jMetal / jMetalCpp

A C++ version of jMetal, a Java framework aimed at multi-objective optimization with metaheuristics.
64 stars 34 forks source link

What are these "problems" to be selected and how to.... #23

Open wellinton-capixaba opened 3 years ago

wellinton-capixaba commented 3 years ago

Hello. I wish to try the Moth Flame algorithm, however I have no clue of what are these "problems" that I have to select. I thought the only thing that could be called a "problem" is the fitness function.

EDITED:

So I need someone explain what these problems means and which fits my needs. What I want to do: Select or implement and then select a fitness function (I want to try all or most from the original paper). Be able to select the parameters (n° iterations, initial limit values for the moths, dimension and n° moths). I think I already located n° iterations and n° moths (I assume it is "swarmSize") on "MFO_main.cpp". I think I also found the initial limit values (-5.12 on "Sphere.cpp"). I still need to know where to find the dim, and more importantly, where the fitness functions are. Then I will try to implement a modified version of the algorithm, called ameliorated MFO algorithm. So I need to know WHERE to edit the original files to add extra fitness functions, change parameters and modify the algorithm to support this new version. For doing all of that, I also need to know how to use it, and hence, what these problems are.

Thanks for the support.

wellinton-capixaba commented 3 years ago

Read the code more. It's very complex, no idea how to implement the necessary fitness functions, due to the excess of auxiliary functions, files and objects, nearly all defined outside the MFO.cpp file, nearly total lack of documentation. All that contribute to make the code useless for me purposes.

yangyongkang2000 commented 7 months ago

I recommend an open-source C++ benchmarking framework for single-objective evolutionary optimization algorithms. You can find it at https://github.com/yangyongkang2000/SEvoBench.