modeling technological innovation

7/27/2019 Modeling Technological Innovation

1/13

Modeling technological innovation risks of an entrepreneurial team usingsystem dynamics: An agent-based perspective

Desheng Dash Wu a,b,, Xie Kefan c, Liu Hua c, Zhao Shi c, David L. Olson d,1

a Reykjavik University, Kringlunni 1, IS-103 Reykjavk, Icelandb RiskLab, University of Toronto, Canadac Wuhan University of Technology, Wuhan, Hubei, 430070, PR Chinad Department of Management, University of Nebraska, Lincoln, NE 68588-0491, United States

a r t i c l e i n f o a b s t r a c t

Continuous technological innovation has been playing a vital role in ensuring the survival and

development of an enterprise in today's economy. This paper studies the problem of

technological innovation risk-based decision-making from an entrepreneurial team point of

view. We identify the differences between this team decision-making and a traditional

individual decision-making problem, where decisions are mainly affected by the decision-

maker's risk and value perceptions, and risk preferences. We create a modeling framework for

such a new problem, and use system dynamics theory to model it from the agent-based

modeling perspective. The proposed approach is validated by a case study of the technological

innovation risk decision-making in a Chinese automobile company.

2010 Elsevier Inc. All rights reserved.

Keywords:

Risk-based decision-making (RDM)

Entrepreneurial team

Technological innovation risk

Agent-based modeling

System dynamics

1. Introduction

Since the nancial crisis, the encouragement of business startups has become the consensus of all walks of life to promote

employment; many major cities in China also have proposed a full businessstrategy, and launched a large number of business

support measures under the direction of macroeconomic policy. Entrepreneurial success mainly depends on three factors: venture

projects, business operations and entrepreneurs. Among these three factors, entrepreneurial activity is the core of success. Since

1977, Cooper and Bruno[1], Thurston[2],Feeser and Willard[3], Doutriaux[4], and Chandler and Hanks[5]have suggested that

venture performance created by an entrepreneurial team is often superior to one created by a single entrepreneur. Lechler [6]also

believes that the average success rate of new enterprises created by teams is higher than that of new enterprises created by

individual entrepreneurs. The entrepreneurial environment has become increasingly complicated, but well-designed and efcient

entrepreneurial teams can quickly analyze, evaluate and predict changes from external environment. At the same time, from the

perspective of entrepreneurial opportunities, entrepreneurial teams have greater capacity for opportunity identication,

development and utilization.In today's economy with ever-changing technology, continuous technological innovation has been playing a vital role in

ensuring the survival and development of an enterprise[7,8]. Technological innovation decisions have become a very important

decision problem that cannot be ignored in the entrepreneurial team decision-making. In this work, we study the problem of

technological innovation risk decision-making in an entrepreneurial team for typical enterprises. Such a problem has two main

differences from traditional technological innovation risk decision-making (DM): rst, the difference between startups and

traditional enterprises; and second, the difference between entrepreneurial team decisions and individual decisions. On the one

hand, compared to a general enterprise, technological innovation in entrepreneurial enterprises has higher motivation of

Technological Forecasting & Social Change 77 (2010) 857869

Corresponding author. Reykjavik University, Kringlunni 1, IS-103 Reykjavk, Iceland.

E-mail addresses:[email protected],[email protected](D.D. Wu),[email protected](D.L. Olson).1 Tel.: +977 402 472 4521.

0040-1625/$ see front matter 2010 Elsevier Inc. All rights reserved.doi:10.1016/j.techfore.2010.01.015

Contents lists available at ScienceDirect

Technological Forecasting & Social Change
http://-/?-mailto:[email protected]:[email protected]:[email protected]://dx.doi.org/10.1016/j.techfore.2010.01.015http://www.sciencedirect.com/science/journal/00401625http://www.sciencedirect.com/science/journal/00401625http://dx.doi.org/10.1016/j.techfore.2010.01.015mailto:[email protected]:[email protected]:[email protected]://-/?-


2/13

technological innovation and lower innovation capability. This suggests an entrepreneurial team may prefer to be risk-seeking in

startups in order to gain technological innovation in the organization. On the other hand, decision-making in an entrepreneurial

team is different from how individuals deal with risks. Entrepreneurial team risk decision-making typically focuses on what action

a group should take. Individual decision-making is mainly affected by individual decision-maker's subjective factors, including

decision-maker's risk and value perceptions, risk preferences, etc. However, the entrepreneurial team includes a number of DM

individuals, where the impacts of a single DM's subjective factors have been signicantly reduced. Instead, the composition of

decision-makers' opinions, the mutual relations among policy-makers and a team's DM system have a greater impact on decision

outcomes. This study describes the general theory of technological innovation risk decision-making in an entrepreneurial team.

Both SDM and ABM have been used in the eld of technological forecasting modeling and risk analysis[911]. We model this

DM problem using a system dynamics model (SDM) from the agent-based modeling (ABM) perspective. SDM uses feedback loops

and stocks and ows to model the behavior of complex systems over time and deals with internal feedback loops and time delays

that affect the behavior of the entire system. SDM is a good tool for modeling aspects of organizational behavior due to its

signicant capabilities for modeling human behavior and DM processes[12]. The strength of System Dynamics lies in its ability to

account for non-linearity in dynamics, feedback, and time delays.

ABM has been treated as a powerful tool for modeling complex adaptive systems with multiple entities reacting to the pattern

these entities create together. Agents in ABM represent autonomous DM entities. ABMs have been employed since the mid-1990s

to solve a variety of business and technology problems. Examples of applications include supply chain optimization and logistics,

modeling of consumer behavior, social network effects, workforce management, and portfolio management. Both ABM and SDM

have a high potential for supporting and complementing each other [12].

