Combining GRA with a Fuzzy QFD Model for the New Product Design and Development of Wickerwork Lamps

Abstract

With the popularization of the concept of sustainability in traditional wickerwork, wickerwork lamps have become the most popular production. When customers purchase wickerwork lamp products, the Kansei consensus has become a key factor influencing the communication between manufacturers and customers. Therefore, the purpose of this paper is to explore the product design solutions for wickerwork lamps that meet the emotional satisfaction of users. Firstly, a three-level evaluation grid diagram driven by user attractiveness through Miryoku Engineering is established. Secondly, this paper uses grey relational analysis (GRA) to extract the priority order and its weight values in the perceptual vocabulary to identify the key user needs in product design. In order to effectively deal with the uncertain product evaluation information, the fuzzy quality function deployment (QFD) is used to construct the “emotional demand-design parameter” transformation model and derive the optimal design parameters in the mapping process, thus effectively reducing the ambiguity and uncertainty in the demand transformation process. Based on the experimental results, it is found that the best combination of Texture light transmittance, Simple wickerwork material, Wickerwork primary colours, Cascaded type and Pastoral style could be preferred by customers, thus this proposed method can effectively reduce the ambiguity and uncertainty in the design process. The results of study enable designers to accurately grasp customers’ perceptions of wickerwork lamp products and obtain the best design parameters for wickerwork lamp products.

1. Introduction

In a fiercely competitive market with an increasing variety of products in the market, companies need to understand and meet rapidly changing customer requirements (CRs), accurately grasp competitive market information, shorten product development cycles, and make sound new product development (NPD) decisions in order to stand out in the competitive market [1]. Therefore, designers need to have a deep understanding of user needs to create design opportunities. Many factors, including product branding, functionality, appearance, and usability, influence customer perceptions, and among these considerations, product appearance remains one of the most important factors in consumer purchase decisions [2]. The development of new technologies has led to significant changes in the traditional product design process and made the analysis of consumer needs and behaviours an important tool for product design and marketing [3], which also intuitively reflects the design strategy of user-centred design (UCD). Furthermore, how to better support designers to acquire design knowledge according to consumers’ needs becomes the key factor to knowledge innovation in the design process [4].

In fact, design is a complex, iterative, and innovative process [5], where the goal of product design is to design a product that is attractive in accordance with user preferences in order to increase market share. In recent years, many companies have been trying to improve the appearance of their products to create a competitive advantage in the market, which is especially true in today’s consumer products [6,7,8]. The composition and arrangement of visual design elements (i.e., shape, colour, texture, etc.) can produce a universally appealing product appearance, which reflects the intrinsic attractiveness perceived by the human senses. According to human perception, certain lines, proportions, and colour combinations are considered aesthetically pleasing, and thus designers often apply aesthetic design principles to the planning and placement of visual design elements [9,10]. However, the process of applying aesthetic design principles relies heavily on the designer’s perception and experience [9], so it can be a challenge for novice designers of product appearance design. In addition, an attractive product is difficult to define in terms of appropriate design elements [11] and aesthetic design principles based on subjective experience [9]. To handle this problem, this study proposes a research framework that combines grey relational analysis (GRA) and fuzzy quality function deployment (QFD) to meet the emotional needs of customers for product design solutions. This study attempted to establish a rational communication channel between the customer’s appeal preferences and the product design elements, which can visually inform designers by modelling the mapping relationship between the product’s design elements and the user, as well as quantifying the customer’s vague appeal needs into specific design parameters.

In recent years, many scholars have successfully applied Miryoku Engineering to the field of product design [12,13,14,15]. The Evaluation Grid Method (EGM) is an important research method in Miryoku Engineering, which can obtain the attractiveness factors between consumer and design elements [16]. In previous studies, the EGM method has been proven to be a promising approach to exploring the attractive factors, but it does have two key problems in its application: 1. The extraction of higher representative perceptual adjectives is mostly based on the importance of the frequency mentioned during the interview process, and the number of mentions itself only represents the user’s subjective feelings without any substantial correlation with the important ones. Some scholars then use Analytic Network Process [17] and Analytic Hierarchy Process (AHP) [18,19] to calculate the weight of each attraction factor in the ranking, but these methods are easily limited by subjective evaluation and the results are prone to bias, while the consistency of experimental results is difficult to be guaranteed due to the complexity of the problem. 2. The mapping link between the upper perceptual factors and the lower specific design factors usually uses statistical analysis methods, which generally follow a linear relationship. Furthermore, it is usually difficult to apply Quantification Theory Type I (QTT-I) and linear regression methods to accurately measure customers’ nonlinear and non-normal emotional images, such as elegant, spirited and other related lexical intentions during the analysis of the upper and lower relations in EGM. Chen and Li [20] obtained specific features of game design based on EGM and used multiple linear regression analysis (MLR) to explore different players’ perceptions of design. Ko et al. [12] combined Miryoku engineering and QTT-I to explore the attractiveness factors of office chair products. However, the linear regression method can only assess the relationship between linear values of length and height [21], but it cannot measure nonlinear emotional variables, and thus the results of linear calculations have some degree of limitation.

In order to handle the first problem, the GRA is used to measure the relative weights and priorities of evaluation items. The GRA have highly advantageous in analysing objects with small sample sizes and dealing with relationships between objects whose potential relationships are unclear [22]. Therefore, this study applied GRA to find the degree of correlation between each factor in the system to determine the importance ranking and weighting of upper-level sentiment words, to help designers prioritise the development of attractive product design solutions with the highest degree of this sentiment need. To address the second problem, the fuzzy QFD was found to have good non-linear matching capabilities to overcome the shortcomings of traditional linear models with missing critical information [23]. The model starts from customer needs, analyses the engineering characteristics associated with user needs, and uses a matrix to reflect the intrinsic link between user needs and engineering characteristics in product development [24,25], which can effectively solve the problem of ambiguity and uncertainty in perception assessment and guide product design practice, thus improving the validity of the calculation results. Therefore, this study uses the fuzzy QFD to establish a functional relationship to convert the language of the customer’s needs into the design parameters required by the designer, which is more beneficial than traditional linear regression in finding the mean value after gradient descent, to achieve the design parametersefficiently.

