System and Architecture Evaluation Framework Using Cross-domain Dynamic Complexity Measures

This work proposes a framework for effective quantification and comparison of system complexity content when architecting systems. Properly interpreted results from complexity analysis help make better informed system architecture selection between competing designs since the increased complexity of a system can lead to increased fragility and more exposure to failures and risks. Therefore the quantification of complexity is important when designing and planning the operation of a complex system. Our prior work on dynamic complexity measures provides the foundation for the framework discussed herein. The scope of this paper is to apply dynamic complexity measures to current, real-world complex systems. This work introduces a multi-step framework to evaluate complex systems and enhance a systems engineer's ability to compare competing systems/architectures. The framework has also proved useful for generating technical risks. A case study is included which utilizes the framework to evaluate an autonomous car architecture being developed by Google. The results demonstrate how the framework can help guide stakeholder decisions. The findings advance system complexity evaluation state-of-the-art by providing a framework using behavioral-based dynamic complexity measures.