To make group decisions in an entrepreneurial team, lots of heterogeneous participating entities can be involved; they not only

interact dynamically with each other, but also adapt or react to patterns generated or forecasted. Traditional optimization

approaches, equilibrium analysis, or other analytical techniques usually fail to handle these complexities. To model

entrepreneurial team risk features and various team interaction and adaptive behaviors, we treat the node representing a single

executive ofcer in the system dynamics model as an Agent. Agents modeling a number of nodes will be used to simulate the

interdependent DM reactions among the business executives. Group DM solutions can be obtained by running the system

dynamics models for the whole entrepreneurial team.

The proposed approach is validated by a case study of technological innovation risk decision-making in a Chinese automobile

company the Automobile Technology Development Ltd. of Wuhan Genpo. The agent simulation and analysis were performed by

an entrepreneurial team of three senior executives. Results show that the agent-based technological innovation risk DM model in

entrepreneurial team is appropriate for start-up enterprises. In addition, SDM was found useful in providing practical guidance to

risk-based decision-making.

Section 1 has presented literature review. Section 2 discusses risk-based decision-making (RDM) in the context of a

technological innovation project.Section 3presents models and analysis.Sections 4and 5 present a case study validating the

proposed models using system dynamics, andSection 6concludes the paper.

2. Individual RDM of a technological innovation project

Technological innovation activity contains uncertain factors in each stage and component, giving it high-risk. The probability of

a successful technological innovation is often less than the probability of failure. Technological innovation risk is mainly due to the

uncertainties of technology, market, innovation benets and institutional environment. Technological innovation projects involve

decision uncertainty, complexity, multiple objectives, and dynamic interactions. We present such an individual RDM problem in

the following modeling framework, which also serves as the basis for the technological innovation RDM in an entrepreneurial

team.

2.1. Modeling

Suppose an entrepreneurial team member intends to make decisions of a technological innovation project, in order to decide

on 1) whether to carry out the project, and 2) if adopted, how much to invest. The project can be either adopted at a high level with

large investment or at a normal level with small investment. Human risk behavior has been researched as individual cognitive

process where individuals collect and treat information to form their actions and decisions [13,18]. Psychological-based

researchers have paid a great deal of attentions to the role of human preference expands the interest of risk management beyond

objective data concerning probabilities to the more complex judgmental forum requiring subjectivity[1921].

In such a technological innovation project, there are three major factors affecting the individual's RDM result: risk perception,

value perception and risk preference. These factors can be seen as three input variables in technological innovation risk decision-

making. Risk perception and value perception have an impact on individual decision-making through the internalization of the

input risk and value information of the decision-making object perceived by the decision-maker. Risk preference is an inherent

property, which directly inuences the decision-making outcome by inuencing the decision-maker. For such a decision-making

problem, we also assume there are three possible outcomes corresponding to the three input factors: reject, agree to proceed and

take the radical scheme with large investment and agree to proceed but take the conservative scheme with small investment. As a

result, these three factors, three possible outcomes and the decision-maker constitute an individual RDM system together, whichis depicted inFig. 1.

858 D.D. Wu et al. / Technological Forecasting & Social Change 77 (2010) 857869


3/13

InFig. 1the decision-maker is the core of the individual RDM system and has the properties of reactivity, autonomy and

sociality. These features are in line with the Agent's basic features in ABM.

Generally speaking, both risk perception and value perception are judgments on decision-making objects by decision-makers;

and their levels are determined by the perception ability of the decision-maker, risk and value of the decision-making objects. Risk

preference is an inherent property of the decision-maker.

Denote the decision-maker (denoted as A)'s risk perception coefcient by r, value perception coefcient w and risk

preference coefcient r, respectively. Risk size w and value size r are properties of the decision-making object P. The risk

perception degree of the decision-makerAcan be computed as fr(r) =rr, and value perception degreefw(w) =ww. To facilitate

calculation, we introduce an adjustment factor tto normalize fr(r) and fw(w) such that 0frt(r),fwt(w),ft(r)3b. Then the

standardized value functions of three adjusted factors are frt(r) = t1rr,fwt(w) = t2ww,ft(w) = t3rrespectively. We can depict

standardized values of three factors to construct a three-dimensional coordinate system, as shown in Fig. 2.

Due to different impacts of risk preference, risk perception and value perception, we introduce three weight factors: risk

preference weight k1, value perception weight k2, and risk perception weight k3. The weights can be determined by expert scoring.

Taking weight into consideration leads to an individual technological innovation RDM function in Eq. (1):

fdfatra;fwtw;frtr= k1fatra+ k2fwtwk3frtr: 1

Fig. 1.Individual technological innovation RDM system.

Fig. 2.Three-dimensional relationships in individual technological innovation RDM.

859D.D. Wu et al. / Technological Forecasting & Social Change 77 (2010) 857869


4/13

As shown inFig 2, the ranges of the standardized value functions of risk preference, risk perception and value perception are

divided into three level sizes. We can describe these three level sizes inTable 1.

