Application of 3D printing and multi-objective genetic algorithm in product design
optimization of auto parts
Jun Zhao*
Zhejiang Industry Polytechnic College, Shaoxing, Zhejiang, 312000, China
*Corresponding author:Jun Zhao
Jun Zhao’s mail: jianpinghej@[Link]
Abstract: In order to solve the problem of poor product design structure in the automobile
industry due to the low matching degree of product genes retrieved by traditional methods, the
author proposes an application of a multi-objective genetic algorithm in the design optimization of
auto parts products. The application first divides the structure of automobile products, and then
proposes the concept of product function to analyze the product design process model of the
automobile industry. After completing the above work, the custom template model is used to
obtain user requirements. In order to obtain the optimal solution in the retrieval, the genetic
recombination process of the genetic algorithm is used to realize the product design of the
automobile industry. In the experiment, a sample was randomly selected, and the design
parameters were obtained according to the triangular matrix, which was used to test the matching
degree of the two methods to retrieve product genes. Experimental results show that: Among the
10 samples of the author's method, the highest matching degree is 0.7%. Conclusion: The multi-
objective genetic algorithm retrieval product has a higher gene matching degree, which meets the
design requirements.
Key words: Automobile industry products; genetic optimization algorithm; design process model;
template model; automobile components; automobile parts; matching degree test
1 Introduction
At present, people have higher and higher requirements for vehicle quality, and the research on
vehicle reliability theory and method also needs to be further discussed [1]. Reliability of auto parts
refers to the ability of auto parts to complete the specified functions within the expected time of
use and the working conditions required by the parts [2]. Although the car is in the generation
process, appropriate parts are designed in order to achieve the specific expected function of the
car, but due to various factors such as theory, materials, and technology, the parts will have
different reliability problems [3]. Therefore, it is necessary to summarize and design a good method
of reliability theory, research the reliability of auto parts.
Among vehicle parts, the reliability of parts is an important factor affecting the reliability of
automobiles [4]. In the actual use of vehicles, the wear rate of parts is related to the driving
conditions of the vehicle, the ambient temperature of the parts, lubrication characteristics, loads
and other factors [5]. Under long-term use, the wear of automobiles gradually increases with the
increase of time, and the size of the worn automobile parts gradually decreases, which will lead to
insufficient conformity with other parts and loss of normal functions [6]. Based on the above
analysis, it can be seen that the damage of automobile parts has a certain degree of gradualness,
after the parts fail, they cannot meet the original requirements. If they are still used, there will be
certain dangers, therefore, the reliability of auto parts is a function of time and exhibits a certain
gradient characteristic [7].
As an effective global search method, genetic algorithm has penetrated into many disciplines,
such as function optimization, combinatorial optimization, production scheduling, automatic
control, machine learning, image processing, artificial life, genetic programming, machine
learning, data mining, etc., at the same time, the application practice promotes the development
and perfection of the genetic algorithm. The rated task that an auto part completes under the
specified time and corresponding conditions is called the reliability of the auto part, whereas the
probability that the part cannot complete the rated task under the specified time and condition is
the unreliability [8]. Before analyzing the gradual reliability evolution of auto parts, the reliability
evaluation index is established, auto parts are an important part of determining the safety of the
car, and its reliability determines the safety of the car, through the reliability index and the fault
type distribution function It plays an important role in calculating the reliability of auto parts [9].
On the basis of analyzing the insufficiency of the traditional auto parts gradual reliability
evolution analysis method, the wear factor, reliability index and genetic algorithm are applied to
the gradual reliability evolution analysis of auto parts [10].
