Energy, matter, and entropy in evolutionary computation

This paper explores general concepts among the paradigms of computational intelligence, especially in evolutionary computation. From the viewpoint of adaptive complex systems, we generalize the characteristics of intelligence as complexity, evolution, sympathy, and adaptation, we also elaborate their corresponding meanings and approaches in artificial intelligence. Furthermore, we introduce the concept of dissipative system from thermodynamics to form three concepts for evolutionary computation: energy fluxion, matter information, and entropy variation. We briefly discuss the adaptive complexity emerging in the frustrated situation produced by the two converse forces approaching order and chaos respectively. Finally, we propose the concept of dissipative machine for natural evolutionary computation