Managing Complex Issues through Evolutionary Learning ...lab.sdm.keio.ac.jp/idc/yasui/papers/j15_2013srbs_bosch...progress is the lack of common understanding and shared vision of

■ Research Paper

Managing Complex Issues throughEvolutionary Learning Laboratories

Ockie J. H. Bosch1*, Nam C. Nguyen1, Takashi Maeno2

and Toshiyuki Yasui21 Systems Design and Complexity Management Alliance, Business School, The University of Adelaide, Adelaide,SA Australia2Graduate School of Systems Design and Management, Keio University, Hiyoshi, Japan

Policy makers, managers and leaders in organizations, governments and business institu-tions are under increasing pressure to make the right management decisions in the face ofa continually changing political and socio-economic landscape. To make matters morechallenging, the complex environmental, socio-economic, business-financial issues thatdecision makers need to deal with tend to transcend the jurisdictions and capacities of anysingle organization. There is a multitude of difficult, long-term global challenges ahead,almost all of which are coupled with the most pressing concerns of different countries atnational and local levels. Despite many efforts to deal with these complex issues facingour society, the solutions so far have seldom been long lasting, because ‘treating the symp-toms’ and ‘quick fixes’, using traditional linear thinking, are the easiest way out, but donot deliver the solutions. This paper describes the processes for unravelling complexitythrough participatory systems analysis and the interpretation of systems structures to iden-tify leverage points for systemic interventions. It further demonstrates the promotion ofeffective change and the enhancement of cross-sectoral communication and collaborativelearning. This learning focuses on finding solutions to complex issues by applying an itera-tive, systems-based approach, both locally—Evolutionary Learning Laboratory (ELLab)—and globally—Global Evolutionary Learning Laboratory (GELL). A generic frameworkand processes for implementing and institutionalizing ELLabs are described, and how thesebecome part of the GELL for managing complex issues is explained. Four case studies areused to demonstrate diverse examples of the application and implementation of the ELLabapproach. Copyright © 2013 John Wiley & Sons, Ltd.

Keywords management; policy making; investment decisions; complexity; systems thinking;participatory systems analysis; Global Evolutionary Learning Laboratory (GELL)

*Correspondence to: Ockie J. H. Bosch, Systems Design and Complexity Management Alliance, Business School, The University of Adelaide,Adelaide, SA 5005 Australia.Email: [email protected]

Received 24 October 2012Accepted 14 January 2013Copyright © 2013 John Wiley & Sons, Ltd.

Systems Research and Behavioral ScienceSyst. Res (2013)Published online in Wiley Online Library(wileyonlinelibrary.com) DOI: 10.1002/sres.2171

INTRODUCTION

Complexity characterizes theworld and all humanendeavours today—in business, government,social, natural, scientific and political spheres.Local and global problems and challenges facingour world today are highly complex in nature,involving decision makers, scientists, NGOs andvarious other stakeholders. These problems andchallenges cannot be addressed and solved in iso-lation and with the single dimensional mindsetsand tools of the past. Collaborative, systemicand integrated approaches are essential to deliverthe sustainable outcomes desired. It has becomecrucially important for decision makers andmanagers involved in the management of anysystem to be equipped with the necessary capabil-ities and skills to make good policy and manage-ment decisions.

In recent years, there has been a growingrecognition of human capacity development as akey lever for sustainable economic, social andecological development. However, recent litera-ture on the success of external actors and agenciesin implementing effective change in developingcountries or regions shows poor outcomes acrossthe board (Umaña 2002; Land et al., 2009; Thomasand Amadei 2010). One of the key barriers toprogress is the lack of common understandingand shared vision of how to address the complexissues facing our world. The lack of cross-functional collaboration leads to fragmenteddecision-making and uncoordinated actions.This is further exacerbated by cross-purposenegotiations, the wasting of public and naturalresources, and a loss of confidence in leadershipand governance. Over time, these all escalate intoa vicious cycle of mediocre performance and pooroutcomes for all concerned. A further importantcontributor to poor outcomes is the fact thatmany of the ways in which problems are beingaddressed are simply ‘quick fixes’ or ‘treatingthe symptoms’. The establishment of a systems-based Learning Laboratory (LLab) has proven tobe an innovative and effective approach (Boschand Nguyen 2011; Nguyen et al., 2011) for dealingwith highly complex and multi-dimensional pro-blems and ensuring that solutions will be foundat the level of the root causes.

In addition, we manage the systems we are partof in a highly compartmentalized structure—organizations, divisions within organizations,business institutions, government departments,university schools, disciplines and so on. Thesestructures help our society to operate in an orderlyway. However, without an understanding thatall these different sectors in life are highlyinterconnected and that there is a strong needfor interdisciplinary, cross-sectoral communicationand collaboration, solutions that effectively addressthemulti-dimensional andmultidisciplinary natureof complexity will remain elusive.

This paper presents the methodology and appli-cation of a ‘new way of thinking’ and radicalapproach to enhancing cross-sectoral and organi-zational communication and collaboration, to dealwith increasing complexity and to promote effect-ive change at local and global levels.

SYSTEMS THINKING

Although systems thinking is an ‘old’ concept(Midgley 2003), it is increasingly being regardedas a new way of thinking to understand and man-age complex problems at both local and globallevels (Bosch et al., 2007b; Cabrera et al., 2008).Maani and Cavana (2007) used the analogy of aniceberg to illustrate the conceptual model knownas the Four Levels of Thinking (Figure 1) as aframework for systemic interventions.

In this model, events or symptoms (those issuesthat are easily identifiable) represent only the visi-ble part of the iceberg above the waterline. Mostdecisions and interventions currently take placeat this level, because quick fixes (treating thesymptoms) appear to be the easiest way out,although they do not provide long-lasting solu-tions. However, at the deeper (fourth) level ofthinking that hardly ever comes to the surface arethe ‘mental models of individuals and organisa-tions that influence why things work the way theydo. Mental models reflect the beliefs, values andassumptions that we personally hold, and theyunderlie our reasons for doing things the way wedo’ (Maani and Cavana 2007, p.15).

Moving to the third level of thinking is a criticalstep towards understanding how these mental

RESEARCH PAPER Syst. Res

Copyright © 2013 John Wiley & Sons, Ltd. Syst. Res (2013)DOI: 10.1002/sres.2171

Ockie J. H. Bosch et al.

models can be integrated in a systems structurethat reveals how the different components areinterconnected and affect one another. Thus,systemic structures unravel the intricate lace ofrelationships in complex systems.The second level of thinking is to explore and

identify the patterns that become apparent whena larger set of events (or data points) becomelinked to create a ‘history’ of past behaviors or out-comes and to quantify or qualify the relationshipsbetween the components of the system as a whole.The systems thinking paradigm and method-

ology embrace these four levels of thinking bymoving decision makers and stakeholders fromthe event level to deeper levels of thinking andproviding a systemic framework to deal with com-plex problems (Maani and Cavana 2007).The application of systems thinking has grown

extensively and encompasses work in many di-verse fields and disciplines such as, to mentionbut a few, management (Jackson, 2003), business(Sterman 2000; Walker et al., 2009), decision mak-ing and consensus building (Maani and Maharraj2004), human resource management (Quatroet al., 2007), organizational learning (Galanakis2006), health (Newell 2003; Lee 2009), commoditysystems (Sawin et al., 2003), agricultural productionsystems (Wilson 2004), natural resource manage-ment (Allison and Hobbs 2006), environmentalconflict management (Elias 2008), education (Hung

2008), social theory and management (Mingers2006), and food security and population policy(Keegan and Nguyen 2011). This paper is the firstto demonstrate how a comprehensive systemsthinking approach, embedded in a cyclic Evolu-tionary Learning Laboratory (ELLab) framework,can be used to deal effectively with complex issuesin a variety of contexts.

