Modelling of value creation based on Emergent ... - Semantic Scholar

CIRP Annals - Manufacturing Technology 57 (2008) 473–476

Contents lists available at ScienceDirect

CIRP Annals - Manufacturing Technology journal homepage: http://ees.elsevier.com/cirp/default.asp

Modelling of value creation based on Emergent Synthesis K. Ueda (1)*, T. Kito, T. Takenaka Research into Artifacts, Center for Engineering (RACE), The University of Tokyo, Kashiwa 277-8568, Japan

A R T I C L E I N F O

A B S T R A C T

Keywords: Emergent Synthesis Service Value creation

With increasing networking and globalization of the market, it becomes increasingly difficult to understand value in production of goods and services. This paper presents a new methodology for modelling value creation based on the concept of Emergent Synthesis. In consideration of interaction among producers, customers, and the environment, the methodology classifies value creation into three models: Providing Value, Adaptive Value, and Co-creative Value. This paper presents multi-agent system simulations of service market to examine the validity of the proposed models, with discussion of the diffusion of new products/services in a society. ß 2008 CIRP.

1. Introduction

2. Recent trends in manufacturing and service industries

Since the end of the twentieth century, many manufacturing companies have been rapidly shifting their attention to service businesses such as solution businesses, consulting businesses, and outsourcing businesses, apparently in response to the rapid increase of networking and market globalization. The commoditization of technologies forces manufacturing companies not only to survive through price competition but also to examine value chains of products, including those products’ use in society. For these reasons, recent manufacturing industries depend not only on technological innovations but also on service innovations. However, it becomes increasingly difficult to predict and control the value of products because products and consumers are involved dynamically in a global information society. A ‘‘de facto standard’’, for instance, is a typical phenomenon by which a certain technique (or standard) becomes dominant in the market in fact, but not by any law or regulation. In such a case, the value of products cannot be determined solely according to their functional dominance or economic advantage; their value becomes apparent through interaction among consumers, products, and producers. Under those increasing complexities of manufacturing environments, it is important to define manufacturing and service activities in terms of value creation. This paper presents a new methodology for modelling value creation based on the concept of Emergent Synthesis. First, we address recent trends in manufacturing and service industries and describe some academic interest. Then, we propose the methodology which classifies value creation into three models: Providing Value, Adaptive Value, and Co-creative Value. Finally, we introduce multi-agent system simulations to examine the values of services in a society.

2.1. Inseparability of manufacturing and service industries

* Corresponding author. 0007-8506/$ – see front matter ß 2008 CIRP. doi:10.1016/j.cirp.2008.03.014

In many developed countries, manufacturing industries’ share of the Gross Domestic Product has been decreasing since the 1980s along with the development of service industries. One reason for this phenomenon is increased outsourcing of operations such as information services, rental and leasing businesses, software development, etc. Therefore, a recent issue confronting manufacturing industries is how to expand their activities into service businesses to increase the value of their products. On the other hand, service industries are expected to increase their productivity because many existing services are thought to be provided less efficiently than manufactured products. For instance, the Japanese Ministry of Economy, Trade and Industry has investigated best practices of actual services and established a commission for academic–industrial co-operation with a view to increasing service-industry productivity [1]. The commission mainly addresses how service engineering can support actual service provision and can contribute to creation of new services. In the discussion, objectives and important technologies can be classified into some topics. Fig. 1 portrays some topics related to the discussion. Recent interest has been devoted to two completely opposite phrases: service-oriented manufacturing and manufacturingoriented service. Those phrases underscore the inseparability of products and services. The authors believe that both problems must be treated comprehensively from the viewpoint of value creation [2]. 2.2. Recent academic interest in manufacturing and service We overviewed precedent studies of manufacturing and services to elucidate the research trends of studies and underlying problems using the Web of Science database [3]. The number of publications including ‘‘manufacturing’’ or ‘‘manufacture’’ in any title or abstract was 74,550; ‘‘service’’ was 127,415 (search dates:

K. Ueda et al. / CIRP Annals - Manufacturing Technology 57 (2008) 473–476

474

Fig. 3. Classification of value models.