2 Literature review
Genetic algorithms simulate the phenomena of reproduction, crossover and genetic mutation that
occur in natural selection and natural genetic processes, in each iteration, a set of candidate
solutions is kept, and the better individuals are selected from the solution group according to a
certain index, these individuals are combined using genetic operators (selection, crossover and
mutation) to generate a new generation of candidate solution groups, and this process is repeated
until a certain convergence index is met. Compared with the traditional heuristic optimization
search algorithm, the main essential characteristics of the genetic algorithm lie in the group search
strategy and the simple genetic operator. Group search enables genetic algorithm to break through
the limitation of domain search, and can realize distributed information collection and exploration
in the entire solution space; The genetic operator only uses the fitness value as the operation index
to perform random operations, which reduces the dependence of the general heuristic algorithm on
human-computer interaction in the search process.
As a non-deterministic quasi-natural algorithm, the genetic algorithm provides a new method for
the optimization of complex systems, it has a wide range of application value in many disciplines
[11]
. But research on algorithms, more attention is paid to the hotspots that have been researched
more deeply, but less attention is paid to potential research hotspots and rapidly developing
research hotspots [12]. Statistics are carried out from two aspects of research content and
application fields, and the results are shown in Figure 1.
Figure 1 Comparative analysis of genetic algorithm research
From Figure 1, the following conclusions can be drawn: From the perspective of research content,
articles involving species diversity, test functions, genetic operators, parameter determination and
other research content occupy a large number; From the perspective of application fields, the
research articles on genetic algorithm in production scheduling and robotics account for the
majority, and the research on automatic control, combinatorial optimization and image processing
also accounts for a large proportion [13]. There are not many articles about genetic algorithms in
function optimization, machine learning, artificial life, data mining and genetic programming [14].
Based on the above comparative analysis, it can be seen that the research of genetic algorithm in
function optimization and combinatorial optimization is decreasing, especially in function
optimization, but the research proportion in the fields of production scheduling and automatic
control has increased significantly, this fully shows that the research of genetic algorithm has
gradually turned from the theoretical aspect to the application field; Robotics and image
processing are also gradually becoming research hotspots [15]. There are not many articles related
to data mining research, but with the wide application of data mining technology, the research of
genetic algorithm in the field of data mining will become a new hot spot [16].
Automotive product design is a complex process, and companies usually use a design process
model to describe the design process [17]. Design process model is a way of product innovation,
many scholars abroad use the concept of functional analysis and biological genetic engineering to
design process models to make the design process more reasonable. Through the analysis of
domestic and foreign automotive industry product design models, the current well-known genetic
optimization algorithm is studied, and the application of genetic optimization algorithm in
automotive industry product design is proposed [18].
In this research, with the discovery of problem solving theory as the leading factor, a product
design method for the automotive industry based on genetic optimization algorithm is constructed
[19]
. During the design process, the structural stage and the parameter stage were designed in detail,
and a new design process was proposed. Realization process: First, apply genetic optimization
algorithm to analyze the evolution mode and evolution route, and obtain the potential needs of
users; Then establish the relationship between user needs and product characteristics, and use the
concept of biological genetic engineering to complete the product design of the automotive
industry; Finally, the feasibility of this method is investigated experimentally [20].
3 Methods
3.1 Vehicle structure design based on genetic optimization algorithm
Using the concept of genetic optimization algorithm to divide the vehicle structure, the definition
is as follows;
The part face set composed of n faces is , when the part Pa is in the face set,
FPa={ f 1 , f 2 ,⋯ , f n }
there will be several functional faces, the functional face is defined as: When the part Pa is in the
face set , the following conditions need to be met:
FPa={ f 1 , f 2 ,⋯ , f n }
1) The contact geometry type can be surface contact and point contact, or plane and curved
surface;
2) There is force transmission between a pair of contact surfaces.
Satisfying the above conditions becomes the functional surface, and Figure 2 shows the design
structure of the automobile.
Figure 2 Automotive design structure
It can be seen from Figure 2 that the structural design of the automobile is formed by multiple
independent functional surfaces. At the product level, the design structure can be a component or a
part. The functional surface is not a solid structure, but needs to have certain functions, so the
functional surface is attributed to the design structure. In the demand analysis stage, all the
information is based on the car virtual product as the description object, in the conceptual design
stage, the description objects include products and institutions.
