System on chip design service e-business value maximization through a novel MCDM framework

(1)

System on chip design service e-business value maximization through a novel

MCDM framework

Chi-Yo Huang

a,⇑

_{, Gwo-Hshiung Tzeng}

b,c

_{, Wen-Rong Jerry Ho}

d a

Department of Industrial Education, National Taiwan Normal University, 162, He-ping East Road, Section 1, Taipei 106, Taiwan

b

Institute of Management of Technology, National Chiao Tung University, 1001, Ta-Hsueh Road, Hsinchu 300, Taiwan

c

Department of Business Administration, Kainan University, No. 1 Kainan Road, Luchu, Taoyuan 338, Taiwan

d

Department of Banking and Finance, Chinese Culture University, 55, Hwa-Kang Road, Yang-Ming-Shan, Taipei 111, Taiwan

a r t i c l e

i n f o

Keywords: Innovation e-Business model e-Commerce

Analytic Network Process (ANP) Silicon intellectual property (SIP) Integrated circuit (IC)

Semiconductor Design service

a b s t r a c t

As the IC (integrated circuit) industry migrates to the System-on-Chip (SOC) era, the SOC design service industry is emerging. Meanwhile, in the past decade, the emergence of Internet has changed the high technology marketing approaches while e-commerce has already become one of the most efficient mar-keting channels. Thus, most leading SOC design service firms tried to leverage novel e-commerce busi-ness models to provide better services including online silicon intellectual property (SIP) sourcing, transactions, integration, etc. to assist customers in enhancing their innovation competences to shorten their time to market and thus, time to money. However, defining appropriate e-business models for com-mercializing new SIPs or SOC design services is not easy for both aspects of technology as well as business development. On one hand, from the aspect of technology, the technical site R&D or production managers are familiar with SOC technologies while do not really understand the needs of customers’ over the Inter-net. On the other hand, from the aspect of business development, the sales or marketing managers may be familiar with online customers’ needs, wants as well as demands while are unfamiliar to SOC technol-ogy developments. To overcome the above mentioned problems, an appropriate e-business model defi-nition framework can overcome this cognitive gap and maximize the value-added of online SOC design services. In this paper, a novel analytic framework based on the concept of design service custom-ers’ competence set expansions by leveraging high technology service firms’ capabilities and resources as well as novel multiple criteria decision making (MCDM) techniques, will be proposed. The framework being proposed can be leveraged by the design service firms to define an appropriate e-business model for possible SIP or design service businesses. Based on the proposed MCDM framework, an empirical study of an SIP commercialization e-business model definition inside an SIP Mall, an SIP e-commerce mechanism being operated by a SOC design service firm, will be provided for verifying the effectiveness of this novel analytic framework. The feasibility of the proposed framework in the real world can be ver-ified by the empirical study. In the future, the novel MCDM framework can be applied to novel e-business model definitions in the SOC design service or other high technology industries.

1. Introduction

In the past decade, the Internet has become an enabling tech-nology in almost any industry and as part of almost any strategy (Porter, 2001). Uses of the Internet and e-commerce have con-verted the traditional way of running business and have thor-oughly changed the channel of enterprise transactions (Shaw, Gardner, & Thomas, 1997). As industries in general, and high tech-nology industries in special, are being reshaped and the nature of competition changes (Raisinghani, Meade, & Schkade, 2007),

decid-ing on an e-business model, a competition strategy for the market-place and a structure of business processes for the entire electronic trade course (Wang, 2001), has become daily important for modern high technology ﬁrms.

As stated byYoung and Johnston (2003), there are a number of traditional business strategy theories that have been used to dis-cuss business-to-business (B2B) e-commerce strategies: transac-tion cost economics, resource-based view, Porter’s market forces theory, and channel theory. However, there currently exists no comprehensive framework linking these theories into a method to rigorously assess value delivery strategies, and in particular to determine how to maximize the impact of the Internet as a value delivery channel (Young & Johnston, 2003). Raisinghani et al. (2007) also mentioned that although the strategy to rebuild a

⇑ Corresponding author. Tel.: +886 936 516698; fax: +886 3 5165185. E-mail address:[email protected](C.-Y. Huang).

Contents lists available atScienceDirect

Expert Systems with Applications

j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / e s w a

(2)

robust e-business model has not been as widely implemented as had been anticipated, it has had a signiﬁcant inﬂuence on company performance.

Moreover, as the IC (integrated circuit) industry migrates to the System-on-Chip (SOC) era, a novel business model, the SOC design services, is emerging. Now, when the Internet is emerging while e-commerce has already become one of the most efficient marketing channels, most leading design service firms tried to leverage novel e-commerce business models to provide better services. The novel design services include online silicon intellectual property (SIP) sourcing, transactions, integration, etc. to assist customers of SOC design services in enhancing their innovation competences (INCs). The INCs, critical capabilities as well as resources for commercial-izing SOCs or SIPs, can shorten the time to market and thus, time to money, of both SOC design firms’ and design service customers’. However, defining appropriate SOC design service e-business mod-els are not easy. On the technical site, R&D or production managers of the SOC design service firms are familiar with SOC technologies while do not really understand the needs of customers’ over the Internet. On the business site, sales or marketing managers of the SOC design service firms are familiar with online customers’ needs, wants as well as demands while are unfamiliar to SOC technology developments. Thus, how to establish a decision support frame-work for defining appropriate SOC design service e-business mod-els to commercialize new SOC/SIP products or services has already become one of the most critical issues for the SOC design service firms. Meanwhile, the proposed framework can also enhance insuf-ficient linkages between traditional business strategy theories and value delivery strategies through the Internet which was men-tioned byYoung and Johnston (2003)as well asRaisinghani et al. (2007).

Therefore, this research aims to establish a novel multiple crite-ria decision making (MCDM) framework, which intends to link be-tween INCs being required by novel SOC design services. The proposed MCDM framework intends to maximize values of both the SOC design service e-business models and end SOC products of SOC design service customers’. The INCs, or evaluation criteria, are ﬁrst summarized using the Delphi method. Then, the relation-ships between the criteria will be derived by DEMATEL (Decision Making Trial and Evaluation Laboratory). The weights of each crite-rion versus the goal of the MCDM problem, maximizing the values of SOC design services e-business models, then will be derived based on the structure of the decision problem by using the Ana-lytic Network Process (ANP). After the criteria are derived, the rela-tionships between the INCs (criteria) as well as e-business models will be derived by using the grey relational analysis (GRA) based on the weights of each criterion being derived by the ANP. Finally, the most appropriate e-business model with the highest grey grades which may compensate the current INCs of SOC design service e-business customers and maximize the value of customer’s prod-ucts and thus, the value of the high technology e-commerce chan-nel, will be selected.

A case study on commercializing a silicon intellectual property (SIP) being developed by an IC design house through an SIP Mall, a web based SIP e-commerce mechanism, being operated by an SOC design service firm will be used for demonstrating the effective-ness of the novel MCDM method. The case study results demon-strated that the IP commercialization model of the SIP Mall which may assist a small-scale IC design house without enough resources to commercialize its SIP products and maximize the SIP value through SIP verification, qualification, marketing, sales, technical supports, etc. will be the most appropriate e-business model to maximize the value of this SIP.

The remainder of this paper is organized as follows: In Section2, the concepts of innovation, INCs, e-business models, e-business model evaluation and INC set expansion are introduced. In

Section3, an analytic framework and methods are proposed for constructing the evaluation criteria and e-business models’ deﬁni-tions. The background of the SOC design service, SIP, SIP market and SIP e-business models will be described in Section4. Then in Section5, a case study follows, deﬁning an e-business model for commercializing an SIP being developed by an IC design house which is in lack of SIP commercialization resources. Discussion will be presented in Section6. Section7will conclude the whole article with observations, conclusions and recommendations for further study.

