrieving the same number later on ... when I compare these two numbers it gives an error difference of .111e^-15 which is although very less but it is affecting the results. All the numbers I m dealing with are decimal numbers (0.54354, 0.44362 applying ceil,floor, round is not working.)...can anyone please help to remove the error.

3 个评论
显示 1更早的评论隐藏 1更早的评论

Stephen23 2017-8-22

编辑：Stephen23 2017-8-22

"I am concatenating a number(say 0.3654).."

Most likely you aren't. You are probably concatenating the closest binary equivalent to the decimal value 0.3654. This is not a bug in MATLAB, this is simply a side-effect of how floating point numbers are stored in binary:

https://www.mathworks.com/matlabcentral/answers/57444-faq-why-is-0-3-0-2-0-1-not-equal-to-zero

https://www.mathworks.com/matlabcentral/answers/41536-why-is-2-24-10-22-4-not-equal-to-0

https://www.mathworks.com/matlabcentral/answers/316889-why-two-equal-numbers-are-not-equal

etc., etc.

"can anyone please help to remove the error."

You need to design your algorithm to take floating-point error into account.

Jan 2017-8-22

编辑：Jan 2017-8-22

@Stephen: This is the correct answer in my opinion. I hesitate to repeat your explanations by posting it as an answer, but I would vote for it, if you move it to the section for answers.

请先登录，再进行评论。

请先登录，再回答此问题。