Advanced Search

Question: Looks like a regression bug in dsolve

Question:Looks like a regression bug in dsolve

Rouben Rostamian 7918 Maple 2023
ode dsolve differential-equations heaviside
February 20 2024
2

Applying Maple 2023's dsolve to the ODE shown below yields the solution y(t)=0 which is obviously incorrect.  Maple 2021 and earlier used to give a nonzero (albeit not very useful) answer.

> 

restart;
> 

F := (t-1)*(t-2)/(t^2+1)^3*(Heaviside(t-2)-Heaviside(t-1));

(t-1)*(t-2)*(Heaviside(t-2)-Heaviside(t-1))/(t^2+1)^3

> 

plot(F, t=0..4);

> 

de := diff(y(t),t,t) + diff(y(t),t) = F;

diff(diff(y(t), t), t)+diff(y(t), t) = (t-1)*(t-2)*(Heaviside(t-2)-Heaviside(t-1))/(t^2+1)^3

> 

ic := y(0)=0, D(y)(0)=0;

y(0) = 0, (D(y))(0) = 0

Huh?

> 

dsolve({de,ic}, y(t));

y(t) = 0

Specifying method=laplace will make that work, but how is an

unsuspecting user to know that what's obtained above is incorrect?

Download dsolve-bug.mw
Please Wait...