||Configurational studies involve the groups of elements fulfilling restrictions defined by the designer in Architectural design. According to its necessities and intentions, the planner distributes the components of the group in a certain tridimensional way, establishing a composition. This operative procedure implies a classification system according to typologies that respond to a bigger system, and this in turn to another, until the whole is obtained. From the beginning the pattern should satisfy form restrictions, as well as dimensional and positional restrictions for each part that conforms the whole. Functional requirements are attended for each object satisfying relationships of connectivity and adjacency among them. In this work the parts are restricted by their relative position to a central element. Evolutionary Algorithms (EA) are used to solve this type of problem. Using evolutionary metaphors they originate concepts such as "genes", "chromosomes", "mutation", "crosses” and " population " (among other), which come closer to one of the solutions looked for by the designer, under combinatory stochastic methods. The most appropriate use of EA corresponds to problems of complexity NP-completeness (for example, problems of generation of cases of composition), allowing an efficient although not exhaustive analysis. Applying this technique to the generation of architectural volumes, some obtained results are exemplified.