Garment mass customization methods for the cutting-related processes

Abstract

In the garment industry, upgrading from mass production to mass customization is imperative, due to ever-changing fashion trends and customers’ increasing personal demands. Based on pattern variations, several practical mass customization methods in terms of custom-fit and co-design are developed in this study for cutting-related processes (i.e. the sizing process and the cutting process). The increment of size number by generating additional sizes and the expansion of size capacity by setting multi-sized darts are two custom-fit methods, while material (fabric) variation by spreading “rainbow plies” and module variation by making a stepwise cutting are two co-design methods. Compared with existing mass customization methods mainly related to the garment design processes, the proposed methods emphasize the garment manufacturing processes in order to resolve the conflict between personalization and cost. A case study on the women’s basic straight skirt is conducted for the performance evaluation (about personalization and cost) of the proposed methods. The experimental results demonstrate the ability of these methods to significantly raise the personalization level with an acceptable price in garment customization. This study provides a reference for garment manufacturers to make proper production strategies, enabling garment mass customization.

Keywords

mass customization garment manufacturing cutting-related processes pattern custom-fit co-design

Craft production mainly targets individuals and economies of scope,¹ and mass production serves a large population and economies of scale.² The former depends on skilled craftsmen creating a perfect fit, allowing any wanted designs with a higher cost,² while the latter features a lower cost but a poorer fit and very limited variants.³ Mass customization is the production mode that integrates mass production and craft production,² first coined by Davis⁴ and further developed by Pine,⁵ which is able to both control cost and satisfy personal needs.¹ Mass customization implies a production process which satisfies the target population by providing customizations, like craft production, and at the same time features a high degree of automation, like mass production. With ever-changing fashion trends and increasing customer demand for personal preferences, mass customization is a promising strategy to support the garment industry,^3,6,7 and improves the personalization level, approaching craft production, and meanwhile controls the manufacturing cost, the production speed, and the product quality, approaching mass production.

There have already been some scientific reports on mass customization in the garment industry,^8–11 most of which deal with the problems in the design stage;^12–17 some others concern the manufacturing process.^18–21 It is the increase of complexity in the manufacturing processes rather than in the design processes which significantly influences the garment cost. As pointed out by Jiao et al.,²² customization increases the number of variants in production, and also decreases the number of items produced per variant, with significant impacts on the cost of the manufacturing process. The fabric cost, namely the material cost, occurs mainly in the cutting process and is related to the cutting-related processes (the sizing process and the cutting process). It is taken as a cutting-related cost.^23,24 The cutting-related costs make up a considerable part of the garment manufacturing cost, as the fabric cost is a major contributor to the total garment cost.^3,25 Thus, the cutting-related processes (the sizing process and the cutting process) bear the brunt of the higher cost caused by mass customization in the whole manufacturing process. However, the cutting-related processes have not received enough attention when considering garment mass customization methods.

The optimization of garment manufacturing systems toward mass customization has not been holistically tackled in industrial and academic practices. General strategies including concurrent engineering,¹⁸ postponement,²⁰ and modularity²⁶ are proposed and administrative systems are built,^19-21 but rarely are specific methods regarding certain manufacturing processes reported in the current literature. Herein, we propose several methods mainly considering the sizing process and the cutting process of garment manufacturing in regard to their performances for personalization and economy. Through the literature study, it is found that custom-fit^27-30 and co-design^31-32 are two main categories of garment personalization.^1,6 Most of the research concerns the design and marketing departments,^1,6,27,31,32 and rarely concerns the manufacturing department.^29,30 As we know, each garment is commonly composed of several patterns. The customization of the garment is indeed the customization of the associated patterns, which are the operation objects in the garment cutting-related processes. Thus, the methods proposed in our study are centered on pattern variations in the two above-mentioned aspects. As reported,^29,30,33 advanced technologies such as virtual reality, 3D body scanning, and computer-aided design/manufacturing are adopted to aid personalized pattern design and the automatic single cut for mass customization, which is not an economic approach for manufacturers. The more inclined to mass production the mass customization is, the more significant the economic benefit it brings. Thus, cost-effective expertise and methods of mass production should be considered, as batch manufacturing with sizing systems is suggested.^28,34 In the discussion of garment mass customization, most research addresses personalization instead of the underlying relationship of personalization and cost. Our research aims to propose some practical garment mass customization methods focusing on the manufacturing processes, more precisely the cutting-related processes, and based on the mass production industrial practice, to reveal and resolve the conflict between personalization and cost. Therefore, the overall strategy of the study is that the mass customization methods are developed from initial classical production practices in mass production to the utmost extent to effectively reduce the manufacturing cost, to help enterprises with the shift from mass production to mass customization. In addition, an analysis of the personalization levels and cutting-related costs when applying the proposed methods is demonstrated in detail for efficiency evaluation, and is a reference for apparel companies to make proper mass customization scenarios according to their own conditions.

In this study, we propose several new practical methods, based on pattern variations, in two different aspects, i.e. custom-fit and co-design, for upgrading the cutting-related processes (the sizing process and the cutting process) in the garment mass customization environment in the next section. The increment of size number by generating additional sizes and the expansion of size capacity by setting multi-sized darts are two custom-fit methods, while the material (fabric) variation by spreading “rainbow plies” and the module variation by making a stepwise cutting are two co-design methods. In addition, the definition of personalization levels, i.e. custom-fit level and co-design level, and the calculations of cutting-related costs are introduced. Following this we demonstrate the implementation of these methods in a case study of the women’s basic straight skirt, where the performance evaluation (regarding personalization and cost) of the proposed methods is made. The fourth section gives the experimental results, demonstrating the ability of these methods to significantly raise the personalization level with an acceptable price in garment customization. The final section concludes the paper and gives remarks for future work.

Methodology

In this section, four mass customization methods, two for custom-fit and two for co-design are demonstrated. Subsequently, the personalization level definition and the calculation of the cutting-related costs are illustrated.

Mass customization methods based on pattern variations

Four mass customization methods regarding custom-fit and co-design are developed. Two of them are related to pattern size, and the other two are related to pattern material and pattern shape, centered on the cutting-related processes in garment manufacturing. The details about these proposed methods are discussed in the remainder of this section.