2. Innovation competence and e-business model assessment Researchers have successfully explored the deﬁnitions of busi-ness models, e-busibusi-ness models, e-busibusi-ness model evaluation as well as innovation, INC, INC set expansion and resource based view. In the following section, the related literature will be reviewed.

2.1. Business model and e-business

The term business model is widely used in business literature.

Lumpkin and Dess (2004) deﬁned business model as a method

and set of assumptions that explains how a business creates value and earns profits in a competitive environment.Tsalgatidou and Pitoura (2001)defined the business model as a logical architecture for product, service, and information flows, including a description of the involved business actors and their roles, as well as sources of revenue.Wise and Baumgartner (1999)mentioned that business models are cases or scenarios.Moore (2003) mentioned that a business model is categorically, a way of making money, the form an offer takes, and the manner in which it is paid for.

Externally, a business model is an implicit contract which a cus-tomer expects and a vendor commits to. Internally, a business model is a platform for execution, a basis for prioritization and trade-offs, and an infrastructure for resource commitments. Final-ly, a business model is the methods of doing business by which a company can sustain itself, that is, generate revenue (Moore, 2003). The basic categories of business models include brokerage, advertising, infomediary, merchant, manufacturer, afﬁliate, com-munity, subscription, and utility (Rappa, 1998).

According toAfuah (2004), a business model is a framework for making money.Chesbrough (2006)also defined the business mod-el to be a useful framework to link ideas and technologies to eco-nomic outcomes. It also has value in understanding how companies of all sizes can convert the technological potential value (Chesbrough, 2006). It is the set of activities which a firm performs, how it performs them, and when it performs them so as to offer its customer benefits they want and to earn a profit. Business models are usually represented by a mixture of informal textual, verbal, and ad hoc graphical representations (Gordijn & Akkermans, 2001). Young and Johnston (2003) defined the strategic options for delivering values from suppliers to customers as specific business models that outline the essential details of how an organization can deliver value to a target customer (Seddon & Lewis, 2003). These business models are a key component to an overall strategy that determines the long-term position of the organization (Porter, 1996, 2001; Young & Johnston, 2003).

As stated byAfuah (2004), a firm makes more money than its rivals if its business model creates and offers superior customer va-lue and positions the firm to appropriate the vava-lue. To perform the activities that enable a firm to offer superior customer value and appropriate the value, a firm needs resources. Resources in and of themselves do not, however, produce customer value and prof-its. Firms must also have the ability or capacity to turn resources

(3)

into customer value and proﬁts. In summary, resources and capa-bilities are the roots of business models.

The rapid deployment of electronic business (e-business) is an economically signiﬁcant issue for today’s business (Raisinghani et al., 2007). An e-business solution is deﬁned as: (1) improving business processes using Internet technologies; (2) leveraging the Web to bring together customers, vendors, suppliers, and employ-ees in ways never before possible; and (3) Web-enabling a business to sell products, improve customer service, and obtain maximum results from limited resources (Turban, King, Viehland, & Lee, 2006).

Wang (2001)defined an e-business model to be a competition strategy for the marketplace and structure of business processes for the entire electronic trade course including marketing and advertising, negotiation, purchasing, logistics of products and/or service delivery, payment with the means of security, post-sales service, and post-sales intelligence. As stated byRaisinghani et al. (2007), although the strategy to rebuild a robust e-business model has not been as widely implemented as had been anticipated, it has had a significant influence on company performance. Meanwhile, Hooshang, Salehi-Sangari, and Engstrom (2006)found that inap-propriate plan and design of an e-business model to be one of the most significant problems causing dissatisfaction with e-business.

Researchers and practitioners use a variety of diagrammatic conceptual models to illustrate new e-business models (Wang, 2001). After creating detailed models, the next step is to evaluate the economic feasibility of an idea in quantitative terms that are based on an assessment of the value of objects for all actors volved. Feasibility of an e-business model means that all actors in-volved can make a proﬁt or increase their economic utility (Gordijn & Akkermans, 2001).

2.2. e-Business model assessment

As advised byGordijn and Akkermans (2001), when developing a speciﬁc business idea, we instantiate our ontology concepts and relations for speciﬁc use cases. Doing so provides a basis for ana-lyzing the characteristics and implications of alternative e-busi-ness models (Gordijn & Akkermans, 2001).

For the development of e-business information systems, three distinct perspectives are important: the value viewpoint repre-sents the way economic value is created, exchanged, and

con-sumed in a multi-actor network; the process viewpoint

represents the value viewpoint in terms of business processes; and the system architecture viewpoint represents the information systems that enable and support e-business processes (Gordijn & Akkermans, 2001).

Moreover, as advised byRaisinghani et al. (2007), a systematic framework for the identification and classification of e-commerce strategy using Internet information, communication, distribution, or transaction channels including: (1) the virtual information space (VIS), which presents new channels for economic agents to display and access-related company product and services information (e.g., marketing and advertising); (2) the virtual communication space (VCS), which includes strategies aimed at monitoring and influenc-ing business-related communications between economic agents operating on the Internet (e.g., negotiations between potential and existing customers partners, government agencies, and com-petitors); (3) the virtual distribution space (VDS), which provides new channels for economic agents to distribute products and ser-vices (e.g., software); and lastly, the virtual transaction space (VTS) provides strategies for economic agents with which to initiate and execute business to business (B2B) or business-to-customer (B2C) transactions, such as orders and payments, are used to access the e-business models.

Young and Johnston (2003)also found that there are a number of traditional business strategy theories that have been used to dis-cuss business-to-business (B2B) e-commerce strategy: Transaction Cost Economics, Resource-Based View, Porter’s Market Forces The-ory, and Channel Theory. However, there currently exists no com-prehensive framework linking these theories into a method to rigorously assess value delivery strategies, and in particular to determine how to maximize the impact of the Internet as a value delivery channel (Young & Johnston, 2003).

2.3. Innovation and INC

Innovation is combinations of knowledge that result in new products, processes, input and output markets, or organizations (Sundbo, 2003) which include not only technical innovations, but also organizational and managerial innovations, new markets, new sources of supply, ﬁnancial innovations, and new combina-tions (Perlman & Heertje, 1991). The literature distinguishes differ-ent types of innovation: incremdiffer-ental, radical, technological, process, product, organizational, operational, managerial, social, or institutional (e.g. Hammer, 2004; Nadler & Tushman, 1999; van Kleef & Roome, 2007). Innovation processes can therefore, be viewed as a sequence of exploration (in which existing products and processes are adapted incrementally or radically through the search for and application of new assets) and exploitation (in which the variety of products and processes decreases while their efﬁciency increases) (van Kleef & Roome, 2007).

Clark and Guy (1998)mentioned that innovation is a critical factor in enhancing a firm’s competitiveness which is generally understood to refer to the ability of a firm to increase in size, mar-ket share and profitability at the firm level (Clark & Guy, 1998). Nooteboom (2000)also argued that innovation is not necessarily related to problem solving, but it is usually related to improving competitiveness and economic success, and it is often pushed by technology. In traditional economic theory, comparative costs of production determine relative competitiveness at firm level – the way to become more competitive is to produce more cheaply: for example, by finding ways to reduce labor costs (Clark & Guy, 1998). However, recent studies have consistently pointed to non-price factors, e.g. human resource endowments, technical factors, managerial and organizational factors, which determine the ability of a firm to attain and maintain a profitable position in the face of changing technological, economic and social environments (Clark & Guy, 1998).

