Predictive Modeling for Infrastructure System Engineering

Civil infrastructure systems emphasize on co-existence of the different structures together as a system that is highly dynamic and with behavior that varies on wide temporal and spatial scales. Furthermore, urban environments can be viewed as a “system of systems” with complex interactions and interdependencies existing among its individual systems. The focus has been comparatively more on the resilient design of the individual infrastructure components rather than how infrastructure components interrelate and interact to make up a dynamic system. To model complex infrastructures systems with time-varying behaviors, system theory is a powerful tool that allows analytical and computational methods to fully describe the complex dynamics of independent and interdependent infrastructure systems. Cyber-physical systems are a new class of sensor-rich engineered systems that entail not only sensors but computing and actuation cohesively integrated for monitoring and controlling of the applications. The vast historical as well as real-time data obtained from such systems and its semantics result in a knowledge repository. Predictive analytics is one of the best practices of extracting information from such a repository in order to determine patterns and predict future outcomes and system behavior. Our earlier work proposed a decision support framework comprising a geologic prediction model, and a construction strategy decision model for tunnel design. We further extend and generalize the framework by including a predictive analytics model for effective engineering of infrastructure component(s) within a civil infrastructure system. The predictive analytics framework is sustainable due to its self-adaptive properties.