A Biological Basis for Architecting Intelligent and Autonomous Software

Since Alan Turing in the 50´s, there have been many biologically-inspired approaches to intelligence and autonomy for machines and software. Commonly, these approaches explored one particular level of organization of living species (e.g. cell, tissue, organ or organ systems) or one particular set of behaviors. Many have arrived at obtaining amazing results. They however experienced conceptual and implementation limitations. One major issue is that while the human-like higher level capabilities such as intelligence are desired for machine and software, there has been no similar foundation, structural and/or functional, that could be established between the complete range of human levels of organizations and those of a human-made system (machine, software or both). In our previous investigations, we have argued that there is a general parallelism between the concept of software continuum (from bits to business ecosystems) and that of the natural continuum (from particles to natural ecosystems). We claimed that the general parallelism suggests possible mappings between the two continua. The mappings in turn suggest that the processes or mechanisms that work in one continuum can be investigated for applicability to the other continuum. This paper introduces an architecture framework for the creation and development of biologically-inspired intelligent and autonomous software species. The development is bottom-up, i.e. in a fashion similar to the creation, growth and development of a human newborn to acquire intelligent and autonomous behaviors. The development of such as software species is needed for future applications with robust intelligence and robust autonomy in different domains such as business, medical, industrial, governmental and military.