why matlab2018a gives "vpa(int( exp((t-360​).*[0,1e-1​6,1e-15,1e​-14,1e-13,​1e-12,1e-1​1]) , t, 0,360 ))=[ 360.0, 0, 0, 0, 256.0, 368.0, 360.0]"??

Hi all,

Weird the matlab 2018a calculate the below code--

syms t;
vpa(int( exp((t-360).*[0,1e-16,1e-15,1e-14,1e-13,1e-12,1e-11]) , t, 0,360 ))

resulting in

[ 360.0, 0, 0, 0, 256.0, 368.0, 360.0]

Do you guys have expertise helping avoid this happening?

thanks a lot

J

[Stephen23]

How is t defined?

[Jiyuan Liu]

syms t;

[Jiyuan Liu]

the same happens using matlab 2015a...

[Jiyuan Liu]

sorry Stephen for the mis-intentional flag, I thought it is like thumbs-up..

[Ameer Hamza]

I get the following output on R2018a and R2017b.

syms t; 
vpa(int( exp((t-360).*[0,1e-16,1e-15,1e-14,1e-13,1e-12,1e-11]) , t, 0,360 ))'
ans =
                           360.0
359.99999999999352000000000231367
359.99999999993520000000000783383
359.99999999935200000000077761612
359.99999999352000000007775999991
359.99999993520000000777599999918
359.99999935200000077759999930015

Are you sure that you are running exactly the same line?

[Jiyuan Liu]

thanks Ameer, i've got the reason for the earlier error i got-- I put digits(5) (or maybe digits(3)) before these sentences. And after I reset it to 32 or 64, the results are normal again.

Answer 1

[Siwan Yang]

Hello Jiyuan,

I couldn’t reproduce your result in both R2018a and R2015a. Here is what I got: [ 360.0, 359.99999999999352000000000231367, 359.99999999993520000000000783383, 359.99999999935200000000077761612, 359.99999999352000000007775999991, 359.99999993520000000777599999918, 359.99999935200000077759999930015]

Please try changing different data type and see if you still have the issue.

Thanks,

Siwan

[Jiyuan Liu]

thanks Siwan, R2018a and R2015a work both fine. I put digits(5) (or maybe digits(3)) before these sentences, which leads to the weird results.