Computer Assisted Construction Plan

8/8/2019 Computer Assisted Construction Plan

1/113


2/113

National Library Bibliothque nationaleduCanadaAcquisitionsand Acquisitions etBibliographie Services services bibliographiques395 Wellington Street 395. rueWdlingtmOttawaON K 1 A W OctawaON K l A Wcanada canada

The author has granted a non-exclusive Licence aliowing theNational Library of Canada toreproduce, loan, distribute or selcopies of this thesis in microform,paper or electronic formats.

The author retains ownenhip of thecopyright in this thesis. Neither thethesis nor substantid extracts fiom itmay be printed or otherwisereproduced without the author'spermission.

L'auteur a accord une Licence nonexclusive permettant laBibliothquenationale du Canada dereproduire, prter, distribuer ouvendre des copies de cette thse sousla fome de microfiche/film, dereproduction sur papier ou sur formatlectronique.L'auteur conserve la proprit du

droit d'auteur qui protge cette thse.Ni la thse ni des extraits substantielsde celle-ci ne doivent tre imprimsou autrement reproduits sans sonautorisation.


3/113

The research presented is entitled Computer-Assisted Construction Planning (CACP) ndit explores the computer-assisted development of initial comprehensive (include elementsof scope, wst, time, performance, and organiuition) construction plans, which the usercan custornize to suit a current project.A thorough investigation into the approaches taken for automated construction plandevelopment was performed exploring the methods to capture and reuse planninginformation and the application of this knowledge to computer-assisted planning. Thedeveloprnent of a common database for storage of al1 project information was completedthrough the selection and refinements of conceptuai sore models of several on-goingmodel development projects. Validation of the developed model was performed whilepopulating the developed database with actuai project information. An inteflace forrepresenting the hierarchical breakdown of project information was developed and a CaseBased Reasoning approach was selected in light of the required fnctionality.The research recognizes and follows the path king established for integratedconstruction management systems that rely on a standard representation of the industry'sinformation requirements. The core information structures were adopted and theirvalidity confirmed by examining a broad application for an integrated system (i-e.,planning beyond time and cost). By examining the comprehensive aspects ofconstruction planning within the fkamework of an integrated construction managementsystem the research demonstrated the useflness of applying sound informationrepresentation structures (Le., part and type relationships). Through the application ofprudent information systems development approaches (i.e., data, process, and interfaceanalyses), and the use of case-based reasoning, the research has advanced the concepts ofplanning tools a s they apply to integrated systems.The research has employed sound information representation principles, explored theimplementation of data management tools and techniques, and followed a rigorousdevelopment process. The resulting contributions to the research efforts in this areainclude, in addition to the data model and prototype deliverables, an integrated datamodel for constniction management information, a prototype interface for an integratedconstruction management system, and a unique approach to assisted planning in supportof a pradicai approach to planning.


4/113

Table of ContentsLik of Tables v iLisi of Figures viiAcronyrns irAcknowledgmenis xChapter 1: Introductiom 11.1. Conceptudizing Integrated Systems 11.2. Goal 51.3. Objectives S1.4. Scope 61 S. Research Methodology 61.6. Reader's Guide 7

Chapter 2: Points of Depmiwe 92.1. Computer Integratcd Construction 92.2. Industry Classification Standards IO2.3. Industry InformationModciing 112.3.1 STEP (Standardfor tbe Exchange ofProductModel Data) 12

2.3.2 Industry Alliance for Intemperability 122.3.3 ProjectModels 122.3.4 PmssModels 132.3.5 ChherModels 13

2.4. AutomatedPlanning 132.4.1 CONSTRUCTION PLANEX 132.4.2 SequencingKnowledge 142.4.3 m e rResearch 142.5. Assisted Planning and Integrated Systems 142.5.1 Integrated Constniaion Pianning Systcm (KPS) 142.5.2 AutomatedBuildingRealizaion System (HISCHED) 15

2.5.3 CADCLMS 152.5.4 ConstructionMetho Driven Schedulcr 162.5.5 Supprt forCo-oa Planning 162.5.6 REPCON Automaied Sctieduling 17

2.6. Case Bascd Reasoning 172.6.1 Design and design selection 182.6.2 Other CBR applications 192.6.3 ConstructionPlanning 192.7. Hiera rch icd Interfaces 192.8. Implemtntation and Ptrcticd Project Management Toob 202.8.1 Generic Phnning S o m pplications 20

2.8.2 Projea Managementfor Iafommtion Systems 21


5/113

2.9. Summary 21Chapter 3: tto ject ManagementData Madd 233.1. Purpose of he Mode1 233 .1 .1 ScopeoftheModcl 233.1.2 Relationshipwitb Intcniational Data Staadards 233.1.3 MudelingMethadology 243.2. Model Developmtnt 243-2.1 Processes Supported 243.3. Adoption ofEristing Data Modeb 273.4. Modeling Issues, Solutions, andApproacbes 28

3 -4.1 Kernel objects 293.4.2 Producs 303.43 Processes 3 13.4.4 Resources 323.1.5 Control 333.4.6 CostInfonnation 343 -4.7 Classifications 353.4.8 Agents 363.49 Records 3 73-4-10 Constraints 383.4.1 1 Attribute sets 39

3.5. Chapter Summary 41Chapter4: Compter-Assisted Co~~~tmct ionafining 42

4.1. Conceptualizing CACP 424.1.1 CACPFeatures 424.1.2 CACPGeneralAtchitecture 434.1.3 Cornputer Assisted vs. Auto- 43

4.2. The Planning Approach 444.2.1 ModelingConpis 441.2.2 The Meuiods Approach 454.2.3 Methods inCACP 464.3. Case Sased Reasoning 484.3.1 CaseB a d Reasoning for CACP 491 . 3 2 Planning with Case-BascdReaSOning 524.3 3 Planning Granularity 544.4. Chapter Summary 54

Chopter 5: Inte$ie and lmpIimentaion 5s5.1. Interface Requircmcnts 555.2. Interface Implemtntation 565.2.1 Browsing 585.2.2 Miting 59

5.2.3 Reasoning 615.3. Implementation 625.3.1 Relational Databax 625.3.2 Data Management 635.3.3 CaseBasedReasoning 63


6/113

5.4. Chapter Summmy 64Chopter 6: Vdidationand TaaVig 65

6.1. Scenu io 1: StartingFr om Scratch 656.1.1 CommentsonRcsults 67

6.2. Scenario 2: ModiQing a Cu rrcn t Projcc 686.2.1 CommentsonResults 69

6.3. Scenario 3: Buildingfrom TypeLibrrries 706.3.1 CommensonWts 70

6.4. Scenario 4: Reasoning About Mdhods 716.4.1 Commens on R e d t s 746-42 Reasonhg About Products 74

6.5. Chapter Summary 74Chpter 7: Contributionsa d ConcIusions 76

7.1. Ov erall Contribution 767.2. Inte gra ted Spccific Con tributions 777.2.1 ConvibutiontoModcl Devtlopmcnt 777.2.2 A SupporthgComponent for an lntcgratedSystem 787.2.3 Managing Existing PlanningMonnation 79

7.3. Conclusions 807.3.1 Implementationand FieldTesting 807.3.2 DataModel 807.3.3 Interface 817.3.4 Reasoning 81

References 82

Appendi r B: CQCPScopeof ImpIementatio~ 95Appendi r C: Case-Barcd Remonhg Disrance CPlrulations 100


7/113

Listof TablesTable 3 .1 : Functionality categ ories forCACP ........................................................... 2 7

..........................................able 4.1: Case-based reasoning implernentation for CACP 5 0...........able 5 .1 : Hierarchy icons the source hierarchy and the concept they represent - 5 7


8/113


9/113

Figure 6.1: A high level product and path to detailed information ................................. 66Figure 6.2 . The product hierarchy for H-piles................................................................67Figure 6.3 A process and its related kernet entities....................................................... 68Figure 6.4.Copying fkom an existing project................................................................ 6 9Figure 6.5 . Selecting from a library of type information ......................................... 71Figure 6 .6 . Set-up for reasoning about methods to select............................................... 72Figure 6.7 . Th e reasoning ptduct and its attributes....................................................... 7 2Figure 6.8. Results of query by similarity...................................................................... 73Figure 6.9. Deciding what information to retrieve......................................................... 73Figure 6.10 . Examining more detail prior to retrieval..................................................74Figure B : Hierarchy of irnplemented Visual Basic forms............................................ 99


