Dynamically predicting simulation fidelity in a self-optimized simulation of a complex system

Abstract

Mechanisms are provided for controlling a fidelity of a simulation of a system. A model of the system is received, where the model has a plurality of individual components of the system. Fidelity values are assigned to models of the individual components of the system and a simulation of the system is executed using the model of the system and the models of the individual components of the system. The simulation comprises a plurality of transactions targeting a receiver component. A history of mismatches between a fidelity value associated with the receiver component and required fidelity values of the plurality of transactions targeting the receiver component is maintained. A prediction of a fidelity value to be assigned to the receiver component based on the history of mismatches is performed. A fidelity value of the receiver component is adjusted based on results of predicting the fidelity value to be assigned to the receiver component.