You have to make your experimental evaluation and compare the performance of your approach to the recently published comparison work by other research.

As you getting experience, you will be getting an ability to write maturely and attractively by knowing what other people want to read and enjoy according to their needs.

Benmesbah Ouissem first, do a good survey of existing algorithm at least last 5 years. And, then put your algorithm and compare your algorithm with those already existing algorithm based on performance measurement parameters.

Dr. Argha Ghosh, in this case, can I compare my genetic algorithm with another algorithm that has used different parameters: differnt objective function, different population size, crossover and mutation probabilities???

This would help you: Article Performance of genetic algorithms and simulated annealing in...

- Hi Benmesbah,

- First of all, you need to define the aims of the problem that you want to solve. You need to research by works with the best exact e heuristics algorithms from the literature used to solve the problem that you are studying, for comparison purposes with your genetic algorithm. You should also take special care to analyze the results if the problem is multi-objective. In this case, you must use statistical analysis methods appropriate to analyze the results, eg analysis of variance and Pareto charts. Finally, I recommend you save and analyze the results of the genetic algorithm like the i) best solution in each generation, ii) the number of the generations, iii) execution time, iv) analysis of convergence (Time to Target Plot chart) and v) best solution obtained (obviously). Use statistical métodos to analyze de the results obtained for a better understanding of the efficiency and performance of the algorithm.

- Good job.

Thank you all for your help!! Your answers are really helpful.

first of all decide your evluation measures and then verify the performance of GA by comparing its results with recebtly published as well as state of the art metaheuristics optimizers

After completed the generation of the solutions for GA problems you can evaluate it results by applied it in real or hypothetical case study which has previous solution by another different methods. Number of generation , values of decision variables ...etc could be considered a good index for evaluation.

thank you very much for your proposition.

For validation of your method you can use it for benchmark function which is considered with many local point.

About second question, I think you should compare your method with other method to validate your problem if you want to write good research article.

thank you very much