Computational Evolutionary Art: Artificial Life and Effective Complexity

Abstract

On the field of Evolutionary Computational Art, artists frequently adopt a top-down process of creation, employing the algorithms only as a mean to express a previously conceived composition. In this sense, the present paper aims to discuss the use of Genetic Algorithms for the development of systems with greater level of emergence, running towards the increase of its effective complexity, understood as suggested by Gell-Mann. In this context it is presented the system Morphogenesis. It was developed as a Multi-Agent Adaptive System, built with Genetic Algorithms to generate movement, feeding, fighting and reproductive behaviors. All these behaviors are programed at the individual level, from which emerge the macro patterns of the groups, simulating the evolutionary process. The system analysis suggests that the fitness function should not be focused at the arrangements of the agents’ genotype, but at the adaptation of the phenotype itself. It is expected that the use of algorithms that allow expressions closer to the evolutionary process has a greater affinity with the aesthetic notion proposed for the field of Evolutionary Computational Art. Hereupon, a qualitative exploratory study was conducted to compare the perception of the high effective complexity arrangements against random arrangements. Preliminary results show that the evolutionary process could be associated with a greater evaluation of intentionality of the compositions and could be also related with a deeper aesthetic evaluation.