10/113

AcronymsAEC -Architecturai, Engineering, ConstructionAl -Artificial IntelligenceBCCM -Building Construction Core ModelCAD-Cornputer-Aided DesignCACP -Cornputer-Assisted Construction PlanningCASE -Cornputer-Aided SoftwareEngineeringCBR -Case-BasedReasoningCIC -Computer Integrated ConstructionCPE -Collaborative Project EnvironmentCS1- Construction Specifications InstituteLAI - ndustry Alliance for InteropenbilityIDEFO- ntegrated Definition for Function ModelingIDEFlX - Integrated Definition for Information Modeling ExtendedIFC - Industry Foundation ClassISO - nternational Organiurtion for StandardkationIT - nformation TechnologyMS - MicrosofiPMDM -Project Management Data ModelPMI -Project Management InstitutePMBOK - Project Management Body of KnowiedgeRAD -Rapid Application DevelopmentRIBA - Royal Institute of British ArchitectsSTEP - Standard for the Exchange of Product Model DataTOPS - Total Project SystemsUofF -Unit of Functionality


11/113

AcknowledgmentsAs is the case with many things undertaken in life, the joumey is often more meaningfulin the end than the goals or rewards that one is purwing. To me, leaniing is one of themost flfilling activitia in life. The completion of this research has been a joumey ofsorts and 1 have learned so much thai 1wish to acknowledge the contributions of themany people who have facilitated this process.Foremost, 1wish to express my gratitude to my supervisor Thomas Froese, who agreed t otake me on as one of his first Ph-D. students. 1 feel very fortunate to have had hisguidance and support throughout this work. His exceptional research provided thestandard to achieve, his perspectives have broadened my thinking, and his well-timedqueries have kept me foaised on cornpleting this work. Lloyd Waugh has unwittinglyserved the role of mentor throughout my graduate studies career, not only in mattersrelated to research and work but aiso in life, thanks Lloyd. One of the finest teachers thatI have had the pleasure of leaming fkom, Alan Russell puts meaning to the challenges thatresearchers face in the construction industry and motivates those that chose to pursuetheir solutions. In my opinion, these three individuals are among the elite in research andteaching performe in the field of Construction Engineering and Management in thiscountry and throughout the world. It is my honor to have been associated with them.1wish to thank members of my exarnining and research cornmittees. They sent me off inthe right direction after my initial twelve-month stay in Vancouver and their contributionshave been invaluable to this work. Thank you to: Tarek Sayed, for impressing the needfor a quality prototype development, Cenon Woo, for his insights on informationmanagement, and Denis Russell for his perspectives on the profession. 1also appreciatethe input provided by the university examiners John Meech and Siegfried Stiemer, andthe extemal examiner Iris Tommelein for their chailenging and constructive reviews.Thanks to my fellow graduate students at both the University of British Columbia and theUniversity of New Brunswick and colleagues at the Construction Technology CentreAtlantic and the University of New Brunswick. 1 have been fortunate to meet a verydiverse group of individuals, collaborate with some, reflect with others, and feam fiomail. An additional thanks to Kirby Ward for the expertise he shared in the implementationof the CBR ools and a special thanks to John Christian for the wntinued support he hasprovided me and the Construction Engineering and Management Group at the Universityof New Brunswick.I also wish to acknowledge the absolute support provided to me by my mother and fatherand thank them for instilling in me a sense of heritage and for being so easy to please. Tomy brother and his family, Deedee, and al1 rny other family and fi-iends, thanks for askinghow things were progressing without asking if 1was ever going to finish.Finally, my most special thanks goes tomy wife Shelley, we agreed to take our adventureto the west-coast four years ago and have been blessed with good fortune since. My son,Keagan Hudson, was born shortly befo n the completion of this research and provided al1the inspiration and necessary distractions require making its conclusion enjoyable.Without their unconditional support 1would not have finished, so it is to them that 1dedicate this work.


12/113

Chapter 1: ntroductionThis thesis discusses research in the field of Computer Integrated Construction (CIC)with a focus towards contributions to the development of an integrated constructionmanagement systern, the Totai Project Sydems (TOPS) roject, as proposed by Froese(1995). In t his regard, the research addresses the voluminous quantities of informationrequired of an integrated construction management system by initially populating aproject plan.The terni "total-project systems" (Froese et al. 1997a, 1997b) is used to refer to aconceptual zed class of construction management computer sy stems that are defined bythe following characteristics: comprehensiveness 4 e y include a suite of applicationsthat suppon a broad range of construction management fnctions, integration - ailapplications contnbute to and draw 6om a s h e d pool of project information, andflexibility - they operate in a highly modular, open, flexible, and distributed fiamework,rather than in a restrictive and prescriptive manner.The constmction industry is characterized as having many players of multiple disciplineswho are brought together at different stages throughout the life cycle of unique projects.The supporting management processes result in a reliance on an enorrnous amount ofinformation produced by many sources at many levels of abstraction and detail. One ofthe primary goals of CIC is to simplie the methods for handling the informationgenerated throughout the life cycle of a project, and thereby promoting more efficient andeffective management processes. This is generally foreseen as a sharing of projectinformation by al1 participants, across the various management supporting computerapplications, applied throughout the construction process.To facilitate the sharing of project information, common modeis of al1 projectinformation (product, process, cost, organizational, resource, etc.) are required that detail,completely, al1 aspects of a project throughout its life cycle. These common models canonly be feasible if standards are established for both their development and structure.One cornerstone of the TOPS approach is to build upon ongoing architecture,engineering, and construction (AEC)data standards efforts to develop common datamodels to support construction management data. Upon widespread acceptance of suchmodels, it will then be fasible to capture and store the knowledge of projects for use onfuture projects. A second major element of TOPS is the development of the variousapplication modules that make up the integrated system. While many of these aretraditional project management applications such a s scheduling, others are moreinnovative, such as the storage and subsequent reux of constmdon planning knowledge.This is important to address TOPS'high data-entry requirements.1.1. Concep ualizing Integmted SystemsIn general, existing construction management software applications are not used to theircurrent potential. This may in fact be partially due to the organizational changes thatmust occur in order to maximize the benefit of their application. It is certainly due, inpart, to the unavailability of complete systems that are applicable to a broad range ofconstruction management functions for a variety of project participants.


13/113

Figure 1.1 depicts the current and conceptuai environment for construction managementtools. In the top half of the figure, current applications fan be descnbed as a twlbox ofproject management tools that are ail usefil in their own specific tasks addressingmultiple views of a project's information and Erequently at various levels of detail (e-g.,scheduling addresses time, estimating addresses cost). The toolbox analogy is f ine fordescribing a tool in tenns of its view, however, it does not account for the degree ofoverlap in the scope of project information that individual tmls address which lead todifficulties in consolidating project infonnation. For example, scheduling tools takeprimarily a time view but dso address cost information as well and usually with adifferent structure and diffrent level of detail than, say, an estimating tool would.Although the information used by aiment construction management tools overlaps,vimially al1 information exchange among these tools today occun by printing the outputof one tool on paper and entering the information in different fonns into another tool.In contrat, t h e bottom half of Figure 1.1 illustrates the conceptuai elements of the TOPSapproach. The concept is for integrated project management systems that rely oninfonnation stmctured according to a shared mode1 base. The integrated tools providemuch th e same functionality as the contents of the twlbox, but with greater informationconsistency.

Project hfanagemrnt Toob(scheduling. cstirnating.mcthds . umtrol, costing, etc.) levelsof deuil)

( rrpoduct a~ nd uscmbly ofinfonnatiambymultiple pt i c ipu i t s )

OComputer-Assisrcd:onStruction Planning

(organizt. manipulate, and Totd-Roject S y mbuildup project information) (modulu. dpai. flexible

and distributalCruacM.>rlt)

Figure 1.1 : Conceptualizing Compter-Assisted Construction Planning in the context ofTotal Project Systems.


