Question #68d98
↳Redirected from
"Suppose that I don't have a formula for #g(x)# but I know that #g(1)
= 3# and #g'(x) = sqrt(x^2+15)# for all x. How do I use a linear approximation to estimate #g(0.9)# and #g(1.1)#?"
The answer is #2ln(e^{x}+1)-x+C#
Start by letting #u=e^x+1# so that #du=e^{x}dx# and #e^x-1=u-2#. Then #dx=1/e^(x) du=1/(u-2)du# and the integral becomes
#\int (e^x-1)/(e^x+1)dx=\int (u-2)/(u(u-1))du#.
Next, use the Method of Partial Fractions to write #(u-2)/(u(u-1))# as #2/u-1/(u-1)#, which can be easily integrated to get #2ln|u|-ln|u-1|+C#.
Now substitute #u=e^x+1#, which is never negative, and use the fact that #ln(e^x)=x# to get
#\int (e^x-1)/(e^x+1)dx=2ln(e^{x}+1)-x+C#.
This can be checked by differentiation:
#d/dx(2ln(e^x+1)-x+C)=2/(e^x+1)*e^x-1=(2e^x-(e^x+1))/(e^x+1)=(e^x-1)/(e^x+1)#
#int(e^x-1)/(e^x+1)dx=2ln(e^x+1) -x+ C #
Given: #int(e^x-1)/(e^x+1)dx#
Insert zero into the numerator:
#int(e^x+ 0-1)/(e^x+1)dx#
In place of the 0 we write #e^x-e^x#:
#int(e^x + e^x-e^x-1)/(e^x+1)dx#
Now we do some clever grouping:
#int((e^x + e^x)-(e^x+1))/(e^x+1)dx#
We combine the first group:
#int(2e^x-(e^x+1))/(e^x+1)dx#
Separate into two fractions:
#int(2e^x)/(e^x+1)-(e^x+1)/(e^x+1)dx#
Please notice that the second fraction becomes 1:
#int(2e^x)/(e^x+1)-1dx#
Separate into two integrals:
#2inte^x/(e^x+1)dx -intdx#
For the first integral we let #u = e^x+1#, then #du = e^xdx#:
#2int1/udu -intdx#
We know these integrals very well:
#2ln|u| -x+ C#
Reverse the substitution:
#2ln(e^x+1) -x+ C#
We can also do this integral as follows:
#int(e^x-1)/(e^x+1)dx=(e^x+1-2)/(e^x+1)dx#
Splitting up the fraction as #(e^x+1)/(e^x+1)-2/(e^x+1)#, this becomes:
#=int(1-2/(e^x+1))dx#
We can integrate the first term easily:
#=x-2int1/(e^x+1)dx#
Now, multiply the integrand by #e^-x/e^-x#. This seems ridiculous, but you'll see why it works in a second:
#=x-2int(e^-x)/(1+e^-x)dx#
Let #u=1+e^-x#, implying that #du=-e^-xdx#. Then:
#=x+2int1/udu#
#=x+2lnabsu+C#
#=x+2lnabs(1+e^-x)+C#
As #e^-x>0# for all Real values of #x#, the absolute value bars aren't needed.
#x+2ln(1+e^-x)+C#
Which can be shown to be equivalent to the other provided answers.