There are 27 combinations of these three levels of the three factors, as shown in Fig. 2. These 27 combinations of factors ultimately

lead to three DM outcomes: reject, agree to proceed with large investment, and agree to proceed but take the conservative scheme

with small investment. Obviously, when the degree of risk preference is risk-loving, the degree of value perception is strong and the

degree of risk perception is weak, the decision has a high probability of choosing radical scheme with large investment. This means in

Fig. 2 whenfrt(r)[0,b],fwt(w)[2b,3b]andft(r)[2b,3b],wehave2bk1+2bk2bk3fd3bk1+ 3bk2 andthe decisionoutcome is

to agree to proceed andtake theradicalscheme withlarge investment. When theweights are different, theset of factor points, which

is in similar or adjacent areas to the left front upper corner of the cube in Fig. 2, can also generate the same outcomes.

Similarly, under risk-aversion, the degree of value perception is weak and the degree of risk perception is strong, the decision is to

choose reject. This means in Fig. 2, whenfrt(r)[2b,3b],fwt(w)[0,b], andfat(ra)[0,b], we have3bk3fdbk1+bk22bk3. The

set of factor points falling in adjacent areas to the right posterior nether corner of the cube in Fig. 2 can also generate the same results.

Thethirddecisionset in Fig.2 corresponds to the casewhenthedecision-making function isfd[bk1+bk22bk3,2bk1+2bk2bk3].

The decision-making outcome in such situations is to agree to proceed but take the conservative scheme with small investment.

From the individual decision-making function, we can see that the individual decision-making is inuenced by many

individual subjective factors including individual risk sensitivity, value sensitivity, and individual risk preference. This leads to a

strong subjectivity in decision-making: when the decision-maker is in too conservative or radical a state, it can easily lead to

wrong decisions. For an extremely radical decision-maker, his sensitivity of the value perception tends to be higher and the

sensitivity of risk perception tends to be lowand he tends to prefer risks in most situations. Measurement of perceptions is rarely

perfectly accurate. The effect is that the decision variations are amplied leading to larger value of the individual decision-

making function. As a result, the decision is to choose a wrong action with the radical scheme using large investment, where the

correct action should be conservative scheme using small investment. Similarly, for an extremely conservative decision-maker,

the effect of such subjective individual decision-making is reject possible project investment that should be termed as agree to

proceed.

Moreover, real decisions are far from symmetric as depicted in Fig. 2; boundaries between adjacent categories are rather vague.

This incurs two main problems for RDM: First, the probability of choosing agree to proceed but take the conservative scheme with

small investment is high. Second, it is likely to cause decision-making errors between two adjacent categories.

Decision-makers are usually embedded in an environment that includes other people and therefore are subject to social inuence

[14]. Even if individual decision-makers can decide independently based on available information exposed to themselves, there are

situations where they areheavily inuenced by other people's dynamic behaviorand actions [15,16]. Wewill considera problem of an

entrepreneurial team RDM and model this dynamic problem using SDM in the next section.

3. Entrepreneurial team RDM

This section discusses entrepreneurial team RDM (risk-based decision-making) for a technological innovation project. We will

discuss risk metrics, causal loop diagrams and system ow chart from the ABM perspective, based on which we will then build

SDM (system dynamics models) of an entrepreneurial team RDM system inSection 4.

3.1. A SDM view

In order to facilitate system dynamics modeling of an entrepreneurial team RDM, we examine our problem from an SDM view.

Several senior executives form a team subject to social inuence. This will be modeled through changing trends of technological

innovation project decision-making and investment willingness of various decision-makers. During the decision-making process,

the entrepreneurial team exchanges material, energy and information with internal and external project risk environments.

Individual executives sometimes are heavily inuenced by other people's behavior and actions, while they usually decide

independently based on available facts and fundamental data from the history, which means in a variety of situations typical

feedback structures exist in the entrepreneurial team RDM system. In accordance with the core idea of system dynamics, the

technological innovation RDM in entrepreneurial team is generated by the cyclic action of the owin the system, which then

forms a feedback loop connecting the interrelated variables, i.e., interrelated decision node variables, and other variables. This willprovide an interrelated and mutually constrained system structure.

Table 1

Different factor levels in individual technological innovation RDM.

The interval of numerical range 0b b2b 2b3b

The degree of risk perception Weak Medium Strong

The degree of value perception Weak Medium Strong

Risk preference Risk-aversion Risk-neutrality Risk-loving



5/13

3.2. Risk metrics

Business exists to cope with risk and earn money. In the entrepreneurial team system, the decision-makers' RDM and

investment willingness of technological innovation projects depend on two main factors.

First, embedded in the internal and external environments, the managers (Agents) collect and process information from different

departments to form their own unique perception of risk and value. There are three typical enterprise risks in the process of

technological innovation: technology R&D risk, research-ndings commercialization risk, and market applications risk. Technology

R&D risk refersto possible risks duringthe stage of thetechnical development. Specically it includes technical risk,nancial risk, and

personnel risk at this stage. Commercialization risk of research ndings refers to possible risks from scientic and technological

development until mass production. The main risk, market risk, includes uncertainties and exposures faced by market players

engaging in economic activities. When the new products are in the market, competitors intervene rapidly, which will lead to a

competitive risk.

Second, the nal decision on the project should also consider external risks resolving method of the enterprise project. This

includes the social risk sharing power, the degree of government support for innovation and other factors, which will affect

decision-making and the input of the various departments within the enterprise and the nal decision-maker.

3.3. Causal loop diagrams of technological innovation RDM

A causal loop diagram (CLD) is a diagram that aids in visualizing how interrelated risk metrics and variables affect one another.