14/113

Computer-Assisted Construction Planning (CACP) is one piece of the integrated projectmanagement puzzle. It supports planning in the context of integrated projectmanagement tools by storing information about construction planning, in general, andp s t projects, in particular, in a format consistent with a project management applicationinformation model. The scope of the planning information, therefore, includes al1 formsof planning such as work plans, scheduleq estimates, organizational plans, documentcontrol and communication plans, etc. The result serves to initially populate the pool ofproject planning information that is then available for use by other specializedconstruction management planning tools. Therefore, it is not a new planning tool butrather an advancement of planning twl concepts in light of an integrated approach. Inthe context of an integrated system, applicable modifications are expected to occur fiomthe beginning of a project's conceptual stage and could continue up to the population ofinformation required for construction planning.To consider the capture and later use of project knowledge, a fiamework is required toallow retrieval of project information at various levels of detail fiom a variety of views.Key components to this fiarnework are both a practical interface and approach. With thisability to retrieve and modify templates of past project knowledge, the planning phase ofproject management is then partially automated. A large body of project informationmust be acquired, maintained, and manipulated in order for an integrated system to beuseful. The computer-assisted aspect of the research requues a tool that helps to easilydevelop initial comprehensive plans, thereby reducing the data handling burden andincreasing opportunities for the development of quality project plans. Of the researchefforts investigated, no one solution addresses al1 the aspects that are felt necessary for acomplete solution to the objectives of computer-assisted construction planning.The concept of an integrated project management system is central to the researchpresented. The results of this effort come at a time when the foundations for integratedsystems with the characteristics described are being established. Two essential elementsof this foundation are being addressed through the development of a data model tosupport the planning steps of construction management and tools to manage the initialinformation requirements of these management processes. To frther the appreciation forreqiiirements of such a system, two simplified conceptual information usage scenarios arepresented in Figures 1.2 and 1.3.This conceptual usage scenario considers the initial entry of project information and thesubsequent steps, in which multiple users (project participants) access portioas of theproject's information for selected management applications (e.g., scheduling, estimating,etc.). Figure 1.2 presents the usage scenario for an integroted system with a sharedproject database and the means to initially populate it. Each box refers to a portion of theoverall project information (although they are not expected to be qua1 in size orcomplexity). The solid boxes indicate information that must be created by componentswithin the system and the dotted boxes indicate where components use projectinformation that was produced by other components within the system. The initialproject information (shaded boxes) is built up using an appliution with the ninctionalityof CACP. Then information is added to the project database by multiple users ofapplications that draw fiom this initial information and apply more detailed planningfnctionality (e.g., resource management algorithrns, economic analysis, etc.).


15/113

application area 1

application area 2

application

initial population

Figure 1.2: Project information with integrated systems.

In a scenario without the complete characteristics of integrated systems (Figure 1.3), theinitial information requirements mua be built up within each application a r e -and thereis the potential for redundancy within each application depending on the tools applied byeach project participant. Although the scenarios presented are rather basic, theramifications should be apparent by comparing the nurnber of boxes representing thedevelopment of project information in order to flfill the planning processes (Le., 19boxes in Figure 1.2 compared to 48 boxes in Figure 1.3). Whether the information is theoutput of an application or simply a reproduction of existing project infonnation, theadditional steps required in the scenario presented in Figure 1.3 increase the burden ofmanaging the project information (e-g., information consolidation from separateapplications, avoiding inconsistencies among redundant data sets, etc.).These scenarios assume that current twls will continue to be implemented in the samecapacity as in today's environment, an issue that must be examined for each applicationin light of expected changes to industry work process and the new fnctionality offered inan integrated implementation. Regardless of the answers for these issues, an envuonmentthat supports integrated project management systems based on standard informationmodels will reduce the information management burden and thus increase themanagement efficiency of the industry.


16/113

appiicab'onarea 2

appicatiori area 3

information#rteqaon

multipk wem rnulbjplr uwm

Figure 1 3 Project information without integrated systems.

7.2. GoalThe topic of this research is the "compirter-assrsted&wIoprnent of initiai comprehensiwconstruction plans irr the context of an integrated colsstrtlction management system.Comprehensive suggests planr thut indu& elemenfs of seope, cost. tirne. per/nnance,and organization."3 ObjectivesThe objectives of this research were as follows:

To contribute to the development of a standard information mode1 for theconstruction and related industries.To develop an effective and escient interface for dealing with large quantities ofinformation.To develop a hierarchical fiamework that facilitates the representation of planninginformation fiom separate views at different levels of abstraction.To demonstrate how existing planning knowledge can be represented and transferredto disseminate the benefit of shared knowledge of a construction project.To explore an approach to managing planning information by applying case-basedreasoning to construction planning.To implement selected components of a system that can generate an initialconstruction plan for use in an integrated construction management system.


17/113

1.4. ScopeThrough fblfilling the objectives listed, the deliverables of the proposed research areessentially the components of the prototype system. Deliverables of the research are asfollows:

A fiamework for representing a hierarchical breakdown of project information andthe ability to manipulate this same information.A system for the assisteci generation of cornprehensive initial plans in support ofTOPS initiatives.

Contributions to the research efforts in this area in addition to the deliverables aresummarized with the following points:A significant effort towards the development of a project management applicationmodel.An example of the application of standard core models in support of integratedsystems.An implementation of integrated systems research through the development of aprototype based on rapid application development (RAD) tools.Support for the tture development of integrated construction management systems.

1.5. Research MethodologyAs a starting point, a thorough investigation into the approaches taken for automatedconstruction plan development was performed. This included an identification ofknowledge requirements through a study of past projects and a more detailedinvestigation of available literature. Particular attention was paid to the manner in whichhistorical planning knowledge is captured and reused and the methods proposed for theapplication of this knowledge to cornputer-assisted planning.The following step was the selection and refinements of conceptual wre models, anddevelopment of models to the expected level of detail required for the overall research.This included an analysis and cornparison of several on-going model developmentprojects and a collaborative effort in the development of an information model in supportof project management with an emphasis towards the supporting structure for planninginforrnation. This step reguired a complete usage scenario for CACP, detailing thesystem's fnctionality, as part of an overall system development plan amplete with asystem quality plan.A contributing effort towards both the CACP and TOPS projects was the development ofa common database for storage of al1 project information. The focus was on theframework for how planning objects within each fiamework is organized in a hierarchicalbreakdown, how stnrcturing of the hierarchy enables mapping between objects, and howthis mapping is represented and wntrolled. Vaiidation of the developed model wasperformed while populating the drveloped database with information fiom an actualbuilding construction project.


18/113

Having structured and developed the datab- the next system development step was tofiilly develop the intedace for representing th e hierarchical breakdown of project(planning) information. This included the exploration of how such an interface can beused to manipulate and modifL project informat on-With the conceptual core model, database, and information interfiace established the nextstep in the research was developing the fnctionality required of the system forgenerating plans with the characteristics discussed. This began by finaiizing the breadthof a plan's development, the approach to plan development, and the structure of thelibraries of planning knowledge. In consideration of these approaches the knowledgesoftware was then selected for completion of the prototype system. However, theselection of the knowledge software was tiom the perspective of "proof of concept,"rather than the system capabilities, as ease of development was weighted high.Having determined the rnost prornising knowledge software methodology, the final stepin the system development was to complete the integration of the database, the interface,and the knowledge application. The last stage in the research was to test the system withthe project information collected previously to evaluate both the system and the completeCACP approach.7.6. Reader'sGuideThese chapter summaries are provided as a synopsis of the contents of each of thechapters to follow.

This chapter summarizes previous research work upon which CACP draws fiomand builds on. The categories of research work include: integrated constructioninformation systems in general, specific components of integrated informationsystems, and approaches to automating and assisting the constructionmanagement tnction of planning.

Chapfer3: Project ManagementDal'aM d fThe steps involved in the development of the Project Management Data Model(PMDM) are described in this chapter. The model, initially based on an existinginformation model is modified, extended, and simplified to tiilfill the scoperequired for CACP.

The approach adopted for CACP is described in this chapter. The approachdescription includes CACP's unique viewpoint fiom an integrated projectmanagement system's perspective and describes the manner in which case-basedreasoning has been adopted.Chqpfer5: Interface and ImplementationThe structure of the 'proof of concept" application is describecl in detail. througha presentation of the module interaction fiom a software perspective.


19/113

Several exa mples ar e presented to describe CACP n action and demonstrate itsuse. This is accomplished with the use of multiple screen captures of theapplication w orking o n sample project data gathered fio m an actual constructionproject.

Chqvfer7: ContributionsandComlw~ollsThis chapter presents the contributions to the current body of knowledge in theresearch of integrated construction management systems and the constructionindustry, specifically in the field of construction inform ation m anagement andconstruction planning. The inal chapter points out that continued application o fCACP and additional prototype systems to current construction projectsattempting an in tegnt ion of team project information is paramount to theunderstanding of ho w these ftt ur e practical tw l s will benefit the industry.