In a fast-changing environment, the competitive advantage or competitiveness of many companies is based on the decision to ex-ploit, to develop the power of knowledge development (Carneiro, 2000). Since knowledge underpins a firm’s ability to offer products, a change in knowledge implies a change in the firm’s ability to of-fer a new product (Afuah, 1998). Thus, based on the extent to which innovation impacts a firm’s capabilities, an innovation can be categorized as radical or incremental (Afuah, 1998). A radical innovation requires the technological knowledge which is very dif-ferent from existing knowledge, rendering existing knowledge ob-solete (Afuah, 1998). Such innovations are said to be competence destroying (Afuah, 1998). A company can decide its competitive advantage or competitiveness as a function of the capability to generate radical change in its processes and technologies and of the flexibility to adapt its resources to the strategic formulation (Carneiro, 2000). If an organization decides to become a fast inno-vator, managers should co-ordinate the ability to formulate a com-petitive strategy and to build advantages against competitors. This ability depends on the capacity of speeding up creative operations to generate innovations (Carneiro, 2000; Page, 1993). Thus, compa-nies strengthen their competence to innovate by developing the

(4)

capabilities of employees within the organization (Hargadon & Sutton, 2000).

Competence is deﬁned afterPrahalad and Hamel (1990), as the learning within the organization how to coordinate diverse pro-duction skills and how to integrate technologies. Capabilities of employees, combined with each other in teams and connected through structures and routines, are the building blocks of compe-tence (van Kleef & Roome, 2007). Competence, therefore, includes the organization of work, the involvement of employees, the com-mitment to working and communicating across boundaries within the organization, and the delivery of value to customers and other stakeholders (van Kleef & Roome, 2007). Competence is seen as the basis of competitiveness; it enables a company to offer products and services of value to customers and to innovate to generate new products and services, while adapting to changing circum-stances faster than competitors (van Kleef & Roome, 2007). 2.4. Competence set expansion

According toYu and Zhang (1989, 1993), Yu (2002), Hu (2003), Huang (2006), for each decision problem (e.g. job selection, corpo-rate stcorpo-rategic definition, conflict resolution, etc.), a competence set consisting of ideas, knowledge, information and skills for its satis-factory solution exists. Companies have to expand its competence set to deepen their knowledge base in their core technologies and to stay ahead of the competition in the current markets (Vanhaverbeke & Peeters, 2005) by investing in research and devel-opment and external technology sourcing (Chesbrough, 2003; Keil, 2002). This challenge derives largely from the fact that the decision to develop new businesses creates a fruitful misfit between the existing competencies and those that are required (Vanhaverbeke & Peeters, 2005). A well-known classification of INCs and search modes is that which distinguishes between exploration and exploi-tation (March, 1991). Tushman and Nadler (1996) are among the earliest authors on innovation to write explicitly about the capabil-ities that contribute to the competence to innovate for purely competitive reasons.van Kleef and Roome (2007)summarized the capabilities that are embedded in the work of leading authors in the field of innovation. Their work mentioned that those capabilities

relate to systems thinking, learning, combining, and integrating, thinking inventively, networking, and coalition building.

Finally, as stated by Young and Johnston (2003), there are a number of traditional business strategy theories that have been used to discuss business-to-business (B2B) e-commerce strategy: Transaction Cost Economics, Resource-Based View, Porter’s Market Forces Theory, and Channel Theory. However, there currently ex-ists no comprehensive framework linking these theories into a method to rigorously assess value delivery strategies, and in partic-ular to determine how to maximize the impact of the Internet as a value delivery channel (Young & Johnston, 2003).

3. Analytic framework and MCDM-based methods for an innovative e-business model deﬁnition

The analytical process for deﬁning e-business models is initi-ated by collecting the INCs needed to develop design service pro-viders’ INC using the Delphi method. Since any INCs to be derived by the Delphi may impact each other, the structure as well as the priorities of every INC will be derived by the ANP. Finally, the GRA will be applied to get the correlation between the INCs and the available e-business models. Based on the grey grades to be derived by the GRA, the most suitable e-business model will be derived. In summary, this evaluation framework (Fig. 1) consists of ﬁve main phases: (1) establishing INCs using the Delphi method; (2) building the structure between INCs by using the DEMATEL; (3) deriving the weights versus each INC by using the ANP; (4) corre-lating the INCs with the available e-business models by using the GRA; (5) deciding the e-business model based on the grey grades versus each e-business model.

3.1. Delphic Oracle’s skills of interpretation and foresight

The Delphi method originated in a series of studies conducted by the RAND Corporation in the 1950s (Jones & Hunter, 1995). The objective was to develop a technique to obtain the most reliable con-sensus from a group of experts (Dalkey & Helmer, 1963). While researchers have developed variations of the method since its introduction, Linstone and Turoff (1975) captured its common

Define an

e-Business Model Define INCs

Establish a Structure of the Decision Problem

Delphi DEMATEL GRA

Correlate INCs and e -Business

Models Derive Weights

versus Each INC ANP

Select Available e-Business

Models

Select the Best e-Business Model

(5)

characteristics in the following descriptions: Delphi may be charac-terized as a method for structuring a group communication process; so the process is effective in allowing a group of individuals, as a whole, to deal with a complex problem. To accomplish this ‘struc-tured communication’, certain aspects should be provided: some feedback of individual contributions of information and knowledge; some assessment of the group judgement or viewpoint; some opportunity for individuals to revise their views; and some degree of anonymity for individual responses (Linstone & Turoff, 1975). The Delphi technique enables a large group of experts to be surveyed cheaply, usually by mail using a self-administered questionnaire (although computer communications also have been used), with few geographical limitations on the sample. Speciﬁc situations have included a round in which the participants meet to discuss the process and resolve any uncertainties or ambiguities in the wording of the questionnaire (Jones & Hunter, 1995). The Delphi method proceeds in a series of communication rounds, as follows:

Round 1: Either the relevant individuals are invited to provide opinions on a speciﬁc matter, based upon their knowl-edge and experience, or the team undertaking the Del-phi expresses opinions on a speciﬁc matter and selects suitable experts to participate in subsequent question-naire rounds; these opinions are grouped together under a limited number of headings, and statements are drafted for circulation to all participants through a questionnaire (Jones & Hunter, 1995).

Round 2: Participants rank their agreement with each statement in the questionnaire; the rankings then are summa-rized and included in a repeat version of the question-naire (Jones & Hunter, 1995).

Round 3: Participants re-rank their agreement with each state-ment in the questionnaire, and have the opportunity to change their score, in view of the group’s response; the re-rankings are summarized and assessed for their degree of consensus: if an acceptable degree of consen-sus is obtained, the process may cease, with the ﬁnal results then fed back to the participants; if not, this third round is repeated (Jones & Hunter, 1995). 3.2. DEMATEL method

The DEMATEL method was developed by the Battelle Geneva Institute (1) to analyze complex ‘world problems’ dealing mainly with interactive man-model techniques; and (2) to evaluate qual-itative and factor-linked aspects of societal problems (Gabus & Fontela, 1972). The applicability of the method is widespread, ranging from industrial planning and decision-making to urban planning and design, regional environmental assessment, analysis of world problems, and so forth. It has also been successfully applied in many situations, such as marketing strategies, control systems, safety problems, developing the competencies of global managers and group decision-making (Chiu, Chen, Tzeng, & Shyu, 2006; Liou, Tzeng, & Chang, 2007; Wu & Lee, 2007; Lin & Wu, 2008). Furthermore, a hybrid model combining the two methods has been widely used in various fields, for example, e-learning evaluation (Tzeng, Chiang, & Li, 2007), airline safety measurement (Liou et al., 2007), and innovation policy portfolios for Taiwan’s SIP Mall (Huang, Shyu, & Tzeng, 2007). Therefore, in this paper we use DEMATEL not only to detect complex relationships and build a network relation map of the criteria, but also to obtain the influence levels of each element over others; we then adopt these influence level values as the basis of the normalization supermatrix for determining ANP weights to obtain the relative importance. To apply the DEMATEL method smoothly, the authors refined the definitions based on above authors and the ones by

Hori and Shimizu (1999), and produced the essential deﬁnitions indicated below.