In addition, wickerwork is one of the ancient traditional crafts in Chinese history and has been listed as a national intangible cultural heritage by the State Council [26]. To promote the revitalization and inheritance of the intangible cultural heritage of Funan wickerwork, the academic community has paid more attention to the redesigned research by implanting cultural genes into the form and structure of cultural creative products and exploring the innovative development of traditional wickerwork in modern life through digital technology [26,27]. To this end, this study takes wickerwork lamps as the research object to promote the revitalization and inheritance of the intangible cultural heritage of Funan wickerwork. How to make the wickerwork lamps effectively meet the expectations of consumers, it is necessary to accurately grasp the psychological feelings brought by the wickerwork lamps to users through the innovation of design methods. What factors in the wickerwork lighting products attract consumers to buy, which is an important issue in the design of wickerwork lamps, and it is also a question worthy of academic discussion. However, there are still relatively few studies combining GRA and fuzzy QFD to design and develop wickerwork lamp products, which has led to a gap in developing user-satisfying wicker luminaire products. Hence, in this study, to fill this gap, this paper proposes a systematic approach to new product development that combines GRA with fuzzy QFD to parse users’ perceptual cognition and discover the artistic characteristics of wickerwork lamps, to generate valuable design solutions with more precise parameters to comprehensively address these barriers and gaps.

Therefore, the main highlight of this paper is to provide an innovative design development method for wickerwork lamp manufacturers, which not only reduces design time but also satisfies the emotional needs of most users. Furthermore, the combination of GRA and fuzzy QFD allows for a more accurate translation of user needs into product design parameters on a small data set, resulting in the design of wickerwork lamp products that meet user appeal. This study takes the revitalized creation practice of the intangible cultural heritage Funan wickerwork craft as a starting point, the main purpose is to assess attractive factors and translate them into design elements to make wickerwork lamp products stand out in the market competition. The QFD is a quantitative method that uses a quality matrix to translate customer needs into engineering features and combined with fuzzy concepts, fuzzy QFD can objectively measure experts’ questionnaires, on the basis of which the most perceptually attractive product design feature elements can be discovered. Accordingly, the results of the study are expected to derive individual design parameters of wickerwork lamps products, which in turn guide the product design and effectively make the transformation of users’ needs clearer and more accurate, thus helping designers to improve design efficiency and promote the design path of traditional handicraft culture to evolve into modern design. In addition, the similarities and differences between the research in this paper and previous studies are briefly compared in

Table 1. A brief comparison between this study and the previous ones.

References	Expert Interview	Kansei Evaluation	Functional Features	Design Innovation	Product Configuration
This paper	EGM	GRA			Fuzzy QFD
Wang [28]		KE	RST		CA, GRA
Ko, Lo and Chen [12]	EGM				QTT-I
Wang [29]		KE	RST	TRIZ
Zhang and Li [30]	EGM				QTT-I
Wang and Zhou [21]	EGM	Kano			IGA
Hsiao, et al. [31]		AHP			QTT-I, GA,
Wang [32]		NLP, GRA			Fuzzy Topsis
Wang and Zhou [13]	EGM	Kano	RST		FWARM
Lin, Yeh and Wang [25]	Focus group	Kano			Fuzzy QFD
Haber, et al. [33]		Kano	AHP		QFD
Ji, et al. [34]		Kano			QFD

QFD: quality function deployment, CA: conjoint analysis, FA: factor analysis, KE: Kansei engineering, EGM: evaluation gird method, NLP: natural language processing, NN: neural network, RST: rough set theory, FWARM: fuzzy weight association rule mining, GA: genetic algorithm, TRIZ: the theory of inventive problem solving, MLR: Multiple Linear Regression, AHP: analytic hierarchy process, IGA: interactive genetic algorithm, NLP: natural language processing.

, and the main contributions of this paper are summarized as follows.

• Applying the EGM to study the charm factor of wickerwork lamps production, and establishing the evaluation construct map with lamps as the sample charm factor.
• Using the GRA to objectively determine the relative importance of the emotional needs of the upper items refined by EGM, to identify the key user needs in the product design, and to obtain people’s key abstract feelings about the style of wickerwork lamps.
• Utilizing the fuzzy QFD model to build a quality house of “emotional needs-design parameters”, and improve the interactivity with users in the mapping process in the form of network interaction to reduce the fuzziness and uncertainty of demand transformation.

This paper is organized as follows. Section 2 reviews the previous related work based on EGM and discusses the application of GRA and QFD. Section 3 briefly introduces our proposed framework and describes each step of the product design study. Section 4 analyses and discusses this study, and Section 5 concludes this study.

2. Review

2.1. Evaluation Grid Method

The evaluation grid method (EGM) is an important research method in Miryoku Engineering [20]. This method explicitly explores the similarities or differences of the subjects by conducting personal interviews on their comparisons and then analyzing the characteristics of the targets to obtain the evaluation elements of attractiveness [12,20,24]. Sanui and Inui [35] improved the psychological concept of the repertory grid method proposed by Kelly and added two steps [30]. Firstly, subjects were asked to answer what they liked or disliked after a two-by-two comparison of the samples. Then, they were asked additional questions about the most preferred reasons and graded according to the meaning or conditions clarified by the questions. In fact, the EGM helps to gain insight into the psycho-cognitive level of the subjects [16], and even abstract psychological feelings and subtle emotional changes that are difficult to capture can be visualized by this method [21,30]. In the EGM, the respondents’ feelings are visualized and divided into three levels: upper, middle, and lower levels [24]. The upper level represents respondents’ abstract emotional phrases, the middle level represents actual attractiveness and original evaluation items, and the lower level refers to specific constitutive conditions or physical characteristics [20], as shown in Figure 1.

