Professional Documents
Culture Documents
Internal states
Product (P)
User 1 a3
Experience
Relation:
User M ' Product
a1 a2 ax aL
N U-U
U-P
Product Product a4
2 N' P-P
User M
User1 User M '
Product ecosystem Ambience
(a) User experience in a product ecosystem
Fig. 1. Entities and relations in a product ecosystem
User i (GA, AF, CG) User i (GA′, AF′, CG′)
E U
i u j | 1 ui , u j 1, u j E U
,i j E U , State of user set EU(i)
ax
New state of EU(i)
State of product set EP(i) New state of EP(i)
E i p j | 3 ui , p j 1, p j E E
P P P , and
(GA: general attribute; AF: affect; CG: cognition)
i u , p (for u u i E U i ; p E P i ). (b) Functional model of an event
Basically, the ambience consists of all the users and Fig. 2. Product ecosystem architecture
products that directly interact with u i , plus the
relationships among these entities. certainty factor (CF) is a value j 0,1 representing the
Typically, user experience is formed as the evolution degree of belief in the truth of the rule.
of the users’ internal states along the chain of events, In the context of affective and cognitive modeling,
where the internal state includes two essential aspects, the fuzzy production rules are developed with respect to
namely the affect and the cognition as a result of human- the events along the chain of user experience. For a
ambience interactions. In view of these considerations, the specific event ax, let denote a set of propositions, which
product ecosystem is re-formulated as in Fig. 2 (a). The are fuzzy concepts consisting of fuzzy variables and their
focal user (e.g. User i) interacts with relevant products in fuzzy values. These propositions are assertions about the
the ambience, which results in a sequence of events status and attributes of users, be it affect or cognition, and
(denoted as hexagons). The series of events that constitute those of the ambience. Next, knowledge about the system
the user experience is defined as A a1 , a2 ,..., aL , where can be codified as the fuzzy rule set, denoted as , where
L is the number of events involved in the sequence. The a fuzzy rule is a mapping from a set of antecedents to a set
input to the event includes the focal user’s affect, of consequences.
cognition, general attributes, and the status of the ambient In view of the above context, the design of the
factors. The event is an information processing unit that product ecosystem involves determining the product
deals with an activity/task of the user. Accordingly, the attributes and the service processes that contribute to
output of the event is the new status of the focal user’s positive of user affect and cognition. Among the entities
affect, cognition, and general attributes, and the new that constitute the product ecosystem, the user set EU is
status of the ambience. The functional model of an event considered to be predefined, i.e. the users’ characteristics
is shown in Fig. 2 (b). can be predicated but not controlled. On the other hand,
the product set E P is controllable and is considered as the
B. Product Ecosystem Design Problem design elements. Moreover, the chain of events
A a1 , a2 ,..., aL in accordance with the user experience
In order to design and evaluate the product is partially controllable, i.e. some events are fixed while
ecosystem, there should be a reasoning engine to predict others are flexible to form the business process.
users’ affects and cognitions that involve immense Additional constraints are exerted in view of the diverse
uncertainties and dynamics. The classical way to tackle relationships among users, the ambience , and the cause
this issue is to use fuzzy production rules. According to and effect relation existing between the ecosystem
fuzzy logic theories, fuzzy production rules are used to configuration and the user affect and cognition. With all
represent the knowledge and reasoning process [4, 5]. A these considerations, the objectives of ecosystem design
general form of fuzzy production rule is given as: are to (1) improve the customer service level by
Rule j: IF d i THEN dk (CF= j ) (1) optimizing the affective and cognitive responses of users,
and (2) reduce the operational cost of the system by
where antecedent d i and consequent d k are propositions maximizing the utilization of the ecosystem capacities.
which may contain one or a few fuzzy concepts; and the Therefore, the product ecosystem design problem is
reduced to an optimization problem of configuration
design, as shown in Fig. 3.
Given: Entities belonging to a product ecosystem: User set EU , and places of a transition ( t j ) is called its postset,
product set E .P
denoted as t j P .
A. CDFPN Modules and Simulation The simulation is first carried out with respect to a
legacy system, whose configuration is specified in Table
The knowledge of how the ecosystem affects I. The performance of the legacy system is evaluated
commuters’ affect and cognition is coded as a set of against the service level and system capacity utilization.
CDFPN modules. For example, the CDFPN module for These results bring about useful implications of system
representing the “read information event” is shown in Fig. design in the following aspects.
4. The module dictates that the user’s valence state
changes due to the visibility of the information board (IB) Service level evaluation
– high visibility leads to pleasurable feeling and low The effectiveness of the system on the commuter
visibility causes displeasure. At the same time, the IB valence is reflected by the average valence of all
usually improves users’ familiarity with the subway commuters, as shown by the solid curve in Fig. 5. Overall,
station. Next, the reasoning logic of the CDFPN is the system produces an average valence of 0.31. Because
translated into a dynamic-link library (DLL) file coded in the initial average valence of all commuters is set as zero,
this result indicates a possible effect of the ecosystem on
the commuter valence. From the commuter stay-time (the 0.9
dashed curve), it is shown that the commuters stay in the 0.8
Capacity utiliation
GTM
shorter than the expected stay-time of 6.5 minutes. This is 0.6
GN
0.5
an indication of acceptable operational efficiency of the 0.4 GW
service process. 0.3 SP
Cost evaluation 0.2 LT
The system capacity utilization is computed to find 0.1 ES
out the cost-effectiveness of the ecosystem. Fig. 6 shows 0
TR
the utilization of the products with relatively high capital
8
0.5
5.5
13
18
23
28
33
38
43
48
10.5
15.5
20.5
25.5
30.5
35.5
40.5
45.5
investment or operational cost. As can be seen from the Time(×103 )
results, the lift (LT), the service person (SP) and the train
Fig. 6. Capacity utilization of individual products
(TR) are characterized by high utilization rates. While a
higher average utilization rate of a product indicates lower
0.5
unit cost per service process, it suggests a possible
bottleneck of the service process, if the users’ valence 0.4
Valence Change
Design improvement 0.2
designer can find useful guidelines for improving the ‐0.1 GTM IB GN GW SP LT ES ST TR
6.45
0.32
As shown in Fig. 8, a Pareto-frontier can be found,
Average Stay‐time
6.4
0.31
6.35
suggesting the optimal trade-offs between capacity
0.3
6.3 utilization and service level. It should be noted that the
0.29
6.25 optima in the Pareto-frontier may not be found in reality
0.28 6.2 because it is impractical to explore the design space
0.27 6.15 thoroughly. The data points in the figure are computed
manually, and are called pseudo-optima.