The diagram consists of a set of nodes representing the risk metrics and variables connected together. The relationships between

these risk metrics and variables are represented by arrows and labelled as positive or negative. Fig. 3 depicts the causal

relationship of technological innovation RDM for an entrepreneurial team, where the qualitative relationship among interrelated

risk metrics and variables are presented and feedback loops are shown. A positive causal link inFig. 3means that the two nodes

move in the same direction, i.e. if the node in which the link starts decreases (or increases), the other node also decreases (or

increases). At negative causal links, nodes change in opposite directions. For example, an increase in the node of market

informationcauses an increase in another node ofthe prot ability of enterprise. An increase in the node ofthe risk preference

of entrepreneurial teamcauses a decrease in another node of the decision power of innovative risk in entrepreneurial team.

FromFig. 3, the causal loop diagrams are divided into three modules: 1) decision-making module, 2) technological innovation

internal risk movement mechanism module and 3) external risk movement module. First, the decision-making module is affected

by internal and external environments of enterprise technological innovation, the perceptions of risk, and value of the

entrepreneurial team are positively associated with the decision-making power of the technological innovation risk of the

Fig. 3.The causal relationship of technological innovation RDM.



6/13

entrepreneurial team, which is negatively affected by risk preference of the entrepreneurial team. Meanwhile, the decision-

making power positively enhances enterprise technological innovation capability, bringing innovation revenue, so it is associated

with the perceptions of risk and value of the entrepreneurial team. Second, the function of the internal risk movement mechanism

module mainly indicates that technological innovation project internal risk can weaken R&D investment desire and determine the

R&D investment amount under the inuence of a certain attractive investment in science and technology. With the increase in the

amount of R&D resources, problems regarding internal funds and personnel increase, the uncertainty of technological innovation

R&D projects also increases, which increases the technological innovation risk within the enterprise. Thirdly, movement regularity

of the external risk of technological innovation RDM in external risk movement module is mainly presented as follows: under the

external market environment of the enterprise, the number and the strength of competitors relate to the strength of corporate

competitiveness. This affects the earning ability of enterprise in technological innovation, and leads to positive impact on

enterprise market size and the uncertainty in the development of the enterprise market. Meanwhile, the risk sharing capacity of

the community outside the enterprise and degree of government support for innovation will work together to affect the strength

of external risk of technological innovation entrepreneurial team. The three modules in Fig. 3 are independent and

interdependent, they constitute the decision-making system of technological innovation risk in entrepreneurial team together.

As a result, based on the analysis of the movement mechanism of internal and external risk of technological innovation projects,

the causal relation of the technological innovation RDM in entrepreneurial team become evident in the above causal loop diagram.

3.4. Systemow chart of technological innovation RDM

Based on the qualitative description of the feedback mechanism of technological innovation RDM in Fig. 3, we draw the system

ow chart in this section in order to have quantitative analysis of the team DM behavior. In this study, every executive in various

elds within the enterprise is deemed as an Agent. Variables (e.g., risk factors) of Agents are dened as state variables to describe

the states of the system elements. Perceptions of risk and value from managers in all areas within the enterprise are selected and

the corresponding project decision-making judgments based on perceptions are dened as the time-varying rate variables to

describe decision-making willingness of the main Agent nodes. In addition, since some variables change over time, the system

should use Time to describe the relationship between the variables and changing situation. For example, we can set bTimeNas

hidden variable to build a two-stage systemow chart of technological innovation RDM of two-agent (two core senior executives)

entrepreneurial team inFig. 4.

Fig. 4describes the simplest type of an entrepreneurial team, where a system including the R&D manager Agent and the team

DM Agent is depicted to show the behavior of technical innovation RDM in an entrepreneurial team. The behavior can be analyzed

as follows. First, the R&D risk and value perception from all levels of R&D managers in the R&D process are affected by lots of

factors such as the complexity of technology, the difculty and complexity of pilot tests, the maturation of the technology, and the

whole ability and strength of project team. Second, the ability of dealing with R&D risk and the risk reference of the manager will

directly affect the judgments of R&D decision and investment. Through the above two aspects, R&D manager Agent generates the

decision and degree of R&D devotion, which will be passed to the team DM Agent.

Based on R&D manager's decisioninformation, the abilityof dealing with the risk fromenterprise, the allocation degree of riskfrom

thesociety andthesupporting power of innovation fromthe government, theprojectteam decision-makermakes hisnaldecisions of

the whole technological innovation project, including the judgments of whether or not to proceed withthe project and the investment

scale.

To consider a more complicated situation with both the internal and external risk environments at the early stage of the

enterprise,Fig. 5 depicts a three-stage system ow chart of technological innovation RDM. In Fig. 5, we add the market factors into

Fig. 4 and have three executive agents: the market manager Agent, R&D manager Agent and team decision-maker Agent. Here, the

Fig. 4.Two-stage system ow chart of technological innovation RDM.



7/13

market manager Agent forms his unique market risk and value perception by taking into account effects from realization degree of

market information, enterprise reputation and popularity, competitors' amount and strength, consumer's demand change and

other factors. Then, through his inherent risk preference and the understanding of the ability of dealing with market risk from

enterprise, market manager Agent makes the decision and market investment scale. According to the degrees of market

investment scale from market manager and R&D manager and the measures to deal with both internal and external risks, the team

decision-maker Agent generates the nal decision based on his risk preference.

Fig. 6.The multi-stage multi-agent system of technological innovation RDM.

Fig. 5.Three-stage system ow chart of technological innovation RDM.



8/13

We now generalize the team with a specic number of persons into the general case with n-person entrepreneurial team.

