Use Riemann sums to evaluate? : #int_0^(pi/2) \ sinx \ dx# ?

1 Answer
Sep 15, 2017

# int_0^(pi/2) \ sinx \ dx = 1 #

Explanation:

By definition of an integral, then

# int_a^b \ f(x) \ dx #

represents the area under the curve #y=f(x)# between #x=a# and #x=b#. We can estimate this area under the curve using thin rectangles. The more rectangles we use, the better the approximation gets, and calculus deals with the infinite limit of a finite series of infinitesimally thin rectangles.

That is

# int_a^b \ f(x) \ dx = lim_(n rarr oo) (b-a)/n sum_(i=1)^n \ f(a + i(b-a)/n)#

And we partition the interval #[a,b]# equally spaced using:

# Delta = {a+0((b-a)/n), a+1((b-a)/n), ..., a+n((b-a)/n) } #
# \ \ \ = {a, a+1((b-a)/n),a+2((b-a)/n), ... ,b } #

Here we have #f(x)=sinx# and so we partition the interval #[0,pi/2]# using:

# Delta = {0, 0+1(pi/2)/n, 0+2 (pi/2)/n, 0+3 (pi/2)/n, ..., pi/2 } #

And so:

# I = int_0^(pi/2) \ sinx \ dx #
# \ \ = lim_(n rarr oo) (pi/2)/n sum_(i=1)^n \ f(0+i*(pi/2)/n)#
# \ \ = lim_(n rarr oo) pi/(2n) sum_(i=1)^n \ sin((ipi)/(2n) )# ..... [A}

If we were dealing with polynomials, we would now utilise standard summation formulas for powers, but as we are dealing with a sine summation those results will not help.

In order to evaluate the sine sum, We consider the following:

# 1 + z + z^2 + ... + z^n = (z^(n+1)-1)/(z-1) \ \ # for z #!= 1#

in combination with Euler's formula by taking

#z = e^(i theta) =cos theta+i sin theta#

Then applying De Moivre's theorem:

# e^(i n theta)=cosntheta+isin n theta #

and equating real and imaginary parts, we eventually find that:

# sum_(j=1)^n sin(jtheta) = (cos(theta/2)-cos(n+1/2)theta) / (2sin(theta/2)#
# " " = (cos(theta/2)-cos(ntheta+theta/2)) / (2sin(theta/2)#
# " " = (cos(theta/2)-(cos ntheta cos(theta/2) - sinnthetasin(theta/2))) / (2sin(theta/2)#

# " " = (cos(theta/2)-cos ntheta cos(theta/2) + sinnthetasin(theta/2)) / (2sin(theta/2))#

So, if we put #theta=(pi)/(2n)# and use this result in the earlier summation [A], we get:

# I = lim_(n rarr oo) pi/(2n) {(cos((pi)/(4n))-cos (pi/2) cos((pi)/(4n)) + sin (pi/2) sin((pi)/(4n))) / (2sin(pi/(4n)))}#

# \ \ = lim_(n rarr oo) pi/(2n) {(cos((pi)/(4n)) + sin((pi)/(4n))) / (2sin(pi/(4n)))}#

# \ \ = pi/4 \ lim_(n rarr oo) 1/(n) (cot((pi)/(4n)) + 1 ) #

# \ \ = pi/4 { lim_(n rarr oo) 1/(n) (1/tan((pi)/(4n))) + lim_(n rarr oo) 1/(n) }#

# \ \ = pi/4 { lim_(n rarr oo) 1/(n) (( pi/(4n))/tan((pi)/(4n))) * (4n)/pi + 0 }#

# \ \ = lim_(n rarr oo) ( pi/(4n))/tan((pi)/(4n)) #

And a standard calculus limit is:

# lim_(x rarr 0) x/tanx =1 #

Utilising this result with #x=pi/(4n)# we get:

# I = 1 #

Using Calculus

If we use Calculus and our knowledge of integration to establish the answer, for comparison, we get:

# int_0^(pi/2) \ sinx \ dx = [ -cos ]_0^(pi/2) #
# " " = -(cos (pi/2)-cos0) #
# " " = -(0-1) #
# " " = 1 #