The Multi-Verse Hyper-Heuristic Optimizer (MVHO) generates a rule-based selection hyper-heuristic, composed of a set of rules that indicate the best heuristic to use at each step when solving a knapsack problem (KP). The multi-verse is filled with a given amount of universes, where each one of the universes represents a plausible hyper-heuristic. Thus, each universe is a set of rules with its own feature conditions and a respective heuristic. The evaluation of a universe is given by solving a training set of KP instances with the MVHO, where a universe's fitness corresponds to the average of the the evaluations obtained for each problem instance in the training set. At each iteration each universe evolves by exchanging objects with other universes through black/white hole tunnels and worm holes with the aim of minimizing its evaluation.