Is it possible to solve a boundary value problem like this?
f1[y_?NumericQ] :=
NIntegrate[
Sqrt[t + (0.001686429/2) t^2]/(1 + Exp[t - y]), {t, 0, Infinity},
MaxRecursion -> 100]
f2[y_?NumericQ] :=
NIntegrate[
t Sqrt[t + (0.001686429/2) t^2]/(1 + Exp[t - y]), {t, 0, Infinity},
MaxRecursion -> 100]
and,
sol = NDSolveValue[{X''[x] +
4*\[Pi]*0.0072971432*
x*(2.28986112*^-8)^3*(((3/(4*\[Pi]))*
HeavisideTheta[
4.16095*10^-8 - x]) - ((Sqrt[
2]/\[Pi]^2)*(0.0037851841)^3*(0.001686429)^(
3/2)*(f1[X[x]/x] + (0.001686429*(f2[X[x]/x]))))) == 0,
X[10^-20] == 0,
X'[0.00002096185] == X[0.00002096185]/0.00002096185},
X, {x, 10^-20, 0.00002096185}];
then,
Plot[sol[x], {x, 10^-20, 0.00002096185}]
{sol[0.00002096185], sol'[0.00002096185]}
But it is not giving a desired answer. can anyone help?
{X[10^-20] == 0, X'[0.00002096185] == X[0.00002096185]/0.00002096185} /. X -> sol
evaluates to{True, True}
What exactly doesn't work and what is the "desired answer"? $\endgroup$Plot[sol[x], {x, 10^-20, 0.00002096185}, WorkingPrecision -> 20]
? $\endgroup$sol
is not a constant, merely linear. If you suspect a precision issue, convert everything to using arbitrary-precision rather than machine precision. Looks like same result. $\endgroup$