The DEMATEL method is based upon graph theory, enabling us to plan and solve problems visually, so that we may divide multi-ple criteria into a relationship of cause and effect group, in order to better understand causal relationships. Directed graphs (also called digraphs) are more useful than directionless graphs, because di-graphs will demonstrate the directed relationships of sub-systems. A digraph typically represents a communication network, or a domination relationship between individuals, etc. Suppose a sys-tem contains a set of elements,S = {s1, s2, . . . , sn}, and particular

pair-wise relationships are determined for modeling, with respect to a mathematical relationship, MR. Next, portray the relationship MR as a direct-relation matrix that is indexed equally in both dimensions by elements from the set S. Then, extract the case for which the number 0 appears in the cell (i, j), if the entry is a posi-tive integral that has the meaning of:

the ordered pair (si, sj) is in the relationship MR;

it has the kind of relationship regarding that element such that-sicauses element sj.

The digraph portrays a contextual relationship between the ele-ments of the system, in which a numeral represents the strength of inﬂuence (Fig. 2). The elementss1, s2, s3and s4represent the factors

that have relationships inFig. 2. The number between factors is inﬂuence or inﬂuenced degree. For example, an arrow from s1to

s2represents the fact that s1inﬂuences s2and its inﬂuenced degree

is two. The DEMATEL method can convert the relationship between the causes and effects of criteria into an intelligible structural mod-el of the system (Chiu et al., 2006).

Definition 1. The pair-wise comparison scale may be designated as 11 levels, where the scores 0, 1, 2, . . . , 10 represent the range from ‘no influence’ to ‘very high influence’.

Definition 2. The initial direct relation/influence matrix A is an n n matrix obtained by pair-wise comparisons, in terms of influ-ences and directions between the INCs, in which aijis denoted as

the degree to which the ith INC affects the jth INC.

A ¼ a11 a12 a1n a21 a22 a2n .. . .. . .. . .. . an1 an2 ann 2 6 6 6 6 4 3 7 7 7 7 5:

Deﬁnition 3. The normalized direct relation/inﬂuence matrix N can be obtained through Eqs. (1) and (2), in which all principal diagonal elements are equal to zero

N ¼ zA; ð1Þ where s₂ 2 3 1 3 s₁ s₄ s3 s₂ 2 3 1 3 s₁ s₄ s3

(6)

z ¼ 1= max

16i6n

Xn

j¼1aij: ð2Þ

In this case, N is called the normalized matrix. Since

lim

k!1N k

¼ ½0_nn:

Deﬁnition 4. Then, the total relationship matrix T can be obtained using Eq.(3), where I stands for the identity matrix.

T ¼ N þ N2þ þ Nk

¼ NðI þ N þ N2þ þ Nk1ÞðI NÞðI NÞ1;

) T ¼ NðI NÞ1; ð3Þ

where k ? 1 and T is a total inﬂuence-related matrix; N is a direct inﬂuence matrix and N = [xij]nn; limk!1 N2þ þ Nk

stands for a indirect inﬂuence matrix and 0 6Pn

j¼1xij<1 or 0 6Pni¼1xij<1, and

only one or somePn_j¼1xijorPni¼1xijequal to 1 for "i, j, but not all. So

limk?1Nk= [0]nn.

The (i, j) element tijof matrix T denotes the direct and indirect

inﬂuences of factor i on factor j.

Deﬁnition 5. The row and column sums are separately denoted as r and c within the total-relation matrix T through Eqs.(4)–(6)

T ¼ ½tij; i; j 2 1; 2; . . . ; nf gm ð4Þ r ¼ ½rin1¼ Xn j¼1 tij " # n1 ; ð5Þ c ¼ ½cj_1n¼ Xn i¼1 tij " # 1n ; ð6Þ

where the r and c vectors denote the sums of the rows and columns, respectively.

Deﬁnition 6. Suppose ridenotes the row sum of the ith row of

matrix T. Then, riis the sum of the inﬂuences dispatching from

fac-tor i to the other facfac-tors, both directly and indirectly. Suppose that cjdenotes the column sum of thejth column of matrix T. Then, cjis

the sum of the inﬂuences that factor i is receiving from the other factors. Furthermore, when i = j (i.e., the sum of the row sum and the column sum (ri+ ci) represents the index representing the

strength of the inﬂuence, both dispatching and receiving), (ri+ ci)

is the degree of the central role that factor i plays in the problem. If (ri ci) is positive, then factor i primarily is dispatching inﬂuence

upon the strength of other factors; and if (ri ci) is negative, then

factor i primarily is receiving inﬂuence from other factors (Tamura, Nagata, & Akazawa, 2002).

3.3. The ANP method

The ANP method, a multi-criteria theory of measurement devel-oped bySaaty (1996), provides a general framework to deal with decisions without making assumptions about the independence of higher-level elements from lower level elements and about the independence of the elements within a level as in a hierarchy (Saaty, 2005). Compared with traditional MCDM methods, e.g. AHP (Analytic Hierarchy Process), which usually assume the inde-pendence between criteria, ANP, a new theory that extends AHP to deal with dependence in feedback and utilizes the supermatrix ap-proach (Saaty, 2003), is a more reasonable tool for dealing with complex MCDM problems in the real world. In this section, con-cepts of the ANP are summarized based on Saaty’s earlier works (Saaty, 1996, 1999, 2004, 2005).

The ANP is a coupling of two parts. The first consists of a control hierarchy or network of criteria and subcriteria that control the interactions. The second is a network of influences among the ele-ments and clusters. The network varies from criterion to criterion and a different supermatrix of limiting influence is computed for each control criterion. Finally, each of these supermatrices is weighted by the priority of its control criterion and the results are synthesized through the addition for all the control criteria (Saaty, 1999, 2004).

A control hierarchy is a hierarchy of criteria and subcriteria for which priorities are derived in the usual way with respect to the goal of the system being considered. The criteria are used to com-pare the components of a system, and the subcriteria are used to compare the elements. The criteria with respect to which influ-ence is presented in individual supermatrices are called control criteria. Because all such influences obtained from the limits of the several supermatrices will be combined in order to obtain a measure of the priority of overall influences, the control criteria should be grouped in a structure to be used to derive priorities for them. These priorities will be used to weight the correspond-ing individual supermatrix limits and add. Analysis of priorities in a system can be thought of in terms of a control hierarchy with dependence among its bottom-level alternatives arranged as a network as shown in Fig. 3. Dependence can occur within the components and between them. A control hierarchy at the top may be replaced by a control network with dependence among its components, which are collections of elements whose func-tions derive from the synergy of their interaction and hence has a higher-order function not found in any single element. The cri-teria in the control hierarchy that are used for comparing the components are usually the major parent criteria whose subcrite-ria are used to compare the elements need to be more general than those of the elements because of the greater complexity of the components.

A network connects the components of a decision system. According to size, there will be a system that is made up of sub-systems, with each subsystem made up of components, and each component made up of elements. The elements in each compo-nent interact or have an inﬂuence on some or all of the elements of another component with respect to a property governing the interactions of the entire system, such as energy, capital, or polit-ical inﬂuence.Fig. 4demonstrates a typical network. Those com-ponents which no arrow enters are known as source comcom-ponents such as C1and C2. Those from which no arrow leaves are known

as a sink component such as C5. Those components which arrows

both enter and exit leave are known as transient components such as C3 and C4. In addition, C3 and C4 form a cycle of two

components because they feed back and forth into each other. C2 and C4 have loops that connect them to themselves and are

inner dependent. All other connections represent dependence between components, which are thus known to be outer dependent.

A component of a decision network which was derived by the DEMATEL method in Section3.2will be denoted by Ch, h = 1, . . . , m,

and assume that it has nh elements (INCs), which we denote by

eh1, eh2, . . . , ehm. The inﬂuences of a given set of elements (INCs)

in a component on any element in the decision system are repre-sented by a ratio scale priority vector derived from paired com-parisons of the comparative importance of one criterion and another criterion with respect to the interests or preferences of the decision makers. This relative importance value can be deter-mined using a scale of 1–9 to represent equal importance to ex-treme importance (Saaty, 1980, 1996). The inﬂuence of elements (INCs) in the network on other elements (innovation competenc-es) in that network can be represented in the following supermatrix:

(7)

A typical entry Wijin the supermatrix, is called a block of the

supermatrix in the following form where each column of Wijis a

principal eigenvector of the inﬂuence of the elements (INCs) in the ith component of the network on an element (INC) in the jth component. Some of its entries may be zero corresponding to those elements (INCs) that have no inﬂuence.