20/113

Chapter 2: Pointsof DepartureToday's construction planning is essentially comprised of scheduling, wst estimating,and technology selection. These processes involve a significant management effort andare al1 supported, at various levels of success, by project management software.Although supporting applications are available, theu total acceptance industry wide anduse in an integrated approach has not yet been practically achieved. This chapterdescribes five areas in which research has been performed or is ongoing to suppon theintegration and acceptance of currently available applications and h u r e constructionmanagement supporting applications. Any criticisms offered in the following discussionare not intended to detract fiom the research perforrned. Instead, the aim is to point outpossibilities for improvements to existing research and oppominities for futurecontributions, while indicating in a broad sense the sirnilarities and differences in theapproaches adopted for the CACP research project. Specific differences b e e n eachapproach and this research are explained where relevant in subsequent chapters. Sections2.6, 2.7, and 2.8 explore the direction of three concepts related to the development ofCACP's prototype application: case based reasoning, current project managementapplications and t heir interfaces. The summary statements provided describe the pointsof departure for this research.2.1. Cornputer lntegrated ConstructionAs stated in the introduction, CIC is viewed as a solution to many of the adversecharacteristics of the construction industry, ofien summed up as a fragmentation of theindustry leading to little improvements in its overall productivity (many players withmany goals). However implementation of new approaches does not occur withoutjustification, and the barriers for successtl integration are significant.Froese and Russell (1995) discuss design challenges facing the medium sized contractor(who accounts for the largest volume of work in the industry) in consideration of CIC.The discussion calls for an increase in the breadth and depth of existing applications, anability to work with expertise (captured industry knowledge) and the integration of al1applications. Requirements of data representation are needed that allow a hierarchicalbreakdown based on generic models leading to more comprehensive and richer data sets.All this contributes towards the "critical ma s " of supporting tools required before suchapplications with their promised advantages will be accepted industry wide. Manyresearchers have proposed hypotheses and strategies for how information technology cansuccessfully contribute towards achieving CIC (Ahmad et al. 1995;Teicholz and Fischer1994; Breuer and Fischer 1994). Researchers continue to examine the need for changesto industry processes in relation to the adoption of information technologies (Hinks et al.1997). in general, al1 approaches cal1 for shared databases of information across theparticipant boundaries of the industry with an accompanying change in the managementstyle of construction projects. Therefore, to date, it is generally accepted that integratedconstruction management systems will be conceptually based on a central informationsource with which integrated applications and the industry participants will interact(O'Brien 1997).


21/113

Summary: This research recognizes the ramijicutions of such strutec changes to theinciustry upon a d o p io ~f CIC. H~ower ,hef m s is net on the orgmizatiomfchangesrequired but rarher on the supporting tools reguiredfor such changes o occur.2.2. Indusfry CkssificationSfandadsIndustry standards are available for the classification of information and the exchange ofstandard documents, whether electronic or otherwise. There are existing standardctassification schemes whose application is norrnally on a national level, whose initiaidevelopment was for the purpose of improving the efficiency of communicationthroughout the construction procesS. Examples of these include the MasterFoxmat (CS11995) or UniFonnat (CS1 1998) produced by the Construction Specification Institute ofthe U.S. and administered by the Canadian Standards Association in Canada, and theCL/SfB Construction Indexing Manual administered by RIBA (Royal Institute of BritishArchitects) in the U.K.For example, MasterFormat, introduced in 1963, was initially based on the nature ofconstruction and the industry at that time. It consists of a single classification tablestructure of information categories (divisions) that paralleleci the disciplines of work (e-g.,Division 1O00 General Rquirements, Division 2000 Site Construction, Division 3000Concrete, Division 4000 Masonry). Within each division subdivisions are available tofurther classify construction information (e.g., 3100 Concrete Forms and Accessories,33200 Concrete Reinforcement, etc.). However, as the industry changes in aspects suchas the structure of its participants, the manner in which projects are delivered, andparticularly in the emergence of increased integration of information and electroniccommunication, these standards are losing there ability to contribute to an efficientcommunication process.In light of the recognized importance of standard classification schemes to moresophisticated management systerns within the industry, there has been a revarnping of thecurrent classitication schemes. Evidence of this trend is found through the work of ISOtechnical comrnittee ISOlTC 59 (ISOTRI4177 1994, 1996) and the British derivative ofthis work, the Unified Classification for the Construction Industry (UniClass 1998). InNorth America the new trend is towards the product-centered scheme entitled UniFonnat.Alt hough UniFonnat also serves the industry better in the definition of performancebased work (the expected origin of its development), it is also an improved scheme forintegrated management system approaches. he Construction Specitications Insitute hasalso embarked on the Information Integration Initiative in order to suppon an integratedindustry approach (Cassidy 1999).These new developments within classification efforts recognize the emerging model-based approaches to information integration and are structuring their schemes in multipletables that address specific types of information groupings. This provides a much moreefficient approach to classifying information by allowing the application of multipletables to provide additional detail for a classification rather than remaining within asingle rigid classification table. For exarnple, classification tables exist for "elements" ofa project while a separate table is available for "processes" allowing the classification ofinformation fiom both these perspecives rather than the single tabled approach describeciabove.


22/113

Summary: The indusfry gui&fims for ck$mt i on of info11114tion me piramount tostmcruring and efi cie nfl y organizing prolect mum gement information in a hierarcicalrepresentation within un integmted infonnation qvprmch.

It is generally accepteci that conceptual core models will provide a consistent approach tomodeling within a speci fic application area and increase the capabilities of informationsharing and integration among such applications within the architecture, engineering,construction (AEC) domain (Froese 19%). Core models serve as a reference toapplication specific or dornain specific information models and by doing ro impute aconsistent stmcture among these specific models that promotes information exchangebetween applications.As discussed in Froese (1996), concephial core information models provide a consistentapproach to modeling across related application areas and increase the capabilities ofinformation sharing and integration. This is the approach taken in many efforts withinthe large-scale information modeling AEC dornain. Figure 1.1 dernonstrates the layeredapproach of core modeling. A domain area (e-g., al1 business within the AEC industry)can be divided into application areas (e.g., siniaura1 analysis, project scheduling, etc.).

Cote Modd

Application Models

Individual ConputerApplicationsFigure 2.1: Ulustration of a layered modeling approach (Froese, et al., 19970).

The representation of information supporting business activities within a domain followsthis same breakdown. The core mode1 at the highest semantic level provides the basisfrom which more application-specific data models are developed. These applicationmodels could be at a level that supports individual cornputer applications within one ormany application areas. The result is a consistent approach to modeling and amechanism to suppon information exchange between application areas within the samedomain-Application models represent the information required for a specific application withinthe core domain. Examples of traditional construction management applications includemodels for scheduling, estimating, and cost control. New applications such as


23/113

technology and methods seledon, project documentation control, and performancesimulation and anaiysis are also examples of applications for which modeling isapplicable. A modeling approach of this nature is emerging to support an overall projectconcept through the accomplishments of some key research efforts.The examples presented in this section are efforts in the estalishment of informationmodels for the purpose of supporting information integration.

2.3.1 STEP (Standardfor the Exchangeof Produd Model Data)The ISO standard 10303 or STEP (Standard for the Exchange of Product Model Data) isa standard for the computer-interpretable representation and exchange of product data,with the objective of providing a neutral mechanism to describe the product data throughthe life cycle of the product independent of any particular system (ISO 1992). OfparticuIar interest is the product data representation of AEC and a subset of M C,building construction. Due to increasing demands fiom industry, the STEP effort ismoving fiom a capability of not only data sharing (exchanging data between systems),to wards data integration (sharing informat ion among different users and applications)(ISO 1995a). To make this transition several of the fundamental changes king proposedare: (1) to develop a r e models for specific domains, (2) to implement templates formodeling styles, (3) the establishment of classification schemes, and (4) the developmentof rules and constraints for mapping purposes. These principles have been applied to thedevelopment of the Building Construction Core Model (BCCM) ISO 1995b).

2.3.2 Industry Alliance for lnteroperabilityCurrent effons within the International Alliance for Interoperability (W),gmupcomprised of industry-based firms, software vendors, and researchers, are workingtowards the development of Industry Foundation Classes (together forming a corereference model) for exchanging data between computer systems within the AECindustry (Autodesk 1995) @Ai 1997). There is in fact quite a synergy betweenindividuals involved with both the STEP and IAI initiatives. Specifically applicable tothis research is the work king produced by the domain cornmittee focusing o n projectmanagement related aspects of the Industry Foundation Classes (Froese and Yu 1998,1999).

2.3.3 Project ModelsVarious research efforts have contnbuted to the research noted above, such as the GRM's(General Reference Model) a predecessor of STEP (Reschke and Teijgeler 1994).Recognized advancements to project model development have also been cited to themodeling work of Froese (1992) and Bjorg (1992).The COMMITT (Construction Modeling Methodologies for Intelligent InformationIntegration) project (Rezgui, et al. 1996) is based on the work of the previous ICON(Information Integration for Construction) project (Ford a al. 1994). The projectproduces individual models to describe di fferent perspectives of project information (e.g.,a construction planning object model) based on sound information technologyframeworks, a slight alternative to a meta-model approach.


