Computers in Industry 61 (2010) 213–222

Contents lists available at ScienceDirect

Computers in Industry
journal homepage: www.elsevier.com/locate/compind

Towards product customization: An integrated order fulfillment system

Linda L. Zhang a,*, Carman K.M. Lee b, Qianli Xu b
a
Department of Operations, Faculty of Economics & Business, University of Groningen, Groningen, The Netherlands
b
Division of Systems and Engineering Management, School of Mechanical and Aerospace Engineering, Nanyang Technological University, Singapore

A R T I C L E I N F O A B S T R A C T

Article history: To stay competitive, manufacturing companies nowadays pursue product customization; and configure-
Received 16 November 2008 to-order (CTO) has been recognized as an ideal business model to implement product customization. Due
Received in revised form 17 August 2009 to the complexities resulting from the large number of customized products, an increased attention has
Accepted 10 September 2009
been paid to the order fulfillment process in CTO. However, most of the solutions delivered either focus
Available online 14 October 2009
on own interested areas without considering the impacts from/on the others or approach order
fulfillment processes from a strategic level. In view of the limitations of existing solutions and the
Keywords:
significance of order fulfillment activity automation and integration, this study considers the entire
Product customization
spectrum of order fulfillment process at an operational level and, proposes an integrated order
Configure-to-order
Order fulfillment process fulfillment system (IOFS). The IOFS is designed to (1) automatically execute order processing, which is
Product configuration conventionally undertaken in a face-to-face manner; (2) configure products and process plans based on
product and process family models; and (3) allow real-time data extraction and sharing of the latest
information. Therefore, it assists companies to quickly respond to diverse customer requirements and
deliver the expected products at low costs. An industrial example of turn unit assemblies is presented to
demonstrate the feasibility and potential of the IOFS for product customization.
ß 2009 Elsevier B.V. All rights reserved.

1. Introduction Fig. 1. Each succeeding stage interacts with its proceeding stage by
utilizing the output of the latter and sending feedback to the latter.
In recent several decades, the diverse customer requirements For example, order planning uses product specifications (i.e., the
along with the short delivery times have pushed manufacturing results of order processing) to configure products and require order
companies to pursue other strategies, rather than the mass processing to clarify or modify specifications so as to better meet
production of standardized products, to guide their daily activities. customer requirements.
Aiming to quickly provide customers with tailor-made products at While each stage has its own focus and, incurs different
low costs, product customization has attracted much attention activities, on receiving of customers’ inquiries or request for
[1–3]. To implement product customization, many companies quotations (RFQs), they collaboratively work together, in attempt-
have changed their business models from make-to-stock to X-to- ing to achieve customer satisfaction with respect to cost, delivery
order, where X stands for design, engineer, configure, etc. [4]. time and expected product. Order processing includes such
Configure-to-order (CTO) has been recognized as an ideal model activities that the communication among people is to articulate
that provides a right amount of product variety and a quick the particular customer needs, to obtain the corresponding
response time to customer orders [5]. product specifications and to estimate the possible cost and
In CTO, final products are configured from a set of predefined delivery time for answering RFQs. In this stage, various partici-
modules and components subject to the constraints among them pants from different companies or from different disciplines in the
[6]. While production in CTO starts after receiving of a customer same companies are involved. They mainly include customers,
order, order fulfillment starts from order processing (also termed salespersons, designers, planners, process engineers, production
as order acquisition in some articles). In general, the activities in engineers and financial staff. Due to the improper communication,
fulfilling a customer order can be grouped into four stages, namely misunderstanding and the limited knowledge owned by the
order processing, planning, execution and delivery, as shown in participants, the traditional (i.e., face-to-face) order processing is
often long (perhaps ranging from weeks to months) and, requires
investigation on a large amount of time and cost [7].
* Corresponding author. Tel.: +31 (0) 50 363 6458. Order planning covers two parts regarding product configura-
E-mail address: [email protected] (L.L. Zhang). tion and process plan generation. (Note, in this study, a process

0166-3615/$ – see front matter ß 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.compind.2009.09.003
214 L.L. Zhang et al. / Computers in Industry 61 (2010) 213–222

