Design and optimization of complex real-time dependable systems

Proposes the RINDO (Rapid Iterative Nested Design Optimization) algorithm for optimizing multi-criteria complex systems. We first use a utility function to formulate multi-criteria stochastic optimization problems. An ordinal optimization algorithm is then applied to the utility function for obtaining a set of good solutions. Finally, the gradient surface method is used to search for the optimal solution iteratively. An especially appealing characteristic of RINDO is that it can identify a set of good designs very rapidly. The key thrust of this approach is its integrated, nested (hierarchical), system-level design capabilities. The framework proposed in this paper is generic and application-independent.