Methods for custom-fit

As the fit is a fundamental need of users, a satisfactory fit in mass customization is of great importance.^27,28 Methods regarding custom-fit are associated with the pattern sizes, the main idea of which is to enhance the fit satisfaction of the targeted customers by adjusting the sizing system. It is difficult to reach a higher level customer satisfaction with fit due to the limited and outdated sizes of ready-to-wear garments.^35-37 It was indicated that 50 percent of women claim that they cannot find apparel that fits, and 50% of garments are returned because of problems with fit.³⁸ To generate appropriate sizes for the target population in mass customization, optimization methods are designed considering the enhancement of number or the capacity of size.³⁹

Increment of size number

Producing mass customization sizing systems (which are generally not proportionally graded) with an increased size number compared with the traditional mass production sizing systems has been addressed in the literature.²⁸ In this case, customers can be better served with more choices and the garment fit is improved. In considering whether to retain the original mass production sizes in the new sizing system, there are two main strategies: (1) retain the original mass production sizes and add additional sizes; and (2) completely abandon the existing sizes and use newly generated sizes. The former updating approach is the focus of the research displayed schematically in Figure 1(a), in which a mass customization sizing system is demonstrated, with additional sizes (gray rectangles) and original mass production sizes (black rectangles).

Figure 1.

Schematic figure of mass customization methods for custom-fit: (a) Addition of pattern size via additional sizes; (b) Expansion of pattern size via expanded sizes.

Expansion of size capacity

Some manufacturing technologies, like elastic waistband attachment⁴⁰ and garment belt utilization,⁴¹ which allow for small variations in fit, offer the potential of expanding the size capacity to improve satisfaction in mass customization. Namely, the size capacity can be also increased to enhance the fit satisfaction, by adjusting the pattern size through a structure processing; for instance, embedding some special structures with high flexibility, including darts, elastic materials, belts, and so forth. As depicted schematically in Figure 1(b), the original mass production sizes (black rectangles) are discarded and replaced by the expanded sizes (gray rectangles), which enables a larger percentage of the target population to be served. However, this method can only bring slight and limited effects on the expansion of size capacity because of its small change range and narrow applicability. Therefore, in practical production, the two strategies of increment in size number and capacity are recommended to be used simultaneously.

Methods for co-design

Apart from the increasing demand for garment fit, customer participation in the design such as garment style, garment detail, fabric type, fabric color, etc., called co-design (collaborative design), is highly desired by customers in the garment industry.³¹ The concept of co-design (participatory design), where the customers provide their co-designs before production, is hard to incorporate into traditional mass production. Traditional mass production is to some extent isolated from co-design, giving this opportunity to mass customization. Thus, we propose two mass customization methods in terms of co-design, regarding the pattern material and the pattern shape.

Fabric variation

“Rainbow spreading” is a typical strategy in garment production,⁴² and it is common to make lays with materials in different colors vertically. The strategy can be extended to endow the same kind of article with greater variability (e.g. color, composition, or texture). However, we need to be aware that not every garment needs that much diversity, so it may bring about overproduction. To avoid this limitation, if we allow multiple fabrics to compose one-ply horizontally during the spreading operation (rainbow ply), the variability can be applied to the different articles. To explain in more detail, Figure 2(a) shows the feasibility of the strategy of combining the “rainbow spreading” with “rainbow ply.” Patterns of the same article are grouped into one section on the marker, so that the single-ply can be composed of different fabric pieces for cutting out. From the figure, we can conclude that the overproduction can be effectively suppressed. Meanwhile, the personal needs of pattern material can be well satisfied by this method. In the future of mass customization, with the fast development of computerization and automation in industry, it is highly likely that the horizontal strategy “rainbow ply” will be widely used, contributing to the progress of mass customization.

Figure 2.

Mass customization methods for co-design: (a) Variation of pattern material via multi-fabric spreading; (b) Variation of pattern shape via stepwise cutting.

Module variation

In addition to fabric variation, the personalization in module variation is also feasible with respect to the pattern shape. The variation of the garment module contains a slight modification of the main module or a type alteration of specific moduli (e.g. collar, sleeve, pocket, waistband). This can be realized through conducting lean cutting with second cuts (first cut in terms of the common outline, then the second cut for cutting out all the variants) or even making extra markers for a wide module variety. When adopting a stepwise cutting, the key point is to generate the markers for stepwise cutting with the common outline of the related variants for the first cuts, see Figure 2(b).

Definition of personalization level

The personalization levels and the cutting-related costs are tightly correlated, because the variety and depth of customizable options result in manufacturing complexity, thereby affecting the costs. For the two mass customization aspects mentioned above, we use custom-fit level and co-design level to specifically represent the personalization level, namely, the level of mass customization.

Custom-fit level

Fit is a complex concept having different definitions as reported in the literature, and the most common criterion noted is the aggregate loss.^43-45 The aggregate loss gives an objective assessment measuring the average distance between the body dimensions of samples and the dimensions of assigned sizes. However, this criterion may not have a precise or even proper presentation of fit. It is noted that it is very hard to normalize the effect of each key dimension on fit to customers, and customers’ subjective feelings of fit are not in a simple linear correlation with distance. Therefore, we propose a new criterion of fit, the custom-fit level, which contains a part of the objective assessment, which measures the average distance between the body dimensions and the dimensions of assigned sizes, combined with a subjective assessment.

Figure 3 demonstrates an example of the custom-fit level definition with three sizes which are presented in different colors. There are five custom-fit levels weighing 0, 2, 5, 8, and 10. D (D_p and D_s) refers to the primary or secondary dimension of the size, and IntD (IntD_p or IntD_s) refers to the corresponding interval value.⁴⁶ The two-dimensional space of the key dimensions is divided into sections by sizes differentiated with specific colors. Samples from the target population whose body dimensions fall into a certain section are assigned with the corresponding size. Each section is further divided into several subsections differentiated with specific shades. Each subsection is marked with an exact weight, reflecting the corresponding custom-fit level, which is determined by the objective difference between body dimensions of the sample and dimensions of the assigned size, simultaneously altered by customers’ subjective evaluation.

Figure 3.

