Dominance Query For Relational Databases With Conditional Preferences

User preferences can guide users' choices in many cases,and questions about preference queries become an increasingly important issue in relational databases.In many applications,qualitative preferences can be applied in a wider range than quantitative preferences.In the existing multi-attribute preference study,the preference attributes do not have dependencies,and CP-nets is a graph model that represents the multi-attribute qualitative preference with dependencies.At present,the processing of the preference query mainly uses the dominance query,and the two outcomes are sequentially compared by the user's preference,and the induced prefernce graph is generated,thereby obtaining the satisfiability sequences and completing the query satisfying the user preference.The induced prefernce graph of CP-nets requires a large number of outcome comparisons,and the classical generation of satisfiability sequences for CP-nets usually leads to exponential complexity.Therefore,this paper obtains the satisfiability sequences through the preference composition,and prunes the flipping sequences according to the pruning techniques,thereby improving the efficiency of the dominance query.This paper mainly conducts the following research:(1)Pareto preference composition: First,we extend the Pareto composition to our model by using the equivalence relation ?,the incomparable relationship ? and the conflict relationship ?,and maintain strict partial order.On this basis,two problems are solved:(a)the generation of satisfiability sequences for CP-nets,and(b)the dominance queries of a relational database with CP-nets preference.For question(a),the CP-net can be induced into multiple tables,so the strong dominance testing between outcomes can be solved by using preference composition instead of using induced preference graph of CPnets.For question(b),we use the concept of Query Lattice to provide natural semantics for answering the block sequence of the preference query,which introduces the algorithm QOCP.From the perspective of the combination of graph model and relational database,these problems have been solved effectively and efficiently.(2)Pruning techniques: The processing of binary CP-nets by dominance query requires a large number of flipping sequences,how to optimize the flipping sequence,and reduce the number of flipping sequences that need to be queried,thereby improving query efficiency.In the research,it is found that each improved flipping sequence corresponds to the directed path in the CP-nets induced graph.We introduce the improved search tree concept to improve the integrity and correctness of the search process.Search trees are pruned to effectively reduce the search space of the database.