In 2A-12, d, part of the explanation in the answer sheet says that "ln(1+h) ~= h". Isn't this the linear approximation? In 2A-12, e, "ln(1+h) ~= h-(h^2/2)". Can someone please explain why the linear approx. is used in 12d, but the quad. approx is used in 12e, when both are asking for a quadratic approximation?
we start with \[ \cos x \approx 1 - \frac{x^2}{2} \] and thus \[ \ln( \cos x) \approx \ln \left(1 - \frac{x^2}{2}\right) \] next, we use \[ \ln(1-y) \approx -y - \frac{y^2}{2} \] with \( y= \frac{x^2}{2} \) we get \[ \ln( \cos x) \approx -\left( \frac{x^2}{2}\right) - \frac{1}{2} \left( \frac{x^2}{2} \right)^2\] notice that the second term is to the fourth power, and thus we drop it from our quadratic approximation, to get the result \[ \ln( \cos x) \approx - \frac{x^2}{2} \] Whoever posted the answer assumed that this process was "obvious to the most casual observer", i.e. "clearly" we need only consider the linear approximation for ln.
Thank you for your quick and helpful reply. To clarify, we know that the linear approximation is enough for ln(1-x^2/2) because it's a squared term (and therefore the quartic term will drop off, so we only need take the linear approximation)? And for 2A-12e, we keep the quadratic approximation, because x=1-h, and h^2 is squared and not quartic (and so will not drop off)?
For 2A-12e \[ x \ln x, \ x \approx 1 \] let x= 1+h (and \( h \approx 0 \)) and in terms of h we have \[ x \ln x \approx (1+h) \ln(1+h) \] use \[ \ln(1+h) \approx h - \frac{h^2}{2} \] to get \[ x \ln x \approx (1+h) \ln(1+h) \approx (1+h) \left( h - \frac{h^2}{2}\right)\] multiplying out, \[ x \ln x \approx h - \frac{h^2}{2} +h^2 - \frac{h^3}{2} = h^2 + \frac{h^2}{2} - \frac{h^3}{2}\] with h near 0, the higher order terms approach zero faster than the linear or quadratic terms, and it's reasonable to drop them. i.e. the quadratic approximation is \[ h^2 + \frac{h^2}{2} \] put in terms of x (from x= 1+h, we have h= x-1): \[ (x-1)^2 + \frac{(x-1)^2}{2} \]
Hijacking this old post, since it concerns the same problem (2A-12e) I got different answers by trying two different methods that (in my mind) should be equivalent. One is the method posted above, using x=h+1. The other is finding the quadratic approximations of x and ln(x) near x=1, using the formula \[f(x) = f(1) + f'(1)(x-1) + \frac{ 1 }{ 2 }f''(1)(x-1)^{2}\] For ln(x), this gives \[\ln x = 0 + (x-1) - \frac{ (x-1)^2 }{ 2 }\] and \[x = 1 + (x-1) = x\] Plugging both of these into the original equation gives: \[x \left[ (x-1)-\frac{ (x-1)^2 }{ 2 } \right]\] Which does not simplify to the answer given by the other method. Did I make a mistake somewhere?
Join our real-time social learning platform and learn together with your friends!