Management science/operations research (MS/OR) and decision support system(DSS) modeling is notoriously error-prone due to the deficiency of semantic formalization of commercial modeling languages and systems. The absence of precise, formal semantic definitions has contributed to the lack of successful modeling environment paradigms in various ways. For example, it has prevented the automatic generation of major components of modeling environments. It has increased the effort and cost of software development and maintenance. Modeling environments have been proposed to support the entire modeling life-cycle with multiple modeling paradigms and solvers. Despite its potential for improving model-based works, developing such an environment is a complicated task requiring a large amount of time and effort. This partially explains why the field of modeling environments is still in its infancy, and practical ones are yet to come. As a result of the complexity, existing environments have suffered from the limitations of flexibility, extensibility, and integration capability. As a way of improving the three capabilities of modeling environments, an approach for generating modeling environments through semantic formalization is proposed in this dissertation. The creation of a new generation of semantics-driven environments where mathematical modeling could be developed in a less errorprone fashion, and where environment tools could be automatically evolved, is one of the most pressing challenges currently faced by the MS/OR and DSS community. The significance of this approach is that it is possible to reduce the amount of time required to develop correct new models and to maintain existing ones. Furthermore, it is possible to automate the generation of the semantic tools and aids for the modeling environments from the formal specification. The benefits of this automation come not only from the time saved during the initial development of modeling environment tools, but f...