We depict in Fig. 6 an Agent group system structure diagram of multi-stage technological innovation RDM for n-person

entrepreneurial team. In this system, the Agent groups, which are composed of the market managers, R&D managers, capital

managers and personnel managers, and the team decision-maker's Agent together constitute a set of multi-component

computable Agents. We will describe this hybrid system structure and discuss the intelligent reaction mechanism and rules of the

Agent in the technological innovation RDM in entrepreneurial team below.

InFig. 6, the reason of introducing the multi-Agent system into the technological innovation RDM in entrepreneurial team is:

there are a number of uncertainties and con

icting information and also some relevant and con

icting objectives between upperand lower layers that need to be dealt with in the entrepreneurial team RDM process. An Agent needs realize his goals subject to

limited perception and behavior capacity. In order to solve the above problems, building a multi-Agent system of technological

innovation RDM in entrepreneurial team has three main benets.

First, within the system, by communicating with other Agents, new planning or decision-making ideas to deal with incomplete

and uncertain knowledge and information can be developed. Second, through the cooperation among a number of senior-

executive Agents, multi-Agent system not only improves the basic function of each Agent, but also further understands

technological innovation RDM behavior in entrepreneurial team from Agents' interaction. Third, it facilitates to organize the

system by use of Modular Style sheets, i.e., specic entrepreneurial team organizational structure and structure system.

3.5. Communication mechanism of multi-stage multi-agent system

In multi-agent system theory, the core problem is to model the cooperation and co-adaptation among various agents. For the

multi-Agent system of RDM of technological innovation in entrepreneurial team, the behavior of the system not only depends onan individual agent, but more depends on cooperation among different Agents. In this section we discuss how to organize agents

into a group, and how to enable them to cooperate effectively to achieve an effective solution for each agent in a collaborative

manner. The key is the communication mechanism of the technological innovation RDM in entrepreneurial team, which includes

three parts: Communication Method, Communication Code and Communication Content as follows.

Communication MechanismbCommunication MethodN, bCommunication CodeN, bCommunication ContentN;Communication

MethodbSpotSpot CommunicationN |b Indirect CommunicationN |bConventions CommunicationN |bMixed communicationN;

Communication CodebNoticeN |bReservationN |bRequestN |bPromiseN |bNotifyN |bRefusedN |bResponseN;Communication Con-

tentbBehaviorN, b Source AgentN, bTarget AgentN, b TimeN, bCauseN, b InformationN.

In the multi-Agent system of technological innovation RDM in entrepreneurial team, the Agent group composed by senior

executivesis affected by technological innovation risks from internal andexternal environmentsthrough sensor functions. Information

is collected regarding three decision-making input variables in different areas: risk perception, value perception and risk preference.

Through event handling and distribution functions of Agent groups, an effective 3C (Communication, Coordination & Cooperation)

scheme can be carried out. Finally the

nal project decision-maker in entrepreneurial team is in

uenced to a great degree.

4. SDM of entrepreneurial team RDM

In the entrepreneurial team, we can integrate the system ow diagram of various senior managers' structure and amounts and

build DM SDM at different stages. TheSDMcan represent actual economic implications of thesystem structureowdiagramby theuse

of state variables, rate variables and auxiliary variables. Mathematical expressions are created to describe relationships among various

internal factors in the technological innovation RDM in entrepreneurial team and their links with external factors[17].

In the market managers agent module, market risk and value perception (mrv) and market investment decision-making (mdd)

are selected as time-varying decision-making rate variables, while consumer's demand change, enterprise reputation and

population (erp), the realization degree of market information (rmi) competitors' amount and strength (cas) and the ability of

dealing with risk (adr) and risk preference are selected as auxiliary variables. Therefore, in the technological innovation RDM in

entrepreneurial team, the dynamics relationship of this module can be expressed as follows:

Lt=Lt1mrvmdd; 2

where

mrv= Dtcdc; erp; rmi; cas; 3

mdd= Dtadr; rp; Time: 4

In the R&D managers Agent module, R&D risk and value perception (rrv) and R&D devotion decision-making (rdd) are selected

as time-varying decision-making rate variables, the maturation and advanced of technology (mat), the difculty and complexity of

technology (dct) and the difculty and complexity of the pilot test (dcp) and the whole ability and strength of project team are

selected as auxiliary variables. Therefore, the module's dynamics relationship can be expressed as follows:

Lt=Lt1rrvrdd 5



9/13

where,

rrv= Dtmat; dct; dcp; ast 6

rdd= Dtadr; rp; Time: 7

In the capital managers Agent module, the capital risk and value perception (crv) and the capital devotion decision-making

(cdd) are selected as time-varying decision-making rate variables, while the project cost (pc) and capital requirement (cr) are

selected as auxiliary variables. Therefore, the module's dynamics relationship can be expressed as follows:

Lt= Lt1crvcdd; 8

where

crv= Dtpc; cr; 9

and

cdd= Dtadr; rp; Time: 10

In the personnel managers Agent module, personnel risk and value perception (prv) and personnel devotion decision-making

(pdd) are selected as time-varying decision-making rate variables, while standard and ability of enterprise personnel (sa) and

enterprise management ability (ma) are selected as auxiliary variables. Therefore, the module's dynamics relationship can be

expressed as follows:

Lt= Lt1prvpdd; 11

where

prv= Dtsa; ma 12

and

pdd= Dtadr; rp; Time: 13

The above system dynamics model is essentially composed rst-order differential equations with time delay, where Dtis the

time interval of the model iteration, which can be set as 0.25, 0.5 or 1 and so on according to different circumstances. For the

functional form between the rate variables and auxiliary variables, we simply set iterative relationship in a general sense. The real

power of SDM is utilised through simulation. We will set explicit functional form in the case study ofSection 5.