24/113

Other research project's have dso recognized a layered modeling concept moving Eromgeneric meta-models, to conceptual models, to industry reference models, and finally tospecific project models. An example is the work performed at VTT Building Technologyfocusing on the development of models and twls for the improvement and re-engineeringof construction processes (Hannus and Philainen 1995, and Hamus 1996).2.3.4 ProcessModels

Existing models (perhaps better characterized as descriptions) of industry processes, suchas the Project Management hstitute's (PMI) ody of Knowledge (BOK) ocument arealso available for reference (PMI 1996). Additional examples a n lso be found in theworks of Sanvido (1990) and Gillespie (1998).2.3.5 Other Models

Effons directed toward otherAEC disciplines such asCOMBINE Computer Models forthe Building Industry in Europe) (Augenbroe 1995), which emphasizes design practices,are also usenil sources of reference. As are the models developed asa result of integratedsystems deveIopment, noted below.Summary: n i s research is itzterested in the representution of the information reqiriredfor a comprehet~s~veroject description (Le.. pradrrcts, procesws, resources. etc.). aswell as, a fonrs on the integrution capabifities oBred by t h qpfication of st-drnodels. Essetztiaffy S E P ' S BCCM "as is " r n d f was the stmting point of modefdevelopment with consideration given to on-going m d f i n g eflorts an q p ro ~ c he swithin IAl.2.4. Automated PlanningAutomated planning is not a new topic. Past approaches have focused on artificialintelligence (AI) solutions (Levitt et al. 1988) and were initially centered on basicinformation objects and low level reasoning to derive project components and to generatea time scaled activity sequence. Although the approaches have been cnticized for theirinability to effectively handle the size and complexity of actual projeas, they have mostdefinitely provided a thorough foundation describing the structure of constructionplanning. The references that follow are prominent efforts in this area of research.

2.4.1 CONSTRUCTION PLANEXCONSTRUCTION PLANEX (Zozaya-Gorostiza et al. 1989; nd Hendrickson et al.1987) s an expert system addressing a11 aspects of construction planning fiom identifiingactivities required for given design elements to selecting technologies and constnictionmethods, and estimating durations and costs. The system requires detailed designinformation descnbing the stnictural elements of a building. Knowledge sources are usedfor selection of construction methods and assignment of durations and costs. Theknowledge is essentially activity based and cost models are described as simple in nature.Precedence relationships are pre-defined and the aggregation capabilities for adivities arel imited. The focus of the work is primarily on the precedence and scheduling aspects ofconstruction planning.


25/113

2.4.2 Sequencing KnowledgeInvestigation into the forrnalizing of sequencing knowledge for construction scheduling isdescribed by Echeverry et al. (1990 and Echeverry 1991). It is proposed that sequencingknowledge can be captured, and divided into one of four groups: physicd relationships,trade interactions, path interference, and code regulations. A nirther breakdown of eachgroup is performed and the resulting application of the knowledge is to serve as an aid forscheduling construction projects.

2.4.3 Other ResearchGHOST (Navinchandra 1988), OARPLAN (Damiche et al. 1988; Winstanley et al.1993), SIPEC (Kmam and Levitt 1990, Kartam et al. 1991), ACP (Waugh 1990). andBuilder (Cherneff et al. 1991) are al1 fiirther research examples of AI approaches to theautomation of construction planning. Included in the next section are some currentextensions to these works.Summary: For ihis research, the importance of autornated scheduling r e s e dpegormed fo &e is the fomaiiza~ionof comtmction plming howIedge and theqprmches laken to capture this knowIedge forjirhrre use. EQuaIiy important are themeth& usedjor groupingamiclss%yig this knowfedgean its interreIationships.2-5. Assisted Planning and lnfegroted SystemsThe following efforts are in line with: a systems approach to planning with a movementtowards integration of applications; a higher or richer level of information representation;and the use of templates or libraries of knowledge structures which can be aggregated,easily accessed and applied. The shift is towards information technologies as opposed toa focus on AI solutions. These research projects are more closely related to the end resultof cornputer-assisted construction planning and do incorporate some but not al1 of theapproaches under consideration.

2.5.1 lntegratedConstruction Plaming System (ICPS)The integrated construction planning system ((CPS) proposed by Yamazaki (1993, 1995)focuses on formalizing the representaion of the planning process as an approach to CICand promotion of interactive cooperation between distinct stages of planning. Theresearch defines building system planning and construction system planning together asan interface between design development and construction planning. Building systemplanning, a fnction of the design process, specifies the fnctional requirements ofbuilding spaces, develops subsystems and components, and defines fnctionalrequirements of the subsystems and components. The constmction system planningevaluates the results of building system planning and proposes constniction processes andmethods for the subsystems and components based on requirements and constraints as theplan progresses through formalized stages of planning (concept, tundamental, andpractical).The research recognizes the differing approaches of a top-dom approach, based initiallyon a product model and deriving processes, as well as a bottom-up approach focusing onthe process model (construction activities). A projec modeling system based on product


26/113

and process models allowing middle-up-anddown approaches (comprehensive planning)is proposed based on an object oriented paradigrn.The ICPS also uses a hierarchical representation of each distinct model identified,however no indications are given as to how the mapping is defined between differentlevels in each model. The research aiso defines distinct planning stages and recognizesthe need to support an evolution of the construction plan. The storing of information in ageneral planning model suggests that some sort of information type Iibrary is considered.

2.5.2 Automated Building Realization System (HISCHED)Various components of an integrated automated building realization system have beeninvestigated at the Israel National Building Resevch Institute and are based on a projectmodel (Warszawski and Sacks 1995, Warszawski and Shaked 1994). One of thecomponents of the system is MSCHED (Shaked and Warszawski 1995), described as aknowledge-based expert system for the construction plan~ngf buildings. The approachof this research is that project models should be tailored to specific building types. Theresearch is described in tenns of design and constniction of onhogonal, multi-storybuildings with uniform floor plans. The project mode1 is composed of a product,resource, and activity models. A structure of building process stages are defined withspecified input requirements and expected output results with the building project modeldata base growing in detail at each stage.The approach begins with the identification of building spaces and their requiredfnctional systerns. Technological solution objects are then provided for each fbnctionalsystem object, and prompt definition of the required work assembly objects and elementobjects. Basic activity objects are then defined to install assemblies and elements. Ahierarchical breakdown requirement is recognized for assembly, element, and activityobjects. As well, mapping issues between different hierarchical levels of objects andduplication of tlfillment of fiinctional requirements by single elements is addressed.The computational solution is described as a "pararnetric assembly" or "adoption oftemplates" solution. The premise is that the templates are stored in libraries outside thecalling program with complete definitions able to serve at any stage due to theirconformance to the project schema and defined hierarchy.The approach requires a project model stmctured in such a way that it can facilitate theaddition of detail throughout the project life cycle, however the manner in which thehierarchical mapping is performed is not very clear. A restriction of parallel majorclasses of space, produa and activity classes is alluded to which may address themapping issue at one level, however other fndamental objects such as cost and resourcesare not discussed. The approach is quite rigid in its stmcture and lirnited in itsapplication.

A more practical focus on the management of product and process information is takenby researchers at the US A m y Corps of Engineers Construction Engineering ResearchLaboratones (USACERL). The system developed, CADCMS (computer aided designand constmction management information system) (Stumpf et al. 1996), links several


27/113

project management sofhvare applications through a relatiod database. Buildingcom pone nts and process information are linked by allowing a grouping o f com pon ents tobe link ed to selected activities, esan tial ly a n activity view. Th e systems go als are Iistedas allowing integration, multiple views and ability to handle changes in information.Th is initiative calls for fiiture research into a structured and efficient information storageto help im pro ve the information transfer.2.5.4 Construction Method Driven Scheduler

CMD Scheduler is a prototype system developed with the intention o f addressing th erequirem ents o f applying knowledge-based scheduling and planning system a s industryapplications (Aalami and Fischer 1996, Fischer and Aalami 1996). Central to thisappro ach is the use o f construction method information, which is captured in a methodmodel a nd re-used in the dwelop rnent o f estirnates and schedules.In the con tex t o f an integrated construction management system, th e researchers n ote thatthere are n o curren t mechan isms that "automatically" and "dynamicaliy" Iink designinform ation to construction planning and scheduling information. With the assum ptionthat th e user has the knowledge to make decisions regarding the constmction methodsapplied, the prototype system addresses this link in support o f the rapid d evelop me nt ofestimates and schedules.A 3-dimensional CAD product model provides input to the model, while processknowledge is available in the form of construction method templates. Th e constructionmet hod temp lates contain information regarding their application doma in, theconstruction activities the method is composed ot, the sequencing of these activities, th eresource requirements, and the objects that each method is applied to. A plan isdeveloped by starting fiom a high level process-product pairing and applying theapplicab le met hods that introduce a refined level of process-product id or m at on. The sesteps a re repeated to the level o f detailed required.

