This paper presents a hierarchical architecture for an innovative self-Adaptive design network for physics-based modeling of complex systems dynamics in order to assess and control its behavior characteristics. Each structural component of the system is simulated by an engineering design agent (EDA) which models the conceptual design parameters of the component in its operational environment. The objective here is to iteratively achieve higher assurance of dynamic system behavior by trading off component design characteristics. A formal model of the design network is formulated in this paper as a finite set of interacting automata. Intelligent agents for design coordination and design supervision are introduced. The introduction of a high order polyadic process algebra-calculus (cost calculus)-Allows the formulation of powerful algorithms for autonomous self-Adaptation of the system design network to achieve high assurance specifications in dynamic and uncertain environments.