Dear all,
I have to know what is the need to run the optimization code for multiple times...
For example,
I m implementing a new variant optimization algorithm, I am setting maximum number of iterations as 100. And number of trails or runs as 25.
What kind of outcome we ll get from number of trails... How to ensure that our proposed work performance is better than other algorithms.
Hi,
running multiple times may have different effects:
- if the method is stochastic (like metaheuristics), multiple runs may produce different results ;
- if you run your method with different starting points, you may obtain different solutions (e.g. if your problem is nonconvex) ;
- if you use multiple exploration strategies (e.g. in a branch and bound framework), you may converge at different speeds.
- etc
Best,
Charlie
Hello Raja Success
It's a very good explanation from Charlie Vanaret. I completely agree with his response. As per my knowledge, different starting points and different iterations lead to different outcomes, that's the reason why we use many numbers of runs and iterations. To conclude which algorithm is better for your specific problem, the algorithm which converges in quickly is considered as a better algorithm for your specific problem. But it may vary for the problem to problem.
Best wishes
PVR
Yes, I do agree with the answer from Charlie Vanaret. Multiple iterations of your code will give you a kind of optimal solution. In each iteration in WSN simulation using simulator give variations in result. So it is better to iterate multiple times.
- Badges
- Science topic
- Similar topics
- Biological Science
- Kinesiology
- Running