Wij¼ wi1j1 wi1j2 wi1j_nj wi2j1 wi2j2 wi2j_nj .. . .. . . . . .. . wi_nij1 winij2 winijnj 2 6 6 6 6 6 4 3 7 7 7 7 7 5 :

After forming the supermatrix, the weighted supermatrix is de-rived by transforming all column sums to unity exactly. This step is very much similar to the concept of the Markov chain in terms of ensuring that the sum of these probabilities of all states equals 1.

Next, the weighted supermatrix is raised to limiting powers, such as Eq.(7)to get the global priority vector or called weights (Huang, Tzeng, & Ong, 2005)

lim

h!1W h

: ð7Þ

In addition, if the supermatrix has the effect of cyclicity, the limiting supermatrix is not the only one. There are two or more limiting supermatrices in this situation, and the Cesaro sum would need to be calculated to get the priority. The Cesaro sum is formu-lated as follows: lim w!1 1

m

Xm j¼1 Ww j; ð8Þ

to calculate the average effect of the limiting supermatrix (i.e. the average priority weights) where Wj denotes the jth limiting

supermatrix. Otherwise, the supermatrix would be raised to large powers to get the priority weights (Huang et al., 2005).

The weights of the kth INCs derived by using the above ANP processes, namely

x

k, k 2 {1, 2, . . . , n}, will be used as inputs for

summing up the grey coefﬁcients of the kth INC in Eq.(12)in the following GRA analysis.

3.4. Grey relational analysis

SinceDeng (1982)proposed Grey theory, related models have been developed and applied to MCDM problems. Similar to fuzzy set theory, Grey theory is a feasible mathematical means that can be used to deal with systems analysis characterized by inadequate information. Fields covered by the Grey theory include systems analysis, data processing, modeling, prediction, decision-making, and control engineering (Deng, 1985, 1988, 1989; Tzeng & Tasur, 1994). In this section, we briefly review some relevant definitions and the calculation process for the Grey relation model. This re-search modified the definitions by Chiou and Tzeng (2001) and produced the definitions indicated below.

GRA is used to determine the relationship between two se-quences of stochastic data in a Grey system. The procedure bears

Goal Criteria

Subcriteria Control Criteria

A possible different network under each subcriterion of the control hierarchy Goal

Criteria

Subcriteria Control Criteria

A possible different network under each subcriterion of the control hierarchy

Source: Saaty (1996)

(8)

some similarity to pattern recognition technology. One sequence of data is called the ‘reference pattern’ or ‘reference sequence,’ and the correlation between the other sequence and the reference se-quence is to be identiﬁed (Deng, 1986; Tzeng & Tasur, 1994; Mon, Tzeng, & Lu, 1995; Wu, Deng, & Wen, 1996).

Deﬁnition 7. The relationship scale also may be designated into 11 levels, where the scores 0, 1, 2, . . . , 10 represent the range from ‘no relationship’ to ‘very high relationship’ between the speciﬁed INC and the e-business models.

Deﬁnition 8. The initial relationship matrix G is an m n matrix, where there are m e-business models and n INCs, obtained by sur-veying the relationships, where gkiis denoted as the relationship

between the kth INC and the ith e-business model.

G ¼ g11 g1i g1n .. . . . . .. . . . . .. . gk1 gki gkn .. . . . . .. . . . . .. . gm1 gmi gmn 2 6 6 6 6 6 6 6 6 4 3 7 7 7 7 7 7 7 7 5 :

Deﬁnition 9. The normalized relationship matrix G can be obtained through Eqs.(9) and (10)

pi¼ 1= max 16k6mgki; X ¼ x11 x1i x1n .. . . . . .. . . . . .. . xk1 xki xkn .. . . . . .. . . . . .. . xm1 xmi xmn 2 6 6 6 6 6 6 6 4 3 7 7 7 7 7 7 7 5 ; ð9Þ Xi¼ piGi: ð10Þ

Deﬁnition 10. Let x0be the reference pattern with n entries (i.e.

dependent variable): x0= (x0(1), x0(2), . . . , x0(n)) and xi, the matrix

containing the normalized mapping information of each e-business

model to the INCs, be one of the m patterns with n entries to be compared with the x0 where xi is written as: when xki= xi(k),

k = 1, 2, . . . , n in Eqs. (9) and (10), then xi= (xi(1), xi(2), . . . , xi(n)),

1 6 i 6 m. The sequence xi generally expresses the inﬂuencing

factor of x0.

Deﬁnition 11. Let X be a normalized factor set of grey relations, x02 X the referential sequence, and xi2 X the comparative

sequence; with x0(k) and xi(k) representing the numerals at point

k for x0 and xi, respectively. If

c

(x0(k), xi(k)) and

c

(x0, xi) are real

numbers, and satisfy the following four grey axioms, then call

c

(x0(k), xi(k)) the grey relation coefﬁcient, and the grade of the grey

relation

c

(x0, xi) is the average value of

c

(x0(k), xi(k)).

1. Norm Interval

0 <

c

ðx0ðkÞ; xiðkÞÞ 6 1;

8

k;

c

ðx0;xiÞ ¼ 1 iff x0¼ xi;

c

ðx0;xiÞ ¼ 0 iff x0;xi2 /;

where / is an empty set. 2. Duality Symmetric

x; y 2 X )

c

ðx; yÞ ¼

c

ðy; xÞ iff X ¼ x; yf g:

3. Wholeness

c

ðxi;xjÞ –

often

c

ðxj;xiÞ iff X ¼ xf iji ¼ 0; 1; . . . ; ng; n > 2:

4. Approachability

c

ðx0ðkÞ; xiðkÞÞ decreases when ððxj 0ðkÞ xiðkÞÞÞj increases:

Deng also proposed a mathematical equation for the grey rela-tion coefﬁcient, as follows:

c

ðx0ðkÞ; xiðkÞÞ

¼min8imin8kjðx0ðkÞ xiðkÞÞj þ fmax8imax8kjðx0ðkÞ xiðkÞÞj ðx0ðkÞ xiðkÞÞ

j j þ fmax8imax8kjðx0ðkÞ xiðkÞÞj

; ð11Þ

where f is the distinguished coefﬁcient (f 2 [0, 1]). Generally, we pick f = 0.5. Source Component Source Component Source Component (Feedback loop) Source Component (Feedback loop) Intermediate Component (Transient State) Intermediate Component (Transient State) Sink Component (Absorbing State) Sink Component (Absorbing State) Intermediate Component (Recurrent State) Intermediate Component (Recurrent State) Outerdependence Innerdependence loop C1 C3 C4 C5 C2 Source Component Source Component Source Component (Feedback loop) Source Component (Feedback loop) Intermediate Component (Transient State) Intermediate Component (Transient State) Sink Component (Absorbing State) Sink Component (Absorbing State) Intermediate Component (Recurrent State) Intermediate Component (Recurrent State) Source Component Source Component Source Component (Feedback loop) Source Component (Feedback loop) Intermediate Component (Transient State) Intermediate Component (Transient State) Sink Component (Absorbing State) Sink Component (Absorbing State) Intermediate Component (Recurrent State) Intermediate Component (Recurrent State) Outerdependence Innerdependence loop C1 C3 C4 C5 C2

Source: Saaty (1996)

(9)

Deﬁnition 12. If

c

(x0, xi) satisﬁes the four grey relation axioms,

then

c

is called the Grey relational map.

Deﬁnition 13. IfCis the entirety of the grey relational map,

c

2C

satisﬁes the four axioms of the grey relation, and X is the factor set of the grey relation, then (X,C) will be called the grey relational space, while

c

is the speciﬁc map forC.

