Discuss the convergence of Determine for which values of the integral converges absolutely, for which it converges only conditionally, and for which it diverges.
Change variables so the weight depends on one variable. The integrand only sees through the denominator, so set , . Then and the Jacobian is . The constraints , become , and becomes , so
Collapse the inner integral with a product-to-sum identity. Using with and gives . Integrating over , the first piece contributes and the second contributes , so
Split the result into two oscillatory tails. Substituting back, The two-dimensional convergence question is now two one-dimensional improper integrals of the same standard shape.
Apply the Dirichlet test to each tail. The partial integrals and stay bounded, so converges whenever decreases to , that is , and converges whenever . The second condition is the binding one. At the second tail is , which oscillates without settling, and for its amplitude even grows, so convergence holds exactly when .
Test absolute convergence separately. Taking absolute values destroys the cancellation: grows linearly, for large , because has a positive average over each period. Hence , which converges exactly when , that is .
Assemble the trichotomy. Absolute convergence needs and plain convergence needs , so for the integral converges but not absolutely, and for it diverges.
Need to solve a different problem like this? Open the solver →