ESTABLISHING A SYSTEMS-BASEDEVOLUTIONARY LEARNING LABORATORY

The LLab is a process, as well as a setting, inwhich a diverse group of participants engage ina cyclical process of thinking, planning, actionand reflection for collective learning towards acommon good. It is an environment where policymakers, managers, local facilitators and research-ers collaborate and learn together to understandand address complex problems of common inter-est in a systemic way (Maani and Cavana 2007).The ultimate goal is to achieve coherent actionsdirected towards sustainable outcomes.

The ELLab is a seven-step iterative process(Figure 2) of group thinking and acting in whichthe participants engage in well-defined activitiesand thus learn together in an ‘experimenting lab’environment about how best to deal with the com-plex multi-dimensional and multi-stakeholder

The Iceberg Approach$$$ for addressing symptoms or events (Quick Fixes)

Symptoms/Events

?

Systems ApproachAddressing fundamental problems to achieve sustainable systems

Mental Models/Mind Maps – People’s Understanding

$$$ for mitigating unintended consequences

$$$ for root causes of issue

Patterns –interactions between

components

Systemic Structures What does system look

like

Figure 1 The iceberg approach versus a systems approach

Syst. Res RESEARCH PAPER


Managing Complex Issues through ELLabs

problems they are facing. Although it builds onevolutionary design principles as described in thework of Banathy (1996) and the concept of evolu-tionary leadership developed by Laszlo (2001),the process of establishing an ELLab (Figure 2)could be regarded as a unique ‘methodology’ tocollaboratively integrate and use existing and fu-ture knowledge to help manage complex issues.It starts at the ‘fourth level of thinking’ with anissues workshop (step 1) and a series of forumswith specialist groups to gather the mental modelsof all stakeholders involved in the issue underconsideration, their perceptions of how the systemworks, what they regard as barriers to successand drivers of the system, and possible strategies(solutions) to overcome these problems.

This is followed by implementing the ‘thirdlevel of thinking’ through follow-up capacity-building (step 2) sessions during which the partici-pants (all stakeholders) learn how to integrate thevarious mental models into a systems structure(step 3). The Vensim software program (Systems2011) is a valuable tool for the development of asystems model (causal loop diagram) of the issueunder consideration. This learning step is ofparticular importance in order for all involved totake ‘ownership’ of the systems model.

Once completed, the participantsmove to the ‘sec-ond level of thinking’ by interpreting and exploring

the model for patterns, how different componentsof the model are interconnected and what feedbackloops, reinforcing loops and balancing loops exist.This step aims to assist relevant stakeholders to de-velop an understanding of their interdependenciesand the role and responsibility of each stakeholdergroup in the entire system. The main barriers anddrivers of the system are discussed in more detail,which provides the stakeholders with an oppor-tunity to develop a deeper understanding of theimplications of coordinated actions, strategiesand policies. Overall, this process provides allstakeholders with a better understanding of eachother’s mental models and the development of ashared understanding of the issue(s) underconsideration.

The interpretation leads to the identification of le-verage points for systemic intervention (step 4). Le-verage points are places within a complex system(e.g. an economy, a living body, a city and an eco-system) ‘where a small shift in one thing can pro-duce big changes in everything . . . leverage pointsare points of power’ (Meadows 1999, p.1). Senge(2006, p.64) also refers to leverage points as the‘right places in a system where small, well-focusedactions can sometimes produce significant, endur-ing improvements’. Identification of leveragepoints greatly assists the devising of systemic inter-ventions (finding systems-based solutions) that willcontribute to the achievement of goals or solvingproblems in the system under consideration.

The outcomes are used to develop a refinedsystems model, which forms at the same time anintegrated master plan (step 5) with systemicallydefined goals and strategies (systemic interven-tions). In order to operationalize the master plan,Bayesian belief network (BBN) modelling (Cainet al., 1999; Smith et al., 2007) is used to determinethe requirements for implementation of themanagement strategies; the factors that couldaffect the expected outcomes; and the order inwhich activities should be carried out to ensurecost effectiveness and to maximize impact.

The process of developing good policies and in-vestment decisions is based on the best knowledge(scientific data and information, experientialknowledge, expert opinions) that is available atany point in time. The systems model can be usedto test the possible outcomes of different systemic

Figure 2 Evolutionary Learning Laboratory for managingcomplex issues




interventions by observing what will happen tothe system as a whole when a particular strategyor combination of strategies is implemented, thatis, before any time or money is invested in actualimplementation.Of particular value is the ability through BBN

modelling to also ‘back-cast’. That is, the goal isset at a 100% probability that it will be achievedand the model back-casts and points out which ofthe components, actions or conditions have themost influence on the achievement of the goal.This is a powerful way of determiningwhere to in-vest time and resources, instead of having only alist of recommendations, without an understand-ing of how they are interconnected, which onesare the most important to invest in and in whatorder the strategies should be implemented to en-sure an efficient and cost-effective plan of action.Once the systemic interventions have been

identified and an operational plan has beendeveloped, the next step for the people who areresponsible for the different areas of managementis to implement the strategies and/or policies (step6) that will create the biggest impact. Targets aredetermined, and monitoring programs are imple-mented to measure and/or observe the outcomesof the strategies and policies. Inmany cases, it onlyrequires an adjustment of existing monitoringprograms to comply with the targets set withinthe ELLab process (e.g. to include factors to bemeasured that were used in the construction ofthe Bayesian management model).Because no systems model can ever be com-

pletely ‘correct’ in a complex and uncertain worldand unintended consequences always occur, theonly way to manage complexity is by reflecting(step 7) at regular intervals on the outcomes of theactions and decisions that have been taken to deter-mine how successful or unsuccessful the interven-tions are and to identify unintended consequencesand new barriers that were previously unforeseen.The iterative process serves as a valuable informal

co-learning experience and leads to new levels ofcapability and performance. Working in this wayas a coalition is the most effective way to deal withcomplex issues; because themethodologies and pro-cesses acknowledge that complex problems aremulti-dimensional and have to involve all stake-holders, they require cross-sectoral communication

and collaborative approaches to resolve, and dealwith many uncertainties that need adaptive man-agement approaches as more knowledge becomesavailable through the iterative process of learningby doing.