An example of custom-fit level definition.

Co-design level

Compared with the custom-fit level, there is hardly any literature concerning the definition of the co-design level. In detail, co-design is associated with the type and complexity of the requested design. Herein, the definition of the co-design level is based on the selected co-design points x, and related to the difficulty and cost of the individual co-design during manufacturing, which can be labeled as complexity coefficients Cx∈[0, 1]. As illustrated in equation (1), the co-design level is evaluated by the sum value of the complexity coefficients of all the participating co-design points, as semi-quantitative characterization.

\begin{matrix} c - design level = \sum C_{x} \times E_{x} \\ E_{X} = \begin{matrix} 1 if x is selected \\ 0 if x is not selected \end{matrix} \end{matrix}

(1)

For instance, for different co-design points, materials (complexity coefficient C_m = α) and pockets like patch pocket, slant pocket, and welt pocket (complexity coefficient C_PP = β, C_PS = γ, and C_PW = δ), if the consumer selects a certain material and patch pockets, the co-design level is α + β, or if they select a certain material and welt pockets, the co-design level is α + δ.

Cutting-related costs

The majority of the garment manufacturing cost occurs in the garment cutting process.^47,48 Costs relative to the garment cutting process^23,24,49,50 can be divided into two groups: one group is caused by consumption of materials (i.e. fabrics and markers), and the other is induced by operations (i.e. spreading, cutting, and sorting and bundling). Labor and equipment costs are included in the calculation of operation costs, and it is worth noting that herein the power cost is classified into the equipment cost to simplify variables. The following equations provide each relationship of key factor(s) and the related cost. In the case study, the exact value of the cost depends on some constants related to production, and presents a certain proportional relationship with certain key factors. Consequently, the use of the values of key factors to represent the corresponding cost makes sense, so as to adapt to the variation of the constants in different production scenarios.

Fabric cost (C_f):

C_{f} = \sum P_{f} \times L_{f}

(2)

where Cf represents the fabric cost, P_f represents the price per unit length of fabric, and L_f the used length. The fabric cost depends on the length of fabrics required for the cutting only.

Marker cost (C_m):

C_{m} = \sum P_{m} \times L_{m} \times α_{m}

(3)

where Cm represents the marker cost, P_m represents the price per unit length of marker, L_m the used marker length, and α_m is the complexity of marker construction. The marker cost depends on the used marker length and the complexity of marker making. Automation and paperless technologies will help to obtain simplicity and low cost in marker making.

Spreading cost (C_s):

\begin{matrix} C_{s} = (P_{o} + P_{sm}) \times T_{s} \\ T_{s} = \sum L_{f} / V_{s} + N_{p} \times T_{sp} \end{matrix}

(4)

where Cs represents the spreading cost, P_o represents the cost per unit time of the operator, P_sm the cost per unit time of spreading machine, T_s the spreading time, V_s spreading speed, N_p ply number, and T_sp the time for per pause during spreading. The spreading cost depends on several elements: the operator, the spreading machine, plies in the lays.

Cutting cost (C_c):

\begin{matrix} C_{c} = (P_{o} + P_{cm}) \times T_{c} \\ T_{c} = \sum L_{c} / V_{c} + T_{cp} \end{matrix}

(5)

where Cc represents the cutting cost, P_cm represents the cost per unit time of cutting machine, T_c the cutting time, V_c the cutting speed, L_c the cutting length, and T_cp the time for pause during cutting. The cutting cost depends on several elements: the operator, the cutting machine, markers of lays, and lays.

Sorting and bundling cost (C_sb):

\begin{matrix} C_{sb} = P_{o} \times S_{p} \times T_{g} \times (1 + α_{sb}) \\ α = N_{ps} / S_{p} \end{matrix}

(6)

where Csb represents the sorting and bundling cost, S_p represents the constant production size (total garment number), T_g the time of sorting and bundling operations of each garment article, and 1 + α_sb the degree of difficulty, where α_sb means the sorting and bundling complexity, which is determined by N_ps the pattern set number and S_p. The sorting and bundling cost depend on the production size, and the sorting and bundling complexity. In general, it is comparatively much lower than the other cutting-related costs.

To comprehensively compare these cost changes, an in-depth analysis of the key factors for each cost is discussed through a case study in the next section.

Case study

To evaluate the performances (i.e. personalization and cost) of the above-mentioned mass customization methods, we present a case study containing six experiments about the production of a women’s basic straight skirt, updated from mass production to mass customization for 451 customers.

Design of experiments

Details of the case study in terms of the objects, contents, and constraints are described in this subsection.

Objects

Mass customization of a basic straight skirt is designed for a target population of 451.

Product Normally, women’s skirts have relatively limited and simple patterns with fewer variants than other types of lower body garments, such as pants and upper body garments. More specifically, the basic straight skirt is commonly used on official occasions, conventionally with the need of large output and further customization. Therefore, the basic straight skirt can be a typical and concise example. Herein, a women’s basic straight skirt, which is represented as a commonly used clinging garment, is selected in our study, and the sketch is shown in Figure 4. The corresponding primary and secondary dimensions are hip girth and waist girth. The experimental object in this work is a straight skirt with simple structures; nevertheless, the methods mentioned in this research are designed with consideration for general use, so they are suitable for the customization of other types of garments as well. The implementations with other types of garments will be completed in future work.

Figure 4.

Prototype of a women’s basic straight skirt.

Anthropometry data Sets of real body-dimension data from the target population of 451 French women, between the age of 25 and 40, have been collected by 3D scanning and prepared as data sources in the experiments.

We designed experiments for both mass production and mass customization with the proposed four methods to evaluate the performances and to make comparisons as well. One of these experiments is conducted in the production mode of mass production (MP), and the five other experiments demonstrate the corresponding production upgrade methods toward mass customization (MC(F_G), MC(F_L), MC(D_FC), MC(D_PT), and MC(D_SL)), shown in Figure 5.

Figure 5.

Flowchart of experiment implementation with mass customization methods.