Due to the strong performance of EGM in extracting design elements, it has been applied to product design by experts in recent years. Zhang and Li [30] used EGM to extract the attractiveness factors of green products and analyzed the influence weights of green product design factors using quantitative theory type I (QTT-I). Kang [36] used Miryoku Engineering to establish a three-level evaluation grid platform and used neural networks to establish a mapping function between key Kansei factors and representative product design elements so that the most perceptually attractive product design was discovered. Wang and Zhou [13,21] applied EGM to establish key attractiveness factors of electric bicycle products and completed the innovative design and development of product form. Liu et al. [24] used EGM to collect and analyze customer needs and design elements, calculated the weight values of customer needs using the AHP, and finally established a fuzzy relationship matrix by QFD, to explore the design element based on user need. Moreover, the specific steps of EGM are as follows [13,21]:

• Firstly, select the more classic experimental samples on the market, printed them out in the size of A4 printing paper, and then displayed them in front of the user.
• The user is asked to pick out their favourite sample pictures, thus dividing the pictures into two categories.
• Ask the user the specific reason why he likes the sample picture, and determine the key reason why the sample attracts the user, that is, the median item of the charming factor of the product.
• To further ask the users about the figurative features of their favourite reason, and the result is the lower morphological feature of the charm factor of the product, and ask the users about the psychological feeling of the specific attractive morphological feature of the product, and the result is the upper perceptual imagery of the attractive factor.
• To survey the upper, middle and lower positions of all preferred sample images, and connected the corresponding items by straight lines to show the hierarchical relationship between Kansei word and modelling elements.
• To merge the same or similar attribute items by the KJ simplification method [37], and repeat the step until no more grouping is possible, to obtain a clear attractive factor of the product.

2.2. Grey System Theory

2.2.1. Grey System Theory and Its Application in Socio-Economic Sustainability

The intersection of the economic and social environment generates many modern complex scientific problems, and there is no doubt that this complexity affects all subsystems and related fields within these systems [38]. The interacting coupled system of urbanisation and the environment can be considered a complex grey system with imperfect information, and its systemic complexity may lead to greater uncertainty [39]. To resolve this considerable ambiguity, it is important to acknowledge the role of uncertainty-based theories and methods.

In 1982, Deng Julong, a Chinese scholar, proposed a statistical analysis method called the grey system theory, which utilizes a multi-factor approach [40]. Grey system theory (GST) is a relatively new approach that focuses on the study of systems involving small samples and poor information [41], and it relies on bodies of knowledge, concepts and methods. Given the complexity and dynamics of socioeconomic systems, GST can greatly assist in the analysis of these socio-environmental systems, thereby contributing to sustainable social development [42]. Huang et al. (2010) listed as many factors as possible, including social, economic and natural factors that may affect the urban heat effect, and further demonstrated the validity of the GST system by conducting a grey correlation analysis on 22 selected indicators and calculating their grey correlation degrees [43]. Wang et al. (2018) used GRA to investigate the relationship between gross domestic product (GDP), income elasticity, social consciousness, the role of urbanisation and renewable energy development [44]. Li et al. [45] quantified the causal relationship between air pollution and social development as a complex grey system. They relied on grey correlation models to present how socio-economic and human activities affect air pollution in urban areas. Shan et al. [46] applied grey absolute correlation to assess the contribution of technology entrepreneurship to national development. Wang and Pei [47] relied on a grey-integrated approach to address the sustainability of urban tourism. They argue that this method can help tourism management to develop policies.

2.2.2. Gray Correlation Analysis in Production Design

The Fourth Industrial Revolution focuses on organizing and controlling the value chain of the product lifecycle to meet the increasingly individualized needs of customers [48]. This presents an opportunity to transition towards a more sustainable and circular production model. The circular economy (CE) is a framework that uses organizational planning processes to efficiently and effectively use ecosystems, economies, and product cycles to deliver products, components, and materials to customers and society while closing the loop on all relevant resource flows [49]. Simultaneously, the value chain is shifting from mass-produced, expert-designed products to mass-individualized, user-driven design products [50]. Therefore, understanding consumer needs and the responses of firms to green products is critical in developing an effective eco-design concept. However, a lack of understanding of consumers’ emotions, perceptions, and innate mental models presents a significant challenge [51]. To address this challenge, researchers have adopted a GST-based approach to product design, which enables the integration of limited, incomplete, and uncertain information into the design process [39], to obtain consumer-related information to gain a better understanding of their needs.

As a method of grey systems theory, grey correlation analysis could address the correlation problem in the field of fuzzy cognition [22]. It is particularly useful for analyzing ambiguous and uncertain decision situations where data lack obvious regularity [52]. The method relies on small sample data to rank the correlation of each characteristic [32]. GRA determines the similarity of the relationship between two sets of random series [53]. It can achieve satisfactory results even with a small amount of data or factors with a large amount of variability [54]. The underlying principle is to determine the degree of association between factors based on the similarity of the geometry of the curves of the system factor series. If the geometry of the curves is similar, the degree of association is higher; On the contrary, the degree of association is lower [32].

Previous studies have used a variety of statistical analysis methods to investigate the relationship between variables, such as factor analysis [55] and multiple linear regression [1], but these methods have certain shortcomings, requiring a large sample size, obeying a specific probability distribution, and a large computational effort. The GRA is based on a ‘small sample, information-poor’ grey system, which makes up for the shortcomings of these statistical methods. At the same time, in the early stages of product development, GRA can accurately analyse data and extract useful information despite the lack of sufficient data or incomplete data. Therefore, it is widely used in the design and development of products. Some scholars have applied GRA to the design and development of products. Wang [32] adopt the GRA to quantify the relationship between coffee machine product form and perceptual image. Quan et al. [56] applied the GRA-TOPSIS method to rank alternative product design solutions, and took an electric drill as an example to describe the specific implementation process of this method. Xue et al. proposed [57] an integrated decision system for optimal product image design based on Kansei Engineering experiments, establishing a grey correlation analysis-fuzzy logic sub-model affecting design elements and then applying the utility optimization model to obtain multi-objective product images. Wei [40] analyzed the shape and colour design of guide signs in public spaces based on the GRA. Chen and Chuang [58] combined GRA with the Taguchi method to optimize the subjective quality of customer satisfaction.

