The main strategy for integration by parts is to pick u and dv so that vdu is simpler to integrate than udv. Sometimes this isn't possible. In those cases we look for ways to relate ∫udv to ∫vdu algebraically, and then use algebra to solve for ∫udv.
Example: Find ∫sin2xdx.
Solution: We set u=sin(x) and dv=sin(x)dx, so v=−cos(x) and du=cos(x). Applying integration by parts gives ∫sin2(x)dx=−sin(x)cos(x)+∫cos2(x)dx.
But cos2(x)=1−sin2(x), so ∫sin2(x)dx=−sin(x)cos(x)+∫dx−∫sin2(x)dx2∫sin2(x)dx=x−sin(x)cos(x)+C∫sin2(x)dx=x−sin(x)cos(x)2+C
This method is especially good for integrals involving products of ex, sin(x) and cos(x). Sometimes you need to integrate by parts twice to make it work.