USING ELLABS TO DEALWITH COMPLEXISSUES IN AVARIETY OF CONTEXTS

As mentioned earlier, the ELLab approach isgeneric and can be used in dealing with anycomplex issue, regardless of its context (e.g.organizational, natural or social systems) or dis-cipline area under consideration (e.g. business,health, engineering, education and marketing).In the following sections, four case studies areused to demonstrate four diverse examples ofthe application and implementation of the ELLabapproach.

Sustainable Development of a UNESCOBiosphere Reserve in Vietnam

Biosphere reserves (BRs) are sites recognizedunder the UNESCO Man and the Biosphere(MAB) program to demonstrate innovativestate-of-the-art approaches to conservation andsustainable development. A comprehensivedescription of the origin and the evolution of theBR concept is presented in a paper (Ishwaranet al., 2008). There are currently 580 BRs in 114countries (UNESCO 2012). UNESCO has recom-mended the launch of pilot projects to use BRsas learning laboratories to address the gap be-tween BR knowledge systems (scientific, experi-ential and indigenous) and the imperative forwider sustainable development. In this regard,the first pilot project, the Cat Ba Biosphere Re-serve (CBBR) sustainability project in HaiphongCity, Vietnam, has been initiated (Nguyen et al.,2011). The project focuses on the interconnected-ness of environment, tourism, livelihood ofpeople and economic benefits, and the adoptionof policies and processes by government andmanagement bodies to ensure that long-termsustainable management will become institutio-nalized and ongoing.




Identify IssuesTwo workshops were conducted in March andOctober 2007 (Bosch et al., 2007a) with a rangeof stakeholders to gather their mental modelson the key issues and challenges that Cat BaIsland is facing. These include waste treatment,pollution, the high number of floating farms,overuse of underground water, strong growth intourism, lack of fresh water and electricity(especially in the summer—tourist season), lackof skilled labour for the tourism industry,uncontrolled tourism development, insufficientinfrastructure, lack of access to suitable marketsfor locally produced products, encroachment onconservation areas, lack of integrated planning,lack of capacity, environmental degradation andpoverty.

Build CapacityA 2-month systems thinking and associatedcapacity-building program was subsequently con-ducted in Australia (October and November 2008)for a group of 10 policy makers, managers andtechnical officers from different levels of govern-ment, across sections of agencies and an NGO,engaged in different capacities in the managementof the CBBR. The process and outcomes of thiscapacity-building program have been reported ina recent paper (Nguyen et al., 2012).

Develop a Systems ModelDuring the capacity-building program, partici-pants worked with the research team to integratethe various issues identified in the issueworkshops into a preliminary systems model.Subsequently, the model (Figure 3) was refinedand validated by various relevant stakeholders(managers and rangers of Cat Ba NationalPark, hotel owners, farmers, local people andofficials from different government depart-ments) in a series of workshops, focus groupdiscussions and in-depth interviews conductedin Haiphong City and on Cat Ba Island atthe end of 2008 and early 2009. This involvementin the evaluation of the model was criticalbecause it led to taking ownership of the modeland enhanced the ability of stakeholders to

understand and carry out future interventionstrategies and actions aimed at improving thesystem for sustainable outcomes.

Figure 3 illustrates the identified interrela-tionships and interdependencies amongst thekey components of the system. The systemsmodel represents a ‘big picture’ of the CBBRsystem and provides a useful platform forlearning, collaboration and decision makingfor relevant stakeholders including policymakers, researchers, managers, practitionersand local people.

Identify Leverage Points and Systemic InterventionsA follow-up workshop was conducted inHaiphong City in May 2009 with the main objec-tive to identify key leverage points and areasfor systemic interventions for sustainability—onthe basis of the systems model of the CBBRand its associated systems archetypes. Systemsarchetypes ‘reveal an incredibly elegantsimplicity underlying the complexity of manage-ment issues. . . [they allow us] to see more placeswhere there is leverage in facing difficultchallenges, and to explain these opportunities toothers’ (Senge 2006, p. 93). Four systems arche-types were identified in the systems model ofthe CBBR—‘limits to growth’, ‘fixes that fail’,‘tragedy of the commons’ and ‘shifting theburden’. These archetypes are discussed in detailby Nguyen and Bosch (2012) and not repeated inthis paper.

The leverage areas require systemic interven-tions that are deemed critical for the long-termsustainability of the CBBR. Those identifiedincluded cross-sectoral collaboration; develop-ment and implementation of government plans;capacity building for decision makers, managersand local people; waste management and treat-ment; people’s awareness; conservation ofendangered species; investment for agriculture;improving the livelihood of commoners; andtourism development. These leverage areas formthe basis for integrated projects and policiescovering multiple aspects of the sustainability ofthe CBBR, including social, economic, culturaland environmental well-being.




Develop Action PlansA series of Bayesian models were constructed todevelop action plans for the identified leverageareas and systemic interventions. An example ofthese models is illustrated in Figure 4.The Bayesian model developed in this study

(Phan 2011) is designed as a decision support toolto assist the management board of the CBBR andCat Ba National Park in developing feasible man-agement and action plans for the conservationand protection of the population of an endangeredspecies (serow—mountain goat) in the CBBR.Short-term and long-term measures for this

endangered species are needed. In the short term,stronger engagement of local people, especiallythe potential poachers, to participate in serowprotection is necessary. Intensifying patrol activ-ities in prioritized conservation areas is needed

to avoid any further loss of individual animals.Simultaneously, more stringent law enforcementby authorities and adopting more severe punish-ment measures for illegal hunting are required.

In the long term, providing opportunities toimprove the financial position of the poorthrough technical support and education is oneof the most important and sustainable solutionsto improve the livelihoods of people on theisland. This would avoid the increasing impactof local people on the resources of the NationalPark. Raising the conservation awareness of localresidents and improving the knowledge andmanagement capacity on biodiversity conserva-tion and conservation planning of managers ofthe Cat Ba National Park are vital to ensure aneffective conservation outcome in the CBBR(Phan 2011).

