Method for enhancing inquiring performance of data warehouse

Abstract

The invention relates to a method for enhancing the inquiring performance of a data warehouse. The method comprises the following steps of: firstly, dividing data in the data warehouse into a fact table and a dimension table according to a logic relation stored in the data warehouse, grouping and gathering the data in the fact table according to the data of the dimension table, generating a view from each group, and adding all the generated views into a candidate view set; and simulating foraging processes of ant colonies belonging to different populations in the natural world by using dual population ant colony algorithm based on pheromone diffusion, searching internal association among the data in the candidate view set according to an inquiring frequency, selecting a generated materialized view, pre-calculating and storing results of the inquiring operation to be subjected to inter-table connection or gathering in a finite storage space so as to enhance the inquiring performance ofthe data warehouse.

Description

technical field [0001] The invention relates to a method for improving the query performance of a data warehouse, in particular to a materialized view selection method using a dual-species ant colony algorithm based on pheromone diffusion. This technology belongs to the field of intelligent information processing and can be applied to knowledge engineering, especially in the fields of information sharing and information integration. Background technique [0002] With the wide application of information technology, information systems have produced a large amount of data, how to extract useful information for decision-making analysis from these massive data has become an important problem faced by decision-making managers. The traditional database system (management information system), that is, On-Line Transaction Processing (OLTP for short), as a data management means, is mainly used for transaction processing, but its support for analysis processing has been unsatisfactory...

AI Technical Summary

Problems solved by technology

Among them, the greedy algorithm for static view selection with a given query frequency has many shortcomings such as too large search space, high time complexity, and does not consider the probability and distribution of queries, and when the source data changes, this change cannot be immediately reflected in the data warehouse. , not suitable for online operation; genetic algorithm (Genetic Algorithm, referred to as GA) has strong global search ability, and shows better solution ability than greedy algorithm and heuristic algorithm, but genetic algorithm has slow convergence speed, and sometimes "Premature" phenomenon

Ant colony algorithm (Ant Colony Algorithm, referred to as ACA) has strong local search ability and positive feedback mechanism applied to the selection of the optimal materialization set, but a single ant colony algorithm has low global search ability and is easy to fall into local optimum

Therefore, the existing materialized view selection methods and tools have certain limitations, and are not suitable for the actual situation of relational databases and online analysis.

Images