Fig. 1. Objectives and key technologies for service engineering.

completely, then the problem is described entirely. However, it is often difficult to find an optimal solution.  Class II – Problem with incomplete environment description: the specification is complete, but information related to the environment is incomplete. The problem is not described completely. Therefore, it is difficult to cope with dynamic properties of an unknown environment.  Class III – Problem with incomplete specification: the environment description and the specification are incomplete. Problemsolving must therefore start with an ambiguous purpose; human interaction becomes important. Emergent synthetic approaches can deal efficiently with those three problems. They include evolutionary computation, selforganization, reinforcement learning, multi-agent systems, and game theory. The authors have demonstrated the efficiency of the approaches in Biological Manufacturing Systems [7]. Traditional approaches such as analytic or deterministic ones might not be adequate to solve Class II and Class III problems. 3.2. Classification of value creation

Fig. 2. Number of articles including some keywords co-occurring with manufacturing ‘‘M’’ or service ‘‘S’’.

1991 to December 2007). Fig. 2 presents the number of articles including some technical keywords that co-occurred with ‘‘manufacturing/manufacture’’ or ‘‘service’’ during 1991–2006. ‘‘Value’’ is an important keyword that has rapidly increased in service studies from the mid-1990s and in manufacturing studies in recent years [4]. ‘‘Optimization’’ and ‘‘complexity’’ show characteristics of academic interest in the complexity of real-world problems. Moreover, ‘‘agent’’ and ‘‘adaptation’’ suggest that the researchers explore not only static optimized solutions but also dynamic and adaptive solutions for changing environments through interactions of service or manufacturing components as agents [5]. In another respect, problems behind those keywords derive from uncertainty related to the products/services environment or their specifications. Regarding design of a product or service, a designer cannot always know the environment completely. Furthermore, customers interact with product and service producers and other customers.

As discussed above, the respective values of products and services are characterized as Emergent Synthesis problems, by which value becomes apparent through interactions among decision-making agents. Fig. 3 presents three value models classified from the viewpoint of Emergent Synthesis: Providing Value Model, Adaptive Value Model, and Co-creative Value Model. As this figure shows, producers, customers, and products/services can be treated as agents. Figs. 4–6 show detailed value models based on the concept of Emergent Synthesis. In the Class I model, product and service producers as well as customers are defined independently of their values. The objectives and environment are clear. The model can be described completely using a closed system. However, in most cases, too many feasible solutions exist, which engenders combinatorial explosion and creates so-called NP-hard problems. Therefore, it is necessary to develop efficient and robust search methods to identify optimal solutions. In the real world, this model can apply to mass-produced products or routine services. In mass-production, a designer determines the specification of a product based on available information about the environment (e.g. consumers’

3. Modelling of value creation In this section, we present a new methodology for modelling value creation based on the concept of Emergent Synthesis. 3.1. Concept of Emergent Synthesis Design, in general, has synthetic aspects; the authors named such design processes ‘‘Emergent Synthesis’’ [6]. Within the concept of Emergent Synthesis, synthesis is classifiable into the following three classes with respect to the incompleteness of information about the environment and/or specifications.  Class I – Problem with complete description: if information related to the environment and specifications are given

Fig. 4. Class I model – Providing Value Model.

K. Ueda et al. / CIRP Annals - Manufacturing Technology 57 (2008) 473–476

475

services as agents which have their own purpose and selforganized internal structures. 4. Multi-agent simulation This section presents a multi-agent simulation of service market. The experiments are demonstrated using the market model of each class, Class I, II, or III, to clarify the differences of diffusion of new services and the variation of the producer’s and customers’ profits. 4.1. Service market model

Fig. 5. Class II model – Adaptive Value Model.