Therefore, in this study, the GRA was introduced into EGM to explore the importance of upper semantic vocabulary based on the semantic evaluation of each Kansei factor, where Kansei factors were randomly set as comparison sequences and additional words were set as reference sequences to finally find the weight values of perceptual words. The GRA calculation process was divided into the following seven steps [28]:

To establish comparison levels: Suppose X denotes an m × n matrix to collect assessments of the performance levels of n-dimensional elements in m alternatives, where x_i(k) denotes element k of alternative I, m is the number of a series of alternatives, and n denotes the number of criterion attributes.

$$\mathrm{X}=\left[\begin{array}{cccc}{x}_1(1)& x_1(2)& \cdots & x_1(n)\\ x_2(1)& x_2(2)& \cdots & x_2(n)\\ ⋮& & & ⋮\\ x_m(1)& x_m(2)& \cdots & x_m(n)\end{array}\right]$$

(1)

The reference series setting: The elements of the reference series are set by their corresponding maximum rating scale, which is also similar to the ideal solution.

$$x_r=\left\{(\max x_i\underset{i}{(j)\left|j\in \widehat{J}\right|},(\min x_i\underset{i}{(j)\left|j\in \widehat{J)}\right|}i=1,2,\dots m\right\}$$

(2)

where $\widehat{J}$ denotes the set of benefit (the-larger-the-better) terms, and $\widehat{J}$ denotes the set of cost (the-smaller-the-better) terms.

The deviation matrix is used to calculate the “degree of greyness” between the target series and the reference series:

$$\Delta =\left[\begin{array}{cccc}{\Delta }_1{}_r(1)& {\Delta }_1{}_r(2)& \cdots & {\Delta }_1{}_r(n)\\ {\Delta }_2{}_r(1)& {\Delta }_2{}_r(2)& \cdots & {\Delta }_2{}_r(n)\\ ⋮& & & ⋮\\ {\Delta }_{mr}(1)& {\Delta }_{mr}(2)& \cdots & {\Delta }_{mr}(n)\end{array}\right]$$

(3)

where ${\Delta }_{ir}(j)=\Vert x_i(j)-x_r(j)\Vert$, x_i(j) and x_r(j) represent the jth element of the target and reference series, respectively.

Calculating the attribute-based grey correlation coefficients:

$$\gamma (x_i(j),x_r(j))=\frac{{\Delta }_{\min }+\xi {\Delta }_{\max }}{{\Delta }_{ir}(j)+\xi {\Delta }_{\max }}$$

(4)

$${\Delta }_{\min }=\underset{\forall i}{\min }\underset{\forall j}{\min }\Vert x_i(j)-x_r(j)\Vert$$

(5)

$${\Delta }_{\max }=\underset{\forall i}{\max }\underset{\forall j}{\max }\Vert x_i(j)-x_r(j)\Vert$$

(6)

where a distinguishing coefficient $\xi \in (0,1)$ is set at the 0.5 value.

The grey correlation degree is the average of grey correlation coefficients:

$$\Gamma (x_i(j),x_r(j))=\frac{1}{n}{\displaystyle \sum _{j=1}^n\gamma (x_i(j),x_r(j))}$$

(7)

2.3. Fuzzy Quality Function Deployment

The QFD was originally developed by Mitsubishi in 1972, and this method is essentially a practical problem-solving technique for performing product design and planning to meet customer needs and expectations [59]. Akao [60] defined QFD as a multilevel analysis method for customer requirements-driven design and development that translates CR into design goals and product quality throughout the production phase. It is a market-user-oriented approach to ensure product quality during product development [59]. The House of Quality (HoQ) is the key point of the QFD [61], which could reveal the relationship between engineering characteristics and customer requirements [62,63], and reflect the intrinsic connection between user requirements and technical information and specifications in product development with a visual matrix [64].

By quantifying and analyzing Customer Requirements (CRs) and Engineering Characteristics (ECs) and the relationships between them, CRs are transformed into ECs [65], and then relationships between ECs and competitive market analysis are added to facilitate the entire QFD process [66]. The structure of HoQ is generally a two-dimensional relationship matrix whose components include the left wall for the input of CRs, the ceiling for the input of ECs, the right wall for the input of competitive analysis, the CR and EC relationship matrix, the correlation matrix between ECs in the roof, the importance of CRs and the importance of ECs in the floor, as shown in Figure 2.

QFD is used to facilitate the transformation of user requirements in design into engineering feature parameters [67], which can effectively lead to product types that meet both market needs and efficient use of corporate costs. Caligiana et al. [68] proposed an innovative sustainable design approach by integrating QFD with the theory of inventive problem solving (TRIZ). Ji et al. [34] proposed a new integrated approach that the qualitative results of the Kano model are integrated into QFD to complete laptop design. Dolgun et al. [69] apply QFD theory to product innovation by converting user needs into design specifications that can be implemented in terms of technical requirements, and using a yoghurt box as an example of an optimal product design process.

The QFD is used to collect and analyze the “voice of the customer” to assess people’s perceptual and linguistic attributes, which inevitably leads to the problem of subjectivity and ambiguity [65]. The ambiguity and imprecision of QFD are caused by two important factors, firstly, there is a lack of formal mechanisms to translate WHAT (voice of the customer) into HOW (design feature). Usually, there are many WHATs used for perception, each WHAT can be transformed into multiple HOWs, and a particular HOW may affect multiple WHATs. Since the relationship between WHAT and HOW is usually vague or imprecise [70], it should be expressed in a more quantitative way and in technical terms [25]. Secondly, due to the uncertainty in the selection process, the data available for the design characteristics are often limited and may be inaccurate. Therefore, a certain degree of ambiguity is often unavoidable [71].

