Citation: Pan, R.; Yu, J.; Zhao, Y.
Many-Objective Optimization and
Decision-Making Method for
Selective Assembly of Complex
Mechanical Products Based on
Improved NSGA-III and VIKOR.
Processes 2022, 10, 34. https://
doi.org/10.3390/pr10010034
Academic Editor: Arkadiusz Gola
Received: 7 December 2021
Accepted: 21 December 2021
Published: 24 December 2021
Publisher’s Note: MDPI stays neutral
with regard to jurisdictional claims in
published maps and institutional affil-
iations.
Copyright: © 2021 by the authors.
Licensee MDPI, Basel, Switzerland.
This article is an open access article
distributed under the terms and
conditions of the Creative Commons
Attribution (CC BY) license (https://
creativecommons.org/licenses/by/
4.0/).
Article
Many-Objective Optimization and Decision-Making Method
for Selective Assembly of Complex Mechanical Products Based
on Improved NSGA-III and VIKOR
Rongshun Pan
1
, Jiahao Yu
1
and Yongman Zhao
1,2,
*
1
Department of Industrial Engineering, College of Mechanical and Electrical, Shihezi University,
Shihezi 832003, China; panrongshun@stu.shzu.edu.cn (R.P.); 20202009046@stu.shzu.edu.cn (J.Y.)
2
Department of Data Science and Big Data Technology, College of Information Science and Technology,
Shihezi University, Shihezi 832003, China
* Correspondence: zhrym@shzu.edu.cn
Abstract:
In Industry 4.0, data are sensed and merged to drive intelligent systems. This research
focuses on the optimization of selective assembly of complex mechanical products (CMPs) under
intelligent system environment conditions. For the batch assembly of CMPs, it is difficult to obtain the
best combinations of components from combinations for simultaneous optimization of success rate
and multiple assembly quality. Hence, the Taguchi quality loss function was used to quantitatively
evaluate each assembly quality and the assembly success rate is combined to establish a many-
objective optimization model. The crossover and mutation operators were improved to enhance the
ability of NSGA-III to obtain high-quality solution set and jump out of a local optimal solution, and
the Pareto optimal solution set was obtained accordingly. Finally, considering the production mode
of Human–Machine Intelligent System interaction, the optimal compromise solution is obtained by
using fuzzy theory, entropy theory and the VIKOR method. The results show that this work has
obvious advantages in improving the quality of batch selective assembly of CMPs and assembly
success rate and gives a sorting selection strategy for non-dominated selective assembly schemes
while taking into account the group benefit and individual regret.
Keywords:
selective assembly; Taguchi quality loss; many-objective optimization; NSGA-III; VIKOR
1. Introduction
In the manufacturing environment of Industry 4.0, it becomes easy to accurately
collect various data during production. Relying on a large amount of data, intelligent
systems can optimize and make decisions in the production process. Most mechanical
products are manufactured by the processing and assembly of components [
1
]. Under the
condition of high requirements for product matching accuracy, due to the limitation of
component processing capacity and manufacturing cost, it is unrealistic and uneconomical
to completely rely on improving the accuracy of the processing process to meet and
improve the product matching accuracy [
2
]. In practice, the demand of producers is to
reduce manufacturing costs as much as possible while ensuring quality to obtain maximum
product competitiveness. Therefore, selective assembly is one of the feasible methods to
achieve high-precision assembly and reduce costs by using data and intelligent systems.
Researchers have made a lot of progress in selective assembly by researching the
grouping of assembly components [
3
,
4
]. However, due to the many limitations of these
methods, researchers have conducted further research on selective assembly methods
from other perspectives. With the development of research, some intelligent optimization
algorithms with better optimization effects have been used in the selective assembly field.
Kannan et al. [
5
] proposed a genetic algorithm for selective assembly. In this research, the
selective assembly process was analyzed, and a better combination was selected through
the use of genetic algorithms to minimize assembly differences. Kannan et al. [
6
] used a
Processes 2022, 10, 34. https://doi.org/10.3390/pr10010034 https://www.mdpi.com/journal/processes
