Semantic knowledge has often been employed to enforce relational database integrity. It also offers the opportunity to transform a query into a semantically equivalent query that is potentially more efficiently processed than the original query. This paper describes a knowledge-based transformation approach that utilizes semantic integrity constraints and transformation heuristics to reduce query processing costs. Implementation in the form of an intelligent system, QUOTA, is also presented.