In the detailed design stage, the information description object should be changed according to
the specific design process. In the design stage of the mechanism, the virtual information of the
automobile describes the mechanism and components of the automobile. In the component design
stage, the core components and parts are mainly described. The above information is unified, and
the design process of automobile structure is obtained.
3.2 Research on car function design based on genetic algorithm
After designing the car hierarchy, in order to obtain the key identification elements of the car
engineering design, follow the steps below: First, in order to effectively obtain customer needs and
accurately define product demand information, a customized template is proposed. The custom
template model is shown in Figure 3.
Figure 3 Custom Template Model
The customer set in Figure 3 is a set of overall demand for a certain type of product obtained
through external and internal data sources of the enterprise. External data sources mainly include
users, books, technical manuals, competitors, networks, movies, etc. Internal data sources mainly
include designers, project managers, and product life cycles. After obtaining the constructive
demand for improvement space, it is necessary to analyze the function, structure, design and
manufacturing of the existing products, inversely find out that the modules with specific functions
can meet the different needs of customers, and according to the obtained results, establish
corresponding space is required.
The customized module realizes the customized expression of product function and structural
parameters, which is easier to understand and describe the product function. After completing the
interactive selection of custom templates, in order to facilitate product customization, configure
the required products. In order to encode production demand genes according to pre-set product
parameters. The demand product is a kind of information descriptive product, which can describe
the customer's demand for the product through the information characteristic value. By analyzing
the relationship between various parameters of customized template products, product function
information tree, structure information tree, technical information tree and management
information tree are generated according to customer requirements. On this basis, the concept of
demand product gene is put forward, and the demand product gene is generated by using the
coding rules of product gene. Generally can be divided into four categories, namely functional
genes, structural genes, management genes and technical genes. The difference is that the demand
gene structure is incomplete. The obtained product gene is the design input of the automobile
industry, and the product gene is retrieved based on the demanded product gene. The specific
retrieval process is as follows: In order to obtain the optimal solution in the retrieval, the genetic
recombination process of the genetic algorithm is used to complete the product design of the
automobile industry. During the execution of the genetic algorithm, the number will gradually
increase exponentially.
In order to obtain the lower limit of the number of individuals produced in the offspring under the
action of transformation, according to the formula of the mode theorem:
m(H , t+1)⩾ m(H , t) (1)
In the formula: t represents the group; H represents the mode; m represents the number of digital
strings.
Functional genes are the core of product design and the index of retrieval. According to the
matching degree function, the functional genes that meet the needs are matched, and the genetic
optimization algorithm is used to complete the design of the product functional gene retrieval
process. The product design process of the automobile industry based on genetic optimization
algorithm is shown in Figure 4.
Figure 4 Design process of automotive industry products based on genetic optimization
algorithm
1) There are three possibilities in the gene bank, one is the most ideal result, which completely
matches the gene at the product level, and the search ends. The other is that there is no perfect
matching product layer gene, but this result is less likely. If the above phenomenon occurs, go to
step 2). The third type is not matched to the product-level gene. The possibility of this happening
is small. If this phenomenon occurs, go to step 2).
2) In the separation-required genes, if no gene fragments are matched, the candidate product
genes in the genes of the retrieval module layer need to be genetically recombined. If the required
gene fragment is not found, you need to go to step 3).
3) Decomposition of module layer genes that were not satisfied in step 2). Based on this, the part-
layer genes are searched, and the retrieved part-layer genes are used for product gene
recombination.
4) After the search is completed, if a satisfactory product gene is not obtained, it is necessary to
innovate the design according to the needs of users. It should be noted that in the retrieval process,
the exact match and the incomplete match are relative, and a threshold needs to be set. If it is
higher than the set threshold, it is a completely matched gene; If it is lower than the set threshold,
there is no matching product layer gene. Each layer of product genes can only be retrieved once,
which can effectively avoid an infinite loop.
