modelling and managing project complexity

Modelling and managing project complexity

Simon Austina,*, Andrew Newtonb, John Steeleb, Paul Waskettb

aLoughborough University, Department of Civil and Building Engineering, Loughborough, Leicestershire LE11 3TU, UKbAMEC Ltd, Stratford-on-Avon CV37 9NJ, UK

Abstract

The Architecture, Engineering and Construction (AEC) industry, like many others, is increasingly aware of the need to improveefficiency and effectiveness to thrive in an increasingly competitive marketplace. A key discovery in their search for improvements isthe benefits of repeatability in both processes and products. However, although the latter has seen significant advances, such as the

adoption of pre-assembly and standardised components and systems, the industry has experienced far greater difficulties identifyingways of capturing, understanding, and replicating work processes. The identification and removal of waste from the process canonly be achieved once the process has been captured. Their repeated use and development, combined with analysis with the Ana-lytical Design Planning Technique, enable the improvement of work practices and culture in terms of integration, decision-making

and reductions in re-work. # 2002 Elsevier Science Ltd and IPMA. All rights reserved.

1. Introduction

To achieve anything more than a superficial under-standing of the building design process, the complexitiesof the design activity have to be identified and repre-sented in an appropriate manner. Graphical models arethe ideal mechanism to achieve this. They allow thedecomposition of complex systems into interrelated sub-elements that can be represented in the form of dia-grams and text that are easier to assimilate. Modellingthe information flows within a particular system orprocess can lead to a greater understanding of thatprocess [1] and these models can then be applied bydesigners to help avoid the careless processing ofincomplete or inaccurate information during the devel-opment of design solutions [2]. This paper describes theapproaches taken by Loughborough University, AMECand other industrial collaborators to modelling andanalysing the building design process. These modelshave taken the form of a high-level description of theentire project process, as well as more detailed studies ofeach of the phases of which it is comprised, combinedwith their analysis with the Analytical Design PlanningTechnique (ADePT).

There are significant differences between the nature ofthe individual phases of the project process during pro-gression from early stage design through to the laterstages. These differences have influenced the choice ofmodelling notation to apply to each phase, with themodels changing progressively from being frameworksfor negotiation and agreement (with little focus on co-ordinated information flow) toward highly co-ordinatedmodels representing structured information transfer(Fig. 1) using sophisticated modelling notations. Never-theless, the use of the project process as the basis foreach of these models has ensured that they can be inte-grated and their interfaces aligned, thus identifying gapsand overlap.

2. The project process

The Generic Design and Construction protocol(GDCPP), being developed by Salford and Loughbor-ough Universities in conjunction with a number ofindustrial collaborators, defines the design and con-struction process as four broad stages, which are thenfurther categorised into 10 discrete phases as shown inFig. 2.The GDCPP was developed from a client perspective,

with the main focus being the uneducated or one-offclient [3]. However, the generic nature of the modelensures that it can be applied by a variety of client typeson a variety of projects, and can be adapted to reflect

0263-7863/02/$22.00 # 2002 Elsevier Science Ltd and IPMA. All rights reserved.

PI I : S0263-7863(01 )00068-0

International Journal of Project Management 20 (2002) 191–198

www.elsevier.com/locate/ijproman

* Corresponding author. Tel.: +44-1509-263171; fax: +44-1509-

223981.

E-mail addresses: [email protected] (S. Austin), andrew.

[email protected] (A. Newton), [email protected] (J. Steele),

[email protected] (P. Waskett).

the internal cultures and working practices of specificorganisations, within the common structure of the gen-eric framework. The GDCPP not only describes thephysical stages of the process, but also addresses itsmanagement. This is an integral component in achievingproject success [4] and the Protocol defines eight keymanagement areas (Development, Project, Resource,Design, Production, Facilities, Health and Safety, Stat-utory and Legal, and Process) involved at each phase.The Protocol is being defined to an increasingly detailedlevel, with 270 level-two activities within the 10 phases(plus standard start-up, on-going and end-of-phaseactivities) now identified.The overriding aim of the map itself is to improve the

collaboration between companies in the traditionallyfragmented construction industry. However, it also

attempts to provide a standard framework for clientsaround which they may enhance the effectiveness oftheir work [5]. It can also contribute to culture changeby improving communication and process managementbetween the fragmented groups within the constructionindustry. In particular, it provides a common languageby which all parties can locate themselves and theirprocesses within the project organisation as a whole. Ithas already been adopted by several major UK con-struction organisations as a vehicle for investigatingtheir processes or addressing the specific requirementsof large projects.The remainder of this paper outlines the development

of the models that represent some of the design stages ofthe GDCPP and how the processes can be improved byapplication of ADePT.

