Exactly characterizable parameter settings in a crossoverless evolutionary algorithm
Levi Koppenhol, Nielis Brouwer, Danny Dijkzeul, Iris Pijning, Joeri Sleegers, Daan van den Berg · Proceedings of the Genetic and Evolutionary Computation Conference Companion · 2022
The number of offspring and the population size greatly influence the performance of the Plant Propagation Algorithm, a crossover-less metaheuristic, when applied to the Euclidean traveling salesman problem. However, with large volumes of experimental data, a clearly characterizable relation emerges, facilitating both an exact parameterization as well as opening up perspectives on a precisely predictable outcome. These results directly question its position between exact, approximation, and PTAS algorithms.