Models are the primary development artifacts used in model-driven software development. Models continuously evolve during the design, development, and maintenance of software systems. Thus, model differencing is an important task to understand the syntactic and semantic differences between model versions. Previous work produced general approaches for syntactic model differencing, but only a few language-dependent approaches for semantic model differencing. Approaches combining syntactic with semantic model differencing by relating the syntactic changes of models to their semantic differences rarely exist. Previous work neglected the development of language-independent approaches abstracting from a concrete model property for detecting the syntactic elements of a model, which cause that the model does not satisfy the property. However, detecting the syntactic model elements causing the non-satisfaction of a property representing a bug facilitates developers in detecting the syntactic model elements causing the bug. This thesis presents a framework for precisely defining modeling languages, including syntax, semantics, and model evolution possibilities as well as syntactic and semantic differencing operators. The operators facilitate developers in understanding the syntactic and semantic differences between models of a language. The framework for precisely defining modeling languages is extended with a modeling language and property-independent framework for automatic model repairs. The framework facilitates developers in detecting the syntactic elements of a model causing that the model does not satisfy a property. To demonstrate its feasibility, the framework is instantiated with four concrete modeling languages.weiterlesen