Fig. 1. The changing nature of the project process.

Fig. 2. The phases of a construction project process defined by the Generic Design and construction process protocol (GDCPP).

192 S. Austin et al. / International Journal of Project Management 20 (2002) 191–198

3. The early design stages—concept and scheme

The early phases of the design process have receivedrelatively little attention, even though decisions madeduring this period have the most far-reaching effects onthe remainder of the project. It is recognised that earlystage design often fails to deliver outputs that meet theexpectations of clients. These failings, which typicallybecome manifest in the need for redesign and poorquality cost advice are, primarily, the result of: (1) poorcommunication between stakeholders; (2) ineffectivecollaboration; (3) little understanding of the complexityof the interdisciplinary nature of design; and (4) weakand unconsidered decision-making. The existing designprocedures that are available to the interdisciplinarydesign team tend to be lists of deliverables rather thanguidance documents providing design teams with anoutline of what to do and by what method it should beachieved. In this respect, there seems to be an over-reli-ance on the experience of designers to ‘know how todesign’. At present, no consistent approach to earlystage design exists within the building industry [6]. Tworesearch projects involving Loughborough Universityand AMEC have addressed this issue through thedevelopment and analysis of generic models of the earlystage design process, with each using very differentapproaches to capturing and representing the processes.In the Mapping Conceptual Design project, under-

taken in collaboration the University of Cambridge, ageneric framework was developed, comprising five pha-ses and twelve activities, for use as a guiding principlerather than a structured plan of work. Fig. 3 shows theframework, which was refined in workshops and subse-quently developed into a generic process model thatclustered the design activities in relation to the mannerin which they were commonly addressed [7]. Addition-ally, this model accounts for the design team’s need tofocus on, and maintain, team performance. In thisrespect, successful collaborative conceptual design is

much more dependent upon the level of negotiation andagreement than the formal co-ordination and transfer ofinformation between team members (Fig. 1).The scheme design stage was modelled in a different

manner owing to the need for improved co-ordinationas the project process advances. It is clear that both theconcept and scheme design stages are primarily con-cerned with information gathering and decision-makingto enable the team to propose a solution to the stake-holders needs. However, as the project progresses intoscheme design the cost of the developing solution mustbe established and the risks involved in its deliveryassessed. Thus, this research developed a model of thescheme design stage centred on decision-making and theresulting transfer of design and cost informationbetween the project team (including the client) as anintegrated process [8]. This was initiated by defining thehigh-level activities undertaken during the schemedesign period and sub-dividing these into their compo-nent parts. This process was repeated a number of timesuntil the lowest level design tasks were identified foreach discipline. In this way a four-level hierarchy ofactivities was produced (an example developed fordetailed design is shown in Fig. 4). In order to developthe process model from this work breakdown structure,information flows between the tasks were captured andrepresented using a structured modelling technique,IDEF0. This notation, which has been used primarily inthe manufacturing and business process re-engineeringdomains, uses boxes to define activities and processeswith arrows denoting information transfer betweenthem. The notation was modified slightly (and renamedIDEF0v) to enable a differentiation between informa-tion transfer within and across disciplines to be repre-sented, thus enabling the building design process to becaptured in a more appropriate and useful manner. Theresulting model, which comprises some 150 tasks and1500 information flows, represents a network of tasksconnected by the flow of information between them.

Fig. 3. An overview of the conceptual design framework.

