Integral involving a partial fraction expansion
Today we show the proof of this integral posted by @integralsbot ∫∞0∞∏k=1k2x2+k2dx=πn2(2n−1) The proof relies on the partial fraction decomposition and some properties of the binomial coefficient.
Proof:
We need the following partial fraction decomposition (inspired by this this exercise ): n∏k=11x+k2=2n∑k=1(−1)k−1k2(x+k2)(n−k)!(n+k)!
Here is the proof:
Since deg1<deg(x+1)(x+22)⋯(x+n2) there exists a partial fraction decomposition: 1(x+1)(x+22)⋯(x+n2)=n∑k=1Akx+k2 Multiplying by (x+1)(x+22)⋯(x+n2) 1=n∑k=1Akn∏i=1i≠k(x+i2) Suppose that x=−m2 with 1≤m≤n Hence 1=Amn∏i=1i≠m(−m2+i2)=Am(−m2+1)(−m2+22)+⋯(−m2+(m−1)2)(−m2+(m+1)2)⋯(−m2+n2)=Am(−(m2−1))(−(m2−22))⋯(−(m2−(m−1)2))((m+1)2−m2)⋯(n2−m2)=Am(−1)m−1(m−1)(m+1)(m−2)(m+2)⋯(m−(m−1))(m+(m−1))((m+1)+m)((m+1)−m)⋯(n−m)(n+m)=Am(−1)m−1[(m−1)(m−2)⋯2⋅1⋅2⋯(n−m)][(m+1)(m+2)⋯(m+(m−1))(m+(m+1))⋯(n+m)]=Am(−1)m−1[(m−1)!(n−m)!][(m+1)(m+2)⋯(m+(m−1))(m+(m+1))⋯(n+m)] Multiplying and dividing by (2m)m!(2m)m! 1=Am(−1)m−1[(m−1)!(n−m)!][m!(m+1)(m+2)⋯(m+(m−1))(2m)(m+(m+1))⋯(n+m)](2m)m!=Am(−1)m−1[(m−1)!(n−m)!][(n+m)!](2m)m! Multiplying and dividing by m 1=Am(−1)m−1m!(n−m)!(n+m)!(2m2)m!=Am(−1)m−1(n−m)!(n+m)!2m2 Therefore Am=2(−1)m−1m2(n−m)!(n+m)!1≤m≤n Hence n∏k=11x+k2=2n∑k=1(−1)n−1k2(x+k2)(n−k)!(n+k)! Back to the integral: I=∫∞0∞∏k=1k2x2+k2dx=(n∏k=1k2)∫∞0∞∏k=11x2+k2dx=2n!2∫∞0n∑k=1(−1)n−1k2(x2+k2)(n−k)!(n+k)!dx=2n!2n∑k=1(−1)n−1k2(n−k)!(n+k)!∫∞01x2+k2dx=2n!2n∑k=1(−1)n−1(n−k)!(n+k)!∫∞01(xk)2+1dx=2n!2n∑k=1(−1)n−1k(n−k)!(n+k)!∫∞01w2+1dx(w↦xk)=2n!2n∑k=1(−1)n−1k(n−k)!(n+k)![arctan(w)]∞0=πn!2n∑k=1(−1)n−1k(n−k)!(n+k)! where we have used the fact that n∏k=1k2=(n∏k=1k)(n∏k=1k)=n!2
Finally, we have to prove that πn!2n∑k=1(−1)n−1k(n−k)!(n+k)!⏟S=πn2(2n−1) Note that S=1(2n)!n∑k=1(−1)k−1k(2n)!(n−k)!(n+k)!=1(2n)!n∑k=1(2nn+k)(−1)k−1k=1(2n)!2n∑j=n+1(2nj)(−1)n−j−1(j−n)(j=n+k)=1(2n)!2n∑j=n+1(2n2n−j)(−1)n−j−1(j−n)((vv−w)=(vw))=1(2n)!m=0∑m=n−1(2nm)(−1)m−n−1(n−m)(m=2n−j)=(−1)n−1(2n)!n−1∑m=0(2nm)(−1)m(n−m)=(−1)n−1(2n)![nn−1∑m=0(2nm)(−1)m−n−1∑m=0(2nm)(−1)mm]=(−1)n−1(2n)![nn−1∑m=0(2nm)(−1)m−2nn−1∑m=0(2n−1m−1)(−1)m]=(−1)n−1(2n)=(vw)+(vw−1))=1(2n)j(vj)=(−1)n(v−1m))=1(2n)![12n−1(2n)!(n−1)!(n−1)!−12n(2n)!(n−1)!(n−1)!]=n2(2n−1)n!n! Hence I=∫∞0∞∏k=1k2x2+k2dx=πn!2n∑k=1(−1)k−1k(n−k)!(n+k)!=πn!2S=πn2(2n−1) Therefore, we can conclude ∫∞0∞∏k=1k2x2+k2dx=πn2(2n−1)