Number of tourists

Tourism revenue

Hotels andRestaurants

Waste

Tourism pollution

Attraction of CBisland

Use ofunderground water

Availableunderground water

Biodiversity

S

S

S

S

S

O S

S

O

S

S

R_T1

B_T1

Livelihood ofCommoner

Misuse of NRNR conservation

GDP per capita

Agriculturerevenue

Investment inagriculture

Information andcommunication

Access to market

Poverty

Health

PopulationEducated

population

S

S

O

O

S

SS

S

S

S

S

O

OS

R_Env

R_Eco1

REco2

REco3

R_S1

B_T3

Life expectancy

S

S

S

B_T2

Governancestructure

Other incomesources

Social evils/crimeS

O

R_S2NGOs

Infrastructure

S

S

S

R_T2

R_T3

ServicesO

S

B_T4

The system is influenced by

Policies

Student populationSS

Cultural valuesS

S

R_S3

Agriculturepollution

S

BEco

Tourismdevelopment People's

awareness

Immigration

S O

S

Food safety

New construction

O S

O

Other pollution sources

S

O

Figure 3 Systems model of CBBR—a platform for collaboration (adapted from Nguyen et al., 2011). Legend: S, same direction;O, opposite direction; R, reinforcing; B, balancing; T, tourism; Eco, economic; Env, environment; S, social. 1, 2, 3 refer to loop

number; for example, R_T1, reinforcing loop no. 1 of tourism




ImplementationA series of strategies are currently being imple-mented to improve the livelihood of the com-moner. A comprehensive model has also beendeveloped for sustainable tourism developmentas a mechanism for improving the livelihoods ofpeople on the island (Mai 2012), whereas modelsfor improvingwastemanagement and agriculturalmarket access are currently being completed.Several small projects and actions have also beenundertaken to address the various leverage pointsand systemic intervention strategies that had beenidentified from the systems model and its asso-ciated systems archetypes. These include buildingthe capacity of the rangers to systemically managethe National Park; conducting a social welfarestudy relating to community development in theCBBR; producing an annual Cat Ba EcosystemHealth Report Card; establishing communitypartnerships in natural resource managementand environmental protection; and relocating thefloating farms away from main tourism areas andout of the national marine protected areas.

ReflectionThe early and consistent involvement of keydecision makers and stakeholders (nearly 200participants to date) has been of paramountimportance for the successful formation andimplementation of an ELLab for sustainabilityin the CBBR. This involvement will be of signifi-cant importance for the seamless continuity andsustainability of the project.

Frequent reflection on the successes and failuresof implemented strategies (systemic interventions)has led to new knowledge and ideas. For example,to enhance awareness of sustainable practicesand increasing employment of locals, a CBBRbrand system has been introduced that is awardedto products (e.g. fish sauce and honey) and busi-nesses (e.g. tourist boat services, recreation parks,hotels, guest houses and restaurants) that complieswith a set of relevant criteria such as businessregistration, water savingmechanisms, employinglocal people, fire safety standards, food safetyand hygiene standards. The collaborative learningprocess has also led to a strong realization that theCBBR management regulations need revision,especially to improve integrated planning andactions across different sectors of society.

Policy Design for Child Safety in Japan

In OECD member countries, more than 125 300children died from injuries from 1991 to 1995,which amounts to 39% of all deaths. Japan wasranked as a medium-risk performer in deaths bydrowning, fire, falls and intentional harm,whereasdeaths due to car accidents were significantlylower than in other countries (UNICEF. 2001).Japanese society often regards parents as the onlypeople responsible for child safety. Japaneseparents tend to feel isolated and frustrated,because there is a clear lack of a coordinatedapproach with other stakeholders in the societyto help prevent injury to their children (Kakefuda

Distance_village0 to 25122512 to 3576>= 3576

33.132.634.3

2820 ± 1300

Distance_station0 to 12941294 to 1998>= 1998

33.132.634.3

1560 ± 750

Total_forest0 to 4848 to 62>= 62

31.735.932.4

49.7 ± 20

Steepness0 to 2828 to 41>= 41

30.932.037.0

33 ± 15

PresencePresentAbsent

41.558.5

0.415 ± 0.49

Elevation0 to 7979 to 125>= 125

33.132.634.3

97 ± 48

Figure 4 Bayesian model of serow occurrence in the CBBR (adapted from Phan, 2011)




et al., 2008). The complexity of this issue warranteda participatory systems analysis approach to createpossible solutions by embedding the systemsmodel in an ELLab context in order to ‘experiment’with potential solutions that could lead to betterpolicies for a safe and secure society.

Identify the IssuesThe mental models of a wide variety of rele-vant stakeholders about the issue were obtainedfrom a focus group meeting (conducted inSeptember 2011) to identify and visualize allfactors related to child injuries.

Build Capacity and Develop a Causal Loop ModelA workshop was held in September 2011 duringwhich various stakeholders collaboratively con-structed a causal loop diagram to identify thecomponents of the system and to explore theinteractions and relationships between them.The facilitator of the group had undergone inten-sive training in systems methodologies, whichmade it possible to structure the mental modelsof the various participants into a model.

Identify Leverage Points and Systemic InterventionsSpecial attention was given to the identificationof reinforcing and balancing loops in order to assistin the identification of possible leverage pointsfor systemic interventions. This was carried outthrough visual observation and discussionsbetween participants on the potential degree ofchange that could be caused by changes to particu-lar components of the system. Seven systemicintervention points were identified (in bold,Figure 5): safer product designs, caring volunteersto support frustrated parents, closer involvementof social workers, more integrated approach bygovernment, more paediatricians, shortening ofthe time between an accident and hospitalization,and better care of students in schools.

Develop Action PlansThe participating stakeholders used the sevensystemic intervention points to structure a BBNmodel for designing policies on child safety(Figure 6). The model was populated by variousstakeholders who jointly used their experientialknowledge to decide on the probabilities of howthe parent nodes would affect the child nodes.For example, what are the probabilities that morescholarships and better insurance policies will

Figure 5 Causal loop diagram and identified systemic interventions points for child safety in Japan




(a)

(b)

Figure 6 Populated Bayesian model for child safety: (a) current conditions and (b) indicating the main leverage points andsystemic interventions that were identified




lead to an increase in the number of paediatri-cians? How would designer training and agovernment that could test the designs changethe probability that the design of products willbe safe? and What is the probability that therewill be less school accidents if there are moreschool volunteers and smaller classes?Through this co-designing process, the stake-

holders recognized that the Bayesian model,populated with information about the currentconditions, indicated that there is only a 19.0%probability that the rate of child injuries will bereduced (Figure 6(a)).A sensitivity analysis of the model indicated

that the most effective parameter to reduce childinjuries was to increase the number of volunteernursing councillors (Figure 6(b)). The model indi-cated that if the number of volunteer nursingcouncillors is set at 100%, the probability thatless child injuries will occur will rise to 46.7%.However, also providing designer training in childsafety, establishing a government board forproduct evaluation, reducing the size of classes inschools and having sufficient numbers of paedia-tricians will increase the probability to have lesschild injuries to about 72%. Therefore, although apolicy to increase the number of volunteer nursingcouncillors would make a big difference, theseadditional four systemic interventions were alsoidentified as important to significantly reducechild injuries (step 5).

ImplementationA change in the policy to increase voluntarynursing staff and implementation of the additionalsystemic interventions have been proposed inorder to experiment how these interventions willaffect child injuries.

ReflectionThe models have been constructed with the bestexperiential knowledge available at the time.These models are therefore embedded in thecyclical process of ‘experimenting’ and reflectingthrough which new knowledge will be created.Strategies will be refined in a co-learning

environment to find the best solutions for thiscomplex problem over time—forming the ELLab.

