I want to solve the following delayed differential equation

$$G'(t)=\Lambda +\omega G(t-\tau)-\mu G(t),$$

when $G(t)=G_0>0$ for $t\in[-\tau,0]$ and $\Lambda,\omega, \mu, \tau>0$. Note that $G_0, \Lambda, \omega, \tau, \mu$ and final time $T_f$ are arbitrary.

I tried to solve it using Wolfram Mathematica 11.0 with code

sol=DSolve[{G'[t]==a+b*G[t-d]-u*G[t], G[t/;t<=0]==G0}, G[t], {t, 0, Tf}]

and I didn't receive the solution.

Can somebody help me?

Thank you very much! Ana



d = 1;
Tf = 1;
DSolve[{G'[t] == a + b*G[t - d] - u*G[t], G[t /; t <= 0] == G0}, G[t], {t, 0, Tf}];

(* {{G[t] -> Piecewise[{{G0, t <= 0},
{(-a + a*E^(t*u) - b*G0 + b*E^(t*u)*G0 + G0*u)/(E^(t*u)*u), 
  Inequality[0, Less, t, LessEqual, 1]}}, Indeterminate]}}*)

DSolve needs a value for d(better exact) and Tf to solve. If you don't gives then MMA can't solve or gives a Error messages.

EDITED -27.04.2018.

Using Laplace transfrom to eq:

eq = G'[t] == a + b*G[t - d] - u*G[t];
eq2 = LaplaceTransform[eq, t, s] 

(* -G[0] + s LaplaceTransform[G[t], t, s] == 
a/s - u LaplaceTransform[G[t], t, s] + 
b LaplaceTransform[G[-d + t], t, s] *)

in the last expression MMA can't find transform of LaplaceTransform(G[t - d], t, s) see page 120 (it's weakness of this function) and puttting initial condition G[0]=G0

eq3 = -G[0] + s LaplaceTransform[G[t], t, s] == a/s - u LaplaceTransform[G[t], t, s] + 
b *Exp[-d*s]*LaplaceTransform[G[t], t, s] /. G[0] -> G0

sol = LaplaceTransform[G[t], t, s] /. Solve[eq3, LaplaceTransform[G[t], t, s]][[1]]
(* (E^(d s) (a + G0 s))/(s (-b + E^(d s) s + E^(d s) u)) *)

Inverse Laplace Transform:

Total@(InverseLaplaceTransform[#, s, t] & /@ {(sol // Apart)[[1]], (sol // Apart)[[2]]})

(* a/u + (E^(-t u) (-a + G0 u))/u + 
b InverseLaplaceTransform[(a + G0 s)/(
s (s + u) (-b + E^(d s) s + E^(d s) u)), s, t]*)

MMA can only find for first experssion.

  • $\begingroup$ I understood your suggestion, but I want to solve the differential equation for an arbitrary $\tau>0$ and for all final time $T_f>0$... $\endgroup$ – Ana Apr 26 '18 at 14:53
  • $\begingroup$ As I said, DSolve can't solve for arbitrary d and Tf.DSolve have some limitations. Putt your questions to math.stackexchange.com maybe will help you there. $\endgroup$ – Mariusz Iwaniuk Apr 26 '18 at 15:31
  • $\begingroup$ I followed your suggestion, putting my question in math.stackexchange.com.I'm going to analyze your previous edition (27-04-2018). Thank you! $\endgroup$ – Ana Apr 27 '18 at 16:57

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.