Deﬁnition 14. Let (X,C) be the grey relational space, and if

c

ðx0;xjÞ;

c

ðx0;xpÞ; . . . ;

c

ðx0;xqÞ

satisfy

c

(x0, xj) >

c

(x0, xp) > >

c

(x0, xq), then we have the grey

rela-tional order: xj xp xq.

When the grey relational coefﬁcient is conducted with respect to innovation policies, we then can derive the grade of the grey relation

c

(x0, xi) between the reference alternative

c

ðx0;xiÞ ¼

Xn k¼1

x

k

c

ðx0ðkÞ; xiðkÞÞ; ð12Þ

where k is the number of INCs,

x

kexpresses the weight of the kth

INC, and

c

(x0, xi) represents the grade of grey relation in xi (the

kth e-business model) correspondence to x0. In this study, we make

the order of the e-business models following the grade of grey relation.

4. SOC design service, SIP and SIP Mall

In the following section, the industry background of SOC design service, SIP, and SIP Mall will be introduced as a background for the empirical study of this paper.

4.1. SOC design service

Entering the 21st century, the global IC industry already has opened up the third industrial revolution under the driving force of the 3C (computer, consumer electronics, and communication) applications and Systems-on-Chip (SOC) (Lai & Liaw, 2007), a sin-gle IC chip that uses computing engines (MPU/DSP), memories, analog blocks (e.g., RF for wireless communication), and some custom logic to ‘‘glue’’ a system together (Tseng, 1999). Several emerging business models are being developed by contenders in this new SOC Olympics (Lu, 2004): SIP providers, design foundries, design service providers, and system design integrators (Lai & Liaw, 2007).

According to James (2005), IC design services suddenly are prominent as the result of a convergence of the following trends: (1) chip design are growing too complex for most companies to handle on their own; (2) SIP is far from plug-and-play, and design teams usually need specialized IP expertise (or even the knowledge and advice of the original designer) to make the IP work correctly; (3) another broad trend, off-shoring, is making design services available to a wider array of companies; and (4) the trend toward design for manufacturing (DFM) also creates more demand for de-sign services.

Design service capabilities of SOCs are the ability to provide customers who finish SOC specification design or SOC circuit de-sign the rest of the procedures of processes required for SOC com-mercialization. The detailed SOC design service procedures include front end design, backend (place and layout), and turnkey (tape out the layout to semiconductor wafer fab for SOC fabrication) ser-vices. The design service capabilities definitely are helpful for SOC design success, since strong design service capability implies strong SIP integration capability, which is a key factor for SOC success.

4.2. SIP

SIP is the subset of intellectual assets that is legally protected. There are five major forms of SIP: patents, trademarks, copy-rights, trade secrets, and semiconductor mask works (Sullivan, Harrison, Keeler, & Villella, 2002). SIP is associated with the own-ership of knowledge, expertise, innovation and resources that went into the creation of a specific hardware core and/or the software and/or firmware program that is required to perform a system function. Examples of SIP are: microprocessor and DSP cores, peripherals, dedicated function accelerators such as MPEG2 decoders and encoders, mixed signal technology, on chip DRAM, and Flash memory technology. Software/Firmware is intellectual property and may be delivered as an indivisible part of the hardware SIP or separately as a necessary system compo-nent. In this report we will differentiate between intellectual property that becomes part of silicon technology and other cate-gories (Baron, 2000).

During the past decade, IC design productivity has failed to keep pace with Moore’s Law, which predicts that the number of electronic devices that can be fabricated on an IC chip will double every 18 months (Moore, 1979). Thus, a ‘design gap’ be-tween IC design complexity increase and productivity increase has emerged (Semiconductor Industry Association, 2002). IC sup-pliers began looking for ways to narrow the gap by designing ICs with reusable SIP that tended to contain increasingly complex functionality (Ratford, Popper, Caldwell, & Katsioulas, 2003). As IC designs have become more complex, a large number of SIP products are being embedded into the designs. The SIP has be-come a key segment of the electronic design process, as it can re-duce IC development costs, accelerate time-to-market, rere-duce time-to-volume, and increase end-product value. The nature of SIP, which can narrow the ‘design productivity gap’, has made SIP critical for the design and implementation of complex systems on chips (SOC), which have become the mainstream solution for realization of electronic system products since the year 2000.

SIP cores can be classified into soft, firm, and hard SIPs. The dif-ference is in the degree of flexibility of the SIP; soft SIP is in the form of RTL code while hard SIP is in GDSII format. Firm SIP is somewhere in between; it is usually presented as a net list with a set of additional views and information pertaining to physical de-sign. Digital designs are commonly defined as RTL code or soft SIP while analog and mixed signal designs come in the form of hard SIP that has been designed and optimized for a specific application and technology (Keating & Bricaud, 1998).

Fabless Semiconductor Association (Ratford et al., 2003) sum-marized business models in the current SIP product marketplace, what is typical to expect from different providers, and consider-ations for determining the economic value of different SIP product types. The discussion of each business model includes the purpose of the particular business model, a deﬁnition of the payment op-tions, typical structure of the fees paid, and most common SIP use scope for the SIP products.

Table 1provides a summary of the principal attributes of busi-ness models for established providers and SIP products. A typical SIP purchase may involve elements of more than one business model.

Another aspect of determining the economic value of an SIP product is related to the different fees for enabling the suc-cessful use of the SIP product. Table 2 provides the typical en-abling components that usually represent a secondary revenue stream for providers. These are often structured as separate fees within the SIP License Agreement itself, or sometimes as a separate Statement of Work (SOW) or contract if there are speciﬁc.

(10)

4.3. SIP Mall and e-business models

As observed byHuang et al. (2007), despite the optimistic fu-ture that exists for SIP, potential problems, including complex laws and regulations pertaining to SIP transactions, negotiations over SIP usage rights, technical support including SIP integrations into SOCs, maintenance, and SIP application engineering based upon current business models, have emerged as roadblocks to successful SIP business. Apparently, SIP transactions and integrations are not easy. Thus, an e-Commerce mechanism for SIP transactions, the SIP Mall – that aims to provide a well-established SIP database, to in-crease SIP transaction efﬁciency, and to provide SOC designers with well-veriﬁed reusable SIPs, design environments, and design ser-vices – can resolve the above-mentioned business, legal, and tech-nical issues effectively, and be most helpful in accelerating SIP and IC industry growth.

The well-established SIP Malls ﬁll the vacancy that exists in the IC industry structure, resolve existing SIP transaction problems, and enable innovations in IC. The upstream electronic system houses and the fabless IC design houses may develop synergies, by leveraging the SIPs and design services provided by SIP Malls and integrating the manufacturing capacities of IC foundries and assembly houses to roll out innovative SOC products at the most advanced process nodes. Thus, the well-established SIP Malls en-hance the nation’s competitiveness in the IC industry.

According toHuang and Shyu (2006), the SIP Mall e-commerce business models for SIP transactions and resolve current problems in SIP transactions include (1) the brokerage model, (2) the distrib-utor model and (3) the commercialization model. The three models are introduced as follows and summarized inTable 3.

(1) The brokerage model: In the brokerage model, the e-com-merce mechanism serves as the broker to introduce buyers and sellers and facilitate transactions. The e-commerce mechanism charges commission as part of the transaction fee; (2) the distrib-utor model: for the distribdistrib-utor model, the e-commerce mechanism assists customers in verifying and certifying SIPs, fabricating test chips for SIPs, providing technical supports, negotiating with SIP end users, charging commission as part of the transaction fee and sharing license incomes; (3) IP commercialization model: The e-commerce mechanism provides a special business model called the commercialization model which helps IC design houses or system houses which have already designed circuit cores, but do not have the resources or capabilities, or INCs, for commercializing the circuit cores.