knowledge; and (2) a smooth transition between any two

consecutive stages and information sharing among all stages
should be realized.
Recognizing the limitations of existing solutions to order
fulfillment processes and the significance of integrating and
automating activities to fulfill a customer order, this paper
proposes an integrated order fulfillment system (IOFS), as shown
in Fig. 2(b), to provide companies with an integrated solution to
product customization. While we envision the applicability of an
enhanced system to other product types, we investigate the
application of the IOFS to electro-mechanical artifacts that have
multiple BOM structures, instead of products with a single level
Fig. 1. The process of fulfilling customer orders.
BOM structure (e.g., printed circuit board assembly products). The
IOFS coordinates the activities in the entire process of order
plan of a product includes both assembly processes for component fulfillment in one platform. It handles customer order acquisition,
assemblies and manufacturing processes for component parts.) It product specification generation, product and process plan
includes a wide range of activities from configuring products configuration, production monitoring and customer order delivery.
through generating the corresponding process plan to schedule Therefore, the single platform and activity integration make order
production. Such activities are carried out by designers, planners tracking easier. This in turn allows companies to be aware of the
and production engineers within a single company or members progress and problems in fulfilling a customer order at any time,
from different locations. Order execution is responsible for the thus taking the corresponding actions. The automatic order
realization of products. It contains such activities as the prepara- processing and product, process configuration help eliminate
tion of materials and manufacturing resources and the setup of the errors in traditional approaches to obtaining customer orders,
production lines. It also includes production and control related product configuration and process plans. Ultimately, the IOFS not
activities. According to the documents (usually in the paper only enables companies to quickly fulfill customer orders but also
format) prepared based on the information provided by the helps gain more potential orders through the online order
previous stage, production personnel carry out these activities on acquisition. In addition, the business process of IOFS is consistent
shop floors. Such paper documents may include bill of materials with the factory-to-business-to-customer model, which aims to
(BOM) reports, routing lists, operations sheets, process plans and realize product customization over the Internet by coherently
flow charts. Finally, order delivery concerns the activities, such as integrating production automation with supply chain manage-
delivering goods to customers. It also addresses the inventory ment and sales support into a collaborative web of interactive
management activities. commerce [21].
Although multiple solution frameworks, methodologies and The rest of the paper is structured as follows. Section 2 presents
methods have been proposed to improve order fulfillment the related work on order processing, product configuration and
processes, articles that explicitly discuss the entire spectrum of order fulfillment processes. The IOFS and its logical process are
order fulfillment processes are relatively limited [8]. Moreover, the introduced in Section 3, by following which, Section 4 describes the
reported studies have their own functional focuses, e.g., quotation system architecture. Sections 5 and 6 detail the system databases
process reengineering [9], cost and delivery time estimation [7,10], and the key system modules, respectively. An industrial example of
product configuration [3,11–13], BOM generation [14], process turn unit assemblies is presented in Section 7; and Section 8
plan generation [15], assembly planning [16]. Along with the provides further discussions regarding order fulfillment require-
advancements in Internet technologies, a number of configuration ments and benefits of the IOFS. Concluding remarks in Section 9
prototypes and systems have been developed to support product end the paper.
customization [3,12,17,18]. The modeling techniques used are
typically based on models such as business-to-business and 2. Related work
business-to-customer. However, without considering other activ-
ities and the impacts associated with them, these modeling As a key business process, the order fulfillment process has
techniques only deal with partial information along the chain of received an increased attention from both industry and academia
order fulfillment (e.g., quotation, order quantity, delivery time). alike. Many studies have been carried out to approach order
Moreover, essentially, they address order planning with an fulfillment processes from different aspects and at different levels.
attempt to obtain product configurations, as shown in Fig. 2(a). Kritchanchai and MacCarthy [8] investigate the connections
The diverse customer requirements, large variety of possible between responsiveness and the order fulfillment process from
configurations, various constraints among modules and compo- a higher strategic level by undertaking field studies. Their purpose
nents together with the resulting complexities in downstream is to specify the concept and meaning of responsiveness so as to
planning and production demand an increased automation in support different types of companies to develop their own
order fulfillment processes for product customization [18,19]. In responsive order fulfillment processes. Turner et al. [22] interpret
addition, as pointed out by Salvador and Forza [20], Yang et al. [3] the order fulfillment process from the reliability engineering
and Yu and Skovgaard [18], the characteristics of order fulfillment viewpoint and suggest a framework involving risk analysis, risk
processes (e.g., the dependency among activities within and assessment and risk management to improve order fulfillment
between stages, the involvement of personnel from different processes. Attempting to assist semiconductor foundry fab super-
disciplines) and the adoption of other processing systems in a visors to easily locate the bottlenecks in fulfilling a customer order,
company raise the importance in integrating the activities of Yu and Huang [23] develop a foundry fab order fulfillment process
different stages, in attempting to achieve a rapid and effective model and apply distributed multi-agents technique to perform
response to customer orders. The implications of such integration simulation. Recognizing the importance of capturing and repre-
and automation are twofold, including (1) the major activities in senting product configuration data in controlling the order
the four stages should be automatically executed by computers fulfillment process, Tseng and Chen [13] propose a binary
rather than the multiple involved personnel with subjective tree algorithm to assist companies to generate better product
 L.L. Zhang et al. / Computers in Industry 61 (2010) 213–222 215

Fig. 2. Positioning of existing work and the proposed IOFS.

