@alguiguilo098 I'm a mathematician and (partly) a physicist. I'd love to make this repo into a package/library/program available for other researchers to use. However, I'm not at all familiar with how to structure code in terms of interfaces, testcases, etc. I know problems which are well-solved with genetic algorithms and yesterday finally the endurance of the material worked.

In the past I had this code working for the labyrinth as well. Howewver, it was not scalable, and so I started learning OOP. At this point I can see how to generalise and develop:

genome generators,
more functions with selection, crossover and even mutation operators,
various metrics for genetic diversification and variation between rival generations.
Other evolutionary programming techniques could be added in the future, because they are useful for optimising parameters of simulations of physical systems on classical computers, like Quantum Key Distribution protocols and some experiments in spectroscopy, in combination with evolutionary reinforced learning.

Once labyrinth works well again, I'll want to work again on BB84 simulators, but it would be great to keep developing the genetic algorithm part independently, for general use too. This is where your help could be critical, as in past weeks.

One thing I was considering is providing users with 2 possibilities: to either import pyqkd.genal and working with it in their own code or just providing code for their simulation/problem, e.g. in terms of genome generator and fitness function, and run them with the genetic algorithm in a GUI (high-level approach, I suppose).

Could you elaborate on splitting the Genetic Algorithm into more sub-classes? How could the interface and testcase look like?

PGA-7: Refactoring class GeneticAlgorithm #35

Description

Metadata

Metadata

Assignees

Labels

Projects

Milestone

Relationships

Development

Issue actions