Currently, there is a wealth of research results on the competing user requirements in QFD, and a more applied treatment of expert experience and evaluation information is the Delphi method [72,73], but this method cannot handle the fuzzy information contained in the process of association determination. To improve judgment accuracy and handle the ambiguity as well as incompleteness and uncertainty in the user information evaluation process, most scholars have widely used theories such as fuzzy theory, grey theory, rough theory, and information axiom to effectively reduce information loss in semantic transformation. Chan et al. [74] used triangular fuzzy functions in the QFD model step to capture human fuzzy semantic evaluation. Vinodh et al. [75] applied fuzzy QFD for the sustainable design of consumer electronics products. Chen [67] develops an integrated approach for identifying risk components, they build a quality function deployment to characterise customer needs under fuzzy assessment semantics, and the failure causality relationships between and within product components are characterised by a directed network model. This literature reveals that there are still few studies that extend fuzzy QFD to product design to accurately translate the diverse perceptual needs of customers into target values of engineering features. In the general industrial design process, designers start sketching and seeking ideas after obtaining requirements, which could ignore the definition of product design parameters. Therefore, this study constructs the fuzzy QFD of “emotional requirements—design parameters” to transform user requirements, and uses the triangular fuzzy numbers instead of clear values to increase the credibility of the transformation process to improve the accuracy of experimental results.

The absolute value weight AI_j of engineering features of QFD can be obtained by fuzzy weight weighting calculation by multiplying the customer demand weight by the correlation sign of the correlation matrix, and the absolute weight of engineering features is calculated and weighted. Based on absolute weight calculation, it is necessary to formalize the calculation of absolute weights to obtain relative weights RI_j, which calculation formula is as follows:

$$\begin{array}{c}A{I}_j={\displaystyle {\displaystyle \sum }_{i=1}^n}{W}_i\otimes R_{ij}\end{array}$$

(8)

$$\begin{array}{c}R{I}_j=\frac{A{I}_j}{{{\displaystyle \sum }}_{j=1}^{m}A{I}_j}\end{array}$$

(9)

W_i is the importance value of the ith customer requirement; R_ij is the relationship value between the ith customer requirement and the jth engineering feature. I = 1, 2,…, n, n denotes the total number of customer requirements. j = 1, 2,…, m, m denotes the total number of engineering features.

For triangular fuzzy numbers, the dependent function distribution can be formed as a triangle containing (l, m, u), where l, m and u denote the minimum possible value, the middle value and the maximum possible value, respectively. In this study, the centre of gravity method is used to obtain the fuzzy weight values. The principle of the centre of gravity method is to represent the collection function by the central value of the fuzzy set. To set A = (l₁, m₁, u₁) as the triangular fuzzy number, its operation is as follows [76]:

$$Defuzzy\left(\tilde{A}\right)=\frac{\left|\left(u_1-l_1\right)+\left(m_1-l_1\right)\right|}{3}+l_1$$

(10)

3. Methods

In this study, the advantages of EGM, GRA and fuzzy QFD are effectively combined to develop an attractive product design framework, as shown in Figure 3. The study is divided into three phases: firstly, the expert group through the EGM interview process to obtain a three-level assessment grid map that captures upper-level abstract causes, middle-level original assessment items, and lower-level specific design conditions. Afterwards, the GRA was used to identify key upper Kansei factors and create corresponding morphological deconstruction maps for the Kansei factors with the highest grey correlation. Finally, the fuzzy QFD is used to establish a mapping model between key Kansei factors and product-specific forms, and the most attractive wickerwork lamp design solution is derived on this basis. To take users’ emotional needs as the starting point based on the fuzzy QFD, which can identify the design parameters constraints that lack competitiveness in the design, and clarify the ideas for innovative product design, to guide the new round of product development.

3.1. Identify the Attractiveness Factors Based on EGM Method

To better conduct EGM experiments, 12 highly engaged subject groups were recruited for experimental interviews in this study. They consisted of six males and six females, aged between 22 and 49 years old, all subjects with more than three years of industrial design experience (

Table 2. The basic information of 12 subjects.

Project	Content	Number	Percentage (%)
Gender	Male	6	50%
	Female	6	50%
Age	22–29	5	41.67%
	30–39	5	41.67%
	40–49	2	16.67%
Education	College	2	16.67%
	University	5	41.67%
	Graduate	5	41.67%
Profession	Student	3	25%
	Teacher	3	25%
	Designer	6	50%

). The results of the frequency analysis of the gender and age of the survey respondents show that this distribution largely meets the needs of the survey: Among the gender survey results that can be seen are 50% male and 50% female; in the age survey, the impact is mainly seen in the 22–29 and 30–39 age groups, with a total of 5 people in each of these groups, each accounting for 41.67%, followed by the 40–49 age group with 2 people, accounting for 16.67%; in terms of profession, there are 3 students and teachers, each accounting for 25%, and the designer group accounts for 50%.

Firstly, by searching Internet resources, consulting magazines and other sources, a total of 80 sample images of wickerwork lamp product designs were collected for this study. After removing pixel blurring and images with large environmental influences (e.g., light, environmental reflections, and shadows), a total of 35 representative and distinctive wickerwork lamp products were selected (Figure 4). The samples of all products were cropped to 10 cm × 10 cm cards using the 2D software (Adobe PS), thus facilitating subsequent subject observation. 12 subjects were invited to compare wickerwork lamp product cases with each other and were interviewed in terms of pros and cons, likes and dislikes, and the original reasons for preference were obtained.