5. An SIP e-business model deﬁnition by using the novel MCDM framework

The authors propose an analytical frame for deﬁning a SOC de-sign service e-business model for an SIP being provided by an SIP provider. SIP experts from the Taiwan Government, industry, and research institutes were invited to evaluate criteria by using the Delphi method. Then the structure between criteria and weights of each criterion will be derived by the novel MCDM framework by the experts. Finally, the relationship between the criteria and available e-business models will be derived through the GRA. De-tail procedures and results are illustrated below.

5.1. Capabilities and resources derivation by Delphi

Sixteen INCs and resources which serve as the criteria for eval-uating the e-business models were collected from interviewing SIP

Table 1

Typical business models for SIP products. Source:Ratford et al. (2003).

Per use Time based Royalty based Access

Purpose Fee for each SIP on deﬁned use scope Multiple uses of SIP over a period of time

Amortize cost of SIP share risk-reward

Fee for SIP portfolio over a period of time

Payments Event based Time based Value based Subscription based

Structure One time fee for a design (ﬁrst or subsequent)

Fee for all designs within a given time

Some or all fees spread across units

Up front fee plus discounted use fee Scope Per design per device Multiple uses per device Percent of unit value per device Multiple SIPS per organization

Table 2

Typical enabling components for SIP products. Source:Ratford et al. (2003).

Maintenance Support NRE Contract service

Purpose SIP updates, bug ﬁxes, and revisions Address speciﬁc customer needs Enable SIP use Enable IC design

Payments % List price license fee Scope based Milestone based Hourly based

Structure Part of initial license agreement (may be included)

Basic package or separate contract Initial fee (SOW) percent milestones

Initial fee (SOW) percent milestones

Scope Changes in spec, process tech, etc. Web, email, telephone, on-site, geography

Modiﬁcations, re-spins, porting, etc.

Tool runs, IC integration, EDA views, etc.

Table 3

e-Business models of the SIP Mall. Source:Huang and Shyu (2006). SIP e-business

model

SIP vendors provide to SIP Mall

SIP Mall’s tasks Distributor Hard core/ﬁrm core/soft core Veriﬁcation

Documentation Certiﬁcation Share license Income Test chip fabrication

Promotion Technical support License agreement Share license income Brokerage Documentation High level veriﬁcation

Simulation models Promotion

Share the commission Share the commission

IP Commercialization Hard core/ﬁrm core/

soft core Veriﬁcation

Share the license income Certiﬁcation Documentation Test chip fabrication Promotion Technical support License agreement Co-owner of the SIP Share license income

(11)

experts from government, industry, and academic institutes by the Delphi procedures. Following are illustrations of the results.

(1) Design capability in digital and mixed signal design: Both dig-ital designs as well as analog and mixed signal designs are the essential capabilities for developing SIP cores. Lack of these design capabilities will deﬁnitely cause fatal errors of the SIP cores in both function and timing of circuits. (2) Time to market capability: Time to market is always the most

important factor for vendors to help customers achieve time to money. As the complexity of SOC increases, the ability of vendors to roll out the solutions in the shortest time is the critical point for maintaining a customer relationship. (3) SIP integration service capability: SIP integration service helps

customers implement designs from concept to production within the shortest time. SIP integration services also help customers narrow the gap between system companies and silicon foundries, enable process technology deployment, and shorten time-to-design, and therefore, time-to-revenue (Global Unichip Corporation, 2005). For system houses and middle to small design houses, SIP integration services help customers ﬁnish their SOCs.

(4) SIP qualification capability: Well-qualified SIP cores are nec-essary conditions for SOC success for both engineering and business aspects. Thus, SIP qualification capabilities of SIP providers are essential for innovation. Typical SIP qualifica-tion includes SIP documentaqualifica-tion and a specificaqualifica-tion check, SIP verification results check, and availability of SIP pilot-run reports.

(5) SIP verification capability: SIP verification is the design proce-dure used to verify the function and timing of SIP cores. Thus, SIP verification capability is as important as SIP quali-fication to guarantee the final success of SIP.

(6) Market leadership and customer education: Technology push is becoming the marketing trend in the IC and SOC industry. Those who have the capability to educate their customers will win at the end. Market leaders can then define specifica-tions for their next generation products. With higher market shares and profits from already developed products, market leaders can also get returns from investment sooner than their competitors. More and more resources can thus be invested into innovation activities to accelerate the time to market of next generation products. Thus, market leaders can confirm their situation for even longer and to a greater degree. (7) SIP e-commerce capability: More and more SIP e-commerce mechanisms are being funded by both governments and pri-vate institutes to shorten the time to market of SIP and enable SIP business. SIP e-commerce capability will be an important issue. SIP Malls have been established in Taiwan. Many design companies, universities, and industrial research labs have launched efforts to develop their own SIP and are attempting to find more opportunities to license or exchange SIPs (Lu, 2004).

(8) SIP design management capability: Major SIP design manage-ment capabilities include project managemanage-ment, database management and revision control, design tool management, design ﬂow management, and security control.

(9) Human resources: The IC industry has become one of the fastest growing and highly invested industries in Taiwan. In order to keep pace with the technology advancement of nanometer semiconductor manufacturing, IC design and SOC techniques are evolving. This evolution is leading to a strong demand from the industry for a large number of well-trained IC design engineers as well as support human resources in business and law (Advisor Ofﬁce of Ministry of Education, 2004).

(10) MPW services: As technology evolves, mask and engineering run costs grow exponentially with each process generation. MPW (Multiple Project Wafer) services aim to reduce IC and SIP developer costs to develop and facilitate faster pro-totyping by sharing the costs of a common mask set and engineering-run. Moreover, this service encourages innova-tion, and offers its customers a great opportunity to prove their design/product and market test samples.

(11) Customer relationship management: Customer relationship management is a strategy used to learn more about cus-tomer needs and behaviors and thereby develop stronger relationships with them. Customer relationship manage-ment is especially important for service-oriented Taiwanese IC companies, including silicon foundries, design services companies, and SIP providers.

(12) Funding capability: SIP designs have become more and more complicated in advanced processes. The most up-to-date SIP design flows and EDA tools are required to achieve higher speed, lower power, and lower cost. MPW blocks in advanced processes are more expensive. How to leverage various funding tools, including loans, venture capital investments, IPOs (Initial Public Offerings), and CBs (Changeable Bonds) has become the most important factor influencing operation efficiency.

(13) Electronic systems know-how establishment: SOC needs a great deal of electronic system know-how for ﬁnal success of SIP applications onto the chips. Establishment of elec-tronic system know-how is also important for SIP success in engineering and business.

(14) Joint development and technology transfer capability: Individ-ual organizations can no longer rely on their own resources to compete in today’s world. Rather, they should look for strategic interactions that will allow them effectively to leverage internal resources by investing in core competen-cies and contracting out other knowledge domains (Sobrero & Roberts, 2002). Various speciﬁcations for embedded processors, peripherals, memories, analog, and mixed signal cores are emerging for future needs of 3C applications. Since there is almost no single company, no matter how professional an SIP provider, fabless IC design house, semi-conductor foundry, or even IDMs that can afford human resources and R&D capabilities to develop every SIP needed, especially at a time when the time to market SOC products is rapidly shortening, a joint development capability (including technology transfer) has become the key aspect for SIP innovations.

(15) SIP porting capability: Moving an SIP to an existing process of other semiconductor Fabs or foundries always involves rede-sign, extensive product development, and often multiple expensive iterations. A better way is to ‘‘port’’ the process, i.e. move it cell by cell without changing the SIP made to use it. By adopting the basic methods used in the silicon industry and paying attention to process details that affect performance, successful porting is faster, much less expen-sive and less disruptive to the flow of the product. Porting also results in multiple qualified fab sites for a given SIP. An added benefit is that multiple geographically separated production facilities make the supply chain immune to dis-ruption by natural disaster or power interdis-ruption. Porting also eliminates the need for customers to re-qualify a prod-uct containing chips from a new source (Williams, Chao, Wang, & Wu, 2002).