2.5.5 IT Support fot Construction PlanningJagbeck (1994, 1998) describes the development of several prototype applications insuppo rt o f construction planning through the incorporation o f a project's productinformation and lays out the conceptual fkarnework for an integrated application thataddresse s construction planning.Again, the overall framework recognizes the need to integnte design information andconstruction knowledge in a planning application (product-process integration). Theprototype application MDA Planner defines "methods" as stores of generic con structionplanning knowledge. A "plan" is developed by adop ting a method in consideration ofexisting preconditions stored in the method template dong with the applicable productsand required resources. A second prototype, PreFacto, focuses on the conversion of"plans" to "methods" and the reapplication of "methods" to products to form new"pIans".The eventual extension to the planning systems desa ibed is a system that supports a linkbetween design and planning through the use o f an ntegrated idor ma tion m odel and thecase based reasoning o f construction planning knowledge stored in template structures.


28/113

2.5.6 REPCON Automated SchedulingT h e evolut on of the research construction planning system REPCON Russell and Wong1993) continues with the conceptual description of a module in support of automatedscheduling. The module is intended to support the creation of dr& plans and schedulesthrough the application of expert systems and large building blocks of predefinedscheduling knowledge (Chevallier and Russell 1998). This research stresses thedeveloprnent of an automated scheduling tool that is fnctional in today's constructionmanagement environment.T h e approach descnbed also recognizes the importance of an integrated produa andprocess view of project information (Russell and Chevallier 1998) and a need to re-usepast construction planning knowledge. The approach expects the user to guide thedevelopment of a plan with the option of applying the support of a simple set of rules indealing with the scale of a project. The rules are applied to attributes such as scale, sitecontext, and contractual requirements in the planning process. Planning knowledge isstored in the form of "activity fiagnets" that consist of the activities requued to constructpart of the defined produa complete with their prtxedence relationships and logic.Sumrnary: An opprouch supporting both aggregation und speciaiizaton of plminginformation s reqtrired Thismust be baced on mi objectdenteiiparadigm (uIlowingarich representation of project injnnation), with O st&d ficatlework (mot&$applicable to o wide range of construction pmjects. Phning occurs a c t o s multiplehieramhicai ievels, an fherefore the upprwch must allow a progression of consmcfionplan development. The concepr of storing cud retrieval of planning knowledge rhroughthe use of templates is the apprrmch adopted2.6. Case Based ReasoningT h e original hype of "case-based reasoning" began with Riesbeck and Schank's (1989)Inside Case-Based Reasoning, a text built on the flourish of Case-Based Ruisoningresearch performed at Yale University in the late 80's. Today case-based reasoning ispresented as a rnethodology for deaiing with large quantities of infonnation (Watson1998) where vanous technologies (nearest neighbour, induction, Nzzy logic, SQL, etc.)are applied to the original algorithm.As applied to planning, people do not construa plans from first principles, rather they tryto find the best previous plan and adapt it to the current situation. A "case-basedreasoner" solves new problems by adapting solutions that were used to solve previousproblems with its elements of a case library and methods of storage, indexing, matching,and adaptation.The basic algorithm that is applied is: 1. input the problem, 2. find the cases similar to thecurrent problem (characterize the input problem, retrieve cases with matching feshires,pick the best match(es)), 3. adapt a previous solution to fit the current problem. Inaddition, concepts are required to support the ever-changing knowledge base that thecase-based reasoning uses. To s u p p o ~hese dynamic memory requirements, commoninformation management concepts of inheritance, aggregation, specialization, andattribute definition are required.


29/113

Applications of case-basd r m n i n g to engineering and construction include design(proportionally much higher), contracting, the application of building regulations,scheduling, and estimating. The following research is noted due to its similarity with theapproach adopted for CACP, while bringing to light the issues involved with thisapplication of case-based reasoning.2.6.1 Design and design selection

Various systems that apply case-based reasoning to support design assistance and designautomation are discussed in Maher and Gomez de Silva Garza (1996, 1997) along withthe difficulties of their implementation. Vol5 (1994) describes a retrieval method for theselection of appropriate designs for building projects in the context of reusable partsrather than entire designs. Sycara et al. (1992) apply CBR to act as a designer's assistantin the design of mechanical devices. While Bilgic and Fox (1996) discuss case-basedretneval as used in engineering design within the context of a n enterprise system.Finally, Flemming et al. (1 994) present case-based reasoning in support of early phasesof building design. Although the area is not mature enough to lead to a generallyaccepted approach, there are many recumng themes noted that are equally applicable toth e application of case-based reasoning to constniction management.In general, the two goals of re-using previously stored design information are to provideaccess to a large memory of past solutions, and act as an aid by providing designers withan initial solution available for editing by the system or user. Addressing these goals forconstruction planning is also the motivation for CACP. The first step is a solution to therepresentation of complex cases, where generalized knowledge and experience must beformalized. The common approach is to represent this information in an object-attributeschema while also maintaining categories for the various attributes (e.g., relation,fnction, behavior, and stmcture). Also, the re u r of solutions is applied across manyphases of a projects design cycle (e.g., conceptual design to detailed design) and thereforeretrieval must be available across a problem that is decomposed in a hierarchicalstructure.Potential solutions must have a dynamic indexing mechanism because the retrievalprocess is iterative (Le., the level of representation of the problem varies). Throughoutthe retneval process the user is quite involved with the pruning process as a designdevelops and additional information is added. Therefore, using a hierarchical approach(relation attributes) allows access to potential sub-solutions.A retrieval process is normally simplified by establishing the context of the problernthrough definition of a set of constraints. The context issue is dealt with up front bymaking it an explicit pan of the query. Then, for each retrieval, a clustering of cases isperformed based on the contexi. The method of retneval and the means of indexingduring retrieval are central to the process. First, the context of the information to beretneved is established, then, the attribute-value pairs that will be used for retrieval areselected.


30/113

2.6.2 Other CBR applicationsCase-based reasoning has also been considered for application to other areas within theAEC industry. Yang and Robertson (1995) provide a fnmework for interpretation ofbuilding regulations, through abstraction hierarchies of legal rules from statutoryregulations and previous relaxcd cases. This example fits into the domain of case-basedreasoning as applied to legal applications. Ng et al. (1998) apply a case-based reasoningtool and address case representation, indexing and retneval, and adaptation for contractorpre-qualification by guiding the user through th e processes of criteria formulation,screening and reviewing, and final assessrnentbase on financial and performance issues.Both examples show the potential of case-based reasoning as an effective aid toapplications within the AEC industry.

2.6.3 Construction PlanningPerhaps closest to the domain of CACP is the research perfonned by Dzeng andTommelein (1995, 1997) towards the development of CasePlan and Tah et al. (1998)with their development of CBRidge, both prototype case-based reasoning system forconstruction planning. Although these prototypes are not based on an integrated view ofproject information, they have considered the basic elements covering a projectsproduas, processes, and resources. Both research effons suppon the concept of appl yingcase-based reasoning to construction planning.Sirmmary.. The kivowledge cornpottent of the CACP prototpe is k e d on a case-basedreasot~ing pproach. As with other applications of case-based reusoning the greotestchallenges are in the prop r representation of i n / m o n tholfonns a case to ennrresrtccessfitl retrieval und azhptaiioa nte application of case-bared rearoning for CACPis upplied o dernomtra~e roof of conceptfor mm~agrhgiored informatiort.2.7. Hierarchical InterfacesConstruction management information is inherently hierarchical in nature, implyinginformation in a tree stnicture. This can be attributed to the unique characteristics of theindustry mentioned previously (many piayen with many goals) which in terms of theinformation requirements translates into many views at a variety of levels of detail. Thehierarchical structuring of information is one of the reasons behind the need for effediveclassification systems discussed previously and it is also a pnmary concem whendeveloping the interface for any construction management application.One of th e most important considerations for a software interface that deals with largehierarchical information structures is th e "focus+context" issue (Lamping et al. 1995).The "focus+context" issue is the term coined to descnbe the trade-off between showingsuficient detail in a hierarchy (focus) while maintaining a sense of where one is in thissame hierarchy (context). 2-dimensional techniques that have been proposed includefish-eye views that allow the user to zoom in on portions of a hierarchical display whilemaintaining a perspective on where they are located in the hierarchy (Sakar and Brown1992). The problem with 2-dimensional approaches is that the growth in the number ofnodes for a single tree is exponential making the layout of any trec very dificult. Toreduce the complexity of a single tree, researchers have looked for ways to conveniently