3.3 Experimental Analysis
Excellent construction machinery companies have strong research and development efforts in
industrial product design, in order to ensure a certain forward-looking, numerical analysis of each
data is carried out during the experiment. In order to prevent users from interfering with the
evaluation value, the traditional method and the design method based on the genetic optimization
algorithm established this time are used to test the gene matching degree of the product.
(1) Experimental samples
In the database, 10 samples were randomly selected, in the experiment, considering the
relationship between the experimental samples, the experimental order should be determined first,
and the product genes with independent functions should be placed in the front, Table 1 shows the
design parameters of the product genes.
Table 1 Product genetic design parameters
serial
DP11 DP12 DP13 DP14 DP15
number
1 0 0 1 0 0
2 1 1 0 0 0
3 0 0 0 0 0
4 0 0 1 0 0
5 0 0 1 1 0
6 1 1 1 0 0
7 0 0 0 0 1
8 1 0 0 1 1
10 0 0 0 0 1
(2) Experimental process
The experimental steps are summarized as follows;
1) Input the product gene, and set the assignment algebra of the randomly generated initial
product gene population to 0;
2) Calculate the fitness function value of each product gene;
3) Obtain the optimal product gene according to Figure 4;
4) Establish an evaluation index set to evaluate the quality of the selected product genes by the
two methods;
5) Using the fuzzy evaluation method, compare the traditional method with the genetic
optimization algorithm-based method for product design in the automotive industry to retrieve the
gene matching degree.
4 Results and Discussion
The relationship between the values is analyzed by means of a line graph, and the experimental
results are shown in Figure 5.
Figure 5 Search product gene matching results
It can be seen from the experimental results in Figure 5 that, the gene matching degree of the
products retrieved by the traditional method is low, the reason for this phenomenon is that the
product structure efficiency of the traditional method is low, and the optimal solution cannot be
obtained. The genetic optimization algorithm adopted by the author is based on the pattern
theorem, and the retrieved products have a higher degree of gene matching. Among the 10
samples of the author's method, the highest matching degree is 0.7%. Thus, it is proved that the
designed method of automobile industry product based on genetic optimization algorithm meets
the design requirements.
5 Conclusion
The author proposes the application of multi-objective genetic algorithm in the design
optimization of auto parts products, aiming at the problems caused by the traditional automobile
industry product design method, the author puts forward the application research of genetic
optimization algorithm in the automobile industry product design. Using the concept of biological
genetic engineering and functional analysis, the research on the product design of the automobile
industry based on genetic optimization algorithm is completed, and the feasibility of the proposed
method is verified. 10 samples were randomly selected from the database, and a triangular matrix
was designed to ensure the rationality of the experimental results. The experimental results show
that the automotive industry product design method based on genetic optimization algorithm
meets the design requirements.
References
[1] Dai, Y. . (2022). Improvement of the quality of curriculum for advancement of the
implementation of "double reduction" — taking curriculum construction of the shangqiao
experimental primary school as an example. Journal of Higher Education Research, 3(3), 280-284.
[2] Laribi, M. A. , Tiebi, R. , Tamboura, S. , Shirinbayan, M. , Tcharkhtchi, A. , & Dali, H. B. ,
et al. (2020). Sheet molding compound automotive component reliability using a micromechanical
damage approach. Applied Composite Materials, 27(5), 693-715.
[3] Krug, A. , Martinez-Turtos, A. , & Verhoeyen, E. . (2022). Importance of t, nk, car t and car
nk cell metabolic fitness for effective anti-cancer therapy: a continuous learning process allowing
the optimization of t, nk and car-based anti-cancer therapies. Cancers, 14(1), 183.
[4] Chen, B. , Liu, Y. , Zhang, C. , & Wang, Z. . (2020). Time series data for equipment
reliability analysis with deep learning. IEEE Access, PP(99), 1-1.
