Hi,

I would like to get some advice about the method that can be used to verify a heuristic method/algorithm that I have designed to solve particular optimization problem.

I have collected the data and compared the results between the exact and my algorithm solutions to verify the method I proposed. If my heuristic method generated an optimum solution that matched the exact solution, my algorithm should be verified to me.

When I submitted my methodology section report to my supervisor, he said I must use a method that has been used by other researchers to verify a new algorithm then cite in my report to support the methodology. He mentioned that self created verification method shouldn't be used in phd research.

Many people has suggested me to use simulation software to verify the algorithm by the available data set so results can be verified by the software which is more reliable. If I am not trying to solve problem as the one who prepared the dataset. How could possibly I use that to verify my algorithm. I am not really comparing how my algorithm perform with other algorithm in solving a common problem.

Please advise whether my method should be okay or I should follow my supervisor's comment. Thanks