Matlab Numerical integral improvement

Hi, I have the integral below:

F_A_I=@(x) besselj(1,x.*3.5).*besselj(1,x.*0.5);
A=integral(F_A_I,0,Inf,'RelTol',1e-6,'AbsTol',1e-12,'ArrayValued',true);

But Matlab said:

Warning: Reached the limit on the maximum number of intervals in use. Approximate bound on error is   1.7e+00. The integral may not exist, or it may be difficult to approximate numerically to the requested accuracy.

while Mathematica can give the answer straightforward A=0.0205664

Could you please help me to improve my code. Thanks

[Niels]

probably a definition gap in your function, integral might converge to inf, in these cases matlab displays -> Reached the limit on the maximum number of intervals in use.

Answer 1

[Karan Gill]

Updated answer for R2017b. Use int and convert the symbolic solution to floating point.

>> syms x
f = int(besselj(1, x/2)*besselj(1, (7*x)/2),x,0,inf)
f =
-(4*(100*ellipticE(1/49) - 99*ellipticK(1/49)))/(21*pi)

>> f_dbl = double(ans)
f_dbl =
    0.0022

>> f_vpa = vpa(f)
f_vpa =
0.0022054352588140668793354496265733

OLD ANSWER from 13-Feb-2017

Try using "vpaintegral": https://www.mathworks.com/help/symbolic/vpaintegral.html#bvaajp9-1

The convert to double using "double".

[Walter Roberson]

vpaintegral() with up to 10000 MaxFunctionCalls complains it cannot reach required precision.

The ratio oscillates a lot.

[Karan Gill]

Updated answer with solution starting R2017b.