configurations by improving collaboration between sales and R&D product data so as to enable product lifecycle management
departments. Essentially, their work presents an alternative and collaborative product design [27]. This study looks at order
approach to configuring products based on customer orders. fulfillment process at an operational level and, develops the IOFS to
While articles looking at the entire order fulfillment process are assist companies to reap the benefits of product customization.
scarce, many literatures have been published with focus on different
areas, order processing and product configuration in particular. To
3. System overview
tackle the complexities resulting from the knowledge and informa-
tion work, Hvam et al. [9] describe how a knowledge-based system
Fig. 3 shows an overview of the IOFS. Internet/WWW are the
can support activities carried out in the quotation process in a
enabling technologies of the IOFS. They also provide a means for
cement company in Denmark. Recognizing the significance of quick
the IOFS to integrate with companies’ legacy systems. As a result, in
and effective response to RFQs, Veeramani and Joshi [7] present a
the iterative order fulfillment processes, the IOFS permits
new way to undertake order processing activities. They also develop
significant information sharing among multiple involved partici-
methodologies for accurately estimating costs for both modified
pants from dispersed locations. It can provide customer needs
standard products and customer-made products. Similarly, Krings-
analysis, and design and manufacturing expertise through the
man et al. [10] discuss a model along with a decision support system
Web-based computing environments. Moreover, the IOFS provides
to deal with customer enquiries in order processing with focus on
participants with the ability to quickly interact by accessing its
cost and lead-time estimation.
built-in expertise. Starting from acquisition of customer needs
To assist companies to quickly figure out products that
(CNs), the IOFS enables requirement transformation, product
can better meet customer requirements, a number of authors,
configuration, process plan generation, and scheduling to be
including Ariano and Dagnino [14], Forza and Salvador [11], Slater
automatically performed based on the expertise. This way of
[12], Tseng et al. [24] and Yang et al. [3], report their solutions
addressing product configuration from different aspects. For
example, Ariano and Dagnino [14] focus on the generation of
dynamic BOMs for office furniture in SmarBOM (i.e., the system
they have proposed); to address customers’ individual require-
ments which cannot be fulfilled by the predefined components and
modules in product configuration systems, Yang et al. [3] put
forward an Internet-based product customization system with
focus on the new design of customized products. To conclude,
while to certain degree, the above methodologies facilitate the
activities involved in order processing and product configuration,
they do not consider the other stages and activities and the impacts
from/on these stages and activities. As stated by Salvador and Forza
[20], Yang et al. [3] and Yu and Skovgaard [18], integrating
activities in the entire spectrum of order fulfillment process is
important for companies to stay competitive.
In general, the Enterprise Resource Planning (ERP) systems and
Product Data Management (PDM) systems are designed to help
companies streamline their business processes at a higher strategic
level rather than deal with order fulfillment activities at a detailed
operational level [25]. Moreover, as pointed by Mabert et al. [26],
one major disadvantage of ERP systems originates from the fact
that they are complex, time-consuming, difficult and expensive to
implement; the focus of PDM is restricted to managing design and Fig. 3. Overview of the integrated order fulfillment system.
216 L.L. Zhang et al. / Computers in Industry 61 (2010) 213–222

Fig. 5. A general three-tier system architecture.

as the programming language and eXtensible Markup Language as

the information exchange file format. The architecture of IOFS, as
shown in Fig. 5, is generally known as a three tier distributed
client/server architecture. In the architecture, there is a clear
distinction between the client and server share of program
execution. While the clients serve as a means for getting user
Fig. 4. Logical process flow of the IOFS.
inputs and displaying the results, the server executes all the data
operations and processing. The database hosts all the data as
offering customized products significantly reduces order fulfill- required.
ment cost and lead-time. The IOFS offers an environment that allows users to access the
By means of a password control mechanism, different users system via any Web browser such as Netscape’s Navigator or the
such as customers, designers, planners, process engineers, pro- Microsoft’s Internet Explorer at different geographical locations.
duction engineers and shop floor personnel have different The data representation tier provides database management
authorities. They are linked up to the Hyper Text Transfer Protocol facilities for product, process, code, order and customer databases,
(HTTP) sever via either Internet or Intranet, and can browse related respectively. Microsoft SQL Server1 is used to establish such a
information stored in databases. Their access rights are decided by database management system. To facilitate communication,
matching user names and passwords stored in the databases. ActiveX Date Object (ADO) can be adopted to interface with the
The logical process flow of the IOFS is shown in Fig. 4. After a relational database through Open Database Connectivity (ODBC).
user (e.g., a customer) logs in the system using his Web browser, he The client tier includes a set of application clients. Via a network
is asked to input his requirements (i.e., CNs). An assessment using Transmission Control Protocol/Internet Protocol (TCP/IP), the
module is about to check out whether or not these CNs are against client computer is connected with the Web server. The only stuff
a set of constraints, which are pre-defined according to company’s needed by client users is a Web browser. Between the data
existing engineering and manufacturing capabilities, business representation tier and the client application, there is a server tier
strategies, performance metrics, etc. If these constraints are (also called business logic tier in some articles). The Internet
satisfied, the input data are transformed into a set of product Information Server (IIS) can be employed as the HTTP server. IIS
specifications. In the meantime, the IOFS prepares the RFQ. After accommodates all Internet services that are needed when
these computations, all the outcomes are displayed on the screen developing Web-based applications. It accepts all the Active
of the customer’s computer. The order placement is conducted Server Page (ASP) requests sent from the client’s browser.
online by filling in the order form if the customer accepts the Microsoft Message Queue Server (MSMQ) can also be installed
displayed information. Or else, he needs to revise either the to provide the technologies that allow the transmission of
requirements input or the order until a satisfied one is reached. messages over network connections and management of these
After the order is accepted based on the evaluation result, which messages.
indicates that it is highly desirable to complete the order within the Among all the issues in Web-applications, these relating to user
target time window, the IOFS automatically configures product, and data security are so important that they cannot be ignored.
generates process plan and schedules the production. The resulting This is due to the fact that the Web, through which data
information is shown on the screens of terminals on the shop floor is transferred and exchanged, is prone to data tapping, data
to guide and control production activities. After delivering the final interception and user impersonation. In the IOFS, system security
products to the customer, system databases are updated. In is achieved through password authentication, data encryption
addition, the new product configurations, process plans and the technology and verification. In this study, password authentication
generated code are stored in the respective databases for resolving involves setting up user accounts and passwords in the database.
problems in the future.
5. System databases
4. System architecture
The IOFS consists of six databases for product, process, code,
To provide the flexibility of interoperability on a variety of customer, order, and user profile, respectively. The product
operating platforms, the IOFS can be developed by employing Java database hosts such data as product IDs, assemblies, parts, their
 L.L. Zhang et al. / Computers in Industry 61 (2010) 213–222 217

