FD1[k_, η_] := η^(k + 1) ;  

eqn = FD1[(d - t)/t, ηs] + FD1[(d - t)/t, ηs - vd] == nd;
(* I want to solve eqn for ηs  where d,t,vd are constants in the limit of very small vd*)

Solve[FD1[(d - t)/t, ηs] + FD1[(d - t)/t, ηs] == nd, {ηs}]
(*First I get an estimate of ηs by solving eqn for vd=0 *)

AsymptoticSolve[eqn, {ηs, 2^(-(t/d)) nd^(t/d)}, {vd, 0, 2}]
(*Then I use AsymptoticSolve to get a solution for ηs in Taylor series of vd, I have only started with 3 terms in the series but I also want to see higher terms dependence*)

(*Now I want to put this value of ηs in another expression J *)
J[η_] := FD1[(d - 1)/t, η] - FD1[(d - 1)/t, η - vd];

(*In the above expression for J, I want to evaluate the J[ηS] which I evaluated from eqn*)

How do I use the expression for $\eta s$ from eqn and susbtitue in the expression for $J$. Even when I do it manually, the expression is very huge, and I cannot simplify it. How can Mathematica simplify the expression. I basically finally want to get the dependence of $J$ on $nd$ once I substitute $\eta s$ from eqn into $J$.


The normal way is simply

FD1[(d - 1)/t, η] - FD1[(d - 1)/t, η - vd]/.AsymptoticSolve[eqn, {ηs, 2^(-(t/d)) nd^(t/d)}, {vd, 0, 2}]
  • $\begingroup$ This may not work as written, because AsymptoticSolve yields a conditional expression in this case. $\endgroup$
    – bbgodfrey
    Mar 22 '20 at 0:42
  • $\begingroup$ So you need to try and anticipate the effect of the conditions. Try to simplify the problem until it can be explained. $\endgroup$ Mar 22 '20 at 10:48

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.