S. Austin et al. / International Journal of Project Management 20 (2002) 191–198 193

4. The late design stages—detailed design and produc-

tion information

The transition from scheme design into detail designbrings with it a shift from negotiation and agreementbeing the principle driver for the design process to theco-ordination of the design activity becoming of greatersignificance to project success. This shift in focus iscommonly recognised within the industry and reinforcesthe importance of effective design management in facil-itating a co-ordinated design, within budget, and ensur-ing the smooth running of projects. To deliver improvedplanning of projects a Loughborough University-basedresearch project developed the ADePT methodology(see later), a component of which involved the con-struction of a model of the detailed design process. Aswith the scheme design model discussed previously, thedetailed design model was derived by first developing ahierarchical breakdown of the activities involved in thedesign process (Fig. 4) before identifying the informa-

tion flows between those activities to generate the model(Fig. 5). However, this model differs from the schememodel in that it is global in nature and is not based on asingle project type. The tasks and information flowscontained within the detailed model can be tailored torepresent the basis of any project, with only minoralterations being required to generate the project spe-cific version.This global model, which was also developed using

the IDEF0v notation, is structured in a manner thatreflects the discipline-based way in which industry cur-rently works (representing architectural, civil and struc-tural engineering, and mechanical and electricalengineering activities). The model comprises some 150diagrams containing 580 design tasks and 4600 infor-mation requirements [9]. In applications to date, theglobal model has been found to contain approximately90% of the tasks required to produce project specificmodels. This figure will increase as the model evolvesthrough further application on a wider range of projects.

Fig. 4. The highest level of the discipline-based work breakdown structure.

Fig. 5. An IDEF0v diagram from the detailed design model.


The effectiveness of this model, and the opportunitiesfor improved planning afforded by the ADePT metho-dology, has also driven the development of models ofthe production information stage of the project process.This work, which has been undertaken as part of theIntegrated Collaborative Design (ICD) research project,has involved both modelling the exchange of informa-tion between the design team and suppliers undertakingthe fabrication activities, and identifying how the pro-cess model and associated analytical techniques(including ADePT) can be used to improve decision-making and activity scheduling.Using the same approach to model development

described previously, the exchange of informationbetween the design team and suppliers undertaking thefabrication and construction activities has been mod-elled [10]. The models have captured the metamorphosisof the intangible design information into tangible con-struction materials (Fig.6).These models have enabled the interfaces between

consultant-based design and supplier-based design to bealigned, allowing the skills and expertise of each to bedove-tailed, and potential duplications and deficits inthe design process to be identified and managed. Assuch, the models have facilitated both the optimisationof the design process to a level beyond that which maycurrently be achievable, and the removal of unnecessaryprojects costs in terms of: reduced prime cost to the cli-ent; higher fee profit for designers; and reduced effortand abortive work.

5. ADePT

5.1. Introduction

Effective design planning requires the application oftechniques that can account for the complexity and non-linearity of the design process. Traditionally, owing tothe successful application of planning techniques suchas the Critical Path Method (CPM) in construction, thedesign process has been planned in a similar manner.

Unfortunately, the iterative nature of the buildingdesign process makes the application of such techniqueswholly inappropriate. The ADePT methodology, whichwas developed in response to this need, provides apowerful, yet simple, means of understanding the inter-dependencies between tasks in the design process.ADePT can take process models, optimise them and

then be used to manage the resulting complexity. Themethodology Fig. 7, comprises three stages. Firstly, amodel of the building design process of a project isproduced, in both graphical and database format,showing the relationship between design activities basedon the flow of information in the process. Secondly,dependency structure matrix (DSM) analysis identifiesan optimum sequence of activities based upon thedependency and availability of design information asdefined in the design process model. Finally, the matrixanalysis is linked to a planning and scheduling packageso that design programmes can be produced whenresources and duration of tasks are allocated to the re-sequenced activity schedule.An example of a DSM, the second part of ADePT, is

shown in Fig. 8.In the matrix tasks are initially listed alphabetically in

the rows of the matrix. The order is mirrored in thecolumns. A mark in the matrix represents a dependencyof the task in the row upon the task in the column, thedependencies being weighted on a three point scale(A,B,C) on the basis of the strength of dependency. Ifdesign is undertaken in the order on the matrix fromtop-left to bottom-right, there is a considerable need foriteration within the process. Fig. 8 also shows thematrix following analysis to determine the optimalsequence of tasks such that iteration is reduced to aminimum. It can be seen that the number of criticalmarks above the diagonal is greatly reduced, as is thescale of iteration within the process which is indicatedby the shaded blocks.ADePT challenges designers to place greater emphasis

on understanding and analysing the process of design.More specifically it offers a means of illustrating to theclient, designers and building contractors, the importance

