We have already discussed in this book how to apply either the objective-based search optimization or Novelty Search optimization methods to the problem of solving a maze. In this chapter, we introduce a modified maze-solving experiment in which we try to combine both search optimization methods using the SAFE algorithm.
We introduce the co-evolution of two populations: a population of maze-solving agents and a population of objective function candidates. Following the SAFE method, we use a commensalistic co-evolution strategy in our experiment. Let's first discuss the maze-solving agent.