Specifically, first, the host will explain the purpose and procedure of the experiment, and then 12 experts will be invited to discuss the collected sample images one by one. In a closed space with no outside distractions, each participant will be interviewed for approximately 40–60 min. During the interview, they can remove samples of wickerwork lamp product examples that they do not like. Next, a deeper interview is made into what customers like about a particular design and the upper-level abstract emotional preferences, middle original design items and lower-level specific design elements are extracted. Finally, the individual evaluation structure was mapped and saved as a document based on the experimental results of the three evaluation items (i.e., upper, middle and lower levels) and the correlations between each level. In fact, the results of the obtained interview items were very large at the end of the interview based on the wickerwork lamp. To avoid too many perceptual factors to burden the designer in the analysis process, the extracted perceptual factors were combined and simplified based on the KJ method, and the obtained results had 5 middle-level evaluation items, 36 lower-level evaluation items, and 14 upper-level items, and the results were connected using straight lines to establish a complete evaluation system to form an evaluation construction diagram, as shown in Figure 5. Specifically, 14 upper-level evaluation terms as shown in

Table 3. 14 representative Kansei words.

K1	K2	K3	K4	K5	K6	K7
Lively	Steady	Smooth	Bright	Ecological	Simple	Warm
K8	K9	K10	K11	K12	K13	K14
Technological	Traditional	Concise	Plain	Personalized	Rhythmical	Vibrant

.

3.2. Using the GRA to Identify Key Kansei Factors

In this phase, we attempt to use GRA to obtain the weight of each Kansei word to assist designers in deciding the key Kansei factors. Firstly, 20 representative wickerwork lamp product design solutions were selected from 35 images, and 146 experts were invited to conduct the perceptual assessment. 146 experts including 60 men and 86 women, aged 23–45, with three years or more industrial design experience, and the mean value of the assessment results were calculated, and the results of the perceptual experiment are shown in

Table 4. The Kansei assessment results.

Kansei Words	1	2	3	4	5	6	…	16	17	18	19	20
K1	2.875	3.250	1.875	2.875	3.500	3.125	…	4.000	3.625	3.250	3.125	2.875
K2	3.375	2.125	4.000	2.625	2.750	3.000	…	2.250	2.875	2.625	2.750	3.250
K3	3.625	3.125	2.250	3.250	3.625	3.625	…	3.125	3.500	3.125	2.750	2.375
K4	3.125	2.750	2.500	2.375	2.750	3.250	…	2.625	3.250	3.125	3.125	2.250
K5	4.375	3.875	3.375	3.375	3.625	3.250	…	4.000	3.000	3.125	3.625	3.125
K6	3.875	3.125	2.875	3.000	3.625	3.250	…	2.500	3.250	3.250	3.125	2.750
K7	4.125	3.500	2.375	2.875	3.500	3.375	…	3.125	3.375	3.125	3.375	3.375
K8	1.875	2.000	1.750	2.250	3.125	2.750	…	2.750	3.125	3.375	2.625	2.250
K9	3.750	3.125	3.750	3.250	2.750	2.375	…	2.875	2.625	2.875	3.000	4.000
K10	3.500	3.125	2.625	3.000	3.625	3.500	…	2.625	3.500	3.500	3.125	3.125
K11	3.875	3.625	3.375	3.125	3.250	3.375	…	3.125	2.875	3.125	3.250	3.000
K12	3.375	3.625	2.250	3.500	4.000	3.625	…	3.375	3.250	3.250	3.375	3.375
K13	3.125	3.375	2.750	3.000	3.500	2.750	…	3.250	3.625	3.375	3.125	2.375
K14	3.250	3.375	2.250	3.625	3.250	3.375	…	3.250	3.500	3.125	3.125	2.750

. Nextly, the word K10 of “Concise” was set as the reference sequence, the other 13 perceptual words were set as the comparison sequence, and the formulas (1)–(7) were used to calculate the degree of grey correlation of these 13 words, and the results are shown in

Table 5. Ranking of GRD.

Kansei Words	GRD	Ranking
K4	0.848	1
K3	0.820	2
K6	0.809	3
K14	0.806	4
K5	0.793	5
K11	0.791	6
K13	0.782	7
K12	0.779	8
K7	0.776	9
K9	0.770	10
K2	0.762	11
K1	0.694	12
K8	0.479	13

.

Based on the results of the EGM experiment, the evaluation construction diagram shown in Figure 5, which has five middle-level design variables: A, the material of the lamp; B, the colour of the lamp; C, the form of the lamp; D, the pattern of the lamp; E, the design style of the lamp. According to the lower-level specific design feature categories shown in the evaluation construction diagram, the morphological analysis method [31,77] was used to deconstruct the appearance features of the wickerwork lamps and to analyze the product design appearance hierarchy, to explore the correspondence between wickerwork lamps product form design and customer attractive factor. The morphological analysis method decomposed the wickerwork lamps into five independent variables: texture, material, colour, shape, and style, each of which was further subdivided into multiple design types. Based on the visual composition to depict the aesthetic features of 35 wickerwork lamp products, the corresponding feature hierarchy categories were obtained, and the specifical experimental results are shown in

Table 6. The form deconstruction of wickerwork lamp production.

Design Projects	Design Features
	Type 1	Type 2	Type 3	Type 4	Type 5
Pattern texture	Light transmittance	Irregular type	Density variation	Curved wave	Organization with regularity
Lamp material	Simple wickerwork	Wickerwork and glass	Wickerwork and stone	Wickerwork and wood	Wickerwork and paper
Lamp colour	Black and yellow	Orange and brown	coffee	Willow primary colour and other colours	Wicker primary colours
Lamp shape	Approximate bamboo shoot type	Approximate flower type	Cascaded type	Vase type	Approximate geometry type
Design style	Chinese style	Pastoral style	Simple style	European style

.

