A novel approach of selection sort algorithm with parallel computing and dynamic programing concepts

Abstract EN

Many research works have been conducted to find out better enhancement for Selection Sort Algorithm, such as bidirectional selection sort "Friend Sort Algorithm" which can position two elements in each round.

We have improved this algorithm by using the concept of parallel computing.

This algorithm is called Min-Max Bidirectional Parallel Selection Sort (MMBPSS).

Also this paper proposes to use dynamic programming (stack) to reduce sorting time by increasing the amount of space.

The basic idea behind using stack is to eliminate unnecessary iteration.

This algorithm is called Dynamic Selection Sort "DSS".

To fuse advantages of "DSS" with advantages of "MMBPSS", we suggest a new third algorithm called Min-Max Bidirectional Parallel Dynamic Selection Sort "MMBPDSS".

It can position two elements: minimum and maximum from two directions using Dynamic Selection Sort algorithm in each round in parallel, thus reducing the number of loop required for sorting.

Results obtained after implementation are provided in graphical form with an objective to show that "MMBPDSS" is saving almost 50% of classical selection sorting time and ensure accuracy.