Parallel and distributed approaches are natural in evolutionary computation and they have been used intensively since the early years of this research field. Evolutionary algorithms, in fact, have often been described as intrinsically parallel computational methods. The reason for this is that many of the main computational tasks characterizing this family of heuristics are independent of each other; thus it is straightforward to perform them at the same time. This is the case, for instance,

more »

... the evaluation of the fitness of the individuals in a population. Furthermore, by attributing a non-panmictic structure to the population, something that also finds its inspiration in nature, genetic operations can be performed independently of each other and thus can also potentially be parallelized. These approaches can be useful even when there is no actual parallel or distributed implementation, thanks to the particular information diffusion given by the more local population structures. But of course parallel and distributed approaches are at their best when the structures of the models are reflected in the actual algorithm implementations. In fact, when compared with other heuristics, evolutionary algorithms are relatively costly and slow. But parallel and distributed implementations can boost performance and thereby allow practitioners to solve, exactly or approximately, larger and more interesting problem instances thanks to the time savings afforded. As mentioned above, these advantages have been known and appreciated for at least three decades; so, what justifies a new presentation of the state of the art in the field? The answer is that significant progress has been made in recent years and these methodologies have become even more important today for at least two reasons. First, M. Tomassini (&)