5. Empirical analysis

In this section, we model the technological innovation project RDM from a Chinese Automobile company called Wuhan Genpo

Automobile Technology Development Co., Ltd. using SDM.

This company is mainly engaged in advanced manufacturing technology, automotive parts and components research,

development, production and business high-tech enterprise. In the domestic automotive brake industry, the company's product

design ideas, product manufacturing processes, product processes equipment, and product testing, testing conditions are among

the best performers of peers. Genpo strengthens R&D mainly using two measures: actively hiring talents, and fruitful cooperation

with domestic key universities.

Since its inception, the company has focused on the design, development, production and sales of the off-road vehicle central

tire ination deation control devices and the off-road vehicle central tire ination deation valve assembly.

After successfully developing a central tire ination deation hand control valve (Product1) almost dominating domestic

market in 2006, the company executive team started in late 2007 a discussion on whether to invest in R&D and production of the

central tire ination and deation valve with a connector assembly (Product2). The company formed a project RDM team mainly

including the general manager, the market manager and the R&D technology manager. By fully understanding the internal and

external risk factors and effective communication within the team, the company needs a rational decision-making scheme to

decide on whether project should be launchedor not.

Compared to the previous project, such a project faces potential large market risk: competitors in the domestic market also

invest in the development and production of similar products. Meanwhile, although a great deal of experience has been obtained

in R&D of Product1 project, the company faces erce competition and has to invest more funds in the new R&D project in order to

maintain the leading position of domestic products This creates a certain degree of R&D risk. In addition, Wuhan has been just

approved as Pilot city for national science and technology insurance. The city government has issued a large number of related

incentive policies to stimulate active innovation in high-tech enterprises. Taking into consideration these factors, the company

executive team is facing a complicated RDM on whether to invest in the technological innovation project and how much to investin.



10/13

5.1. Parameter values and initial conditions of SDM

Before using the SDM approach, parameter values and initial conditions need to be set to reect opinions from the

entrepreneurial team. Genpo's corporate management team members use expert knowledge to assign different weights to

different risks of the model inFig. 5. Using these expert knowledge leads to explicit functional form in Eqs. (14) and (15).

mrv= 0:2 erp + 0:3 rmi + 0:3 cdc+ 0:2 cas; 14

rrv= 0:3 dct+ 0:3 mat+ 0:3 dct+ 0:1 ast; 15

where notations are dened inSection 5. Different weights are also assigned for factors including personnel devotion decision-

making (pdd), the supporting power of innovation from government (spg), and the allocation degree of risk from society (ars) in

Eqs. (16) and (17):

mdd= rdd= 0:5 adr+ 0:5 rp; 16

pdd= 0:4 rp + 0:3 spg+ 0:3 ars: 17

Variable initial values represent the initial states of investment willingness of three executives in the team before the project

launched. They have to be set depending on the actual initial condition and then updated in the simulation to reect the

changing trends of the executives' investment willingness. Here the initial values of investment willingness of executives were setto 0. In order to express the relationship between the auxiliary variable and the rate variable more directly, we set an interval

constraint for intangible variables such as various risk factors of the market and development.

5.2. Model validity check

A validity check of system dynamics model of technological innovation RDM is necessary before running the simulation. There

are two testing methods: theory test and history test. Theory test is mainly based on the rationality of the model boundaries, the

authenticity of the relationship between model variables, dimension consistency and the rationality of the exogenous variables

and parameters, etc. History test compares model results with historical data in order to test the degree of goodness between the

model and the objective system. From the SDM perspective, the correctness of model structure is more important than the choice

of parameter values, so the test should mainly focus on consideration of the validity, consistency and adaptability of the model

structure, which means theory test should be adopted.

For the purpose of validity check of SDM of technological innovation SDM, we employ previous initial values of variables andGenpo investment willingness for project decision-making in Eq. (18).

DEC= INTEGfpdd+ 0:5Market Agent+ 0:5R & DAgentABSTime0:01g 18

where DEC denotes resulted investment willingness value based on which decisions are made, INTEG denotes an integral operator

and ABS(X) denotes the absolute value ofX. Eq. (18) indicates that our SDMs are essentially rst-order differential equation

systems with time lag.

Fig. 7.Investment willingness in Genpo.



11/13

Using Eq. (18), we calculate investment willingness values of technological innovation RDM in entrepreneurial team for

different periods.Fig. 7depicts investment willingness values for six typical project time points, where the initial investment

willingness value is set as 0. As similar products appear and competition intensity increases, on the one hand, Genpo started to

invest in R&D in order to maintain a competitive advantage of company products. However, Genpo faced lots of risks and

uncertainties in R&D projects. Such risks and uncertainties made the Genpo entrepreneurial team pessimistic about investing in

future R&D and production projectin almost one and a half years. As a result, the investment willingness values kept negative from

time 0 to around time 18(unit: month) in Fig. 7. On the other hand, the former research project on central tire ination and

deation valve with a connector assembly has been very successful from market introduction stage to mature stage, which

accumulated a great deal of experience in technology research. Such experience in a successful project made the senior

management team gradually perceive the project's market and R&D value. Meanwhile, as Genpo went through the early stage, the

entrepreneurial team has a strong awareness of various types of project risks. This leads to increased investment willingness after

the project was proposed for 10 months; and positive investment willingness after 18 months.

