Abstract:

The paper explores a problem of MV selection in the technology of automatic MV creation. An algorithm of query grouping on the basis of comparison of abstract syntax trees was proposed, which makes it possible to reduce the number of created MVs and reduce the total amount of physical resources required for its servicing. To solve the problem of MV selection out of the set of similar queries, a genetic algorithm was applied, which made it possible to distinguish the groups, for which a query execution efficiency increase by using MVs would be maximum while the maintenance cost would remain minimum. The objective function was proposed, which takes into account the ratio of the query execution efficiency increase by using created MVs to their maintenance cost. It helps to define which groups require MV creation and which of them should be created as virtual, as well as helps to define the queries within one group, which will form the next central query, on the basis of which the final MV can be created. Experimental data demonstrated that by using the proposed algorithm it is possible to obtain such a set of MVs, at which the maximum query execution efficiency at the lowest physical resources consumption for the servicing of these MVs is achieved. Author Biographies Kateryna Novokhatska, Odessa National Polytechnic University Shevchenko ave., 1, Odessa, Ukraine, 65044 Postgraduate student Department of System Software

Keywords: