Home » Ask & Discuss » Mathematics. » Differential Calculus « Back to Discussion
Differential Calculus
Comments (12)
Let
be a function that is differentiable infinitely many times and all its derivatives are also continuous. If
satisfies the differential equation
for all
and
, find
.
Apply Rolle's theorem for the interval [0,2009]
so for some p we have f'(p) = 0 since f(2009)=f(0)=0
Now apply mean value theorem for the interval [p,2009]
hence for some q we have
f''(q) = f'(2009) - f'(p) / 2009-p
= f'(2009)/ [2009-p] since f'(p) = 0
Now
f(x) = f ' (x) + f '' (x)
f ' (x) = f '' (x) + f ''' (x)
f '' (x) = f '''' (x) + f '''' (x)
.........
For x = 2009
f(2009) = f ' (2009) + f '' (2009)
but since f(2009) = 0
f ' (2009) = - f '' (2009)
f '' (2009) = -f ''' (2009)
f ''' (2009) = - f '''' (2009)
.......
Hence we have f ' (2009) = f ''' (2009) = f ''''' (2009) = ......... = - f '' (2009) = - f '''' (2009) = ......
But f '' (2009) = f ''' (2009) + f '''' (2009)
and also f ''' (2009) = - f '''' (2009) we have f '' (2009) = 0
Therefore f ' (2009) = f ''' (2009) = f ''''' (2009) = ......... = - f '' (2009) = - f '''' (2009) = ...... = 0
Hence f '' (q) = 0
But f(q) = f ' (q) + f '' (q) = f ' (q)
differentiate above eq. to get f ' (q) = f '' (q) = 0
hence f(q) = 0
Now we can prove that f(q) = f ' (q) = f '' (q) = ...=0 just as we did for x = 2009
Now apply mean value theorem between [p,q] to get a point r such that f(r) = f' (r) = f '' (r) = ..... = 0 just as we did for point q
Same thing can be done with the interval [0,p]
Thus by induction we can prove that all the point between 0 and 2009 are zero
Thus f(x) = 0 for x = [0,2009]
Hence f(1004) = 0
Ankit... your solution is correct but a bit lengthy. Indeed, we prove that
for all
. To see this we first note that if
is constant, it obviously has to be zero identically.
So, let
be non-constant. Then it must attain a maximum and minimum in
.
Claim 1:
cannot have a positive maximum. If it had so, then at the point of maximum,
, and
so that
, a contradiction. Hence
for all
.
Claim 2:
cannot have a negative minimum. If it had so, then at the point of minimum,
, and
so that
, again a contradiction. Hence,
for all
Together, Claim 1 and Claim 2 imply that
for all
.
absolutely right sir, this method must have striked b4 the lengthy one
sometimes previous problems do affect the thinking (in the last prob. intermediate value theorem was used so my thinking was diverted)
such thinking (the one in ur solution) does help in many cases specially when we deal with problems based on parity and odd/even
just prove that one term cannot be odd and even at the same time just as in the above prob. required to prove that f(x) cannot be =ve and -ve aat the same time and hence it ought to be zero.
nice solution short and sweet (lol) :D :D :D
hmm I thought for x=2009 f ' (x) = - f '' (x)
so diff. both sides to get f '' (x) = -f ''' (x)
but now I get it if two fuctions are equal at some point (or if we have some relation between them) then the equality may not be true for its derivative
plz. help sir, how would u prove it using Rolle's and lagrange's mean value theorem
Now that Bhatt sir has pointed it out, the line of action followed by Ankit is incorrect ( probably I got carried away with the flow).
However, we can also apply brute force. We first solve the differential equation (which is easy to do):
The auxiliary equation is
. So the eigenvalues are
and
So the general solution is
To find the solution satisfying the given boundary conditions, we see that
and
, since the quantity in the bracket is non-zero. And so
. As such
for all ![xin [0,2009]](http://alt1.mathlinks.ro/latexrender/pictures/7/5/7/757dd7aa36a4177fcc560dfeebe079d789d531df.gif)












Wait a minute sir, posting the solution