If ∫ 0 1 − 1 + x 1 + y − y d x = π , find y 2 − 7 8 y .
This section requires Javascript.
You are seeing this because something didn't load right. We suggest you, (a) try
refreshing the page, (b) enabling javascript if it is disabled on your browser and,
finally, (c)
loading the
non-javascript version of this page
. We're sorry about the hassle.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 |
|
In each iteration, your algorithm potentially evaluates
quad(f,0,1,args=(mid,))[0]
at least twice (except in the final iteration), and potentially three times, which must slow your routine down somewhat. Surely it would be better to assign this calculated value to a variable, and then test this variable (is it "equal to" π , less than π , or greater than π )?
Your algorithm goes from (lower,upper) to (mid,upper) if the integral is less than π , and to (mid,upper) if the integral is greater than π . In other words, you are assuming that the integral is a increasing function of y . Why? Given the exact answer, this is not at all obvious, although it is true.
Problem Loading...
Note Loading...
Set Loading...
Let us assume that y > 0 .
The substitutions u 2 = x 1 + y − y and v = u − 1 gives A = ∫ 0 1 − 1 + x 1 + y − y d x = 2 ( 1 + y ) ∫ 1 ∞ ( u 2 + y ) 2 u u − 1 d u = 2 ( 1 + y ) ∫ 0 ∞ ( ( v + 1 ) 2 + y ) 2 ( v + 1 ) v d v Defining v on the cut plane ∣ v ∣ > 0 , 0 < A r g v < 2 π , and integrating around a suitable keyhole contour gives A = 2 π i ( 1 + y ) ( R e s v = − 1 + i y ( ( v + 1 ) 2 + y ) 2 ( v + 1 ) v + R e s v = − 1 − i y ( ( v + 1 ) 2 + y ) 2 ( v + 1 ) v ) Now the function v ↦ f y ( v ) = ( ( v + 1 ) 2 + y ) 2 ( v + 1 ) v has a double pole at each of v = − 1 ± i y , and hence R e s y = − 1 + i y ( ( v + 1 ) 2 + y ) 2 ( v + 1 ) v = d v d ( v + 1 + i y ) 2 ( v + 1 ) v ∣ ∣ ∣ v = − 1 + i y = − 8 y ( 1 + y ) − 1 + i y ( y − i ) and, similarly, R e s y = − 1 − i y ( ( v + 1 ) 2 + y ) 2 ( v + 1 ) v = − 8 y ( 1 + y ) − 1 − i y ( y + i ) With this form of the cut plane, we have − i + i y = a + i b and − 1 − i y = − a + i b for some a , b > 0 , and we deduce that A = 2 y π ( b y − a ) We now calculate that a = 2 1 + y − 1 b = 2 ( 1 + y − 1 ) y obtaining, finally, A = π 8 y ( 1 + y ) ( 1 + y − 1 ) Putting t = 1 + y , we need to solve the equation ( 1 + y ) ( 1 + y − 1 ) t 2 ( t − 1 ) ( t − 1 ) ( t 2 − 8 t − 8 ) = 8 y = 8 ( t 2 − 1 ) = 0 Thus we must have t = 1 , 4 ± 2 6 . Since y > 0 we must have t > 1 and hence t = 4 + 2 6 . Thus y = 3 9 + 1 6 6 , and hence y 2 − 7 8 y = 1 5 .
Note that although t = 1 is a solution of the final cubic equation, y = 0 is not a solution to this problem. The formula for A is not valid for y = 0 , and indeed the limit of A as y → 0 is 4 1 π , which is the correct value of the integral when y = 0 , as an x = sin 4 θ substitution readily shows). By its definition, t must be positive, and so the 4 − 2 6 value must be discarded from this analysis.
That said, there is a second solution for y , where − 1 < y < 0 . If we put y = − p 2 where 0 < p < 1 , then the poles for f y ( v ) are − ( 1 + p ) and − ( 1 − p ) , both real and negative. The same contour integration can be attempted, and in this case we obtain A = π 4 p 1 − p 2 ( 1 + p − 1 − p ) and we obtain A = π when p 4 + 7 8 p 2 − 1 5 = 0 , or y 2 − 7 8 y − 1 5 = 0 , and hence we obtain a solution for y with − 1 < y < 0 , namely y = 3 9 − 1 6 6 . This second solution still gives y 2 − 7 8 y = 1 5 , of course.