31/113

partition a single tree into multiple trees based on the properties of nodes (leaves on atree) (Koike and Yoshihara 1993).The advances in this research area of hierarchical representation and manipulation arebeing made with 3-dimensional solutions that are employing techniques such as conestructures or box structures to represmt the third dimension of a hierarchy. However,solutions that require only modest computational needs are only now emerging (NVision1999) and therefore it is more appropnate to look at current solutions and examples to 2-di mensional tree representation.A combination of a tree view and detailed information window is the most commonexample of an effective 2-dimensional hierarchical interface which allows the user toexpand o r collapse branches of the tree structure while clicking on a tree node to viewdetailed information relevant to each individual node. Microsoft Windows Explorer isan example of this approach with its folder and file tree stnicture and detailed window offile information.Summary: The uppIicution's infer/ce wilI fo hw the pend of m e n t 2~me nsi ona lrepresentations of hierarchical infonnaion while considering rne~hods f reducing thecompfexitysuch aspartitioning a d he combined use of a tree view and rietaiied view.2.8. lmplemenfatjon and Practical Pmjecf Management ToolsAn investigation into the current state of industry-generic project management software(e.g., MicrosofYs Project, Primavera Systems' Sure-Trak, Scitor's Project Scheduler,etc.) and its level of support in the storage and reuse of past planning knowledge wascompleted. This was a result of sevaal consulting projects undertaken by theConstruction Technology Centre Atlantic Inc. (CTCA), and lead by the author. Theintent of the investigation was not to compile an exhaustive list of the produas availablein suppon of project management fnctions but rather to gain an ovexview of existingcapab l ities and contribute to defining the functionality required of CACP. Conceptsrelated to the capture and reuse of planning knowledge continue to be exarnined underongoing consulting projects.

2.8.1 Generic Planning Software ApplicationsIn summary, the general trend among the project management applications exarnined,was a concened effort towards: increased capabilities in support of a multi-projectplanning and control, integration with complementing communication applications, and ashifl towards a clientlserver type environment. The reuse of past planning information, ifsupported at all, is generally limited to a simple storage and recall of schedulingtemplates. The most sophisticated of the applications examined, makes use o f a "wizard"that leads the user through a brainstorming session to initially identie project phases,goals, resource, assignments, and tasks-which are selected from an organized collectionof ternplate knowledge. At the end of a "wizard" session the user has the b e g i ~ i n g s f aplan. However, th e pnmary perspective for these applications is scheduling (timeplanning) only.


32/113

2.8.2 ProjectManagement for InformationSystemsIn addition to the genenc project management applications examined, several plandevelopment applications for the Information Systems (1s) domain were aiso examined(i-e., System House's Transform, LBMS's Process Engineer, and Xpert Corporation'sPlanXpert). A common trait among these planning applications is the reuse of planningknowledge that is stored based on industry methodologies. Using an analogy to theconstruction industry, these methodologies cwld be thought of as different contractingapproaches (e.g., fast tracked design build construction). From a xheduling perspective(a hierarchy of tasks), these 1s twls offer additional information such as taskdeliverables, resource requirements, and precedence relationships. More detailedinformation-which might be considered a step towards the capture of planningknowledge-includes guidelines for estimating task duration based on suggestedresources and definable project parameters (e.g., size and complexity of the project).These parameterized estimates and the results also serve as benchmarks for fiinireprojects upon completion.For the IS tools, a plan is also supported by information (in a variety of multi-mediaformats) attached to appropriate tasks, that provides an explanation of how the task is tobe completed. For example, if the tasic is to define the data requirements of aninformation systern, then this task may be supported by an explanation on how to createan entity-relationship diagram and perhaps even contain a link to the software toolsupporting its creation. It is these types of advanced support twls ' concepts that areproposed for total project systemsSummary: Tne trends in cucunent softwme mpporting the planning process me to increaseintegratio~iand communication with other project management took while directiynrpporting the pl~u~ttit~groses ~hroughhe provision of g u i h c e (some in the f o m ofremplates) throt@out the plnningprocess. hese trends also match with the objectivesof CACP.

2.9. SummaryThe foilowing statements describe the starting point for this thesis and point the way forthe research thrusts undertaken.This research recognizes the ramifications of such strategic changes to the industry uponadoption of CIC. However, the focus is not on the organizational changes required butrather on the supporting tools required for such changes to occur.The industry guidelines for classification of information anparamount to smicturing andefficiently organizing project management information in a hierarchical repnsentationwithin an integrated information approach.This research is interested in the representation of the idonnation required for acomprehensive project description (Le., produas, processes, resowces, etc.), as well as, afocus on the integration capabilities offered by the application of standard models.Essentially STEP's BCCM "as is" model is the starting point of model development withconsideration given to on-going model ing efforts and approaches within I N .


33/113

For this research, the importance of automated scheduling research performed to date isthe formalization of construction planning knowledge and the approaches taken tocapture this knowledge for fture use. Equally important are the m e s sed forgrouping and classifying this knowledge and its intemlationships.An approach supporting both aggregation and specialization of planning information isrequired. This must be based on an objectoriented paradigm (allowing a nchrepresentation of project information), with a standard fiamework (model) applicable to awide range of construction projects. Planning occun across multiple hierarcbical levels,and therefore the approach mu* allow a progression of construction plan development.The concept of storing and retrieval of planning knowledge through the use of templatesis the approach adopted.The knowledge component of the CACP prototype is based on a case-based reasmingapproach. As with other applications ofwe-based reasoning, the greatest challengesarein the proper representation of information that forms a case to e m r e successfil retrievaland adaptation. The application of case-based reasoning for CACP is applied todemonstrate proof of concept for managing stored information.The app!ication7s interface will follow the trend of current 2-dimensional representationsof hierarchical information while considering methods of reducing the complexity suchas partitioning and the combined use of a tree view and detailed view.The trends in current soAware wpporting the planning process are to increase integrationand communication with other project management tools while directly supporting theplanning process through the provision of guidance (some in the form of templates)throughout the planning process. These trends also match with the objectives of CACP.


34/113

Chapter 3: ProjectManagementData ModelAl1 effective information technology management sy stems m u a be based on a welCdefined da ta model. Th e importance of such a data model to this research has beenmentioned several times in the previous chapter. This chapter explains in detail th erelevance o f the d ata model, describes the steps take n during its development, and definesthe model in its entirety. Preliminary steps in this data model development, suc h asidentiQing existing models for comparison, were perforrned in collaboration with otherTOPS esearchers a t th e University of British Columbia. Th e author then completed eachdata m odel dev elopm ent step in the context o f this research.3.1. Purpose of the ModelThe foundation of the TOPS approach is basing systems on cornmon data models ofconstruction project information. As stated in Froese e t al. (1997), the significance of th ecornmon data model is that it provides the primary mechanism: for structuring datawithin TOPS pplications; for exchanging information am ong TO PS omponents; an d forinteract ing with oth er computer applications throug h internat ional da ta standards.

3.1.1 Scope of the ModelTh e intended sco pe o f the data model covers the d iscipline of construction m anagement,and perhaps the broader scope of project management, as applied to AEC projects.However, the depth required to support specific applications such as cost estimating orresourc e managem ent is not the intent.The reason for this breadth of scope stems fiom a need to support not only areas ofproject management that are currently well-supporteci by computer tools but alsoapplications within TOPS hat are intended to address new areas of computer supportwithin project manag ement. Currently, docu men ts relating to virtually al1 o f projectmanagement can be tracked within com puter systems, requiring at least some informationabou t the docu men ts (if not al1 of their contents) to be structured in the d ata model.The approach is to develop al1 project management application areas within one overallproject management application model for the present. Th is is an alternative toseparating the information model into distinct application models. A larger modelpromises easier integration across applications areas within project management, butleads to greater difi cu ltie s in developing and maintaining the model. Partitions to th eoverall mode1 are still possible in fiiture revisions, while still maintaining the layeredmodeling approach. A certain amount o f depth is added fo r the specific application o fCACP.

3.1.2 Relationship with International Data StandardsThe PMDM adopted emerging international data model standards that were availableduring this development, such as STEP's Building Construction Cor e Model (BCCM )and the I N Industry Foundation Classes where they already address the appropriatebodies o f information. These models cover much o f the breadth o f project managem entfnctions, bu t a s expected o f cor e models, they fall short of developing the necessary


35/113

depth within specific application areas. The Industry Foundation Classes have a sectionthat covers project management, and several application areas are under developmentsuch as scheduling and estimating (Froese and Yu 1999). A major objective of theproject management data model for TOPS, beyond supponing the TOPS system itself, isto contribute to these development efforts, an ongoing process.