Enhancing the Reputation of a UniversitySchool in Japan

The Graduate School of Systems Design andManagement (SDM) at Keio University in Japanwas established in 2008. This school is rapidlybecoming a focus point in the Asia-Pacific regionfor its mission to educate students who can solvecomplex and large-scale problems in any systemranging from social (human dimensions) tohighly technological issues. The school is build-ing its foundation on systems and design think-ing and has a strong focus on industry andcommunity needs, while taking into account thatall problems are embedded in a complex web inwhich environment, security and safety, healthand welfare, economics, politics and culture areall highly interconnected. What makes the schoolparticularly unusual is the fact that it attracts stu-dents for masters and PhD programs from all dif-ferent disciplinary backgrounds (Figure 7), whichcreates a collaborative learning environment forthe evolution of creative and innovative thinkingand systems design.

In April 2011, SDM decided to revisit its initialvision and strategies in order to develop a‘clearer and more committed operation’ and tobe recognized as a world-class institution in thearea of systems design. Because of the complexityof this task and the intention of the school to findlong-lasting solutions, rather than quick fixes,SDM decided, as part of this process, to establishthe school as an ELLab.

Identify the Issues and Build CapacityThe first step was to hold a workshoprepresented by a number of students and staffmembers who were all trained (step 2) in thedevelopment and interpretation of systemsmodels. The participants’ mental models onhow they believe the school can improve its repu-tation, the drivers and barriers in achievingthis and possible solutions to overcome thebarriers were collected.




Develop and Interpret a Bayesian BeliefNetwork ModelIn this particular case, the mental models wereintegrated by directly structuring them into aconceptual inference diagram, which formed thebasis for a BBN model (Figure 8).

The probability tables were populated with theexperiential knowledge (mental models) of the par-ticipants to form a first draft model that describedthe main components of the system and how theyare related to each other (step 3). Testing of differentscenarios by changing different components of themodel and combinations of components facilitatedan evaluation of how well the model reflects thereal situation.With this information, the probabilitytables were revisited and refined until the modelprovided a realistic description of the current schooland the system in which it operates.

Identifying Leverage Points and SystemicIntervention StrategiesPatterns and relationships were exploredby changing each of the components in the ‘whatcan we do’ or ‘action’ nodes of the model indivi-dually to observe how such a change affects theend goal of SDM to be recognized as a world-renowned school with a reputation of excellence.

Appointing or consulting a competitive intelli-gence professional that can provide appropriate

intelligence for different audiences (e.g. industryand potential students) formore effective promotionand marketing of SDM [the probability for this tooccur changed from 54% to 90%—comparingFigure 8(a, b)] will have the largest single effect onachieving the goal to become a world-class school,increasing the probability from 64% to 72%. Otheroutcomes that will improve the probabilities forachieving the end goal include an increase in thenumber of applications (from 58% to 86%) and theprobability that more high-quality professors willbe attracted to SDM (from 57% to 85%). A furtherimprovement of the relationships that SDM alreadyhas with industry will have the second largest effecton the goal. This will lead to the probability toincrease the budget of SDM from 56% to 80%; forstudents to have access to better research facilitiesfrom 63% to 90%; and the ability to fund languagetraining from 52% to 72%.

Implementing both the aforementioned actionswill lead to an increase in the probability toachieve the end goal from 64% to 76%. This prob-ability can further be increased to 80% by review-ing the criteria for entry to SDM. More stringentcriteria will lead to a higher probability of high-quality students; and if they have good commu-nication skills (through language training) andwork under the supervision of high-quality pro-fessors (who are attracted by good promotion),the probability for high-quality research willincrease from the current 61% to almost 80%.

Science and Technology, 25

Engineering, 21

Law, 7Environment, 5

Economics, 4

Politics, 3

Literature, 3

Pedogogy, 2

Commerce, 2

Agriculture, 2

Others, 21

Figure 7 Diverse backgrounds of students in the 2012 postgraduate class




ImplementationIn summary, to achieve SDM’s goal of being recog-nized as a world-class institution, investmentshould first be in appointing or consulting a com-petitive intelligence professional and in furtherenhancing its relationships with industry. A com-bination of these two actions will have the biggesteffect on the end goal. Other actions that could beimplemented, but would not significantly

contribute to achieving the end goal, include theprovision of language training and more stringentselection criteria to ensure high-quality studentswith good communication skills.

The school is consulting an expert in the area ofcompetitive intelligence (one of its staff members)to develop effective marketing and promotionmaterial and mechanisms for different types ofaudiences (e.g. large companies, potential students

(a)

(b)

Figure 8 First draft BBN model to enhance the reputation of the school: current situation (a) and with systemic interventionsimplemented (b)




and government departments). Stronger collabor-ation with industry is being established throughthe selection of real issues in different companiesand government agencies for student assignmentsand masters projects (e.g. Toshiba, NEC and Yoko-hama City).

ReflectionThe effects of changes in those parameters ofthe model that were identified during model de-velopment as being affected by the actions under-taken are being monitored (e.g. increase in theschool budget, language competency, number ofstudent enrolments and availability of quality re-search facilities in industry). The outcomes of thesewill be used to refine the first draft model—start-ing the cyclic process of experimenting and adapt-ing of the SDM ELLab.

Managing Tree Density in the Rangelands ofNorthern Queensland, Australia

Much of Australia’s grazing land is composed ofwoodland. Trees and native pastures coexist inthese ecosystems, where they compete for water,nutrients and sunlight. However, there is also amutually beneficial relationship between treesand pastures, provided that the balance is right.When a favourable tree–grass balance exists,trees provide shade and shelter for livestockand support biodiversity. They also carry outkey ecosystem functions, such as water and soilnutrient cycling, and contribute to healthy landcondition by preventing erosion and salinity,storing carbon and enhancing soil condition(Liedloff and Smith 2010).

There is an increasing recognition of the rolethat trees play in grazing systems, which hasled to a demand for sustainable woodlandmanagement. Of particular importance is theman-agement of tree cover thickening in the tropicalsavannas, which has the potential to change catch-ment hydrology (Krull et al., 2007), carbon stocks(Burrows et al., 2002; Henry et al., 2002), pasturebiomass available for grazing animals (van Lange-velde et al., 2003) and wildlife habitat (Tassickeret al., 2006). Tree thickening is therefore an

important issue to many stakeholders, includingpastoralists, conservationists, land managers andthose interested in carbon markets, each with awide range of opinions and vested interests in theprocess (Bosch et al., 2007b).

The demand for better management of the com-plex interactions between different factors andcomponents of the tree thickening system has ledto the establishment of an ELLab for sustainablewoodland management.

Identify IssuesSeveral workshops were held during 2005 indifferent localities in the rangelands of NorthernQueensland. Graziers, researchers and extensionofficers discussed the tree thickening problemand identified the factors that they believedwould influence tree density. Possible manage-ment actions and non-manageable factors thatmight influence density were also identified anddiscussed.

Build Capacity and Develop a ModelThe knowledge of the workshop participants wascaptured by mapping out an influence diagram.The process allowed for the integration of thedifferent mental models of the stakeholders(varying perspectives and divergent views). Whiledivergent views occur, the appreciation of eachother’s views gained through ‘mapping the sys-tem’ helped stakeholders to develop a commonunderstanding of the management system.