First, the experiment MP is conducted with an existing size chart of this skirt type from a garment company in the real market on the s.Oliver website (https://www.soliver.eu/sizetables/size-table-women-GENERAL_SIZETABLE_WOMEN.html). Afterward, the other five experiments are performed for each evolutionary process of mass customization, of which two are for custom-fit and three for co-design. The main target of the case study is to implement the proposed four mass customization methods and evaluate the corresponding performances, so the focus is on mass customization experiments, which are further described and discussed. The mass production experiment is the basic line for comparison.

For the custom-fit part, the first mass customization experiment for custom-fit is defined as MC(F_G), and we supplement a set of additional sizes by a shift of the existing sizes, illustrated in the “General fit improvement” section, with size number doubled from existing mass production sizes. Compared with creating a novel set of sizes for mass customization, the strategy has the advantages of simplicity, reflected in a slight change in pattern development and product manufacturing. In addition, it exhibits the benefit of sustainability, namely, an increase of pattern utilization (dart variations are based on existing patterns, no new patterns are created), and flexibility in the switch between mass production and mass customization. The second experiment for custom-fit is an optimization of the experiment MC(F_G) by dart modification for mass customization, defined as MC(F_L), where we apply darts with three sizes, namely small, medium, and large, to expand the capacity of existing sizes, illustrated in the “Local fit improvement” section.

Regarding co-design, customizations are conducted on the pattern material or the pattern shape. Pattern shape variation contains variations of the main moduli, i.e. front and back pieces, and specific small moduli, e.g. waistband, pocket, and vent. We select fabric type for the pattern material variation, pocket type for the pattern shape variation with small module, and garment length for the pattern shape variation with main module—three typical design components for garment customization according to the concept of increasing the options for the customer with a minimum of production adjustments,^6,51 and especially for the garment type of skirt.^32,52 In the study, three detailed experiments are designed, with the variations of fabric color, pocket type, and skirt length evolved from the experiment MC(F_L) and defined as MC(D_FC), MC(D_PT), and MC(D_SL), respectively. The related progressive relationship can be observed in Figure 5. In addition, all the details related to the experiments are summarized in Table 1.

Table 1.

Details of experiments applied with mass customization methods

Experiment		Description	Method
Mass production	MP	no customization	use existing size chart from a real market
Mass customization	MC(F_G)	general fit improvement	generate additional size, retaining sizes in experiment MP
	MC(F_L)	local fit improvement	apply multi-sized darts to sizes in experiment MC(F_G)
	MC(D_FC)	design of fabric color	allow multi-fabric spreading with sizes in experiment MC(F_L)
	MC(D_PT)	design of pocket type	distribute pocket patterns in markers with sizes in experiment MC(F_L)
	MC(D_SL)	design of skirt length	use superimposed pattern outline for maker making with sizes in experiment MC(F_L)

To further conduct the experiment, we simulate one order of customized products for 451 people of the database. In the custom-fit experiments, the most suitable size in the defined sizing chart according to the body dimensions of the consumer is assigned. In the co-design experiments, an equal division of design options is set. As shown in Figure 6, random distributions of single-point co-design selection in experiments MC(D_FC), MC(D_PT), and MC(D_SL) are made for the 451 samples. In this case, the assignment of samples to each co-design element is mostly uniform. In detail, for fabric, there are 34% of the consumers for fabric FB001, 31% for fabric FB002, and 35% for fabric FB003; for pocket type, 37%, 32%, and 31% for patch pocket, slash pocket, and without pocket, respectively. For skirt length, the percentages of the three skirt lengths (short, original, and long) are almost identical, approaching 33%.

Figure 6.

Random distributions of co-design selection in MC(D_FC), MC(D_PT) and MC(D_SL).

Constraints

The production constraints which have been considered in this case study are listed below:

To realize the concept of “manufacturing on-demand,” customized products are produced for individuals in mass customization, and excess products are valueless. To keep the constraint consistent, no excess production is considered in the experiments, or for the MP experiment.

To produce the exact quantities on demand, lay plans are made with step lays, where the layout of the articles on one marker is sequenced in ascending or descending order of ply number.

The cutting scope is set as follows:

Determined by specific cutting equipment in the industrial production, the maximum length of fabric is 3 m, and the maximum height of lay is 30 mm. The fabric thickness is set as 1 mm. We assume that each layer of ply is tightly stacked, so the maximum number of ply is 30. The maximum garment sizes contained in the marker is three, and its maximal length is not allowed to exceed the maximum length of fabric (3 m). Thus, the maximum number of sizes assigned in each lay is set as three.

In industrial practice, allocating mixed combinations of small sizes and large sizes in the same lays is strongly preferred due to an even multi-size distribution to markers, which would contribute to the marker efficiency and finally benefit the reduction of fabric cost.⁴⁸ Herein, the size matching strategy is adopted.

Implementation of experiments

Mass customization methods, i.e. generating additional sizes, embedding various sizes of darts, spreading rainbow plies, and operating second cuts are applied in experiments MC(F_G), MC(F_L), MC(D_FC), MC(D_PT), and MC(D_SL) in the case study on a women’s basic straight skirt to cater for customers’ personal demands.

General fit improvement (MC(F_G))

The experiments make up a series of experiments in which the latter is updated from the former, illustrated in Figure 5. Therefore, in the experiment about size number increment MC(F_G), the new sizes are designed with enough interval to the original sizes instead of being bound, for the flexibility of dart variation of the following experiments. Afterward, in the experiment about the expansion of size capacity MC(F_L), multi-sized darts will fill up the reserved interval to make them interconnected.

The first step of fit customization is to generate a set of additional sizes through a shift of the existing mass production sizes in the mass customization experiment MC(F_G), as shown in Figure 7. For the existing mass production sizes, the minimum interval of D_s is defined as Min (ITV). According to the relative positions of the regression curve line (blue line) and original sizes (gray rectangles), additional sizes are determined by an upward shift of original mass production sizes with the distance of 3 times of Min (ITV), leaving the blank areas of 2*Min(ITV) for the dart setting of the next experiment. For instance, by shifting the original medium size M, its primary dimension, and secondary dimension (MD_p, MD_s), the corresponding additional size is generated as M* (MD_p, 3*Min(ITV) + MD_s).

Figure 7.

Size distribution in mass customization by adding sizes (MC(F_G)).

