Identification of Systems With Similar Chains of Components for Simulation Reuse

Simulation is an essential tool to evaluate a complex system’s behavior. Simulation reuse can potentially improve simulation quality, cost, and delivery. However, identifying reusable simulations is a difficult task, of ten manual and based on limited information. This paper presents a method to facilitate the reuse of specific parts of past simulations. The system to simulate is compared to systems which have already been simulated. The comparison, which permits to identify similar chains of components in the systems’ block diagrams, is formalized as inexact graph matching. The comparison takes into account standardized tags as well as block properties defined by a name, a value, and a unit. Similar systems are identified with a limited computational cost. When two systems are similar, a mapping between their components and interactions can be obtained at a higher computational cost. A software prototype is implemented to perform the necessary computations, visualize the results, and accordingly select the simulation parts to reuse. The software prototype is tested with the block diagram of an autonomous electric car. Similar chains of components are successfully identified in the powertrain of a non-autonomous electric car represented in a past simulation. The corresponding part of the past simulation can then be selected for reuse.