13
18
23
28
33
38
43
48
10.5
15.5
20.5
25.5
30.5
35.5
40.5
45.5
3
8
0.5
5.5
Aggregated Capacity Utilization
Design of product ecosystems suggests a new
perspective of product design with expanded scopes and
Fig. 8. Aggregated service level vs. aggregated capacity utilization
objectives. On the one hand, it signifies more added-value
to the products or service processes owing to the
improvements of the service level. The other extreme can differentiation enabled by the inclusion of user’s affect
be found at points with high capacity utilizations together and cognition. On the other hand, it fosters the
with the plummet of the service levels. This means that development of new methodologies and technologies to
the cost saved from lower system investment is eroded by tackle the unprecedented complexities related to the
the worsening service level, causing it to be unacceptable ecosystem design problem. In this regard, a few
to the users. Highlighted in the chart are two points constructive observations are made based on the outcome
indicating the performances of the benchmark system and of the study. First, the ambience is a useful and promising
a newly designed pseudo-optimal system. As can be seen concept in design for user experience, with its rich
from the data, the new system outperforms the benchmark implications, such as the multiple entities, multivariate
system by providing higher service level at lower system relations and event-driven processes. Second, the CDFPN
cost (indicated by the higher system capacity utilization). is suitable to capture the uncertainties of the ecosystem
and enable consistent reasoning with respect to user’s
affect and cognition. Third, the proposed modeling
IV. DISCUSSION framework provides decision support to product design
owing to its capacity to construct and simulate user
The case study shows that the roadmap towards experience in relation with product ecosystem
designing complex, high-value added systems is viable configurations. In practice, the concept of affective-
through the design for user experience accommodating cognitive modeling sheds light on industrial applications,
users’ affective and cognitive needs. The CDFPN model such as cabins of airplanes, trains, yachts, subways,
is capable of describing the fuzzy causal relations exhibition halls, shopping malls, and the like.
between the products and the user experience.
(1) Formulation of product ecosystems: The product
ecosystem formulation introduces a few important REFERENCES
concepts such as the ecosystem architecture, the [1] R.E. Clark, and F.Estes, “Cognitive task analysis”, Int. J.
ambience, and the event-driven user experience. By Edu. Res., vol. 25, no. 5, 1996, pp. 403–417.
integrating products, users, and the business process into a [2] H.M. Khalid, and M.G. Helander, “Customer emotional
coherent model, the product ecosystem expands the scope needs in product design”, Concur. Eng. Res. App., vol.14,
of traditional product and service process design, which no.3, 2006, pp. 197–206.
facilitates consistent design decisions. [3] J. Jiao, Q. Xu, J. Du, Y. Zhang, M. Helander, M. Khalid, P.
(2) Affective-cognitive modeling: With emphasis on Helo, and C. Ni, “Analytical affective design with ambient
the affective and cognitive states of the users, the system intelligence for mass customization and personalization”,
is expected to elicit pleasurable, consistent user Int. J. Flex. Manuf. Sys., vol. 19, 2007, pp. 570–595.
[4] C.V. Negoita, Expert Systems and Fuzzy Systems,
experience, and ultimately high customer satisfaction. At Benjamin/Cummings, Massachusetts, 1985.
the same time, the systematic viewpoint allows the [5] S. M. Chen, “A new approach to handling fuzzy decision-
designer to inspect the system development cost along making problems”, IEEE Trans.Sys. Man. Cyber., Vol.
with performance requirements of the system. As SMC-18, no. 6, 1988, pp. 1012–1016.
evidenced by the subway station case, the designed [6] H.J. Zimmermann, Fuzzy Set Theory and Its Applications,
system can achieve high performance on the one hand, Kluwer-Nijhoff Publishing, Boston, 1985.
and low cost on the other. [7] W. Pedrycz, Fuzzy Control and Fuzzy Systems (2nd ed.),
(3) CDFPN formalism: CDFPN suggests itself to be a Research Studies Press Ltd., Somerset, England, 1989.
powerful tool to model the complex, dynamic process of [8] A.B. Raposo, A.L.V. Coelho, L.P. Magalhaes, and I.L.M.
Ricarte, “Using fuzzy petri nets to coordinate collaborative
the user experience. First, the CDFPN is flexible in the activities”, in Proc. Joint 9th Int. Fuzzy Syst. Assoc. World
sense that the color scheme allows multiple rules to be Congress & 20th North American Fuzzy Information
handled in a single structure. Second, the CDFPN Processing Society Inte. Conf., Vancouver, Canada, IEEE
accommodates the dynamic, temporal properties of the Press, 2001, pp. 1494–1499.