From the area enclosed by a solid gray line to the area enclosed by a red dotted line, it is indicated that a larger portion of the target population is accommodated after the application of additional sizes (red rectangles). By adding size numbers to engulf more points (samples), we can conclude that the coverage range can be expanded and global garment fit can be significantly improved.

Local fit improvement (MC(F_L))

After realizing a general optimized coverage of the customer body dimensions, the second step of fit customization is to extend the interval of waist girth via darts in multiple sizes in the experiment MC(F_L). We set the darts with the consideration of the shape stabilization of the garment and the maximal coverage range at the same time. First, dart sizes should be appropriate and not produce a disturbing effect on the garment shape; second, we take full advantage of multi-sized dart to bring a maximal coverage range. Therefore, based on the experiment MC(F_G) shown in Figure 7, as the skirt has four darts, in this case, Min (ITV)/4 is set as the value of the interval of the dart. We set the new dart sizes as follows:

\begin{matrix} D_{size} + = D_{size} + Min (ITV) / 4 \\ D_{size} - = D_{size} - Min (ITV) / 4 \end{matrix}

(7)

where D_size represents the original/medium dart size, D_size + the large dart size, and D_size- the small dart size.

To visualize the upgrading of the system via the second experiment concerning the custom-fit, Figure 8 exhibits the size distribution of mass customization by multi-sized darts (MC(F_L)). In total, there are three dart sizes: small, medium (original), and large dart. From the figure, it can be concluded that the application of multi-sized darts combined with the increment of size numbers results in further enhancement of the accommodation rate, and intuitively a higher proportion of the population with a perfect fit.

Figure 8.

Size distribution of mass customization by multi-sized darts (MC(F_L)).

Design of fabric color (MC(D_FC))

To provide a customization of the pattern material for the skirt, multi-fabric spreading is applied to support the fabric variation with rainbow plies. The principle of this method is trying to keep the same fabric in one-ply, and otherwise permit rainbow plies containing multiple fabrics, which can be observed in Figure 2(a). Shown in Table 2 is one of the spreads within the lay planning result of the experiment MC(D_FC). It is noticed that Ply 13–14 are different from other plies, which contain two different types of fabrics FB001 and FB002 for two different sizes XXXL and XXL*, respectively. The arrangement is exactly realized by the “rainbow ply” strategy.

Table 2.

One spread contained in the lay planning result of MC(D_FC)

Ply NO.	Fabric	Size
Ply NO.	Fabric	XXS*	XXXL	XXL*
1–2	FB001	2	2	2
3–4	FB002	2	2	2
5	FB003	1	1	1
6–9	FB001	–	4	4
10–12	FB002	–	3	3
13–14	FB002	–	2	0
13–14	FB001	–	0	2
15–18	FB003	–	4	4
19–20	FB001	–	–	2
21–25	FB003	–	–	5
Number of plies		5	18	25

Design of pocket type (MC(D_PT))

For the straight skirt used in this study, the main moduli comprise the front piece and the back piece, while specific moduli include the waistband, the pocket, the vent, and so forth. Experiments regarding co-designs of a main module (MC(D_SL)) and a specific module (MC(D_PT)) are carried out through the variation of skirt length and pocket type, respectively, in the following.

For the co-design of pocket type (MC(D_PT)), attempts are made to occupy the spare room of lays with the small patterns of pockets first for a maximal cost-saving. Considering the material shade problem, the bundling problem, a separate marker is seldom made for small pattern pieces, such as collars, pockets, or cuffs.⁴⁸ Advanced fabric dying technology improves the color consistency along the fabric so that patterns of the same article can be placed separately.^53,54 With the development of item tracking technology,^55,56 small patterns can be well arranged during their transportation process. It can be inferred that this will no longer be a limiting condition for the garment production with a series of development. So far, there are few works in the literature that demonstrate the efficiency of marker making if a separate marker is permitted. Therefore, the following discussion is based on this assumption.

The main strategy is to make the groupage of the small patterns form a close arrangement. It is worth noting that pattern set sequences should be considered based on pocket types for second cuts of the front piece. Arrangement of the appropriate locations of pocket patterns on the existing markers (the areas which are highlighted with dotted lines in Figure 9) is a challenging job. If no sufficient place is left in the existing lays for the groupage of patterns, an extra marker should be adapted for the rest of the patterns.

Figure 9.

Layouts of pocket patterns on an existing marker for MC(D_PT).

Design of skirt length (MC(D_SL))

For the co-design of skirt length, the purpose of lay planning is to figure out the differentiated edge of various patterns for second cuts that are used in marker making, shown in Figure 2(b). As shown in Figure 10, a marker is made with a superimposed outline of patterns that are adapted for three different skirt lengths. For instance, the areas marked by a dotted line, a gray shadow, and a solid line represent patterns of the same size 3XL* with skirt lengths of 50 cm, 60 cm, and 65 cm, respectively. The garment articles in lays are sequenced by skirt length, so that the operations of second cuts are conducted accordingly.

Figure 10.

Marker with superimposed outline of patterns for MC(D_SL).

Results and discussion

As mass customization methods have been demonstrated in the experimental part as mentioned above, the corresponding results regarding personalization levels and cutting-related costs are demonstrated and discussed in detail in this section.

Results on personalization level

Distributions of custom-fit levels and co-design levels of the experiments are displayed in the following to evaluate their performances of custom-fit and co-design.

Results of custom-fit The distributions of custom-fit levels in the experiments are displayed in Figure 11. It is shown in Figure 11(a) that both experiments MC(F_G) and MC(F_L) perform well in the improvement of garment fit. As a result, the proportion of low custom-fit levels (i.e. unwearable and moderate) is gradually reduced and almost eliminated, and the higher levels gradually dominate. In detail, the unwearable rate evolves from 26% to 15% in the experiment MC(F_G) and from 15% to 10% in the experiment MC(F_L). In addition, the moderate rate is changed from 25% to 4% in the experiment MC(F_G) and from 4% to 1% in the experiment MC(F_L). In contrast, mass customization methods significantly increase the fraction of the target population with custom-fit levels of good and perfect, for which Figure 11(b) shows the trend. It can be observed that a significant increase from 48% to 81% is achieved in the experiment MC(F_G), and subsequently, further optimization is obtained via multi-sized darts in the experiment MC(F_L), toward 89%. The global fit improvement experiment MC(F_G) plays the role of dramatically decreasing the rate of low custom-fit levels, and the local fit improvement experiment MC(F_L) offers a further optimization to make sure that the target population has a dominant fraction in a perfect fit (83%). In summary, an integration of additional sizes and multi-sized darts efficiently contributes to a global and high custom-fit level.

