Multiple sums are documented with two or more iterators, for example:

Sum[1/(j^2 (i + 1)^2), {i, 1, Infinity}, {j, 1, i}]

however the same answer can be obtained with two nested sums:

Sum[Sum[1/(j^2 (i + 1)^2), {j, 1, i}], {i, 1, Infinity}]

Is there a specific example where a properly formatted single Sum[] command with two iterators will give or should give a different answer from two nested Sum[] commands with the same limits?

I am interested in mathematical reasons, like possible convergence issues, for the moment I am not interested in evaluation speed, neither in ill-formatted code. This is my context: I am the main author of an old Mathematica package for quantum mechanics Quantum, and I have been working during several months on a totally new version, from zero. In the package I define my own versions of several Mathematica commands, which will work on kets and operators instead of complex numbers.

In the old package a single QuantumSum with two iterators was always transformed to two nested QuantumSum commands, but I have been always worried that they way it is implemented in Mathematica implies that they are not always the same. If so, do you have a specific example, maybe one relevant to quantum mechanics?

  • 2
    $\begingroup$ There is a mistake in the subject line, the first iterator is for the outer sum when expanded into nested sums. Your code examples do this correctly. Either a Sum with multiple iterators or expressed with nested sums should be equivalent. If not, I believe that there would be a bug in Mathematica's implementation. $\endgroup$
    – Bob Hanlon
    Commented May 17, 2016 at 0:06
  • $\begingroup$ Thank you @BobHanlon I corrected the subject line. $\endgroup$ Commented May 17, 2016 at 0:45
  • 2
    $\begingroup$ Shouldn't be a problem for finite sums (or more precisely, if the inner sum is finite); stuff like the sum for the Madelung constant are a different matter, tho. $\endgroup$ Commented May 17, 2016 at 0:56


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.