Modeling and quantifying uncertainty in the product design phase for effects of user preference changes
| Published date | 19 October 2015 |
| Pages | 1637-1665 |
| Date | 19 October 2015 |
| DOI | https://doi.org/10.1108/IMDS-04-2015-0163 |
| Author | Hamid Afshari,Qingjin Peng |
| Subject Matter | Information & knowledge management,Information systems,Data management systems |
Modeling and quantifying
uncertainty in the product design
phase for effects of user
preference changes
Hamid Afshari and Qingjin Peng
Department of Mechanical Engineering,
University of Manitoba, Winnipeg, Canada
Abstract
Purpose –The purpose of this paper is to quantify external and internal uncertainties in product
design process. The research addresses the measure of product future changes.
Design/methodology/approach –Two methods are proposed to model and quantify uncertainty in
the product life cycle. Changes of user preferences are considered as the external uncertainty. Changes
stemming from dependencies between components are addressed as the internal uncertainty. Both
methods use developed mechanisms to capture and treat changes of user preferences. An agent-based
model is developed to simulate sociotechnical events in the product life cycle for the external
uncertainty. An innovative application of Big Data Analytics (BDA) is proposed to evaluate the
external and internal uncertainties in product design. The methods can identify the most affected
product components under uncertainty.
Findings –The results show that the proposed method could identify product changes during its life
cycle, particularly using the proposed BDA method.
Practical implications –It is essential for manufacturers in the competitive market to know their
product changes under uncertainty. Proposed methods have potential to optimize design parameters in
complex environments.
Originality/value –This research bridges the gap of literature in the accurate estimation of
uncertainty. The research integrates the change prediction and change transferring, applies data
management methods innovatively, and utilizes the proposed methods practically.
Keywords Product design, Big data analytics, Agent-based modelling, Uncertainty quantification
Paper type Research paper
1. Introduction
A product life cycle includes several stages from intangible conceptual design to used
product at the end of its life time. Managing the product life cycle requires finding
solutions for uncertain changes and unpredicted events. Studies showed that more than
half of initial user requirements will be changed before a project completion (Kobayashi
and Maekawa, 2001; Ramzan and Ikram, 2005). Improper management of requirement
changes imposes negative consequences to a system or product such as increased
complexity (Chen, 2006), data loss (Morkos et al., 2010), and wasted time and money
(Morkos and Summers, 2010; Morkos et al., 2012). However, if probable changes and
uncertainties are predicted in advance, the chance of design fail (e.g. customer’s
dissatisfaction) can be reduced. Therefore, it is essential to evaluate uncertainties in the
product life cycle. Industrial Management & Data
Systems
Vol. 115 No. 9, 2015
pp. 1637-1665
©Emerald Group Publis hing Limited
0263-5577
DOI 10.1108/IMDS-04-2015-0163
Received 30 April 2015
Revised 25 August 2015
2 September 2015
Accepted 13 September 2015
The current issue and full text archive of this journal is available on Emerald Insight at:
www.emeraldinsight.com/0263-5577.htm
The authors wish to acknowledge that this research has been supported by the Natural Sciences
and Engineering Research Council (NSERC) of Canada and UMGF Grant from the University
of Manitoba.
1637
Quantifying
uncertainty in
the product
design
Uncertainty is inevitable in engineering systems. Any lack of data or lack of trust in
identified customers’needs is considered as uncertainty (Wynn et al., 2011; Afshari and
Peng, 2014). The research (Eckert et al., 2009) showed that “customers’need”is a
dominant driver of changes in the product life cycle. Uncertainty in the customer need
affects the design solution. Customers may update their needs and preferences during
the product life time. Such uncertainty affects product development (PD) in term of
cost, adaptability, and time.
It is proved that decisions in the design stage contribute to 70-85 percent of the total
product cost (Ullman, 1997; Besterfield et al., 1995; Cao et al., 2008). In terms of
sustainability, these decisions would impact 80-90 percent of the final performance of a
product during its life cycle (May et al., 2012). Therefore, if a designer could identify
future changes of a product in the design stage, a proper decision can be made to
minimize cost and environmental impacts of the product. Hypothetically, effects of the
design stage decisions can be extended to other measures and indexes (e.g. product
quality, durability, adaptability, etc.).
The existing research methods in the product change mainly study the propagation
of changes into product components and functions. In other words, the propagation of
changes within product structure is discussed regardless of the source of changes (e.g.
Martin and Ishii, 2002; Yang et al., 2014). The change of customers’preference in a
product life cycle is a significant uncertainty for product design. Despite the variety,
current qualitative and quantitative methods for the change of preferences (e.g.
interview with customers and experts, questionnaires, QFD, marketing research, and
engineering methods) have limitations. For example, the change propagation methods
do not provide a metric for comparing design alternatives in different scenarios. Thus,
two methods are proposed in this research to bridge the gap of literature. Both methods
use innovative mechanisms to capture and transfer changes into the product design.
The goal of this research is to quantify the changes of customers’preferences during
the life time of a product. By quantifying the changes, a designer will be able to provide
appropriate solutions in product design stage. The research question proposed in this
paper is to find ways to measure future changes of customers’needs in the design
stage. If the quantified changes of customers’preferences are provided to designers,
product components to meet functional requirements (FRs) and design parameters
related to the changes can be considered to meet the changing need.
The proposed agent-based model (ABM) simulates changing events and interactions
in a product life cycle. An ABM consists of a set of elements (agents) characterized by
some attributesto interact each other through definedrules in a given environment. The
Big Data method is proposedfor further improvements of the presented ABM in termof
social and technical factors, and the study scope. Big Data improves deficiency of other
methods to quantifyexternal and internal uncertaintiesin the product design process. In
other words, Big Data analytics (BDA) uses real data instead of predicted or simulated
data in other methods.Big Data is a buzz word used in academia andindustries recently.
The application of the Big Data is growing for the better data-driven decision making
(Obitko et al., 2013). The Big Data provides a cost-effective way to obtain users’
information for a knowledge economy. As a result of the age of information, a lot of user
and product data are available in the internet for analysis of the interaction between
users and producers.Using BDA, product and user data can be easilycollected to be used
for product improvement. Among discussed types of BDA including descriptive,
predictive, and prescriptive data analysis, this research develops a prescriptive analytics
for product design process. In this type of BDA, not only past trends are used to mine
1638
IMDS
115,9
To continue reading
Request your trial