The influence diagram (Figure 9) provided astructure through which stakeholders could ex-press and discuss their understanding of the causeand effect relationships between managementactions, controlling factors and resource manage-ment outcomes or goals. The diagram also assistedthe stakeholders in identifying how their know-ledge contributed to a better understanding ofthe overall management system and to appreciatehow other stakeholders understand the links be-tween management actions and outcomes(providing a mechanism for externalizing and in-ternalizing knowledge). This co-learning process(capacity building) consists of individual stake-holders who are socializing and externalizing




their nowledge within a group, combiningthis knowledge, and learning from each other(internalization) (Nonaka and Konno 1998).The influence diagram was used as a

framework for the development of a BBN model(Figure 10) through which it was possible tointegrate experiential knowledge, scientific dataand models to populate the BBN model. Thisprocesses ensured that the knowledge createdby scientists became integrated with the under-standing of systems by land managers, conserva-tionists and other stakeholders.Figure 10 shows a completed BBN systems

model for tree density management. Each nodehas two or more states, and arrows represent thecausal relationships between nodes. Conditionalprobability tables (CPTs) specify the relationshipsbetween the nodes. Bosch et al. (2007b, Table 1)described the CPT in an example of how fuelbuild-up and fire season influence fire intensity.The first row represents the scenario wherefuel build up is high (>1800 kg/ha) and the fireseason (time of fire) is ‘late_dry’ (October/November). ‘Under this scenario there is a100% chance that fire intensity will be hot.By completing the probability table for eachnode in the BBN, available data, informationand experiential knowledge are integrated in asystematic way. The result is a knowledgebase and a dynamic systems model that can

assist stakeholders (particularly managers) indecision-making through analysing differentscenarios’ (Bosch et al., 2007b, p. 220).

Identify Leverage Points and Systemic InterventionsAn evaluation of the model and identification ofleverage points and systemic interventions that willaffect the goal (avoid thickening of tree density)was done by testing model behavior with stake-holders through applying different managementscenarios and predicting the possible outcomes.Back-casting was also used to identify whichactions and factors would have the largest effecton the goal, providing or confirming the systemicinterventions identified during scenario analysis.

The incidence of fire and the factors that deter-mine the nature of fires were identified as themost important leverage point for controllingthickening of trees. This conclusion was verifiedby scientific data and models (Liedloff andSmith 2010) and experiential knowledge of landmanagers. It was mentioned that where fire hasbeen a regular feature within the landscape, theremoval of fire will often lead to woodlandthickening. Grazed woodland ecosystems evolvedwith fire, which suppresses tree thickening.Without a disturbance such as fire, many landtypes will have a higher tree density.

Figure 9 Influence diagram of issues related to managing tree density (adapted from Bosch et al., 2007b)




Develop Management Plan and ImplementationFrom the BBN model, it was clear that themost economic and environmentally sustainableway to control tree cover thickening is with fire,provided conditions such as fuel load are satisfac-tory. The BBN model served as a tool to identifypossible management scenarios before actualimplementation.

ReflectionThe approach of stakeholder involvement andsystems thinking described earlier led to a modelthat represents the mutual understanding ofstakeholders and their current knowledge basefor decision making. However, this knowledgebase is rarely perfect because natural systems arecomplex, and their management takes placeagainst a background of continuous and

unpredictable change in environmental, economicand social conditions. Because of this, theuncertainties in achieving the desired resourcemanagement outcomes remain high. However,new knowledge about management systemsbehavior is continuously generated throughobservation (monitoring) and the evaluation ofoutcomes of implemented management strategies.Embedding the BBN model in the cyclic processof the ELLab allowed for continuous improvementof the knowledge base, and its usefulness formanaging natural resources under uncertain andvariable conditions.

Reflecting on management outcomes empha-sized the importance of fire as a management tool.It became clear that tree density and structure areconstantly changing because of climatic variationand the use of fire. In many regions, a thickeningof trees occurred during higher rainfall periods

Figure 10 BBN model for tree density containing alternative scenario (adapted from Bosch et al., 2007b)




and thinning during drought. Where fire has beena regular feature within the landscape, its removaloften led to the thickening of tree cover. Inextensive systems where thickening has occurred,mustering costs have increased by up to 30% andproduction has suffered as pastures compete forscarce resources. It has become clear that the mosteconomic and environmentally sustainable way tocontrol woody thickening is with fire, providedconditions such as fuel load are satisfactory.This finding during the reflection stage has led tothe development of a more detailed model thatfocuses on the influence of management and non-management drivers on woody vegetation change(Liedloff and Smith 2010).

THE GLOBAL EVOLUTIONARY LEARNINGLABORATORY

Once an ELLab has been established in eachparticular region or country, it will operate as amanagement tool for the reform and sustainablemanagement of complex issues in their respective

systems. As described in the above case studies,management strategies and policies are imple-mented and the ELLab runs ‘Reflection’meetings(step 7) to discuss the outcomes (successes andfailures) and decide how to change the manage-ment or how to adapt a policy. These reflectionmeetings will lead to new levels of learning andenhanced management performance in the differ-ent sectors of the system as a whole.

Each individual ELLab will also become part ofthe Global Evolutionary Learning Laboratory(GELL) (Figure 11) and continually share the les-sons it has learned with ELLabs (and other similarinnovations) in other parts of the world, throughthe lenses of different political systems, culturesand so on. GELL is currently being enhancedwith advanced e-technologies that will help it toserve as a platform for continuous sharing andco-learning, leading to new levels of learning andperformance at regional and global levels. It willalso help individual ELLabs to learnmore and per-form better in their own countries, organizations,businesses and communities.

Evolutionary Learning Laboratories for e.g.(ELLabs)

4. Identify Leverage/Systemic Interventions

3. Develop OrRefine SystemsMaps Or Models

1. IdentifyIssues

7. REFLECTION

Environmental

Eco

nom

ic Social

New Levels of Learning and Performance at LOCAL Level

5. Develop OrAdaptIntegratedStrategic &Action Plan

2. Build Capacity

6. ImplementActionsStrategies

& Policies

StakeholderMental Models

SystemsStructure

Patterns &Relationships

Cultural Values

Global Evolutionary Learning Labor(GELL)

Creating a collaborative intercultural learning

environment

GELL

New Levels of Learning and Performance at GLOBAL Level

- Sustainable Management of Cat BaBiosphere, Vietnam

- Managing Poverty Reduction

- Transforming the Graduate School forSystems Design and Management, KeioUniversity ,Japan

- Integrated Systemic Governance,Hai PhongCity, Vietnam

- Developing Intercultural Tools forSystems Transformations, CentralAustralia and South East China

- Managing Balance between ProductionEconomics & Biodiversity Goals, China

- Improving Organisational performance

- Improving Child Safety, Japan

- Managing human relationships inorganisations

- Systems Education

SHARINGREFLECTIONS

Figure 11 The Global Evolutionary Learning Laboratory (GELL)




CONCLUSION

Globally effective researchers, as well as existingand future leaders and managers, will need tounderstand complexity and how to deal with itin multi-stakeholder scenarios. Systems thinkingis therefore the underlying paradigm and researchapproach. This paper has described the applicationof systems thinking in the establishment of ELLabsfor managing complex issues through enhancingcross-sectoral communication and collaboration,and promoting effective change. Each ELLabdevelops uniquely because of the political andcultural systems of each country, organization orbusiness. The GELL can greatly enhance ourcapacity to address globalized issues and servesas a global knowledge hub.

