I'm trying to implement a variation of the Iterative Structure-Based Algorithm (ISBA) concept to create a structure-based GP in GPLearn. Specifically, I'm trying to implement a penalty for excessive similarity to past populations in the fitness function, by calculating the relative distance between the current program's structure and saved past structures. However, to do this, I need to be able to access the individual whose fitness is being evaluated in my custom fitness function. Said custom fitness function is being implemented in accordance with the documentation. Is there any way to get a representation of the individual from GPLearn in a custom fitness function?
Thank you in advance for any help!
Update: upon further research, I found a GitHub issue mentioning a similar situation. Unfortunately, it seems that getting the formula expression of each individual is not supported as of the 24th of November 2022, making using GPLearn as a library impossible for this particular structure-based GP implementation.