Artificial immune systems are metaheuristic algorithms that mimic the adaptive capabilities of the immune system of vertebrates. Since the 1990s, they have become one of the main branches of computer intelligence. However, there are still many competitive processes in the biological phenomena that can bring new advances for many applications. The Germinal Center reaction is one of these competitive processes that had not been fully modeled until now, and that was the inspiration to design the

more »

... vel optimization algorithm that we present in this work. Our proposal implements a competitive-based nonuniform distribution to select particles to be mutated, which can be interpreted as an implementation of temporal leadership in population-based metaheuristics. We model the dark-zone and light-zone of the Germinal Center and their competitive processes like clonal expansion, T-cell binding and life signal decay. We also propose the combination of this selection method with the use of one Differential Evolution-based strategy to substitute the somatic hypermutation process. To show the performance, we include a benchmark with the comparison of our approach versus some of the state-of-the-art bio-inspired optimization algorithms. We show that the proposal has a statistically significant improvement over the other algorithms for low dimensionality problems.