Taxonomy of bio-inspired algorithms
share
In recent years, a great variety of nature and bio-inspired algorithms have been published in the literature. These algorithms simulate some natural or biological processes such as natural evolution, to optimize complex problems. Last years, the accumulation of this type of algorithms has achieved a number difficult to manage, so it is needed a good taxonomy that allows researchers to organise existing and future meta-heuristics in well-defined categories. In this work, we have reviewed more than two hundred nature-inspired and bio-inspired algorithms, and proposed two taxonomies that group them in categories ans subcategories, considering two different criteria. The first category consider the source of inspiration, while the second one consider only the behavior of each algorithm. It is shown the proposals in each one of the categories, and the summarise of each category. Later, results of these two taxonomies are compared, and obtained the most similar classic algorithm for each reviewed papers. We have obtained that there is a poor relationship between the natural inspiration of an algorithm and its behavior, and that, their similarities in behavior are greater than expected. Even more, more than half of reviewed proposals are versions of classic algorithms. Although there are many proposals based in different concepts of reality, the majority of them are actually more similar than expected, so authors should focus more on the behaviour. Also, we give several suggestions to continue improving this growing research field.
READ FULL TEXT
