Cognition Evolutionary Computation for System-of-systems Architecture Development

Abstract The evolving nature of system-of-systems requirements and corresponding architecture, and the complex causal relationship between the architecture of system-of-systems and its emergent behavior make the problem of system-of-systems architecture development a great challenge. As a tentative effort in meeting this special challenge, a new evolutionary computation paradigm–named Cognition Evolutionary Computation(CEC) is proposed, which models the creative cognition process of divergent and convergent thinking, adaptation and innovation, that drives the co-evolution of problem space, knowledge space and solution space. The optimization algorithm for CEC uses causal probabilistic network as the knowledge representation mechanism. A theoretical framework for CEC based system-of-system architecture generation, evaluation and optimization is discussed.