Randomized Optimization of Object Oriented Queries in a Main Memory Database Management System

This thesis investigates the behavior of different algorithms when optimizing Object SQL-queries in an object oriented database management system called WS- IRIS (WorkStation-IRIS). Query optimization is a combinatorial optimization problem which makes exhaustive search impossible when the query size grows. This has led to a new approach to query optimization, namely the use of rand- omized algorithms. In this thesis a number of randomized algorithms are applied to optimize queries in WS-IRIS. A comparison of the algorithms has been done and the combination of the Iterative Improvement algorithm and the Sequence Heuris- tics algorithm showed the best performance. This report also shows the importance of good cost estimates for query optimization and some improvements to the WS- IRIS cost model are presented.