systems. As a result, the precise response to RFQ can be quickly

achieved, which is important for companies to attract more
customers and gain more orders.
After filling in the user name and the password, the user is
prompted with the corresponding Graphic User Interface (GUI)
based on his profile stored in the database. If the user is a potential
customer, he needs to answer the displayed questions about basic
characteristics of the product that he is interested in. These
questions are designed according to constraints among product
items (i.e., according to the product family model). After the
Fig. 6. Product data hierarchy in the product database.
answers are confirmed by the customer, the assessment sub-
module is to conduct a general check to determine whether or not
interconnections, etc. By tracing the parent–child relationships these input answers exceed the company’s available engineering
among product items, a common product structure underlying a and manufacturing capabilities. This check requires incorporating
product family (i.e., a product family model) can be identified. Also the current shop floor and inventory status. If the capabilities are
hosted are the quantities of assemblies and parts as required by violated, a warning popup window is displayed on the customer’s
their immediate higher level parent items. Design parameters of screen. In this case, the customer may need to re-answer the
intermediate items and the variety parameters coupled with questions. After the questions are accepted by the assessment sub-
possible values of parts and raw materials at the lowest levels in module, the transformation sub-module and the RFQ sub-module
the common product structure can be retrieved from this database begin to work on these requirement inputs. First, these needs are
as well. To enable the system to properly work, the product transformed into FRs, which are not displayed on the customer’s
database is organized into a hierarchy, as shown in Fig. 6. The screen as the final result. However, designers are allowed to trace
process plan related information, including operations IDs, types and modify them for better design. Then, these FRs are mapped to
and sequence, machines, tools and fixtures, etc., can be extracted the corresponding design parameters and their possible values.
from the process database. Same as product data in the product Such design parameters and values are combined in pairs, forming
database, process data in the process database is also modeled as a a set of product specifications. Last, the results including a list of
hierarchy, as shown in Fig. 7. The code for design parameters, their specifications, estimated product prices and delivery date ranges
possible values, and other product, process related data is located are displayed on the customer’s screen.
in the code database. The connections between product and If the customer accepts the displayed information, he proceeds
process data are established through these codes. In other words, to online place the order by filling in and submitting the order
the same set of code is used by a particular product configuration form. In case the customer is not satisfied with either the product
and the corresponding process plan. Therefore, the three databases specifications or product price or delivery date, he can get the
contribute to integrated and automatic product configuration and different results by changing the answers to questions or
process plan generation, thus being the core of the IOFS. While the modifying the required delivery date. Till this point, order
customer database provides customer related information, includ- processing covering CNs acquisition, general requirements assess-
ing IDs, names, addresses, contact phone numbers and purchased ment, CNs transformation and online order placement is com-
products information such as IDs and quantities, in the order pleted. The next stage deals with the automatic product
database, the data ranging from order IDs and status (open or configuration and process plan generation based on the obtained
closed) to the IDs of products to the IDs of customers, etc., are product specifications.
stored. At last, for the security issues, the user names and
passwords are saved in the user profile database. 6.2. Product configuration and process plan generation
All the above databases can be relational databases. Foreign
keys are used to establish links among databases; and primary keys On receiving an online customer order, a salesperson starts the
are applied to identify the corresponding data tables in one evaluation module, which is to carry out a thorough examination
database. to determine the degree to which the order can be completed
within the delivery time. Differing from the requirement assess-
6. System modules ment module, it conducts the evaluation by computing current
available materials and manufacturing resources, checking open
The function of the IOFS is collaboratively performed by four orders, inventory and shop floor status, comparing the required
main system modules: order processing, product configuration items with the available counterparts, and performing simulation.
and process plan generation, production execution and product Similarly, some of the data and information required for order
delivery. evaluation can be obtained from the legacy systems. The final
result, either order acceptance or order rejection, is displayed on
6.1. Order processing the salesperson’s screen. If an order is not acceptable, a prompted
dialogue window with the possible reasons is popped up on the
The order processing module consists of three sub-modules: an screen of customer’s computer. Such reasons can be ‘‘the product
assessment sub-module, a transformation sub-module and an RFQ price is low’’ or ‘‘the delivery date is tight’’. Also displayed in the
sub-module. The assessment sub-module is to check the validity of popup window are the new product price range and delivery time
customer inputs (i.e., CNs). The transformation sub-module period.
automatically transfers the CNs to a set of functional requirements After an order is accepted, the configuration module starts
and further to design parameters and their feasible values; the RFQ searching in the product, process and code databases for the
sub-module is responsible for making an accurate estimation of possible features and options to fulfill the product specifications
product costs and delivery dates. The real-time data and latest transformed. In this study, the coding system in [28] is adopted to
information required in the estimation process is obtained from organize the code database. Based on the selected items and the
the legacy systems, such as, inventory management, shop floor product family structure, the set of all feasible product config-
control, financial, material and capacity requirement planning urations and process plans are automatically generated. Without
218 L.L. Zhang et al. / Computers in Industry 61 (2010) 213–222