The establishment of ELLabs and the GELLis an ongoing process. The research so farhas achieved various active engagements atspecific levels, local and global, including localcommunities, national park staff, local andnational governments, the national MAB com-mittees in different countries and the UNESCOMAB program. UNESCO/MAB has alreadyacknowledged this approach as best practice forpotential applications to more than 580 BRsglobally (Nguyen et al., 2011).

The research has helped to build the capacityof various people (relevant stakeholders) indifferent places where ELLabs are being estab-lished. The stakeholders are closely involved inall the different steps of the establishment of theirrespective ELLabs. This close involvement hasenabled a shared vision amongst stakeholdersand helped them to understand complexity andbe able to identify the root causes of problems,rather than merely treating the symptoms. It hasalso helped them to develop solutions collabora-tively over time, experimentwith themand be ableto adapt when required through knowledgesharing and discussions with others. In addition,the close involvement has enabled the relevantstakeholders to take ownership of the ELLab andto know how to operate it.

Having a ‘champion’ is another importantlesson learned through the research. The authorshave been fortunate to work with a champion(a key person in a leading position, who

understands and supports the approach) in everysite where an ELLab has been established. This isessential for the successful implementation andoperation of the ELLab.

The key challenge in this research is securingfunding to address the identified leverage pointsand systemic interventions. It is common fordonors and funding agencies to provide fundingfor treating the ‘symptoms’ with quick fixes, inorder to see (and show to the world) immediateresults from their funding efforts. However, it couldtake several years for a systems-based approach toachieve long-lasting sustainable outcomes by solv-ing the root causes of problems. Finding the fundsfor a process with often non-tangible outcomes (asapposed to tangible outcomes such as a bridge, aschool or a road) has proved to be a majorchallenge, especially for developing countries.

A further important challenge is the ‘silo’structure of ministries and organizations in everycountry, which makes ‘collaboration’ a foreignconcept. A paradigm shift is needed to moveaway from this kind of structure. Furtherresearch to institutionalize the ELLab concept,leading to the use of collective intelligence indecision making across sectors and organizationsand effective collaborative governance, hasbecome a high priority.

Computer-based modelling systems can beuseful tools to explore and make managementaction decisions that are more systemic than thedecisions produced by traditional approaches.Of particular importance is their ability to beused within a participatory process, to enableknowledge capturing, testing and refinementof multi-stakeholders. Used in this way, acomputer-based modelling system (such as aBBN) can (i) provide a flexible modelling envir-onment, (ii) allow uncertainty in knowledge tobe expressed using probabilistic relationships,(iii) allow biophysical, economic and social vari-ables (either quantitative or qualitative) to berelated, (iv) enable a graphical (flow chart) inter-face that is easily understood and facilitatescommunication between stakeholders and (v) beeasily updated as new knowledge emerges, with-out the need for specialist computer skills (i.e.nodes added or removed, links changed andprobabilities updated).




In summary, a new way of thinking can changethe effectiveness of government departments,businesses, organizations and communities inmany ways:

• better mutual understanding of the diversemental models of different stakeholders;

• moving away from traditional linear thinkingthat leads to quick fixes and treatment of thesymptoms, to long-lasting systemic solutionsthat address the root causes;

• ability to collaboratively identify leveragepoints and systemic interventions to underpinsystems-based master and strategic plans;

• deep understanding of the interconnectednessbetween possible actions in order to develop effi-cient and cost-effective management strategies;

• working knowledge of cutting edge systemstools to test the outcomes of strategies, includingidentification of unintended consequences—be-fore actual implementation;

• ability to use back-casting to identify thosefactors that will have the most influence onthe achievement of goals (knowing whereand when to invest in the system); and

• using the ELLab as an ongoing process forcontinuous co-learning and refinement of man-agement strategies.

REFERENCES

Allison HE, Hobbs RJ. 2006. Science and Policy inNatural Resource Management: Understanding SystemComplexity. Cambridge University Press: UK.

Banathy BH. 1996. Designing Social Systems in aChanging World (Contemporary Systems Thinking).Springer: New York, USA.

Bosch O, Maani K, Smith C. 2007a. Systems thinking—language of complexity for scientists and managers.Proceedings of the Conference on Improving theTriple Bottom Line Returns from Small-scale Forestry.S. Harrison, A. Bosch and J. Herbohn. 18–21 June, 2007,Ormoc, the Philippines.

Bosch OJH, King CA, Herbohn JL, Russell IW, Smith CS.2007b. Getting the big picture in natural resourceman-agement—systems thinking as ‘method’ for scientists,policy makers and other stakeholders. Systems Re-search and Behavioral Science 24(2): 217–232.

Bosch OJH, Nguyen NC. 2011. Establishing the GlobalLearning Laboratories NET for managing complex

problems (Working Paper). Brisbane, Australia, Schoolof Integrative Systems, The University of Queensland.

Burrows WH, Henry BK, Back PV, et al. 2002. Growthand carbon stock change in eucalypt woodlands innortheast Australia: ecological and greenhouse sinkimplications. Glob. Change Biol. 8: 769–784.

Cabrera D, Colosi L, Lobdell C. 2008. Systems think-ing. Evaluation and Program Planning 31(3): 299–310.

Cain J, Batchelor C, Waughray D. 1999. Belief networks:a framework for the participatory development of nat-ural resource management strategies. Environment,Development and Sustainability 1: 123–133.

Elias AA. 2008. Towards a shared systems model ofstakeholders in environmental conflict. InternationalTransactions in Operational Research 15(2): 239–253.

Galanakis K. 2006. Innovation process.Make sense usingsystems thinking. Technovation 26(11): 1222–1232.

Henry BK, Danaher T, McKeon GM, BurrowsWH. 2002.A review of the potential role of greenhouse gasabatement in native vegetation management inQueensland’s rangelands. Rangeland J. 24(1): 112–132.

Hung W. 2008. Enhancing systems-thinking skillswith modelling. British Journal of Educational Technol-ogy 39(6): 1099–1120.

Ishwaran N, Persic A, Tri NH. 2008. Concept and prac-tice: the case ofUNESCObiosphere reserves. Int. J. En-vironment and Sustainable Development 7(2): 118–131.

Jackson MC. 2003. Systems Thinking: Creative Holism forManagers. John Wiley & Sons: Chichester, UK.

Kakefuda I, Yamanaka T, Stallones L, Motomura Y,Nishida Y. 2008. Child restraint seat use behaviorand attitude among Japanese mothers. Accident Ana-lysis and Prevention 40: 1234–1243.

