I think he knows that he was for looking for a method for the ones that do have solutions.

]]>That is the thing,I want a method that will work for any n.

Those two sites are saying the mathematical method does not yet exist. If it did this would not be an open question. It can be done by computer though.

There are solutions if you can work with positive and negative numbers. You have not told me what constraints there are on the problem.

]]>If x>0 and y>0 and z>0 then as far as I know there are only computer solutions. It is an open problem

http://en.wikipedia.org/wiki/Erd%C5%91s … conjecture

http://en.wikipedia.org/wiki/Egyptian_fraction

whether there is a solution for all n.

Sometimes it can be solved, for instance when

then a solution is:

]]>