3.3. Using the Fuzzy QFD to Establish the Mapping Model

According to the experimental calculation results (Table 5), the most representative semantic words of wickerwork lamp products were obtained. Based on the degree of grey correlation calculation results, the top five Kansei words were selected in descending order, and the Kansei words were placed on the left side of the QFD as the user’s emotional needs, and the design features derived from the EGM evaluation construction diagram were placed in the upper layer of the QFD as engineering features. The fuzzy QFD was used to explore the association degree between the design elements and perceptual vocabulary.

First, we again invited 22 experts to form a focus group, and most of all the subjects were aged between 20 and 29 years old, accounting for 40.91%; 11 were male and 11 were female, accounting for 50% of the total, respectively. The main education level was a bachelor’s degree or above, accounting for 90.91%; the occupational distribution was more even, with most of them being designers, including 4 students, 10 designers, 6 teachers, and 2 corporate product managers, accounting for 18.18%, 45.45%, 27.27%, and 9.09% of the total number, respectively (

Table 7. The basic information of the 22 subjects.

Item	Content	Number of People	Percentage (%)
Gender	Male	11	50%
	Female	11	50%
Age	20–29	9	40.91%
	30–39	8	36.36%
	40–45	5	22.73%
Education	College	2	9.09%
	University	8	36.36%
	Postgraduate	12	54.55%
Occupation	Students	4	18.18%
	Teachers	6	27.27%
	Designers	10	45.45%
	Product Manager	2	9.09%
Design Experience	Three years	10	45.45%
	Five years	6	27.27%
	One year or less	6	27.27%

). The focus group members, through full communication and discussion, determined the strength of the connection between customer requirements and engineering features with the correlation matrix symbols _, △, ◯, ◎ [78], which were correspondingly converted to the corresponding triangular fuzzy numbers according to the strength of the correlation, as shown in

Table 8. The correlation symbols and converted fuzzy numbers in fuzzy QFD.

Symbols	Definition	Fuzzy Numbers
_	No relationship	(0, 0, 0)
△	Weak relationship	(1, 3, 5)
◯	Medium relationship	(3, 5, 7)
◎	Strong relationship	(5, 7, 9)

. The final fuzzy QFD is calculated by Equations (8)–(10), and the results are shown in

Table 9. Incidence matrix table of wickerwork lamp products in fuzzy QFD.

ECs		Patterns					Material					Colour					Shape					Design Style
CRs		1	2	3	4	5	1	2	3	4	5	1	2	3	4	5	1	2	3	4	5	1	2	3	4
Bright	0.208	◎	△	◯	△	△	◯	◎	△	△	◯	△	△	△	◎	◎	◯	◯	◎	◎	◎	◎	◎	◎	◎
Smooth	0.201	◯	◎	◎	◎	◎	◯	△	△	◯	◯	_	_	_	_	_	◯	◎	◎	◯	◯	△	◎	△	△
Simple	0.198	◯	◎	◯	△	◯	◎	◯	◯	◎	◯	◯	◯	◯	△	◎	◯	△	◯	◎	◎	△	◯	◎	△
Vibrant	0.198	◎	◯	◯	◎	△	◎	△	△	◯	◯	△	△	△	◯	◎	◎	◎	◯	◯	△	◯	◎	△	△
Ecological	0.195	_	_	_	_	_	◎	◯	◎	◎	◎	△	△	△	◯	◎	_	_	_	_	_	◯	◎	◎	△
Absolute weight		4.839	4.410	4.430	4.012	3.618	6.182	4.618	4.175	5.370	5.389	2.793	2.793	2.793	4.013	5.592	4.423	4.428	4.846	4.840	4.445	4.617	6.603	5.404	3.832
Relative weights %		0.045	0.041	0.041	0.037	0.033	0.057	0.043	0.038	0.050	0.050	0.026	0.026	0.026	0.037	0.052	0.041	0.041	0.045	0.045	0.041	0.043	0.061	0.050	0.035
Rank		9	16	13	19	21	2	10	17	6	5	22	23	24	18	3	15	14	7	8	12	11	1	4	20

ECs: Engineering Characteristics.

.

According to Table 9, the five design features with the highest degree are pattern 1 type (4.46%), material 1 type (5.70%), colour 5 type (5.16%), shape 3 type (4.47%), and design style 2 type (6.09%).

4. Discussion

To review previous studies, most of them focus on studying consumers’ emotions or perceptions of electronic products, office supplies and transportation, and few studies explore consumers’ behaviours and perceptions by focusing on the specific design features of cultural and creative products. The EGM of Miryoku Engineering, as an interview and analysis method, can logically organize users’ perceptions of things under fuzzy information sessions in consumer perception studies. Therefore, this study uses EGM to explore consumers’ perceived needs for attractiveness factors of wickerwork products, and analyze the logical relationships between them. The combination of GRA and Fuzzy QFD proposed in this study enables a quantitative weight ranking of the upper Kansei factors and constructs a nonlinear mapping function between upper emotional factors and lower specific conditions, which solves the problem that traditional Miryoku engineering can only provide qualitative relationships between customer preferences and design elements. Specifically, this study first performs the operational process of EGM for Miryoku engineering to obtain 14 upper-level abstract Kansei factors that express users’ appealing and emotional experiences on wickerwork lamp product design, and discovers 5 original design items and 36 specific evaluation items to derive specific design features of wickerwork lamp products. Obviously, these items are defined as the factors of the attractiveness of wickerwork luminaire products to consumers, and the conclusion suggests that these 36 factors should be fully considered in the product design process. To distinguish the importance of these emotional elements, GRA was used to analyze the importance of perceptual words. The top three sentiment words were found to be Bright (0.848), Smooth (0.820) and Simple (0.809). These results show that the most important design style feature of wickerwork lamp products is the bright Kansei need, which could remain the soul of the design. Secondly, the smooth lamp product design is easy to attract users’ attention and will bring graceful feelings to users, designers should give priority to this emotional preference. Furthermore, because of modern life, interior decoration style to pursue the effect of simplicity, in the design of lamps also further highlighting the style of simplicity, to fit the indoor environment. Finally, the importance of these Kansei words was introduced into the QFD model to propose new product design concepts for wickerwork lump and suggest the best design parameters for styling design solutions of wickerwork lamp products, the results obtained are the combination of multiple design forms, which is the combination of pattern 1 type, material 1 type, colour 5 type, shape 3 type, and design style 2 type. As a result, designers can focus on the attractive factors of customer preferences obtained from the research results and plan new product development strategies.