Keegan M, Nguyen NC. 2011. Systems thinking, ruraldevelopment and food security: key leverage pointsfor Australia’s regional development and popula-tion policy. Migration Australia (launch issue) 1(1):50–64.

Krull E, Bray S, Harms B, Baxter N, Bol R, Farquher G.2007. Development of a stable isotope index to as-sess decadal-scale vegetation change and applica-tion to woodlands of the Burdekin catchment,Australia. Glob. Change Biol. 13: 1455–1468.

Land T, Hauck V, Baser H. 2009. Capacity changeand performance: capacity development: betweenplanned interventions and emergent processesImplications FOR development cooperation (PolicyManagement Brief No. 22). Maastricht, ECDPM.

Laszlo KC. 2001. Learning, design, and action: creatingthe conditions for Evolutionary Learning Commu-nity. Systems Research and Behavioral Science 18(5):379–391.

Lee, A. 2009. Health-promoting schools: evidence for aholistic approach to promoting health and improv-ing health literacy. Applied Health Economics andHealth Policy 7(1): 11–17.

Liedloff AC, Smith CS. 2010. Predicting a ‘tree change’in Australia’s tropical savannas: combining different




types of models to understand complex ecosystembehaviour. Ecological Modelling 221: 2565–2575.

Maani K, Maharraj V. 2004. Links between systemsthinking and complex decision-making. System Dy-namics Review 20(1): 21–48.

Maani KE, Cavana RY. 2007. Systems Thinking, SystemDynamics: Managing Change and Complexity. PrenticeHall: Auckland, NZ.

Mai TV. 2012. Sustainable tourism—systems thinkingand system dynamics approaches: a case study inCat Ba Biosphere Reserve of Vietnam. School ofAgriculture and Food Sciences. The University ofQueensland: Australia. PhD Thesis.

Meadows D. 1999. Leverage points: Place to intervenein a System. Hartland, VT, USA, The SustainabilityInstitute.

Midgley G, Ed. (2003). Systems Thinking (Volumes 1–4).Sage: London, UK.

Mingers JC. 2006. Realising Systems Thinking: Know-ledge and Action in Management Science. Springer:New York, USA.

Newell D. 2003. Concepts in the study of complexityand their possible relation to chiropractic healthcare: a scientific rationale for a holistic approach.Clinical Chiropractic 6(1): 15–33.

Nguyen NC, Bosch OJH 2012. A systems thinking ap-proach to identify leverage points for sustainability:a case study in the Cat Ba Biosphere Reserve,Vietnam. Systems Research and Behavioral Science InPress (DOI: 10.1002/sres.2145; first published online11 October 2012).

Nguyen NC, Bosch OJH, Maani KE. 2011. Creating‘learning laboratories’ for sustainable developmentin biospheres: a systems thinking approach. SystemsResearch and Behavioral Science 28(1): 51–62.

Nguyen NC, Graham D, Ross H, Maani K, Bosch OJH.2012. Educating systems thinking for sustainability:experience with a developing country. Systems Re-search and Behavioral Science 39(1): 14–29.

Nonaka I, Konno N. 1998. The concept of ‘Ba’: buildinga foundation for knowledge creation. CaliforniaManagement Review 40(3): 40–54.

Phan TD. 2011. Optimizing conservation effort forserow, Capricornis milneedwardsii, in Cat Ba Archi-pelago, Hai Phong, Vietnam. School of Geography,Planning and Environmental Management, TheUniversity of Queensland. Master of EnvironmentalManagement.

Quatro SA, Waldman DA, Galvin BM. 2007. Develop-ing holistic leaders: Four domains for leadershipdevelopment and practice. Human Resource Manage-ment Review 17(4): 427–441.

Sawin B, Hamilton H, Jones A. 2003. Commodity SystemChallenges: Moving Sustainability into the Mainstreamof Natural Resource Economies. Sustainability Institute:Hartland, USA.

Senge PM. 2006. The Fifth Discipline: The Art andPractice of the Learning Organization (Revised andUpdated). Random House, Inc: New York, USA.

Smith C, Felderhof L, Bosch OJH. 2007. Adaptivemanagement: making it happen through participatorysystems analysis. Systems Research and Behavioral Sci-ence 24(1): 567–587.

Sterman JD. 2000. Business Dynamics: Systems Thinkingand Modeling for a Complex World. Irwin McGraw-Hill:Boston, USA.

Systems V. 2011. Vensim program, Ventana SystemsUK. from http://www.ventanasystems.co.uk/.

Tassicker AL, Kutt AS, Vanderduys E, Mangru S. 2006.The effects of vegetation structure on the birds in atropical savannawoodland in north-easternAustralia.Rangeland J. 28(2): 139–152.

Thomas E, Amadei B. 2010. Accounting for human be-havior, local conditions and organizational constraintsin humanitarian development models. Environment,Development and Sustainability 12(3): 313–327.

Umaña A. 2002. Generating Capacity for SustainableDevelopment: Lessons and Challenges. UNDP,UNDP Choices Magazine.

UNESCO. 2012. Biosphere Reserves—Learning Sites forSustainable Development. Retrieved 30th March 2012,from http://www.unesco.org/new/en/natural-sciences/environment/ecologicalsciences/biosphere-reserves/.

UNICEF. 2001. A League Table of ChildDeaths by Injuryin Rich Nations. Innocenti Report Card, Issue No.2.

van Langevelde, F, van de Vijver CADM, Kumar L,et al. 2003. Effects of fire and herbivory on the stabil-ity of savanna ecosystems. Ecology 84(2): 337–350.

Walker GH, Stanton NA, Jenkins DP, Salmon PM. 2009.From telephones to iPhones: applying systemsthinking to networked, interoperable products. Ap-plied Ergonomics 40(2): 206–215.

Wilson J. 2004. Changing Agriculture: An Introductionto Systems Thinking. QLD, Australia, Print on De-mand Centre, University of Queensland Bookshop.



http://www.ventanasystems.co.uk/http://www.unesco.org/new/en/natural-sciences/environment/ecological-sciences/biosphere-reserves/http://www.unesco.org/new/en/natural-sciences/environment/ecological-sciences/biosphere-reserves/http://www.unesco.org/new/en/natural-sciences/environment/ecological-sciences/biosphere-reserves/http://www.unesco.org/new/en/natural-sciences/environment/ecological-sciences/biosphere-reserves/http://www.unesco.org/new/en/natural-sciences/environment/ecological-sciences/biosphere-reserves/http://www.unesco.org/new/en/natural-sciences/environment/ecological-sciences/biosphere-reserves/http://www.unesco.org/new/en/natural-sciences/environment/ecological-sciences/biosphere-reserves/http://www.unesco.org/new/en/natural-sciences/environment/ecological-sciences/biosphere-reserves/http://www.unesco.org/new/en/natural-sciences/environment/ecological-sciences/biosphere-reserves/

Documents

Managing Complex Issues through Evolutionary Learning ...lab.sdm.keio.ac.jp/idc/yasui/papers/j15_2013srbs_bosch...progress is the lack of common understanding and shared vision of