Hi.
I have been trying to reproduce the experiments done in the original: "Firefly Algorithm for multimodal optimization" (linked in the question) so far: unsuccesfully. For the moment being I'm okay if anyone point me to the right direction.
I wrote the algorithm as specified in the paper in C++ programming languaje (I also downloaded several other implementations from internet for comparation purpouses) and used the very same parameters as specified in the paper (a random steep of 0.2, an initial light intensity of 1.0 and a light decay coefficient of 1.0, a population size of 40). I used the two bright update ecuations given and for De Jung test function (as for example) a number of dimensions of 256 in a search domain in [-5.12, 5.12] as refered in common optimization literature and in paper.
In the paper the algorithm converges very quickly, as can be expected since this is a very simple test function, however, neither my implementation nor any code I have downloaded converges with that parameters.
My final questions are:
1. Am I doing something wrong with the experimental methodology or am I using wrong parameter settings (may be something different than the original paper)?
2. Do anyone knows where can I find a code sample of Firefly Algorithm that I can use to reproduce the experiments of the mentioned paper?
Please notice that there may be a lot of variations of this algorithm that can produce better results, but right now I'm only intrested in reproduce the experiments of the so-called paper.
https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=2&cad=rja&uact=8&ved=0ahUKEwiAjIfH_4XSAhUKORoKHTz6CAAQFgglMAE&url=https%3A%2F%2Farxiv.org%2Fpdf%2F1003.1466&usg=AFQjCNEvGZqL7uwqPNvSjt1tnfHRQfQo7A&sig2=q6uJgxCsfALbXe0LiLn-sQ&bvm=bv.146496531,d.d2s
Hi Iago,thanks for your answer. Yeap it was a typo. The correct search domain I'm using is [-5.12, 5.12]. Do you say that you have obtained similar results but with other parameters different than in paper? Is possible that the author didn't explain what parameter setting he used?
Is confussing to me the fact that the algorithm performs that well for different test functions (even different dimensions) using the same parameters. Is there any paper or documentation that I can follow to validate my implementation or to do some benchmarking?
It may also be important for you to know that the Firefly Algorithm is a meta-heuristics that that thrives on randomness. As a result, it is not in all cases that you can reproduce the same results even if you use the same parameters
This is generally disconcerting though. One issue of great concern in this field is reproducibility. I would think that you should be able to generate statistically similar results, but you will have to apply the appropriate tests, etc. to figure that out.
Results may vary based on the author's implementation as well - which is why you should always share the code publicly for any published research paper!
As far as the algorithm, you can see the link below.
One of the main problems in such algorithms is that the authors claim to be more efficient or converging faster, but it is not clear what all the parameter were. And I'm not talking about the parameters of the algorithm, but also the specifications of the computer used for the runs, the stopping criteria used etc.
Until today there is no established method to test the performance of these algorithms, but only to test them in real world problems, where the solutions can be visualized (the values of decision variables and the cost they give).
https://dk.um.si/IzpisGradiva.php?id=50116&lang=eng
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Algorithms