Fig. 7. Process data hierarchy in the process database.

considering the current manufacturing capabilities, cost, time, etc., documents that are used to instruct and control production
a number of product alternatives are able to achieve the same activities, the IOFS provides online production control. Therefore,
functions for the given product specifications. Therefore, con- both the operators’ training time and production engineers’ time to
strained by the above influencing factors, one out of the group best prepare the documents, such as Manufacturing Process Instruc-
fits itself to the current production situation and satisfies customer tion, Operation Sheet and Process Flow Charts, are reduced.
demands. In the IOFS, the family of all possible configurations is Meanwhile, operation accuracy can be increased and the failure
generated first. rate can be reduced thanks to the animation of operation activities
After all configuration alternatives are generated, the following shown on terminals’ screens. Rather than all information, the IOFS
step starts evaluating these alternatives to determine a better one. only allows the display of the necessary information as required at
In this study, the evaluation is performed subject to multiple a particular workstation for a particular job.
objectives, including the minimization of production costs and the For collecting the real-time data from the shop floor, the IOFS
maximization of customer perceived values. Once the evaluation requires operators to key in the beginning and ending time of
is completed and product configuration and process plan are setup, operation and machine breakdown, respectively. Operators
determined, the scheduling module, called scheduler, starts also need to select reasons for these machine stops/breakdowns
scheduling the production. To ensure that the schedule can be from the drop-down menu, which lists all the possible causes. All
completed within the due date, the scheduler extracts the information is then recorded and displayed on the screens of the
necessary data from the legacy systems. The final results including corresponding engineers’ computers for them to monitor and take
the BOM list, drawings for part and subassemblies, routing list, proper actions. At any point in time, the status of order fulfillment
flow chart, and production scheduling, etc., are displayed on the can be viewed online. This is coupled with the due date, time
screens of different users’ computers according to their accessi- elapsed from order acceptance, time left to due date, number
bility to the IOFS. A work order is automatically released online of products produced and the left ones to be completed. This
according to the schedule. Rather than a new ID, the work order has information is then used to determine whether or not a
the same ID as the accepted customer order, which is assigned by reallocation of manufacturing resources and materials is needed
the salesperson at the time when it is accepted. The use of same IDs or if particular bottleneck operation can be avoided.
is to provide the clarity and traceability not only inside a company
but in all participating companies. In the scenario where 6.4. Product delivery
unexpectations, such as machine breakdown for a long time and
defective parts/products increase, occur after the release of work The IOFS is designed to allow the connection with existing
orders, the scheduler re-schedules the production according to the inventory management systems, which record the types, quan-
latest information. tities and status of finished goods. According to the information
Furthermore, the IOFS saves the configurations of past provided by online orders including customers’ names, IDs,
produced products and retrieves them for the same products addresses, contact phone numbers, ordered products and quan-
ordered later. In this case, the process plans are still regenerated. tities, and due dates, the shipment department sends the right
However, they may not be the same as the previous ones due to the products to the right destinations, and changes orders to the closed
different production situations. status after the customers accept products.
With the close of an order, all system databases are updated.
6.3. Production execution The new product configuration, process plan, code, order, customer
are added to their respective databases for problem solving in the
Essentially, shop floor control systems are responsible for the future.
timing and execution of production activities to materialize
products. Enabled by the advancement in IT/communication 7. An industrial example
technologies and manufacturing equipment, these systems are
connected with other systems (e.g., inventory management To demonstrate the potential of the proposed IOFS, a family of
systems) to share and exchange the real-time data. Such real- turn unit assemblies, which are typical CTO products, is adopted in
time data exchange and sharing contributes to maintaining this study. A turn unit assembly is a device used at a conveyor belt
production to be as smooth as possible. By connecting with these to turn locomotives or other rolling stock onto one of several
legacy systems, the IOFS displays all necessary information for radically arranged tracks. It consists of several major components,
production on terminals’ screens on the shop floor. The displayed such as a head module, a motion module and an enclosure module.
BOM and routing lists, assembly plan, etc., can guide the operators Each module can be further decomposed into several child
to prepare materials, machines, tools, fixtures, jigs and to set up components. Fig. 8 shows a turn unit assembly and the common
production lines accordingly. Instead of the traditional paper product structure and process structure of the turn unit assembly
 L.L. Zhang et al. / Computers in Industry 61 (2010) 213–222 219

Fig. 8. A turn unit assembly and the common product and process structure.