Fig. 6. A model of the changing nature (from information to material) of exchanges.


of timely release of information, appropriate quality ofinformation and fixing of design, and the resultingimplication for cost, design flexibility and risk. It alsoensures that the appropriate informationis exchangedbetween members of the design team and that the pro-blem of information overload is minimised. Variationscan be assessed rapidly, allowing objective decisions to

be made about the resulting changes to project dura-tion, resource levels and engineering economics [11].Some of the practical application of ADePT and asso-ciated benefits are described later.

5.2. Improving the design process.

The integration of stages of a project and team mem-bers within each stage changes to the ways a project ismanaged and team members behave and interact.Where the design team may be co-located or expected todevelop the design through a series of workshops, thissuggests a change to the way complex co-ordination isapproached. The blocks of interdependent design activ-ity require a concerted management effort, rigorousreview strategy and a strong link to the client’s decision-making and approval process. They also highlightwhere the concurrent, collaborative working strategy isappropriate for the team members, who must liase clo-sely in all decisions, understand each others’ designrequirements and constraints, and have confidence ineach others’ commitment to the achievement of a com-mon aim.The graphical nature of a matrix allows the impact of

changes and variations to be envisaged quickly andeasily, by moving tasks within the matrix (usually downthe order) to simulate them being undertaken followingthe change. The tasks that must then be re-examined areFig. 7. An overview of the analytical design planning technique.

Fig. 8. A simple example of dependency structure matrix analysis.


clearly indicated by the matrix. This is a particularlyuseful feature where the work of one design discipline isaffected by the decisions of another, or where the designin general is delayed by the decisions of the client.A further area is the co-ordination of work between

the design phases, to ensuring that adequate designdevelopment is undertaken in each discipline to providethe required cost certainty and confidence to the client.The GDCPP should provide a means of identifying thetimely introduction of suppliers into the design process,a benefit that is beginning to be seen during the latterstages of a design project from the implementation ofADePT.

5.3. Integrating design and construction

Scheduling the design process with ADePT identifiesthe optimal sequence of tasks to satisfy the deveopmentof a design solution. In practice, it is unlikely that thissequence will be realistic because of the productionconstraints put on the process by the need to deliver aproject in as short a time-scale as possible. However,comparison with a view of the ideal constructionsequence (which is relatively easy to determine with theuse of readily available project planning tools), providesa good starting point to intergrate design within thewider project process (Fig. 9).This integration is not straight-forward, as the two

processes do not fit together comfortably. In order thatthey are integrated, the constraints that each processputs on the other must be considered. The schedule isproduced through the analysis of the constraints on thedesign process: the cost of fixing or estimating informa-tion within the design can be compared against the riskof not doing so, thereby allowing the engineering eco-nomics in design to be assessed and logged in a riskregister. As such, ADePT can act as a tool to compli-ment risk management. It identifies areas of designwhere risks are present, illustrates the scales of risk inthe design process itself (in a similar way to evaluatingthe effects of change) and contributes to the develop-ment of a legacy risk register.Having established an approach to undertaking the

design and an agreed procurement strategy, each con-tract can be examined to determine who in the supplychain is best placed to undertake the design. The matrixanalysis stage of ADePT also provides a means ofassessing the impact of each package of work upon theothers, and the need for co-ordination between them.This is in accordance with the UK construction indus-try’s call for integration within the project supply chain,and the application of ADePT to the fabrication designstage (production information) of a project has beenundertaken through the ICD project [12]. This project isdetermining strategies for integrating contractors andsuppliers into the consultants’ design process in a man-

ner that is both timely and that allows the design co-ordination and contracts to be effectively managed. Thekey to this approach is that participants should beintroduced into the project early enough to allow theirdesign to be co-ordinated with other participants of theproject, and as late as possible such that their design tobe co-ordinated with other parts of the project, and aslate as possible such that their design is not constrainedby decisions made by the consultant.