After consulting with the Genpo senior board, we conclude that our above agent-based SDM analysis re ects the reality of

Genpo company life cycle and product life cycle. Genpo senior board was willing to take advice suggested from our agent-based

SDM analysis. We then conduct detailed SDM simulation in the next section.

5.3. Simulation

The real power of SDM is gained through simulation. In this section, we conduct agent-based simulation using the popular SDM

software Vensim PLE. We will adjust risk parameters to see how risk-based investment willingness is affected. Initial conditions

are set as follows: INITIAL TIME=0, FINAL TIME=24 (Month), TIME STEP=0.5.

We ran the SDM simulation for three practical scenarios: scenarios 1, 2 and 3. Scenario 1 means Genpo does not take any risk

control measures. Scenario 2 corresponds to a situation where Genpo was suggested to buy insurance to cover possible

technological risks; here risks are controlled through adjusting the supporting power of innovation from government (spg), and

the allocation degree of risk from society (ars) in Eq. (17). In Scenario 3, our entrepreneurial team suggests Genpo to manage

possible technological risks from the beginning of the investment project; here risks are controlled through adjusting the company

ability of dealing with risk (adr). Fig. 8depicts technology innovation RDM investment willingness in entrepreneurial team in

these three scenarios.

To see some detailed values of three scenarios, we present investment willingness results in Table 2. Scenario 1 is actually

presented inSection 5.1.

Table 2

Investment willingness results.

Scenario 0 month 0.5 month 10.5 months 18 months 18.5 months 24 months

1 0 0.5 7.04134 0.556219 0.422031 16.7854

2 0 0.48 6.59558 0.402524 1.42912 18.5183

3 0 0.48 6.62134 0.163782 1.16203 17.7454

Fig. 8.Technology innovation RDM investment willingness.



12/13

Using such SDM results generates insights in two aspects. First, investment willingness is quantied and represented in real

values that can be compared. Statistic values are then generated to understand investment willingness of the entrepreneurial team

at different stages and make appropriate technology innovation RDM. Second, we can easily perform scenario analysis and

sensitivity analysis by adjusting various parameters during technological innovation RDM in the project. We can design different

mechanisms to realize preferred projects, where the designed measures in the form of model parameters can be recommended to

management.

From bothFig. 8andTable 2, we can see that in both Scenarios 2 and 3, Genpo will be better off if some measures are taken to

mange technology innovation risks. Interestingly enough, Scenario 2 dominates Scenario 3, which means measures using thesupporting power of innovation from government (spg), and the allocation degree of risk from society (ars) is better than that of

using the company ability of dealing with risk (adr). This coincides with the government practice: since the end of 2006, Wuhan

City strengthens the implementation of technology insurance to provide an opportunity for science and technology enterprises to

effectively control risks. Based on this analysis, we recommend to management:

During the early stages of R&D innovation project, Genpo should participate in the science and technology insurance; and

explore effective measures to circumvent the technological risks, enhance risk awareness of the team decision-maker and

change individual subjective decision-making to risk-based team decision-making.

Regarding technological innovation in domestic entrepreneurs, we can see that government support and the community's

concern are most effective factors for enterprises to upgrade their abilities of dealing with risks at present. In recent years, the

Wuhan municipal government has enhanced support efforts to high-tech enterprises in technological innovation. Obviously,

better external innovation environment has enhanced the technological innovation risk perceptions of Genpo and increased

investment willingness in technological innovation. In the technological innovation project of central tire ination deation valvewith a connector assembly, the process of technical innovation risk investment decision-making by management team re ected

the coordinated cooperation and co-adaptability among various agent-based systems. To some extent, it also reects the

intelligent reaction mechanism in technical innovation RDM in an entrepreneurial team.

6. Conclusions

Continuous technological innovation has been playing a vital role in ensuring the survival and development of an enterprise in

today's economy. We have studied the problem of technological innovation risk-based decision-making from an entrepreneurial

team point of view. We have built individual risk-based decision-making model with respect to the decision-maker's risk and

value perceptions, and risk preferences. We then create a modeling framework for such a team decision-making problem, and use

the system dynamics model to model it from the agent-based modeling perspective.

To validate the proposed approach, we conducted a case study of the technological innovation risk decision-making in aChinese automobile company called Genpo. Simulation of the system dynamics model generates outcomes that are consistent

with both the government and company practice. We have also performed scenario analysis and sensitivity analysis by adjusting

various parameters during technological innovation RDM in the project. Using scenario analysis, we can design various

mechanisms to achieve preferred projects. One of the key suggestions is that Genpo will be better off if some measures are taken to

manage technology innovation risks.

Acknowledgements

Project No. 70772076 supported by NSFC. We would also like to thank all the anonymous referees for their helpful comments

and suggestions, which led to an improved version of this manuscript.