The service market model used in this simulation consists of one service producer and 400 customers. The producer produces services and each customer decides, based on its own value concept and demand, whether to use them. The detailed models of a service, customers, and a producer are described below. 4.1.1. Service Let each service, Sm, comprise two functions: Sm ¼ Sm ð f 1 f 2 Þ; where f1, f2 = 0, 1, 2. Here, fi (i = 1, 2) is the level of each function. The price of the service Sm is given as follows:   g1 ; (1) P m ¼ Sm G ¼ ð f 1 f 2 Þ

g2

where G(g1 g2) denotes the unit prices of f1 and f2.

Fig. 6. Class III model – Co-creative Value Model.

average demand or production costs) in advance. Consequently, the designer treats the information as a complete one. In a routine service such as a fast-food service, the service should be always provided in the same way. In the Class II model, the objective of the customer is defined completely. However, environments are changing and unpredictable. Therefore, the model is an open system. In our models, the environmental changes can occur in two types: changes attributed to customers (e.g., diversity of preferences or societal influence) and changes attributed to producers (e.g. changes of technologies and resources). In this class problem, approaches based on learning and adaptation, such as reinforcement learning or adaptive behaviour based methods, are feasible to resolve this class of problems. This model is applicable to customer-oriented products or services such as semi-order-made goods and recommendation systems of books based on collaborative filtering. Adaptive strategies are necessary to respond to the diversity of customers’ preferences. In the Class III model, along with the lack of environmental information in advance, the customer objectives are ambiguous. The producers and customers are mutually inseparable from the viewpoint of value creation. Consequently, the producers are involved mutually with customers to co-create the value. In the real world, open source software such as Linux, knowledge databases, and doctor–patient medical services might correspond to this kind of value model. In such cases, it is usually difficult to control the value that emerges through the interaction between producers and customers. A de facto standard can also be treated as a Co-creative Value. In such a case, network externalities can play important roles. Network externalities are defined as externalities by which a consumer’s utility depends on the number of users who consume the same product [8]. From the viewpoint of synthesis, the authors believe that one solution to create and control the Class III values is to treat products or

4.1.2. Customer Each customer Cn(n = 1, 2, . . ., 400) has a demand level Dn(dn1 dn2) (dn1, dn2 = 0, 1, 2) and her or his own reservation value Vn(Vn1 Vn2), which expresses how much (s)he can pay for one level of each function. The initial value of Dn is set as (1 0), (1 1), (1 2), or (2 1) in each of four groups of 100 customers. RPn is the reservation price, i.e., the price Cn is willing to pay for the service, given by:   v (2) RPn ¼ Dn V n ¼ ðdn1 dn2 Þ n1 : vn2 Customer Cn makes a decision of whether to buy and use the service Sm or not. When it uses Sm, it gains the utility Un = RPn  Pm. 4.1.3. Producer The producer produces services with the unit cost T(t1 t2). The profit the producer gains when she provides Sm is defined as X P ¼ ðPm  Sm TÞ  Nm (3) m

Here, Nm is the number of those customers who are using Sm. The producer makes the decision as to which service to produce at every step, with the intention of producing a service whereby the total profit is maximized. The market is modelled as a learning system; the service by which the producer’s profit is maximized would be achieved through interactions among agents at every step, based on the SLA algorithm. Three cases are demonstrated as described below. (Case I) Dn and Vn are fixed during a simulation. This case is expected to correspond to the Class I model. (Case II) Vn is stochastically variable at each trial (average rate of change is 20%). This case might correspond to the Class II model. (Case III) Vn reflects network externality. In this simulation, only v2 is modelled as a function which can be influenced by other customers: the more customers use a service, the more highly the customer evaluates it: n N vn2 ¼ vn2 þ b  (4) N

K. Ueda et al. / CIRP Annals - Manufacturing Technology 57 (2008) 473–476

476

ment such as Case I, a mass-production strategy might secure the service producer’s profit, while it often undermines consumers’ profit. Moreover, as in Case II, the producer’s adaptive strategy to service environment changes might enhance customers’ profit. Furthermore, as in Case III, a win–win like relationship between the producers and customers can sometimes be achieved through dynamic interaction among customers, producers and services. The mechanisms underlying a ‘‘de facto standard’’ or ‘‘Brand value’’ are probably also related to factors such as network externalities like those demonstrated in Case III. 5. Concluding remarks

Fig. 7. Transition of producer’s profit during simulation.

