Use cases for genetic algorithms
Based on the material we covered in the previous sections, genetic algorithms are best suited for the following types of problems:
- Problems with complex mathematical representation: Since genetic algorithms only require the outcome of the fitness function, they can be used for problems with target functions that are hard or impossible to differentiate (such as planning and scheduling), problems with a large number of parameters (such as image reconstruction), and problems with a mix of parameter types (such as hyperparameter optimization).
- Problems with no mathematical representation: Genetic algorithms don’t require a mathematical representation of the problem, so long as a score value can be obtained, or a method is available to compare two solutions. This can be useful, for example, when solving reinforcement learning tasks or optimizing the architecture of a deep learning model.
- Problems involving a noisy environment: Genetic...
