f(x)=x^(2)*(sin(1/x)) is it differentiable at x=0
a function being differentiable at a given point means the derivative is continuous at that point. the derivative of f is 2xsin(1/x)-cos(1/x). it's fairly straightforward to check continuity at 0.
thanks cnknd!! but i have doubt that does cos(1/x) also wobbles between 0 ,-1 and 1 like sin1/x
ok, lets break that thing into 2 parts first: 2xsin(1/x), is this continuous at 0?
yes
2nd part, cos(1/x): is this continuous at 0?
remember what it means to be continuous: the left and right limits exist and are equal
cos1/x is similar to sin1/x as x approaches to 0 i.e it keeps oscillating between 0 ,1 ,-1 as x approaches 0
well the answer here is that neither the left nor right limits exist for cos(1/x) as x goes to 0 (showing why this is true is a bit more involved, let me know if u want to go through the rigor)
yes please that will be helpful
ok, so let's look at the right limit of cos(1/x) as x -> 0 i'll first make a change of variables: let y = 1/x
so this limit now becomes: limit of cos(y) as y -> positive infinity
well, i havent specified how y is going to infinity. perhaps it goes as the following sequence: 1, 2, 3, 4, ....; or this sequence: 3, 5, 7, 9, 11, ...
if i can identify 2 ways (2 different sequences) for y to approach infinity, but the 2 paths would give different limits of the function cos(y), then it doesn't make sense to say that the limit of the function exists
following me so far?
i dint get that 2 sequences
a sequence here just means how y is going to infinity.
so here's one way for y to approach infinity: y = 0, 2pi, 4pi, 6pi, ..... and so the value of cos(y), as y approaches infinity in this path, goes as: cos(y) = 1, 1, 1, 1, ... so without going further, have i convinced you that the limit of cos(y) is 1?
yes i got that
well here's another sequence for y: y = pi, 3pi, 5pi, .... so we have: cos(y) = -1, -1, -1, ... and it looks like the limit here (for the same function) is -1. the limit can't possibly be -1 and 1 at the same time. therefore it's undefined
y dint u take npi/2 for various values of n in consideration?
i could, and that would give me a limit of 0
still the same conclusion though, the limit can't be -1, 0, and 1 at the same time, so it does not exist
ok thanks cnknd!!! that means f(x) is not differentiable at x= 0
that is correct.
in one of the class notes i saw that lhl, f'(0) =limf(0-h)-f(0)/h=h*sin(1/h) =0 ;h --> 0
and rhl, f'(0)=lim f(0+h) - f(0)/h =h* sin1/h = 0
lhl= rhl thereby showing that f'(x)is continuous at x= 0 , i cant find the mistake
i have no idea how they did this step: (f(0+h)-f(0))/h = h*sin(1/h)
they have shown f(0+h)= h^2 * sin(1/h) aand f(0)= 0 and then divided by h
o.o sorry i think i was wrong... the wikipedia page has this function as an example of differentiable but not smooth: http://en.wikipedia.org/wiki/Differentiable_function i apologize for misleading you
but even in your approach i didnt find anything wrong!! i am really confused
oh i went wrong in the definition: i thought it was asking for smooth (i.e. differentiable, and derivative is continuous). where as it was simply asking for differentiability
being differentiable and derivative is continuous? are they both different?
the distinction here is not a place in mathematics that I've looked at in detail. but wikipedia says a function can have certain kinds of discontinuity in the derivative, and still be differentiable: http://en.wikipedia.org/wiki/Differentiable_function#Differentiability_classes
so to test differentiability what should be my approach for any given f(x)
ur lecture notes would be correct for simple differentiability i.e. use the definition of the derivative.
i still dont understand that if we take derivative of f(x) and try to check its continuity it is discontinuous....and yet it is continuous using the defintion of derivative why two different answer?
oh the definition of the derivative doesnt say that the derivative is continuous
the limits in the definition of the derivative is telling you whether or not your approximations of the derivative is continuous (as they get more and more accurate to the true derivative)
but y two different answer...because even the former approach seems to be correct
well it's not asking for whether or not the derivative is continuous, it's just asking for whether or not the derivative exists
ok to test differentiability i should use limf(x+h)-f(h)/h as h --> 0
yes
ok thanks cnknd!!
Join our real-time social learning platform and learn together with your friends!