family. The detailed information of the assembly process of the 8. Further discussions
product is shown in Table 1.
A prototype IOFS is developed using ASP.Net programming, 8.1. Requirements of order fulfillment for product customization
where MS Access1 is used for databases management. According
to the input information from a user, such as user name and Product customization is essential for companies to survive the
password, the relevant Web-pages are displayed; and different fast changing business markets which are characterized by various
levels of access to the system are authorized to the user. For a individual customer requirements. The fall of barriers to interna-
customer, he needs to provide the answers to questions capturing tional trade and communication leads to the increasingly intense
his requirements on the turn unit assembly. Thus, the IOFS can get global competition. Such intense global competition together with
customer requirements which will be used to derive the technical individual customer requirements force companies to reduce order
definitions of the expected turn unit assembly. Suppose the CNs are fulfillment cost and time and to improve product quality, thus
within the coverage of company’s engineering and manufacturing achieving product customization. Rapidly fulfilling diverse custo-
capabilities, the IOFS automatically transfers these needs to FRs mer orders at low costs is a complicated process that involves
and product specifications. Meanwhile, the RFQ sub-module will numerous activities and participants. Consequently, there are
compute the product price and possible delivery period. Then these some inherent problems in order fulfillment:
estimated numbers and the product specifications are displayed on
the customer’s screen showing information about the expected
product, such as price, receiving date and technical solutions. By
filling in and submitting the order form, an order is placed online
by the customer in case he accepts the product outlined.
According to the product specifications transformed, the
configuration module first generates a family of feasible product
configurations and process plans, then evaluates and determines
the optimal ones. For the turn unit assembly in Fig. 8(a), its product
configuration and process plan are shown in Figs. 9 and 10,
respectively.
With all necessary information, such as BOM and process plan,
production personnel can carry out production activities. The
production information reflecting the real-time shop floor status is
displayed on the screens of concerned persons’ computers and the
shop floor terminals. Such updated information and production
data enables the right decision to be made and the effective actions
to be taken. Moreover, the IOFS can extract the real-time product
data, which enabling the accurate estimation of cost and delivery
time in processing orders. After completing production, the final
products are delivered to the customers. Fig. 9. Product configuration of the turn unit assembly in Fig. 8.
220 L.L. Zhang et al. / Computers in Industry 61 (2010) 213–222

Table 1
Process data of turn unit assemblies.

Operation Description Input Output Resource

Machine Cycle time

A0 Final assembly Head, motion, enclosure Turn unit assembly M-A0 10.5  0.6
A1 Assembly of head module Head box, cable holder Head module M-A1 4.2  0.5
K1 Knitting of head box Head box Head box M-K1 2.1  0.3
A2 Assembly of motion module M0 and conveyer rotary mount Motion module M-A2 9.6  0.9
A2.1 Assembly of M1 and motor bracket M1, motor bracket M1 and motor bracket assembled M-A21 3  0.2
(denoted as M0)
A2.2 Connecting motor with gear box Motor, gear box Motor and gear box assembled M-A22 2.2  0.5
(denoted as M1)
A2.3 Knitting of conveyer rotary mount Conveyer rotary mount Conveyer rotary mount M-A23 1.6  0.45
A3 Assembly of enclosure M2 and enclosure cover Enclosure module M-A3 2.4  0.5
A3.1 Assemble of motor enclosure Motor enclosure, base plate Motor enclosure and base plate M-A31 0.9  0.2
assembled (denoted as M2)

(1) The acquisition of CNs is a time-consuming and error-prone manufacturing environments nowadays. Therefore, the pro-
elaboration process and requires the investigation of huge duction plan and schedule made is usually hard to follow.
amount of time and cost. This is due to the fact that customers
usually do not clearly know the technical details of a product. To conclude, the solutions to fulfilling customer orders should
(2) Although in CTO, most of the customized products are designed be capable of aiding the explicit acquisition of CNs and the
by combining and/or modifying existing components, it is still conversion of these needs to rigorous engineering terms. More-
difficult for designers to choose the right items just based on over, the downstream fulfillment activities are to be considered as
their intuition when the number of component items is very well. Such solutions that only facilitate single stages are not
large. The increased complexity of products aggravates this adequate by considering the entire order fulfillment process.
situation. In spite of their great numbers, in general, the design Therefore, developing solutions which allow a seamless naviga-
tools provide point solutions by nature. Therefore, they cannot tion, information and data sharing across the entire spectrum of
be effectively adopted to obtain a good solution. order fulfillment process has been put on the agenda of research
(3) The major headache troubling production personnel is the community [18,20]. Considering the distributed nature of order
frequent process changeovers, which are impossible to predict fulfillment, these solutions should be designed to be employed by
and prepare if there is no proper tools to provide the necessary various participants regardless of their dispersed geographical
information in advance. locations. In addition, to lessen the burden of developing such
(4) As far as the supporting tools for planning and scheduling are solutions and to utilize the existing systems as more as possible,
concerned, they often designed based on such assumptions these solutions should be designed to be able to integrate with the
that are meaningless in the dynamic and complicated legacy systems.

Fig. 10. Process plan of the turn unit assembly in Fig. 8.

