Math Is Fun Forum

Fistfiz

Member

Registered: 2012-07-20

Posts: 33

Wronksian determinant in 2nd order linear DE

Hi guys,
I'm studying some demonstrations about 2nd order differential equations of the form:
y''+2by'+ay=f(t)
where a,b are constants.
Suppose that u,v are linearly independent solutions. Now, in several demonstrations, it's needed that the Wronksian determinant of u,v it's different from zero.
I see from Abel's identity (http://en.wikipedia.org/wiki/Abel's_identity) that if this is true for some t0 value, then it's true for all t. Provided this, can I always say that the Wronksian of u,v is always non-zero??

---

30+2=28 (Mom's identity)

scientia

Member

Registered: 2009-11-13

Posts: 224

Re: Wronksian determinant in 2nd order linear DE

Yes, if you can show that

for some

then

for all

.

Fistfiz

Member

Registered: 2012-07-20

Posts: 33

Re: Wronksian determinant in 2nd order linear DE

Of course but what i meant was: can I avoid to include W(t0)!=0 for some t0 in my hypotesis? In other words, if I have two linearly independent solutions u and v, can I automatically say W(u,v)!=0 for all t?

---

30+2=28 (Mom's identity)

scientia

Member

Registered: 2009-11-13

Posts: 224

Re: Wronksian determinant in 2nd order linear DE

I see. Well, if

are differentiable, then linear independence implies

for all

. If they are not both differentiable, then it is possible that they are linearly independent yet

. See http://en.wikipedia.org/wiki/Wronskian# … dependence for an example.

Fistfiz

Member

Registered: 2012-07-20

Posts: 33

Re: Wronksian determinant in 2nd order linear DE

Thank you, this remark:
"Peano (1889) observed that if the functions are analytic, then the vanishing of the Wronskian in an interval implies that they are linearly dependent."
opened my eyes

---

30+2=28 (Mom's identity)