6. Challenges and benefits

Through the development and application of designprocess models the design team can make more con-sidered decisions, as they are aware of all factors relat-ing to the design task at hand and the other activities itinfluences. This enables risks to be identified and trans-ferred into the risk management process, thus allowingeffective control measures to be introduced. In analysingthe process models as part of ADePT, the tasks withinthe model can be programmed optimally to deliverimproved efficiency in the design production process,savings on design fee expenditure, and benefits in theform of the improved co-ordination with construction(resulting in improvements in cost, programme perfor-mance, and predictability).The use of the process models within ADePT also

improves project team performance by fostering trustand encouraging collaborative working. In order toimprove and maintain both efficiency and effectivenessintegrated teams must achieve a collaborative, con-tinuous-improvement culture of ‘right on time, firsttime’ over the course of a number of projects. Designersand constructors must improve their understanding ofthe process, in conjunction with their roles and respon-sibilities within it, if this is to be achieved. Capturingand representing these complex processes in the form ofmodels, and analysing them using the ADePT metho-dology provides a mechanism to achieve this, in addition

Fig. 9. A schematic of the integration of design and construction

process.


to improving the effectiveness and efficiency of thedesign planning process. The testing and application ofADePT has demonstrated that it is a viable techniquewith which to plan, manage and control design workand aid integration of the design and construction pro-cesses. The technique is being further developed andmade available to practising planners, project managersand designers through an Internet-provided softwareapplication called PlanWeaver.Through the use of process modelling, DSM analysis

and the production of design programmes, the planningof complex design projects can be approached in a moresystematic, informed, and efficient manner comparedwith current practice.

References

[1] Austin S, Baldwin A, Newton A. A data flow model to plan and

manage the building design process. Journal of Engineering

Design 1996;7(1):3–25.

[2] Kraol L. 1983. Implementation experiences of design applica-

tions. Proceedings of the international conference on computers

in architecture ‘83’. On-Line Conferences Ltd:65-80.

[3] Kagioglou M, Cooper R, Aouad G, Hinks J, Sexton M, Sheath

DM. A generic guide to the design and construction process

protocol 1998. University of Salford.

[4] Pugh S. Design activity models – world-wide emergence and

convergence. Design Studies 1986;7(3):167–73.

[5] Sheath DM, Woolley H, Cooper R, Hinks J and Aouad G. 1996.

A process for change—the development of a generic design and

construction process protocol for the U.K construction industry.

In CIT’96. Sidney, Australia.

[6] Austin S, Steele J, Macmillan S, Kirby P, Spence R. Mapping the

conceptual design activity of interdisciplinary teams. Design

Studies 2001;22(3):211–32.

[7] Macmillan S, Steele J, Austin S, Kirby P, Spence R. Development

of a generic conceptual design framework. Design Studies 2001;

22(2):169–91.

[8] Baldwin A, Austin S, Pendlebury M. 1997. The interface of early

stage design and cost advice in the building design process. InProcs

of the 13th ARCOM annual conference, Cambridge., vol. 2: 395–404.

[9] Austin S, Baldwin A, Li B, Waskett P. Analytical design planning

technique: a model of the detailed building design process. Design

Studies 1999;20(3):279–96.

[10] Hammond J, Choo HJ, Tommelien I, Austin S, Ballard G. Inte-

grating design planning, scheduling and control with deplan,

Proceedings 8th annual IGLC conference, Brighton, July 2000.

[11] Austin S, Baldwin A, Li B, Waskett P. Integrating design in the

project process. Proceedings of the ICE, Civil Engineering 2000;

138(4):177–82.

[12] Austin S, Baldwin A, Hammond J, Murray M, Root D, Thom-

son D, Thorpe A. Design chains—a handbook for integrated

collaborative design. London: Thomas Telford, 2001 (In press).


modelling and managing project complexity

Documents