Multivariable calculus
Double Integral of xy e^(-x^2) Over [0,1] by [0,2]
The double integral of x times y times e to the minus x squared, over x from 0 to 1 and y from 0 to 2, equals 1 minus 1 over e, about 0.63212. The integrand factors, and the x integral is the substitution u = x squared whose du the loose x supplies, giving half of 1 minus 1 over e, times 2.
Numerically 0.632121, confirmed by quadrature on every build.
Factor, then substitute in x
The integrand is times , so it factors, and the interesting half is the integral.
Substitute , so and . The limits move from to .
The iterated version, if you prefer to see the layers
Freeze . Everything except is a constant multiplier, so the inner integral is a one line power rule.
The factor 2 from the integral is exactly what the substitution wanted, so the outer antiderivative comes out with no fractions at all.
The mistake: assuming e to the minus x squared can be integrated on its own
The loose is doing all the work. Remove it and has no elementary antiderivative at all: it is the error function, and no substitution will produce one.
- , because is half of .
- is not elementary, so a problem that asks for it in closed form has a typo.
- Writing is the other frequent slip: differentiate it and the product rule leaves an extra term.
Check the size before you move on. On this rectangle the integrand peaks at about 0.858, at and , and the area is 2, so the answer must sit between 0 and about 1.72. The value 0.632 fits.
Frequently asked questions
Why does the substitution need the x in front?
Because forces . The integrand must supply that , or the substitution leaves a leftover that cannot be written in terms of . That is the difference between an easy integral and a non-elementary one.
What is the value if y runs only from 0 to 1?
The integral becomes instead of 2, so the answer is . Shrinking one side scales the whole answer by the ratio of the two integrals.