Results of co-design The following experiments are concerned about the co-design part, based on the completion of the aforementioned custom-fit experiments. Compared with the serial design of the two experiments of custom-fit, the co-design experiments are conducted in parallel. One co-design point is selected among the three: fabric color, pocket type, and skirt length. The difficulties in realizing these designs in manufacturing are hard to precisely normalize. Thus, the corresponding co-design levels of each co-design point for all the population are assumed as the same in this study, not avoiding making the semi-quantitative comparison.

Figure 11.

Custom-fit level distributions in experiments MP, MC(F_G), and MC(F_L).

Results on cutting-related cost

The five cutting-related costs (fabric, marker, operations of cutting, spreading, and sorting and bundling) are not directly calculated but are represented by corresponding main factors. Based on the earlier “Cutting-related costs” section, it can be concluded that the key factors have a strong positive correlation with the cutting-related costs. Therefore, the fabric length stands for the fabric cost (C_f), marker length for marker cost (C_m), spreading length and ply number for spreading operation cost (C_s), cutting length for cutting operation cost (C_c), degree of difficulty for sorting and bundling cost (C_sb).

Figure 12 shows cutting-related costs in experiments under different mass customization circumstances where the costs in the experiment MP are set 100% for comparison. The material-related costs are composed of fabric cost and marker cost. There is almost no change in the fabric length, indicating that the fabric cost is not altered significantly. Compared with the fabric cost, the marker cost is significantly impacted by the incorporation of the mass customization. The increase is mainly reflected in the lift in the marker length, approaching over 120%. In addition, it is indicated that if the material cost is relatively stable once the first experiment of mass customization upgrade is carried out.

Figure 12.

Cutting-related costs in different experiments.

The operation-related costs include three parts: the spreading cost, cutting cost, and sorting and bundling cost. The spreading cost is determined by the fabric length as well as the number of plies. There is almost no change in fabric length, while the number of plies is increased to some extent over 110%, mainly due to the application of “rainbow plies,” leading to more time in the spreading pause. As a result, the benefit of the fabric variability will bring about a slight increase in the spreading cost. Similar to the material cost, the spreading cost retains the same standard once the first mass customization experiment is conducted. However, the cost of cutting and cost of sorting and bundling exhibit different trends, of which the related cost is gradually enhanced. In detail, the two main steps, custom-fit and co-design, have a distinct influence upon the related cost. Regarding the cutting process, the custom-fit experiment and co-design of fabric color are on the same level of influence, and afterward, the co-design of pocket type and skirt length will introduce more difficulty into the cutting process, of which the cost is over 160%. Regarding the sorting and bundling cost, the custom-fit experiment has no significant impact on the degree of difficulty, while the co-design has a visible influence which elevates the related cost. Generally speaking, the spreading cost mainly originates from the material usage, while the other two operation costs mainly rely on the complexity of the garment, which corresponds with the mass customization updating.

Relationship between personalization levels and costs

Based on Figure 12, the cost growth ratio (cost growth of each step divided by the cost of the previous step) in each step is summarized in Table 3. To clarify the important parameters determining the total cost, the growth ratio which is over 5% is listed in the flowchart of the upgrading process in Figure 13. In this section, the relationship between personalization levels (custom-fit, co-design) and costs will be discussed.

Figure 13.

Flowchart of upgrading experiments with cost growth ratios.

Table 3.

Comparison between production modes by costs

Cost	Corresponding factor		Growth ratio (%)
Cost	Corresponding factor		F_G-MP	F_L-F_G	D_FC-F_L	D_PT-F_L	D_SL- F_L
Fabric	fabric length	a	0	–1	2	1	3
Marker	marker length	b	29	–2	0	2	2
Spreading	spreading length	c	0	–1	2	1	3
Spreading	number of plies	d	13	1	4	0	0
Cutting	cutting length	e	33	0	1	22	22
Sorting & bundling	degree of difficulty	f	2	7	19	19	19

In the aspect of custom-fit, additional sizes in the experiment MC(F_G) expand the coverage range and improve the global garment fit. Due to the doubling of size numbers, the cost increment lies in operation cost in cutting (33%) followed by spreading cost (13%), as well as material cost in marker (29%). In the following experiments, most of the cost growth occurred in the operation part rather than the material part. Using multi-sized darts in the experiment MC(F_L) is more powerful in local improvement of custom-fit level, reflected in a transition of custom-fit level from good to perfect. Regarding the cost in this experiment compared with the experiment MC(F_G), the cost is not altered significantly and the extra cost mainly exists in the sorting and bundling part, at 7%. This proves that the two methods concerning custom-fit improvement will cause varied cost growth in different parts and extent. Briefly, the MC(F_G) method significantly ameliorates the custom-fit level of garments, but simultaneously brings about the issue of cost growth in material and operation. In contrast, the MC(F_L) provides a further improvement in custom-fit level without significant cost growth. Additional sizes with a support of multi-sized darts lead to a better performance for the balance of fit and cost.

Regarding co-design, all the co-design experiments cause a 19% cost growth in sorting and bundling. This can be ascribed to the increment of the complexity accompanied by the increase of the co-design level. In addition, for the experiments MC(D_PT) and MC(D_SL), there is a 22% cost growth in the cutting operation. This originates from the second cuts for the complex objects. Furthermore, some extra markers in the experiment MC(D_PT) are also needed for holding the pocket patterns, which also raises the cost of the cutting operation. Compared with the two experiments, the experiment MC(D_FC) concerning the fabric color does not have an obvious effect on the cutting operation cost. Similarly, the increment of the co-design level also brings about some cost growth, mainly reflected in the operation part. Different strategies for co-design improvement have different impacts on cost as well.

