Abstract
Compact algorithms are optimization algorithms belonging to the class of Estimation of Distribution Algorithms (EDAs). Compact algorithms employ the search logic of population-based algorithms but do not store and process an entire population and all the individuals therein, but on the contrary make use of a probabilistic representation of the population in order to perform the optimization process. This probabilistic representation simulates the population behaviour as it extensively explores the decision space at the beginning of the optimization process and progressively focuses the search on the most promising genotypes and narrows the search radius. In this way, a much smaller amount of parameters must be stored in the memory. Thus, a run of these algorithms requires much more limited memory devices compared to their corresponding standard population-based algorithms. This class of algorithms is especially useful for those applications characterized by a limited hardware, e.g. mobile systems, industrial robots, etc. This chapter illustrates the history of compact optimization by giving a description of the main paradigms proposed in literature and a novel interpretation of the subject as well as a design procedure. An application to space robotics is given in order to show the applicability of compact algorithms.