 L.L. Zhang et al. / Computers in Industry 61 (2010) 213–222 221

8.2. Analysis of the benefits of the IOFS configure a product. However, due to their inherent limitations,
they cannot resolve some problems encountered in the fulfillment
(1) CNs acquisition. Based on the product family model, the work process (e.g., the access to the same data describing same objects)
for explicitly elaborating CNs is simplified to a large degree. In [18]. Furthermore, the claimed benefits of these solutions may be
the IOFS, the only thing for acquiring CNs is to online answer off set by the poor execution and results of other relevant activities
several questions at the customer side. Compared with and processes.
traditional CNs acquisition, which often requires the presence This study presented an integrated system to fulfill customer
of multiple people and usually involves a tedious and long orders in CTO – the IOFS – along with some preliminary results
process, the time taken in IOFS is significantly reduced. from an industrial example. The IOFS allows automation and
(2) Customer order processing. In the conventional order processing, integration of activities in fulfilling customer orders at an
the process lead-time is long and often ranges from weeks to operational level. Thus, it enables companies to quickly react to
months. This long lasting processing may cause companies unceasing customer requirement changes. Thanks to the connec-
to lose their customers. Meanwhile, the resulting product tions with companies’ legacy systems, the IOFS can extract the
specifications and the response to RFQ are usually not accurate real-time data and the latest information. Based on the predefined
and precise enough. Such vagueness (especially in product product and process family models, the IOFS can generate
specifications) may increase order fulfillment cost and reduce appropriate product configurations and process plans. In addition,
the profit, or even make the negative profit. As opposed the IOFS can retrieve the past product configurations for new
to this, the order processing in the IOFS is carried out by two orders. However, the IOFS is limited to the manufacturing
sub-modules – the transformation sub-module and the RFQ environment, where the individual products are achieved by
sub-module. They perform their responsibilities by following a configuring from a set of pre-defined components. For a new
rigorous way, thus eliminating the human misunderstanding, product design, it is not applicable. In this regard, one possible
improper communication and the waiting time both in queue avenue for future research is to design an extended system to
and for processing. As a result, not only is the time required accommodate new product design and the associated order
reduced but also the accuracy is increased. fulfillment activities. In this study, discussions of using the IOFS
(3) Product and process design. Although in current family-based are developed based on the application to turn unit assemblies,
production, both product and process can be designed by which are not very complicated products. In this respect, the
adopting the similar product and process structures, without research efforts may be directed to scale up the system so that it
the support of an automatic tool, designers and process can accommodate more complicated products.
engineers still are troubled by a lot of time-consuming but less
value-added work. For example, the process engineer may need References
to modify the entire set of process plan only for a small change
in design. However, such trivial activities in traditional design [1] B.J. Pine, Mass Customization: The New Frontier in Business Competition,
Harvard Business School Press, Boston, 1993.
cannot be avoided in order to satisfy customers. The IOFS [2] M. Spring, J.F. Dalrymple, Product customization and manufacturing strategy,
allows the automatic generation of a better product config- International Journal of Operations and Production Management 20 (4) (2000)
uration and the corresponding process plan without the 441–467.
[3] Y. Yang, X. Zhang, F. Liu, Q. Xie, An Internet-based product customization system
intervention of human being. Therefore the designers and
for CIM, Robotics and Computer-Integrated Manufacturing 21 (2005) 109–118.
process engineers can shift their time and focus from the light [4] W. He, X. Ming, Q. Ni, W. Lu, B.H. Lee, A unified product structure management for
or non value-added activities to more value-added ones (e.g., enterprise business process integration throughout the product lifecycle, Inter-
innovative product development activities). national Journal of Production Research 44 (9) (2006) 1757–1776.
[5] F. Cheng, M. Ettl, G. Lin, D.D. Yao, Inventory-service optimization in configure-to-
(4) Planning and production. The real-time data and the latest order systems, Manufacturing & Service Operations Management 4 (2) (2002)
information provided by the legacy systems ensures that the 114–132.
production planning and scheduling are rational and stable. [6] S. Mittal, F. Frayman, Towards a generic model of configuration tasks, in: Pro-
ceedings of the 11th International Joint Conference on Artificial Intelligence,
Consequently, the best use of the limited materials and Detroit, MI, (1989), pp. 1395–1401.
manufacturing resources can be realized. Moreover, the frequent [7] D. Veeramani, P. Joshi, Methodologies for rapid and effective response to requests
changes to production, such as the changes to work centers, for quotation (RFQs), IIE Transactions 29 (1997) 825–838.
[8] D. Kritchanchai, B.L. MacCarthy, Responsiveness of the order fulfillment process,
machines, tools, fixtures and production line re-layout, are International Journal of Operations & Production Management 19 (8) (1999) 812–
avoided; and thus the smooth production is maintained. In 833.
addition, benefiting from the sharing of real-time data, the [9] L. Hvam, M. Malis, B. Hansen, J. Riis, Reengineering of the quotation process:
application of knowledge based systems, Business Process Management Journal
proactive actions (e.g., the check of machine before they 10 (2) (2004) 200–213.
breakdown, the reallocation of resources before the bottleneck [10] B. Kringsman, L. Hendry, A. Mercer, A. de Souza, Responding to customer enquiries
is formed) can be taken to prevent the occurrence of such events in make-to-order companies problems and solutions, International Journal of
Production Economics 46–47 (1996) 219–231.
which will reduce productivity and increase the scrap rate.
[11] C. Forza, F. Salvador, Managing for variety in the order acquisition and fulfillment
process: the contribution of product configuration systems, International Journal
of Production Economics 76 (2002) 87–98.
[12] P.J.P. Slater, Pconfig: a Web-based configuration tool for configure-to-order
9. Conclusions products, Knowledge-Based Systems 12 (1999) 223–230.
[13] H.-E. Tseng, C.-C. Chen, Coordinating product configuration in the order fulfill-
In CTO, fulfilling customer orders involves many participants ment processing: an approach based on the binary tree algorithm, International
Journal of Computer Integrated Manufacturing 19 (7) (2006) 716–726.
from different disciplines, the interactive activities and iterative
[14] M. Ariano, A. Dagnino, An intelligent order entry and dynamic bill or materials
processes. The high variety of products and the resulting large system for manufacturing customized furniture, Computers and Electronic Engi-
volumes of product and production data make the order fulfillment neering 22 (1) (1996) 45–60.
[15] L. Zhang, J. Jiao, Modeling production configuration based on nested colored
process more complicated. Thus, an integrated and automated
object-oriented Petri nets with changeable structures, Journal of Intelligent
system is important for companies to efficiently articulate Manufacturing 20 (4) (2009) 359–378.
customer requirements, configure and produce the right products [16] J. Gao, N.W. Bowland, R. Sharma, A product-configuration-driven system for
and make the final delivery. On the other hand, to some degree, the assembly planning within a product data management environment, Interna-
tional Journal of Production Research 40 (9) (2002) 2041–2051.
solutions in the order fulfillment and product configuration studies [17] A. Haag, Sales configuration in business processes, IEEE Intelligent Systems 13 (4)
may help companies, for example, reduce the time and cost to (1998) 78–85.
222 L.L. Zhang et al. / Computers in Industry 61 (2010) 213–222