References

[1] A.C. Cooper, A.V. Bruno, Success among high-technologyrms, Business Horizons 20 (1977) 20.[2] P.H. Thurston, When Partners Fall Out, Harvard Business Review, NovDec, 1986.[3] H.R. Feeser, G.E. Willard, Incubator and performance: a comparison of high and low growth high techrms, Journal of Business Venturing 4 (1989) 429442.[4] J. Doutriaux, Emerging high-tech rms: how durable are their comparative start-up advantages? Journal of Business Venturing 7 (1992) 303 322.[5] G.N. Chandler, S.H. Hanks, An Investigation of New Venture Teams in Emerging Businesses, Babson College, Wellesley, MA, 1998.[6] T. Lechler, Social interaction: a determinant of entrepreneurial team venture success, Small Business Economics 16 (2001) 263278.[7] D. Wu, D.L. Olson, Supply chain risk, simulation, and vendor selection, International Journal of Production Economics 114 (2008) 646655.[8] D. Wu and D. L. Olson, 2008b. Enterprise risk management: coping with model risk in a large bank. Journal of the Operational Research Society (accepted on

Sep 15,08)[9] M.P. Hekkert, R.A.A. Suurs, S.O. Negro, S. Kuhlmann, R.E.H.M. Smits, Functions of innovation systems: a new approach for analysing technological change,

Technological Forecasting and Social Change 74 (2007) 413432.[10] M.Y. Wang, W.T. Lan, Combined forecast process: combining scenario analysis with the technological substitution model, Technological Forecasting and

Social Change 74 (2007) 357378.[11] J. Shin, Y. Park, Brownian agent-based technology forecasting, Technological Forecasting and Social Change 76 (2009) 10781091.[12] J. Sterman, Business Dynamics; System Thinking and Modeling for Complex World, Mc Graw-Hill Companies, 2000.

[13] T. Fenzl, T. Brudermann, Risk behavior in decision-making in a multi-person-setting, Journal of Socio-Economics 38 (2009) 752

756.[14] D.A. Harper, Towards a theory of entrepreneurial teams, Journal Of Business Venturing 23 (2008) 613626.



13/13

[15] D. Wu, D.L. Olson, Introduction to the special section on optimizing risk management: methods and tools, Human and Ecological Risk Assessment 15 (2)(2009) 220226.

[16] D. Wu and D. L. Olson, 2009b. Enterprise risk management: a DEA VaR approach in vendor selection. International Journal of Production Research (acceptedand in print).

[17] I. Simona, M.C. Philip, The structure and evolution of industrial clusters: transactions, technology and knowledge spillovers, Research Policy 35 (2006)10181036.

[18] A. Tversky, D. Kahneman, Judgment under uncertainty: heuristics and biases, Science 185 (Sept. 27 1974) 11241131.[19] B. Fischhoff, Informed consent in societal riskbenet decisions, Technological Forecasting and Social Change 13 (May 1979) 347357.[20] D. Kahneman, A. Tversky, On the psychology of prediction, Psychological Review 80 (July 1973) 237251.[21] P. Slovic, B. Fischhoff, S. Lichtenstein, Facts and fears: understanding perceived risk, in: R. Schwing, W. Albers Jr. (Eds.), Societal Risk Assessment: How Safe Is

Safe Enough? Plenum, New York, 1980, pp. 181216.

Desheng Dash Wuis the afliate professor at RiskLab of University of Toronto and director of RiskChina Research Center at University of Toronto. His researchinterests focus on enterprise risk management, performance evaluation, and decision support system. He has published more than 40 journal papers appeared insuch journals asRisk Analysis,Decision Support Systems,International Journal of Production Research,European Journal of Operational Research,Expert Systems with

Applications,Socio-Economic Planning Sciences,International Journal of Production Economics,Annals of Operations Research,Journal of Operational Research Society,IEEE Transactions on Knowledge and Data Engineering,Computers and Operations Research,International Journal of System Science, et al. He has coauthored 3 bookswith David L Olson. He has served as editor/guest editors/chairs for several journals/conferences. The special issues he edited include those for Human andEcological Risk Assessment (2009, 2010), Production Planning and Control (2009), Computers and Operations Research (2010), International Journal ofEnvironment and Pollution (2009) and Annals of Operations Research (2010). He is a member of PRMIA (the Professional Risk Managers' InternationalAssociation) Academic Advisory Committee and steering committee member.

Kefan Xieis a professor and the deputy dean of Management School, Wuhan University of Technology. He worked at University of Kyoto as a Post doctor from2000 to 2003. In the recent several years, he has published over 170 papers and over 9 books, and has presided more than 30 research projects, including 4 fromNSFC (National Natural Science Foundation of China). He was the winner of Huo Yingdong educational funds, excellent young teachers' subsidizing project fromMinistry of Education of PRC, and Twilight Program of Wuhan. And he has been awarded more than 11 ministerial level and provincial prizes. His principalresearch interests include risk management, S&T (Science & Technology) management, strategic management, independent innovation, and industryuniversityalliances.

Hua Liu received his MS degrees in Management Science and Engineering from Wuhan University of Technology in 2007. He is currently a PhD candidate inFinance Engineering and Management in Wuhan University of Technology. His major research interests include scienti c and technological insurance, scienticand technological nance, and risk management. He has published over 20 papers, and has taken part in over 10 projects from enterprises or government.

Shi Zhao received his BS degree in Project Management from Wuhan University of Technology in 2008.He is currently a second grade postgraduate inManagement Science and Engineering, Wuhan University of Technology. His major research interests are in Science & Technology (S&T) management, groupdecision-making, and risk decision-making. He has published 8 papers in international conferences or domestic journals. He has taken part in over 10 projectsfrom national governments or international partners or enterprises.

David L. Olsonis the James & H.K. Stuart Professor in MIS and Chancellor's Professor at the University of Nebraska. He has published research in over 100 refereedjournal articles, primarily on the topic of multiple objective decision-making. He has authored or coauthored over 20 books. He is a member of the Association forInformation Systems, the Decision Sciences Institute, the Institute for Operations Research and Management Sciences, and the Multiple Criteria Decision MakingSociety. He was named Best Enterprise Information Systems Educator by IFIP. He is a Fellow of the Decision Sciences Institute.


modeling technological innovation

Documents