This is a problem from "A Course of Pure Mathematics" by G H Hardy. Find the limit limx→0xsin(sinx)−sin2xx6
PS: The answer is 1/18 and can be easily verified by a calculator by putting x=0.01
Answer
Preliminary Results:
We will use
sin(2x)−2sin(x)tan(2x)−2tan(x)=2sin(x)(cos(x)−1)tan2(x)tan2(x)⏟1−tan2(x)2tan3(x)⏟=sin−2sin3(x)cos(x)+1sincos(x)cos(2x)2sin3(x)=−cos(x)cos(2x)cos(x)+1
Therefore,
limx→0sin(x)−2sin(x/2)tan(x)−2tan(x/2)=−12
Thus, given an ϵ>0, we can find a δ>0 so that if |x|≤δ
|sin(x)−2sin(x/2)tan(x)−2tan(x/2)+12|≤ϵ
Because limx→0sin(x)x=limx→0tan(x)x=1, we have
sin(x)−x=∞∑k=02ksin(x/2k)−2k+1sin(x/2k+1)
and
tan(x)−x=∞∑k=02ktan(x/2k)−2k+1tan(x/2k+1)
By (3) each term of (4) is between −12−ϵ and −12+ϵ of the corresponding term of (5). Therefore,
|sin(x)−xtan(x)−x+12|≤ϵ
Thus,
limx→0sin(x)−xtan(x)−x=−12
Furthermore,
tan(x)−sin(x)x3=tan(x)(1−cos(x))1x3=sin(x)cos(x)sin2(x)1+cos(x)1x3=1cos(x)(1+cos(x))(sin(x)x)3
Therefore,
limx→0tan(x)−sin(x)x3=12
Combining (7) and (9) yield
limx→0x−sin(x)x3=16
Additionally,
sin(A)−sin(B)sin(A−B)=cos(A+B2)cos(A−B2)=1−2sin(A2)sin(B2)cos(A−B2)
Finishing Up:
xsin(sin(x))−sin2(x)=[(x−sin(x))+sin(x)][(sin(sin(x))−sin(x))+sin(x)]−sin2(x)=(x−sin(x))(sin(sin(x))−sin(x))+(x−sin(x))sin(x)+sin(x)(sin(sin(x))−sin(x))=(x−sin(x))(sin(sin(x))−sin(x))+sin(x)(x−2sin(x)+sin(sin(x)))
Using (10), we get that
limx→0(x−sin(x))(sin(sin(x))−sin(x))x6=limx→0x−sin(x)x3limx→0sin(sin(x))−sin(x)sin3(x)limx→0(sin(x)x)3=16⋅−16⋅1=−136
and with (10) and (11), we have
limx→0sin(x)(x−2sin(x)+sin(sin(x)))x6=limx→0sin(x)xlimx→0x−2sin(x)+sin(sin(x))x5=limx→0(x−sin(x))−(sin(x)−sin(sin(x))x5=limx→0(x−sin(x))−sin(x−sin(x))(1−2sin(x2)sin(sin(x)2)cos(x−sin(x)2))x5=limx→0(x−sin(x))−sin(x−sin(x))+sin(x−sin(x))2sin(x2)sin(sin(x)2)cos(x−sin(x)2)x5=limx→0sin(x−sin(x))x32sin(x2)sin(sin(x)2)x2=16⋅12=112
Adding (13) and (14) gives
limx→0xsin(sin(x))−sin2(x)x6=118
Added Explanation for the Derivation of (6)
The explanation below works for x>0 and x<0. Just reverse the red inequalities.
Assume that x>0 and |x|<π/2. Then tan(x)−2tan(x/2)>0.
(3) is equivalent to
(−1/2−ϵ)(tan(x)−2tan(x/2))≤sin(x)−2sin(x/2)≤(−1/2+ϵ)(tan(x)−2tan(x/2))
for all |x|<δ. Thus, for k≥0,
(−1/2−ϵ)(2ktan(x/2k)−2k+1tan(x/2k+1))≤2ksin(x/2k)−2k+1sin(x/2k+1)≤(−1/2+ϵ)(2ktan(x/2k)−2k+1tan(x/2k+1))
Summing (17) from k=0 to ∞ yields
(−1/2−ϵ)(tan(x)−limk→∞2ktan(x/2k))≤sin(x)−limk→∞2ksin(x/2k)≤(−1/2+ϵ)(tan(x)−limk→∞2ktan(x/2k))
Since limk→∞2ktan(x/2k)=limk→∞2ksin(x/2k)=x, (18) says
(−1/2−ϵ)(tan(x)−x)≤sin(x)−x≤(−1/2+ϵ)(tan(x)−x))
which, since ϵ is arbitrary is equivalent to (6).
No comments:
Post a Comment