Although the main difference between the relational and the object-oriented data model is the possibility to define object behavior, query optimization techniques in object-oriented database systems are mainly based on the structural part of objects. We claim that the optimization potential emerging from methods has been strongly underestimated so far. In this paper we concentrate on the question of how semantic knowledge about methods can be considered in query optimization. We rely on the algebraic and rule-based approach for query optimization and present a framework that allows to integrate schema-specific knowledge by tailoring the query optimizer according to the particular application's needs. We sketch an implementation of our concepts within the OODBMS VODAK using the Volcano optimizer generator.>