SUMMARY

Designing and developing complex engineering systems is a collaborative effort. In Model-Based Systems Engineering (MBSE), this collaboration is supported through the use of formal, computer-interpretable models, allowing stakeholders to address their particular concerns of interest using well-defined modeling languages. However, because concerns cannot be separated completely, implicit relationships and dependencies among the various models describing a system are unavoidable. Given that models are typically co-evolved and only weakly integrated, inconsistencies in the agglomeration of the information and knowledge encoded in the various models are frequently observed. The challenge is to identify such inconsistencies in an automated fashion. In this research, a probabilistic approach to abductive reasoning about the existence of specific types of inconsistencies and, in the process, semantic overlaps (relationships and dependencies) in sets of heterogeneous models is presented. The basis for the approach is Bayesian probability theory. A prior belief about the manifestation of a particular type of inconsistency within a specific context is updated with evidence, which is collected by extracting specific features from the models by means of pattern matching. Pattern matching across heterogeneous models is enabled by translating the information and knowledge encoded in models to a common, graph-based representational formalism. Results of the inference procedure are then utilized to improve future predictions by means of automated learning. The primary focus of the investigation is the development of a mathematically sound framework as a basis for a formal computational method. The effectiveness and efficiency of the approach is evaluated through a theoretical complexity analysis of the underlying algorithms, and through application to a case study. A prototypical, semantic web inspired implementation of supporting software tools was developed as a basis for performing the necessary accompanying measurements. As a case study, randomly generated sets of disparate, heterogeneous models of railway systems are considered. These generated sets of models are algorithmically injected with inconsistencies, and with features representing the result of human error and incompleteness.