PSO-Based General Fragments Allocation Method for Join Query in Distributed Database

Abstract

The distributed database is a logically interconnected set of databases that, from the
user's point of view, appears as a single database. The distributed database has many
benefits such as reduced communication costs, more processing power, scalability, and
reliance on a server device. With the advancement of computer network technology and the
increase in database size, decentralization of databases has led to the development of
distributed databases on multiple devices, in which database distribution is invisible to
users. The problem of fragments allocation for join query in distributed database is
inherently NP-Hard, and solving it is a difficult problem. Numerous research has been done
on finding a suitable algorithm for achieving the optimal solution, especially when in the
high dimensional database. The PSO algorithm, due to features such as positive feedback,
distributed calculations, and combination with other evolutionary algorithms, can increase
the accuracy of fragments allocation to join the query in the distributed database. In this
paper, an improved PSO algorithm with mutation operator is proposed for fragments
allocation in distributed databases. The experimental results showed the appropriate
performance of the proposed method both in terms of accuracy and in terms of
computational complexity