3.1.3 Modeling MethodalogyThe development methodology used for the PMDM draws nom several modelingmethodologies. The main technique is to carry out some fonn of ninctional analysis thatidentifies the fnctional requirements for the data model, then to devise the data topicsrequired to meet these fnctional requirements, and finally to fully detail the objectdefinitions that descnbe each data topic.This approach was initially applied with the w p e of TOPS in rnind and later refinedspecifically for CACP. The fnctional analysis consists of developing a hierarchicallisting of fnctional capabilities required of the systems that the data model wouldsupport. Because the goal is to develop a wmmon data model that serves as a uniQingcore of many individual applications, this is a more abstract process than if the firnctionalanalysis were being performed for a single specific cornputer application. The initialanalysis consisted of defining the fiinctional requirements in the form of fnctionalcategories (topics) that the system and thus the data model should be capable of workingwith (a topdown approach) (Table 3.1). A subsequent analysis define the fiinctionalrequirements fiom the perspective of the specific documents that are commonly used inthe construction industry (a bottom-up approach). The result was categories of data typesrequired to support the fnctional requirements. These data topics (categones)correspond to the units of functionality in the STEP methodology. The collection of al1of the data topics makes up th e scope of the data model. The resulting data model wasthen used for development of the PMDM for CACP.3.2. Mode1 DevelopmentAs mentioned, CACP takes a broader view of planning than what has been typicallytackled by construction plan generating twls. The scope of planning supponed by CACPis defined in terms of the Project Management Institute's Guide to the ProjectManagement Body of Knowledge (PMBOK) (PMI 1996)and its definition of knowledgeareas for describing project management processes. As a hinher refinement to the sopeof CACP, the project phases to which the planning will apply begin with the conceptualphase and end before the development of a detailed construction plan. Again, CACP sused to initially populate a construction plan with information based on p s t experience.Its purpose is not to replace existing detailed project management applications, but ratherto provide a stepping Stone for subsequent detailed planning and controlling applications.

3.2.1 Processes SupportedThe supported planning processes are defined using the PMBOK's tenninology. CACPis intended to support planning from the perspectives of: scope, tirne, cos& performance(quality), and organization. Therefore, support is extended past the npresentation of core


36/113

planning processes of Jcope, time, cost by adding knowledge that is usually implicitlyincluded in the planning processes (i.e., pe do nn an ce and organization). The results ofthis analysis of the planning processes supported by CACP are demonstrated in Figure3.1.A process rnodeling structure similar to the format applied in the DEFo (MST 1993a)methodology is used to describe the planning processes. As is the case in IDEFo, theoutputs, inputs and constraints are identified as flows leading to and 6.om a givenprocess. Me chanism s however, are usudly synonymous with the rewurce requirementso f the process, and in the case of construction management processes, identify thepersonnel involved. It was found more usefl t o identify th e potential tech niqu es tha t theprocesses may use. A h , n the case of construction man agem ent sub-processes, ther e isnot necessarily a sequential execution and, therefore, it is not as useful to disp lay thern inthe common IDEFo nod e form at (i-e., expanding a process into sub-process an d sh owin gtheir intemal flows). An additional modification is the concern with the source off l o w s s p e c i f i c a l l y, wh et her they o riginate from, or contnbute to, a general industryinformation source or a specific project information source. In th e case of CACP, thegeneral information corresponds to past captured planning knowledge while projectinformation corresponds to the current project being pl am ed .Figure 3.1 shows the highest level process of construcfionpfanning.Using the modifiedIDEFo notation the figure shows that, at an overview level, the inputs originate fiom thegeneraf infionnation (e.g., historicat information, work bre akdow n structure, etc.) and th eplanning process produces outputs contributing to the project infonnation (e-g., activity1 st, resource requirements, etc.). Comtraints initial1y originate fiom the generafinformation, but they also draw on the project infonnation as a plan is defined whenknowledge about the planning environment is added and the constraints becorne morefocused. Figure 3.1 alro shows the sub-processes within the ocope of CA CP . Theoutputs identifjc the potential intemal flows within construction planning; how theseinterna1 flows add to the project information; how they are used in other sub-processes;and how they add to th e constraint knowledge.Not al1 mechanisms and techniques shown are necessarily supported by CACP (e-g.,deco mpo sition and templa te use are, while mathematical ana lysis and simu lation ar e not).However, the d ata requirements for al1 techniques m u a be supported at least at anabstract level of detail. The resulting fiinctional categories of this planning processanalysis are displayed in Table 3.1. These finctional categories served as the startingpoint for th e elaboration of each information object towar ds development o f th e PMDMrequired for CACP. The modeled objects went through several iterations and weredetined and organized into units of fnctionality as presented in the n e a section.


37/113

Figure 3.1 : Planning processes supported by CACP.

26


38/113

Table 3.1: Functionality categories for CACP.GeneralClassifications

ListshdexesMet a-Mode1Object HistoryRoot and KemelScopeDrawing (Document)ProductProduct DesignProduct RepresentationCost CostCost InformationCost TypeEstimate (Document)Estimate Item

TimeActivityActivity SequenceConstraint sProcessSchedule (Document)Timing and DatesTime (Resource)ResourceResource AllocationResource Availabi lityResourceCOSResource ProductivityResource RequirementResource Type

Time (Methods)Constmction ConditionsMet hodMethods Plan

Performance (Documents)Codes of PracticeConformanceRegulationsSpeci icationsStandardsOrganizationOrganizational StructurePeople& OrganizationsParticipant RolesResponsibil tyDocumentationContractsControlDrawingsEstimatePurchasesRequirementsSchedule

3.3. Adoption of Existing Data ModelsThe development of the data model began with the adoption of the Building ConstructionCore Mode1 (BCCM) version Str411 (ISO 1995b) and followed several versions of thiseffort including versions Tl00 and T200 (ISO 1996). As the foais of those involved inth e development of the BCCM shifted to efforts within International Alliance forInteroperability, so did the tracking of changes to the core modeling efforts. The finalmodels examined for adoption of modeling constmcts was International Alliance forInteroperability's Industry Foundation Classes @;Cs) versions 1 5 and 2.0 ( I N 1997,1998).As its name suggests, the BCCM is intended as a core model for building constructionand therefore has a narrower scope than that of the hdustry Foundation Classes, whichare intended for the entire AEC industry. The Industry Foundation Classes also containseveral layers which include not only a layer for defining the core model but also layersthat support the "core layer" (the "resource layer defines basic concepts such as datatypes and units of measure) and expand the "core layer" in areas to support specificapplications (the "interoperability layer" and "application layer"). Therefore, detail wasadded to the PMDM nitially based on the BCCM, while not al1 Industry FoundationClass constmcts were detemineci to be relevant or the cornplexity provided not rq u i rd ,due to the specific applications they support. Where simplifications are implemented, itis for the most part the adoption of a defauh Iist provided in place of a more complicatedobject structure. For example, a default list of "product systems" is used rather thandefining these systems as objects within the model. Distinctions between the two base


39/113

models (Building Construction Core Mode1 a d Industry Foundation Cluses) and theresultingPMDM re provided in the following section.The PMDM model also benefited fkom a precursor project that consisteci of thedeveiopment of a data model to support a commercial Facility Management TurnoverSystem (Froese, et al. 1997a) in which the author was a principal in developing. Thismodel was also based on the latest developments within IN.3.4. Modeling Issues, Solutions, and AppmachesThis section describes the issues that have beenaddressed throughout the development ofthe current version of the PMDM. Each issue corresponds to a "Unit of Functionality7'(UotF), the term adopted to define the partitioning of the data model for the purpose ofmore easily describing and defining it. The information structure adopted within eachUofF is shown in the figures that follow each section along with an explanation of theintended solution it offers and how this solution compares with other modeling efforts.Although the model is representative of an object model the model was developed with acustomized CASE tool built in Microsoft (MS) Access and later implemented in MSAccess. Therefore, the diagrams are displayed using IDEF1X (NIST 1993b), a semanticdata modeling technique that has a simple representation and is convenient to producefrom an MS Access environment. Produced with Visio Professional, the diagramsinclude standard attribute notation to identify the relationships between entities (Figure3.2). The notation (FK) s used to indicate a "foreign key" (identifies the primary keyattribute of another entity wntributed by a relationship), while (IE) denotes "inversionentry" (a non-unique access identifier for an entity).

attribute 1-1 (IE)atir'bute 1-2 prmw key 2 (Fma t e ibu ie 1-3 a t i i t eA- 1

O b i a t 2

attnite2- (IE)amite 2-2mite 2-3

one-wmany relationshipFigure 3 2 :Legend for D E F l X diagrams.

Refemng to Figure 3.3, each of the information objects is represented as a table (box)with the primary key listed first (e-g., project object id). The ines that cm~ectach boxrepresent the one-to-many relationships between tables with the nodes representing themany. Foreign keys denote a link

Computer Assisted Construction Plan

Documents

Transcript of Computer Assisted Construction Plan