[5] Tekkouk, N. , Arbaoui, M. , Abdi, S. , & Rezzoug, A. . (2020). Transition between severe and
mild wear of 2024a-t4 anodized aluminum alloy under severe wear conditions. Journal- Korean
Physical Society, 76(10), 899-903.
[6] Pucker, A. D. , Steele, K. , Rueff, E. , Franklin, Q. X. , & Walline, J. J. . (2021). Contact lens
adaption in neophytes. Optometry and Vision Science, 98(3), 266-271.
[7] Zeng, S. Y. , Wu, R. , Truong, K. N. , & Chevalier, F. . (2020). Evaluating the gradual
delivery of knowledge-focused and mindset-focused messages for facilitating the acceptance of
copd. Proceedings of the ACM on Interactive Mobile Wearable and Ubiquitous Technologies,
4(4), 1-23.
[8] Lomakin, M. I. , Sukhov, A. V. , Dokukin, A. V. , & Niyazova, Y. M. . (2021). Assessment of
reliability indicators of space vehicles under conditions of incomplete data. Cosmic Research,
59(3), 199-203.
[9] Chao, M. A. , Kongyou, W. U. , Pei, Y. , & Tianran, L. I. . (2022). Quantitative analysis of
tectonic characteristics and evolution in the eastern junggar basin. Journal of Geomechanics,
25(3), 313-323.
[10] Wu, L. Y. , Wang, Z. , Ma, D. , Zhang, J. W. , Wu, G. , & Wen, S. , et al. (2022). A
continuous damage statistical constitutive model for sandstone and mudstone based on triaxial
compression tests. Rock Mechanics and Rock Engineering, 55(8), 4963-4978.
[11] Cosma, O. , Pop, P. C. , & Zelina, I. . (2021). An effective genetic algorithm for solving the
clustered shortest-path tree problem. IEEE Access, PP(99), 1-1.
[12] Chen, L. , Kong, J. , Deng, S. , & Shen, C. . (2021). Analysis on overview and hot spots of
polar science based on bibliometrics. Chinese Journal of Polar Research, 33(2), 294-306.
[13] Wang, J. , Du, H. , Xing, J. , Qiao, F. , & Ma, Y. . (2020). Novel energy- and maintenance-
aware collaborative scheduling for a hybrid flow shop based on dual memetic algorithms. IEEE
Robotics and Automation Letters, 5(4), 1-1.
[14] Pr, L. , Jinny, S. V. , & Mate, Y. V. . (2021). Early prediction model for coronary heart
disease using genetic algorithms, hyper-parameter optimization and machine learning techniques.
Health and Technology, 11(1), 63-73.
[15] Latifa, N. B. , & Aguili, T. . (2022). Optimization of coupled periodic antenna using genetic
algorithm with floquet modal analysis and mom-gec. Open Journal of Antennas and Propagation,
10(1), 15.
[16] Ren, X. , Li, C. , Ma, X. , Chen, F. , Wang, H. , & Sharma, A. , et al. (2021). Design of multi-
information fusion based intelligent electrical fire detection system for green buildings.
Sustainability, 13(6), 3405.
[17] Shriram, S. & Jaya, J. & Shankar, S. & Ajay, P.. (2021). Deep Learning-Based Real-Time AI
Virtual Mouse System Using Computer Vision to Avoid COVID-19 Spread. Journal of Healthcare
Engineering. 2021.
[18] Zhao, X. L. , Liu, X. , Liu, J. , Chen, J. , Fu, S. , & Zhong, F. . (2019). The effect of ionization
energy and hydrogen weight fraction on the non-thermal plasma vocs removal efficiency. Journal
of Physics D Applied Physics.
[19] R. Huang, X. Yang, "Analysis and research hotspots of ceramic materials in textile
application," Journal of Ceramic Processing Research, vol. 23, no. 3, pp. 312–319, 2022.
[20]Liu, Q., Zhang, W., Bhatt, M. & Kumar, A. (2021). Seismic nonlinear vibration control
algorithm for high-rise buildings. Nonlinear Engineering, 10(1), 574-582.