Fig. 8. Averaged producer’s profit, consumers’ profit and their total profit during the last 100 trials.

In that equation, nN/N represents the percentage of customers who use the same service as Cn, of customers in the market. This case is expected to correspond to the Class III model.

This paper presents a new methodology for modelling value creation based on the concept of Emergent Synthesis. First, it describes recent trends and problems in manufacturing and service industries, and analyzes academic interest. Then, it proposes a methodology which classifies value creation into three models: Providing Value, Adaptive Value, and Co-creative Value. It presents multi-agent system simulations of a service market to examine the validity of the proposed three models. The simulation results show the variation of profits of the producer and customers. The producer can gain the greatest profit under a predictable service environment in the Class I model. The adaptation to changes of each customer’s value perception results in increasing customers’ profit in the Class II model. For the Class III model, the total profit of the producer and customers is highest of all the cases. This occurs due to the increased value perception by the customers because of the network externalities. Recent concerns in manufacturing and services are minimizing the cost and maximizing the value. The problems of minimizing costs under a consistent and predictable environment can be treated as a Class I value creation problem. The same problems under changing and unpredictable environments could be treated as a Class II value creation problem. For the purpose of maximizing the value, it is important to pay attention to the Class III value creation problem. Co-creation is a promising concept not only to enhance the value but also to create a new value in a society.

4.2. Results and discussion Acknowledgement Fig. 7 depicts the transition of the producer’s profit for each case during the learning process. This figure indicated that, after a sufficient number of trials, the system successfully learned to select a service by which the producer can gain the greatest profit. In this case, 200 customers use the service. In Case II, although the same service can be selected through a learning process, the producer’s profit is less than that of Case I, according to the decrease of the number of the service users (average users of the last 100 trials is approx. 180). However, as Fig. 8 shows, the customers’ averaged profit is higher than in Case I, relating to changes of each customer’s reservation value. In Case III, the profit of the producer is higher than that of Case II, corresponding to the increased number of service users (average users of the last 100 trials are approx. 193). Moreover, the customers’ profit is higher than in other cases. This phenomenon can occur with increased value perception by customers because of network externalities (influences of other customers). Consequently, the total profit of the producer and customers is the highest of all the cases. Although the cases of the introduced simulations are simple and limited, their results can qualitatively explain real-world business activities. In a statistic and predictable service environ-

The authors express their gratitude to Mr. Kosuke Fujita, the University of Tokyo for his support in the simulation.

References [1] Japanese Ministry of Economy, Trade and Industry, http://www.meti.go.jp/ english/report/downloadfiles/0707ServiceIndustries.pdf/. [2] Ueda K, Takenaka T (2007) Classification of Service Model and Value Creation. Proceedings of the 9th IEEE International Conference on E-Commerce Technology, Tokyo, Japan, 497–498. [3] Web of Science, Thomson Scientific, Philadelphia, PA, http://scientific.thomson.com/. [4] Arai T, Shimomura Y (2005) Service CAD System; Evaluation and quantification. Annals of the CIRP 54(1):463–466. [5] Kwon OB (2006) Multi-agent System Approach to Context-aware Coordinated Web Services Under General Market Mechanism. Decision Support Systems 41(2):380–399. [6] Ueda K (2000) Emergent Synthesis. Artificial Intelligence in Engineering 15(4):319–320. [7] Ueda K, Markus A, Monostori L, Kals HJJ, Arai T (2001) Emergent Synthesis Methodologies for Manufacturing. Annals of the CIRP 50(2):535–551. [8] Katz ML, Shapiro C (1992) Network effects, Software provision, and Standardization. The Journal of Industrial Economics 40(1):85–103.