[18] B. Yu, H.J. Skovgaard, A configuration tool to increase product competitiveness, customization, production reconfiguration, supply chain management, integrated
IEEE Intelligent Systems 13 (4) (1998) 34–41. product family development and business process reengineering. On these areas,
[19] P. Czapiewski, R. Ganesan, D. Volkert, B. Weber, Integrated computer-aided order she has published a number of articles in international refereed journals, such as
fulfillment, IEEE Computer 32 (4) (1999) 99–102. IIE Transactions, IEEE Transactions on Engineering Management, Decision Support
[20] F. Salvador, C. Forza, Configuring products to address the customization- Systems, International Journal of Production Research, etc.
responsiveness squeeze: a survey of management issues and opportunities,
International Journal of Production Economics 91 (2004) 273–291.
[21] M.M. Tseng, T. Kjellberg, C.-Y. Lu, Design in the new e-commerce era, Annals of the Dr. Lee is currently an Assistant Professor of Mechanical
CIRP 52 (2) (2003) 509–519. and Aerospace Engineering at the Nanyang Technologi-
[22] T.J. Turner, K. Mendibil, U.S. Bitici, P. Daisley, T.H.J. Breen, Improving the reliability cal University. She received her bachelor’s and her
of the customer order fulfillment process in a product identification company,
doctorate degree from The Hong Kong Polytechnic
International Journal of Production Economics 78 (2002) 99–107.
University. Before joining NTU, Carman was the post-
[23] C.-Y. Yu, H.-P. Huang, Development of the order fulfillment process in the foundry
doctoral fellow in the Hong Kong Polytechnic University.
fab by applying distributed multi-agents on a generic message-passing platform,
IEEE/ASME Transactions on Mechatronics 6 (4) (2001) 387–398. Her current research areas include RFID, logistics
[24] H.-E. Tseng, C.-C. Chang, S.-H. Chang, Applying case-based reasoning for product information systems, product development and supply
configuration in mass customization environments, Expert Systems with Appli- chain management. She focuses on applying computa-
cations 29 (2005) 913–925. tional intelligence such as genetic algorithm and case
[25] O. Balogun, H. Hawisa, F. Tannock, Knowledge management for manufacturing: based reasoning for supply and demand chain manage-
the product and process database, Journal of Manufacturing Technology Manage- ment. Her research results have been published in
ment 15 (7) (2004) 575–584. various international journals including Expert Systems
[26] V.A. Mabert, A. Soni, M.A. Venkataramanan, Enterprise resource planning: com- with Applications, International Journal of Production Economics and International
mon myths versus evolving reality, Business Horizons 44 (3) (2001) 69–76. Journal of Production Research.
[27] T. Bilgic, D. Rock, Product data management systems: state-of-the-art and future,
in: Proceedings of the 1997 ASME Design Engineering Technical Conferences and
Computers in Engineering Conference, Paper No. DETC97/EIM-3720, Sacramento, Dr. Xu is currently a research fellow in the School of
CA, USA, 1997. Mechanical & Aerospace Engineering, Nanyang Tech-
[28] L. Zhang, J. Jiao, S. Pokharel, Internet-enabled information management for nological University, Singapore. He obtained his BE and
configure-to-order product fulfillment, Asia Pacific Management Review (Special ME degrees from Tianjin University, China, in 1999 and
Issue on e-Business and Management) 9 (5) (2004) 851–876.
2002, respectively, and PhD degree from the Depart-
ment of Mechanical Engineering, National University of
Dr. Zhang is currently an Assistant Professor in Singapore in 2007. His major areas of interest include
Department of Operations at University of Groningen, product family design, intelligent products and manu-
Groningen, The Netherlands. She obtained her BE and facturing systems, and mass customization and perso-
PhD degrees in Industrial Engineering in 1998 and nalization. He is a member of ASME, IEEE Technology
2007, respectively. Her research interests lie in mass Management Council, and IIE.