4.1. QFD and Other Intelligence Technology in Production Design Process

With the development of artificial intelligence techniques, such as BP neural networks, RST and GA have been applied to measure user fuzzy Kansei knowledge. However, the main difficulty of BP neural networks is that the modelling process requires human intervention to select control parameters such as the number of layers and the number of neurons per layer [79]. RST is an effective method for dealing with fuzzy data based on setting boundary regions, but it is limited to dealing with discrete variables in its calculations [13]. The genetic algorithm has been widely and successfully used in engineering optimisation problems, but they are more complex to compute and requires more parameters to be adjusted. Compared to traditional parametric statistics and neural network methods, QFD has the advantage of being non-linear, computationally simple, and can take the user’s emotional needs as a starting point, allowing for rapid discovery of competitive functional requirements and design parameters in a design, thus shortening the product development cycle and improving design efficiency.

Table 10. Comparison of fuzzy QFD and other mapping algorithms in new product design.

Algorithms	Calculation Time	Sample Size	Inferring Complexity	Fuzzy Expression	Loss Rate	Validity of Results
Fuzzy QFD	Low	Low	Low	Medium	N/A	High
BPNN	Medium	High	Medium	High	Low	High
ARM	Low	Low	Low	Low	N/A	Medium
DCNN	High	High	High	High	Low	High
GA	Medium	Medium	High	High	Low	High
QT-I	Low	Medium	Low	Low	High	Low
RST	Low	Low	Low	Medium	High	Low
SVR	Medium	Low	Medium	Medium	Low	High

shows how QFD compares with other methods in terms of Kansei mapping.

Moreover, in this study, we analyze each attraction factor of the wickerwork lamp product, in turn, to find out which design elements are effective for attraction factors. A nonlinear mapping model between design elements and emotions was established by applying a fuzzy QFD model. Based on the experimental results, pattern 1 type, material 1 type, colour 5 type, shape 3 type, and design style 2 type were the best parameters for the product. Therefore, these design elements are key points in the product development process based on attractive factors. Thence, in the design process of lamp production, the designer should emphasize the aesthetic characteristics of simple design, accompanied by exquisite details of shape and pattern expression, so that they make the lamp bright and simple, thus highlighting the perfect design taste.

4.2. Future Directions of This Study

With the rapid development of the global economy, lamp manufacturers are gradually focusing on the sales process rather than the development of the function of the lamp itself. Major manufacturers are marketing unique lamp products at various exhibitions, sales platforms and other promotional activities. To this end, this paper selects a representative sample for qualitative and quantitative analysis, evaluates the attractive features of the marketed wickerwork lamp product designs, and explores the quantitative relationship between customer abstract perceptual factors and specifical design elements. Moreover, the most innovative wickerwork lamp was designed on this basis. These parameters are the result of effective derivation based on fuzzy QFD, and, thus, have greater market value for improvement and upgrading and as a reference for the development of next-generation products.

This study could be further developed in the future. Firstly, this study only considers the use of traditional interviews to obtain demand, in the future, it should consider the combination of natural language and text mining to obtain customer attractiveness factors when customer demand is changing more rapidly. Secondly, the Chinese cultural and creative products should go global, and although the study sampled Chinese people as its subjects, in the future, the study should be expanded to include participants from other countries to obtain more comprehensive results. In addition, the fuzzy QFD results are based on known results from interviews, and only some classic design details have been presented, in the future, designers will need to be more imaginative and inspired to break away from traditional designs and create more unique products in a variety of forms and shapes based on advanced digital technology.

5. Conclusions

Economic globalization has accelerated the pace of companies implementing mass customization, and improving customer satisfaction has become a necessary means for companies to achieve competitive advantage, a good product design of wickerwork lamps can distinguish their products from many similar competitive products and significantly increase product sales. In fact, the user needs analysis plays a rather important role in the product design process, the designer needs to identify target users before the product conceptual design, analyze and summarise users’ emotional needs, to identify relatively important demand items. The main objective of this study is to develop an attractive product design for wickerwork lamps that combines GRA and fuzzy QFD on the evaluation structure chart platform established by EGM of Miryoku Engineering. Firstly, we use the EGM of Miryoku Engineering to obtain 14 representative upper-level Kansei factors, 5 original causes, and 36 lower-level specific design conditions based on a survey of 12 expert groups. Secondly, the GRA of artificial intelligence technology is used instead of the traditional frequency summation method to obtain the key Kansei factors using the calculation for the degree of grey correlation. Finally, the fuzzy QFD was used to establish the mapping relationships between the key Kansei factors and the corresponding design elements of the wickerwork lamp shapes, and to enhance the attractiveness of traditional wickerwork lamp designs.

This study has some limitations and needs continuous improvement in the future. First, although the hierarchy of expert opinions extracted by EGM is detailed and refined, it may be difficult to comprehensively record subjective emotions using only adjectives because each perceptual word cannot fully state the semantic space of the product. Second, the research data in this paper was selected at a fixed time period, and the interviews were static customer needs. However, in different time periods, customer demand will change accordingly, how to build a dynamic customer demand acquisition model to guide the product dynamic design will become the next research direction. Finally, this paper only considers the morphological changes of luminaire design, and future research should be extended to the size, structure and brand. Specifically, we can attempt to study a certain brand of products, it can help companies to develop new products effectively.