(16) Close relationship with foundries: SIPs, especially mixed signal and analog SIPs from professional SIP providers, always get major foundries, for example, TSMC, UMC, or IBM, to tar-get major foundry SOC customers. Close relationships with

(12)

foundries are essential for professional SIP providers; so they can have an SIP veriﬁcation, qualiﬁcation and marketing channel.

5.2. Decision problem network relation map structuring by DEMATEL Since the inter-relationships between the 16 INCs summa-rized through above Delphi process seem too complicated to be analyzed, the decision problem structure will be deducted with the DEMATEL method introduced in Section 3.2. At first, the direct relation/influence matrix A is introduced as shown in Fig. 5. After that, the direct relation/influence matrix A is normalized based on Eq. (1) and the normalized direct relation/ influence matrix N is shown inFig. 6. Finally, the direct/indirect matrix is deducted based on Eq. (3) and shown inFig. 7 where major relationships were deducted by setting the threshold value as 0.280. If the threshold value is lower than 0.280, the relation-ships or linkages in the casual diagram of total relationrelation-ships will be too many to be analyzed. On the other hand, if the threshold value is higher than 0.280, the total relationships being derived could be too few to demonstrate proper number of relationships. The total relationships being derived will references for calculat-ing weights between INCs in the followcalculat-ing ANP processes.Table 4 summarizes symbols for the INCs being used in the empirical study of a SOC design service firm. The ri+ ci and ri ci values

calculated from the direct/indirect matrix T are demonstrated in Table 5.

5.3. Calculating the weights of INCs by ANP

By setting selecting an appropriate e-business model for com-mercializing the SIP as the goal, pair wise comparisons of the INCs were executed based on opinions being collected from SIP experts. Following are the illustrations based on the ANP.

First, based on the decision problem structure which was de-rived by DEMATEL in Section3.2, INC 4, 6, 8, 13, 14, 15 and 16 were selected as the elements of the major component where the ele-ments belonging to this component impact the goal and other INCs signiﬁcantly based on ri ciand the casual diagram. Other INCs

including INC 1, 2, 3, 5, 7 and 9–12 were selected as the elements of the minor component where elements belonging to the minor component are mainly impacted by INCs in the major component. The inter-relationships between the Goal, the Major component, and the Minor component are illustrated inFig. 9, where the direc-tions of arrows mean the direcdirec-tions of inﬂuences.

Finally, the pair wise comparison results and the decision prob-lem structure serve as inputs for the ANP. With the aid of the Super Decisions (Creative Decisions Foundation, 2006), a software which is used for decision-making with dependence and feedback by implementing the ANP, the limit super matrix W is calculated and shown inFig. 10. Weights corresponding to each INC (Table 6) are derived accordingly which will be used for calculations of grey grades in Section3.4.

The GRA was used so as to derive the relationships between INCs and e-business models. The initial relationship matrix G, a

0.000 5.833 5.667 6.000 7.167 5.333 2.333 5.667 6.333 2.833 3.833 6.333 4.500 5.667 5.333 6.000 5.167 0.000 5.333 4.500 4.500 6.000 4.167 5.500 4.833 4.667 5.667 4.667 3.333 5.000 5.167 4.667 6.167 6.333 0.000 5.333 6.167 5.333 1.333 5.000 5.000 4.000 4.667 5.333 4.000 5.000 2.500 5.000 6.000 5.500 6.833 0.000 7.833 5.000 2.167 5.000 4.333 4.500 3.667 5.167 3.167 4.167 5.000 6.667 6.000 5.500 5.333 6.000 0.000 4.667 1.500 5.000 4.500 3.667 3.833 5.000 3.333 3.833 5.000 6.000 4.333 4.333 5.833 6.000 5.833 0.000 6.333 6.000 7.833 5.333 6.333 7.667 5.667 6.000 4.333 6.833 1.167 2.167 1.000 0.833 1.167 4.667 0.000 2.667 2.500 1.667 3.167 5.167 2.667 2.167 0.667 2.000 6.833 6.500 6.500 6.500 6.833 5.667 2.667 0.000 7.167 3.167 2.833 5.500 3.500 4.333 4.500 3.333 5.000 5.333 4.500 4.500 4.500 5.500 2.333 5.667 0.000 1.167 2.000 4.000 1.667 1.500 1.500 1.833 3.500 3.000 4.500 3.000 4.667 2.333 2.500 2.833 3.333 0.000 3.500 3.333 1.333 2.500 3.333 6.667 3.500 6.000 3.333 3.833 3.500 5.000 2.667 2.667 2.333 3.000 0.000 2.167 3.500 6.167 3.000 5.000 5.500 5.167 4.000 4.667 4.167 6.667 3.833 4.667 5.833 3.833 3.833 0.000 2.500 4.667 2.500 4.667 5.833 4.500 4.833 4.000 5.000 5.000 4.167 5.000 4.333 4.333 4.833 4.667 0.000 3.833 2.833 4.000 5.667 6.333 6.000 5.333 3.500 5.000 1.833 4.000 3.500 4.000 5.500 5.000 4.667 0.000 3.333 4.667 7.333 5.000 6.000 7.167 7.167 3.667 3.000 4.333 3.500 4.167 2.667 3.000 3.167 2.667 0.000 6.167 6.167 6.833 5.167 3.833 4.333 5.833 2.667 3.167 4.667 5.667 5.167 5.167 2.833 5.000 6.667 0.000

A =

Fig. 5. The direct relation/inﬂuence matrix A.

0.000 0.066 0.064 0.068 0.081 0.060 0.026 0.064 0.071 0.032 0.043 0.071 0.051 0.064 0.060 0.068 0.058 0.000 0.060 0.051 0.051 0.068 0.047 0.062 0.055 0.053 0.064 0.053 0.038 0.056 0.058 0.053 0.070 0.071 0.000 0.060 0.070 0.060 0.015 0.056 0.056 0.045 0.053 0.060 0.045 0.056 0.028 0.056 0.068 0.062 0.077 0.000 0.088 0.056 0.024 0.056 0.049 0.051 0.041 0.058 0.036 0.047 0.056 0.075 0.068 0.062 0.060 0.068 0.000 0.053 0.017 0.056 0.051 0.041 0.043 0.056 0.038 0.043 0.056 0.068 0.049 0.049 0.066 0.068 0.066 0.000 0.071 0.068 0.088 0.060 0.071 0.086 0.064 0.068 0.049 0.077 0.013 0.024 0.011 0.009 0.013 0.053 0.000 0.030 0.028 0.019 0.036 0.058 0.030 0.024 0.008 0.023 0.077 0.073 0.073 0.073 0.077 0.064 0.030 0.000 0.081 0.036 0.032 0.062 0.039 0.049 0.051 0.038 0.056 0.060 0.051 0.051 0.051 0.062 0.026 0.064 0.000 0.013 0.023 0.045 0.019 0.017 0.017 0.021 0.039 0.034 0.051 0.034 0.053 0.026 0.028 0.032 0.038 0.000 0.039 0.038 0.015 0.028 0.038 0.075 0.039 0.068 0.038 0.043 0.039 0.056 0.030 0.030 0.026 0.034 0.000 0.024 0.039 0.070 0.034 0.056 0.062 0.058 0.045 0.053 0.047 0.075 0.043 0.053 0.066 0.043 0.043 0.000 0.028 0.053 0.028 0.053 0.066 0.051 0.055 0.045 0.056 0.056 0.047 0.056 0.049 0.049 0.055 0.053 0.000 0.043 0.032 0.045 0.064 0.071 0.068 0.060 0.039 0.056 0.021 0.045 0.039 0.045 0.062 0.056 0.053 0.000 0.038 0.053 0.083 0.056 0.068 0.081 0.081 0.041 0.034 0.049 0.039 0.047 0.030 0.034 0.036 0.030 0.000 0.070 0.070 0.077 0.058 0.043 0.049 0.066 0.030 0.036 0.053 0.064 0.058 0.058 0.032 0.056 0.075 0.000