In summary, an increase of custom-fit and co-design can be achieved by a well-arranged multi-fabric spreading at a reasonable extra cost, which is less than 33% for each item. Based on the inference, the garment company can make a better decision to choose the targeted upgrading route, in order to control the cost to meet the demanded levels of custom-fit and co-design. The cost growths in mass customization experiments mainly lie in the operation part. However, with the development of automation, the proportion of the operation cost to that of material will be gradually decreased in the future, which is beneficial for the cost control of mass customization.

This provides strong evidence that in the garment manufacturing domain, mass customization is very promising as a tool for garment manufacturers to balance the challenge between personalization and cost, in the fierce market competition resulting from customers’ increasing demand for customized products.

Conclusion

An upgrading solution for garment mass customization was demonstrated in this study in the context of cutting-related processes (the sizing process and the cutting process). It is mainly aimed at custom-fit and co-design customization, and a case study on women’s basic straight skirt is selected for personalization level and cost evaluation. The results prove that it effectively makes the tradeoff between personalization and cost.

Both custom-fit methods behave well with controllable extra costs. Supplementary sizes and multi-sized darts improve the custom-fit level globally and locally, respectively. The cost growth differs with the two methods, which are recommended to be simultaneously utilized.

Regarding co-design optimization, co-design of the material (fabric) “rainbow plies” has no obvious increase in the cutting-related cost, while that of the shapes (pocket type and skirt length) brings about further lift due to second cuts or even extra markers. It is interesting that the cost growth does not lie in the fabric cost, but the marker and operation-related costs. It can be expected that with the aid of highly automated devices and intelligent computing technologies, mass customization methods, especially the co-design part, can provide full economic advantages in the future.

This is a pioneering work of developing garment mass customization methods concerning the manufacturing process. The customization levels are demonstrated and the source of the extra cost is calculated in detailed items. It helps enterprises to conduct precise customization expectation and cost control, and finally make a proper production strategy to accomplish the upgrading task of garment mass customization.

Potential future work includes: Other typical garment types (e.g. shirt, jeans) can be studied; the performance evaluation criteria need to be further developed; and more advanced computer-aided technologies (e.g. computer algorithms) can be introduced.

Footnotes

Acknowledgements

Thanks to the Chinese Scholarship Council (CSC) for the financial support to this research. Also, Doctor Xiang Yan for fruitful brainstorming, and Professor Christopher Fuhrman for checking of English writing.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship and/or publication of this article: The financial support is provided by the China Scholarship Council (CSC), File No. 201606920026.

ORCID iD

Yanni Xu

References

Yang

Kincade

Chen-Yu

. Types of apparel mass customization and levels of modularity and variety: Application of the theory of inventive problem solving. Clothing Textiles Res J 2015; 33(3): 199–212.

Dong

Jia

, et al. Implementing mass customization in garment industry. Syst Eng Proced 2012; 3: 372–380.

Nayak

Padhye

Wang

, et al. The role of mass customisation in the apparel industry. Int J Fashion Des Technol Educ 2015; 8(2): 162–172.

Davis

. Future perfect 1987; Vol 106 Reading, MA: Addison - Wesley.

Pine BJ. Mass customization: The new frontier in business competition. Boston: Harvard business school press; 1993.

Anderson-Connell

Ulrich

Brannon

. A consumer-driven model for mass customization in the apparel market. J Fashion Market Manage 2002; Sep 1.

Fogliatto

Da Silveira

Borenstein

. The mass customization decade: An updated review of the literature. Int J Prod Econ 2012; 138(1): 14–25.

Lee

Kunz

Fiore

, et al. Acceptance of mass customization of apparel: Merchandising issues associated with preference for product, process, and place. Clothing Textiles Res J 2002; 20(3): 138–146.

Fiore

Lee

Kunz

. Individual differences, motivations, and willingness to use a mass customization option for fashion products. Eur J Market 2004; Jul 1.

10.

Alptekinoğlu

Corbett

. Mass customization vs. mass production: Variety and price competition. Manuf Serv Oper Manag 2008; 10(2): 204–217.

11.

Senanayake

Little

. Mass customization: points and extent of apparel customization. J Fashion Market Manag 2010; May 11.

12.

Ulrich

Anderson-Connell

. Consumer co-design of apparel for mass customization. J Fashion Market Manag 2003; Dec 1.

13.

Dai

Han

, et al. An interactive web system for integrated three-dimensional customization. Comput Ind 2006; 57(8–9): 827–837.

14.

Lee

Park

. Online service personalization for apparel shopping. J Retailing Consum Serv 2009; 16(2): 83–91.

15.

Satam

Liu

Lee

. Intelligent design systems for apparel mass customization. J Textil Inst 2011; 102(4): 353–365.

16.

Vogiatzis

Pierrakos

Paliouras

, et al. Expert and community based style advice. Expert Syst Appl 2012; 39(12): 10647–10655.

17.

Landon

Segonds

, et al. A decision support model in mass customization. Comput Ind Eng 2017; 114: 11–21.

18.

Kincade

Regan

Gibson

. Concurrent engineering for product development in mass customization for the apparel industry. Int J Oper Prod Manag 2007; May 29.

19.

Wang

Chen

, et al. The development of an intelligent system for customized clothing making. Expert Syst Appl 2010; 37(1): 799–803.

20.

Qin

. On capacity allocation model of partial postponement strategy. Proced Eng 2011; 15: 4342–4346.

21.

Watcharapanyawong

Sirisoponsilp

Sophatsathit

. A model of mass customization for engineering production system development in textile and apparel industries in Thailand. Syst Eng Proced 2011; 2: 382–397.

22.

Jiao

Zhang

Pokharel

. Process platform planning for variety coordination from design to production in mass customization manufacturing. IEEE Trans Eng Manag 2007; 54(1): 112–129.

23.

Jacobs-Blecha

Ammons

Schutte

, et al. Cut order planning for apparel manufacturing. IEEE Trans 1997; 30(1): 79–90.

24.

Wong

Leung

. Genetic optimization of fabric utilization in apparel manufacturing. Int J Prod Econ 2008; 114(1): 376–387.

25.

