Software module clustering using grid-based large-scale many-objective particle swarm optimization

Amarjeet Prajapati

doi:10.1007/s00500-022-07182-w

2022-05-29

PRE

Amarjeet Prajapati *

DOI：10.1007/s00500-022-07182-w

原文链接

原文求助

摘要

En 中文

There are huge numbers of real-world optimization problems, which often contain a large number of decision variables (n > 100) and objective functions (m > 3). Such optimization problems are generally regarded as large-scale many-objective optimization problems (LSMaOPs). Although a variety of search-based optimization algorithms have been proposed to solve various types of synthetic and real-world LSMaOPs, the problems of producing a well-distributed approximation of the Pareto front remain challenging. In this work, we propose a grid-based large-scale many-objective particle swarm optimization, namely GLMPSO, for solving the LSMaOPs, i.e., large-scale many-objective software module clustering problems (LMSMCPs). To balance the convergence and diversity of the GLMPSO, a grid-based ranking strategy and angle-based selection strategy are employed at different stages of the selection process. To balance the exploration and exploitation of the solution space, and avoid GLMPSO getting stuck in local minima, we use the center-based velocity computation. To test the effectiveness of the proposed GLMPSO, it is applied over nine LMSMCPs of software module clustering and the obtained results are compared with five existing approaches. The comparative results demonstrate that the proposed approach is more effective and has significant advantages over existing approaches.

Keyword:

Many-objective optimization

Grid-based criteria

Software clustering

Two-archive optimization