Naveed

Hussain

Zhong

. Reducing fabric wastage through image projected virtual marker (IPVM). Textil Res J 2018; 88(14): 1571–1580.

26.

Wang

Chen

Zhao

, et al. Modularity in building mass customization capability: The mediating effects of customization knowledge utilization and business process improvement. Technovation 2014; 34(11): 678–687.

27.

Ding

, et al. A hybrid neural network and immune algorithm approach for fit garment design. Textil Res J 2009; 79(14): 1319–1330.

28.

Mpampa

Azariadis

Sapidis

. A new methodology for the development of sizing systems for the mass customization of garments. Int J Clothing Sci Technol 2010; Mar 2.

29.

Tao

Chen

Zeng

, et al. A customized garment collaborative design process by using virtual reality and sensory evaluation on garment fit. Comput Ind Eng 2018; 115: 683–695.

30.

Lee

. Heuristic misfit reduction: A programmable approach for 3D garment fit customization. Comput Graphics 2018; 71: 1–3.

31.

Teichmann

Scholl-Grissemann

Stokburger-Sauer

. The power of codesign to bond customers to products and companies: The role of toolkit support and creativity. J Interact Market 2016; 36: 15–30.

32.

Chen

. A model of an e-customized co-design system on garment design. Int J Clothing Sci Technol 2018; Sep 3.

33.

Kuzmichev

Moskvin

Surzhenko

, et al. Computer reconstruction of 19th century trousers. Int J Clothing Sci Technol 2017. Aug 7.

34.

Duray

. Mass customization origins: Mass or custom manufacturing? Int J Oper Prod Manag 2002; Mar 1.

35.

Ashdown

. An investigation of the structure of sizing systems. Int J Clothing Sci Technol 1998; Dec 1.

36.

Pei

Park

Ashdown

, et al. A sizing improvement methodology based on adjustment of interior accommodation rates across measurement categories within a size chart. Int J Clothing Sci Technol 2017. Sep 4.

37.

Xia

Istook

. A method to create body sizing systems. Clothing Textil Res J 2017; 35(4): 235–248.

38.

Loker

Ashdown

Schoenfelder

. Size-specific analysis of body scan data to improve apparel fit. J Textil Apparel Technol Manag 2005; 4(3): 1–5.

39.

Gill

. Sizing in clothing: Developing effective sizing systems for ready‐to‐wear clothing. J Fashion Market Manag 2008. Sep 19.

40.

Price EC, Dasher PB, Pate M, inventors; Atlanta Attachment Co, assignee. Elastic waistband attachment system. United States patent US 5,743,202. 1998, Apr 28.

41.

Wilson

inventor

Wilson

Barry

assignee . Garment belt. United States patent US 5,758,366. 1998; Jun 2.

42.

Farley

. Mathematical programming models for cutting-stock problems in the clothing industry. J Oper Res Soc 1988; 39(1): 41–53.

43.

Ashdown S (ed) Sizing in clothing. Elsevier: Woodhead Publishing Limited; 2007.

44.

Zakaria

Gupta

. Anthropometry, apparel sizing and design, Cambridge: Woodhead Publishing, 2019.

45.

Zakaria

. Clothing for children and teenagers: Anthropometry, sizing and fit, Cambridge: Woodhead Publishing, 2016.

46.

BSI B. 13402-2: 2002, Size Designation of Clothes-Part 2: Primary and Secondary Dimensions. British Standards Institute, London, 2002.

47.

Degraeve

Vandebroek

. A mixed integer programming model for solving a layout problem in the fashion industry. Manag Sci 1998; 44(3): 301–310.

48.

Vilumsone-Nemes

. Industrial cutting of textile materials, Cambridg: Woodhead Publishing, 2018.

49.

Elomri

Morel

Pun

, et al. Solving a cutting problem based on existing patterns. Eur J Oper Res 1994; 77(1): 169–178.

50.

Nascimento

de Figueiredo

Mayerle

, et al. A state-space solution search method for apparel industry spreading and cutting. Int J Prod Econ 2010; 128(1): 379–392.

51.

Ashdown SP. Creation of ready-made clothing: The development and future of sizing systems. In: Designing apparel for consumers. Cambridge: Woodhead Publishing, 2014, pp.17–34.

52.

Mok PY, Xu J and Wu YY. Fashion design using evolutionary algorithms and fuzzy set theory–a case to realize skirt design customizations. In: Information systems for the fashion and apparel industry. Cambridge: Woodhead Publishing, 2016, pp.163–197.

53.

Huda

. Dyeing conditions and their effects on mechanical properties of polylactide fabric. AATCC Rev 2003; 3(8): 56–61.

54.

Yang

Naarani

. Effect of steaming conditions on colour and consistency of ink‐jet printed cotton using reactive dyes. Color Technol 2004; 120(3): 127–131.

55.

Preradovic

Balbin

Karmakar

, et al. Multiresonator-based chipless RFID system for low-cost item tracking. IEEE Trans Microw Theor Tech 2009; 57(5): 1411–1419.

56.

Ngai

Chau

Poon

, et al. Implementing an RFID-based manufacturing process management system: Lessons learned and success factors. J Eng Technol Manag 2012; 29(1): 112–130.

Garment mass customization methods for the cutting-related processes

Abstract

Keywords

Methodology

Mass customization methods based on pattern variations

Methods for custom-fit

Increment of size number

Expansion of size capacity

Methods for co-design

Fabric variation

Module variation

Definition of personalization level

Custom-fit level

Co-design level

Cutting-related costs

Case study

Design of experiments

Objects

Contents

Constraints

Implementation of experiments

General fit improvement (MC(FG))

Local fit improvement (MC(FL))

Design of fabric color (MC(DFC))

Design of pocket type (MC(DPT))

Design of skirt length (MC(DSL))

Results and discussion

Results on personalization level

Results on cutting-related cost

Relationship between personalization levels and costs

Conclusion

Footnotes

Acknowledgements

Declaration of conflicting interests

Funding

ORCID iD

References

General fit improvement (MC(F_G))

Local fit improvement (MC(F_L))

Design of fabric color (MC(D_FC))

Design of pocket type (MC(D_PT))

Design of skirt length (MC(D_SL))