Dynamic Change Management for Fast-Tracking Construction Projects

7/27/2019 Dynamic Change Management for Fast-Tracking Construction Projects

1/9


2/9

2

construction sequence by triggeringsubsequent changes on other tasks, which oftencontributes to unanticipated schedule delaysand cost overruns in fast-tracking construction.For these reasons, to effectively handle fast-tracking change iterations involved in fast-tracking need to be identified, and the dynamic

behavior of construction resulting from thosechange iterations must be dealt with in asystematic manner.

2. CONSTRUCTION CHANGES

Non-value adding iterations in construction aremainly associated with construction changes.Accordingly, reducing wasteful constructioniterations requires effective changemanagement, which should start with theunderstanding of different characteristics andbehavior types of construction changes.

2.1 Types of Changes

Normally, construction changes refer to workstate, processes, or methods that deviate fromthe original construction plan or specification.They usually result from work quality, workconditions or scope changes. Meanwhile,changes that have been already made (denotedas Changes as Result in Figure 1) can be thesource of subsequent changes in other tasks

(denoted as Changes as Source in Figure 1).For example, changes in the design work thathave been made by mistake can causesubsequent changes in construction. In thiscase, the design changes are a result to thedesigner, while they can be a need for changesto the construction crew. In addition, changecan be also seen as an action of making achange (denoted as Change as Behavior inFigure 1), which is further categorized intounintended change and managerial change.Unintended changes occur without theintervention of managerial actions. The arrows

labeled E, F, and G in Figure 1 illustrate theunintended change process. Meanwhile,managerial changes are made by managerialdecisions during quality management orproject monitoring and control. As illustratedin Figure 1, once changes occur duringconstruction (A and B), changes result in eithersubsequent changes (C) or rework (D),depending on managerial decisions.

2.2 Differentiating from Rework

Both change and rework are done in the formof either adding, deleting or replacement(deleting and adding). However, given thesame problem, they have different behaviorpatterns, since change and rework have

different characteristics, as summarized inTable 1. For example, in Case I on Figure 2,given the problem (a hump on the concretesurface), rework would be done by deleting theproblem, while change would be done byadding some more concrete. In addition, inCase IV where floor tiling has been finishedwith less than the required height, althoughboth change and rework have the samebehavior pattern (replacement) in solving theproblem, the object would be the problem areain rework, while the previous work would bethe object in change.

2.3 Tradeoffs

In construction, the change option is moregeneral. Since construction has a physicalmanifestation, construction rework is usuallyaccompanied with the demolition of what havebeen already built, which normally has abigger direct impact on the constructionperformance than the change option. Byadopting the change option, it is possible toavoid rework on problematic tasks that may

require more resources. However, aspreviously discussed, changed tasks can alsobecome a change source that can cause othersubsequent changes, which might have moreimpact on the construction performance thanthe rework option in certain conditions. Forexample, the increased concrete height in CaseI and Case III on Figure 2 may triggersubsequent changes in succeeding tasks, i.e.,reducing the size of ventilation ducts. Inaddition, in Case V on Figure 2 where some ofpiles have not been correctly positioned, it maybe possible to proceed with the superstructure

without correcting the position of the piles bychanging the position of columns. However,this change option may necessitate unplannedcantilever construction in order to keep theoriginal floor layout, which needs to beevaluated as compared to re-driving the piles.Consequently, a decision on the change optionneeds to be carefully made based on a goodunderstanding of how changes evolve to non-value adding iterations, which can create


3/9

3

unanticipated and indirect side effects of thedecision.

3. DYNAMIC PROJECT MODEL

The dynamic project model to be presented hasbeen developed taking into consideration

effective change management and operationlevel construction policy making. To developthe model feedback processes involved in fast-tracking construction were identified focusingon how they can trigger non-value addingiterations in the form of construction changes.Having identified feedback processes, thegeneric construction process, which constitutesthe skeleton of the project model, was modeled.

3.1 Feedback Processes in Construction

Normally, construction involves feedback

processes represented in the causal loopdiagram on Figure 3-a, 3-b, 3-c, and 3-d. Whentasks and resources are available, first, theupstream work, based on which the availabletasks will be carried out, is reviewed beforecommissioning resources for the tasks. Duringthe review process, problems made in theupstream work can be discovered. Once theyare found, depending on managerial decisions,workers may request the upstream worker tocorrect the problematic work. More upstreamhidden changes can cause more requests forthe upstream work reprocess, which results in

more pending tasks (A) and schedule delays (B)in the downstream work. Otherwise, workersconstruct tasks not having problems in theassociated upstream tasks, with givenresources. Once tasks are completed, theconstruction performance on the tasks isperiodically monitored or inspected to seewhether or not the target quality is met and theintended functions are achieved. Through thisquality management process, a decision onwhether releasing the completed tasks or notcan be made.

Unintended changes resulting from low workquality, bad work conditions or frequent scopechanges can cause managerial changes (C),rework (D), or hidden changes (E), dependingon managers willingness to adopt the changeoption and quality management thoroughness.The more construction is delayed the moreoften the change option tends to be adopted (F),in order to avoid rework, which is normally

perceived to have a bigger impact on theschedule performance. However, suchmanagerial efforts can create unplanned and/orindirect side effects. As a result of feedbacksinvolved in the processes (F, G, H, I, J),managerial changes can trigger further delaysas well as rework. As diagramed in Figure 3-a,

managerial changes trigger reprocess iterationsof tasks that have been already released (referto the definition of managerial changes inTable 1), while rework delays the constructionprogress by creating reprocess iterations oftasks that have not been released.

In addition, delays also may make qualitymanagement efforts less thorough (K), whichresults in more hidden changes (L). During thedownstream review process, hidden changesreleased from the upstream work can bediscovered. Once they are found, depending onmanagerial decisions, downstream workersrequest the upstream worker to correct thehidden changes. As a result, more hiddenchanges can cause more correction requestsfrom the downstream (M), which also candelay the construction progress as a result ofsubsequent feedback processes (N, I, J)diagramed in Figure 3-b.

Furthermore, increased willingness to adoptmanagerial changes also can increasesubsequent changes in the downstream work

(O), which delays the downstream workprocess. Consequently, reprocess requestsfrom the downstream work are also delayed(R), which again impacts the scheduleperformance of the activity that has originatedchanges (N, I, J). Meanwhile, lowered qualitymanagement thoroughness creates morehidden changes (L). Increased hidden changescan deteriorate the work quality of thedownstream work, which creates morereprocess iterations of the downstream tasks.This also impacts the upstream scheduleperformance through (R,N, I, J). All of these

feedback processes can impact the constructionperformance, combined with resourceavailability, construction policies, and peoplereactions to work conditions and policies.

3.2 Model Description

Based on feedback processes and relationshipsamong construction variables in the causalloop diagrams on Figure 3, the quantitative


4/9

4

representation of the generic constructionprocess has been modeled using systemdynamics modeling techniques. In addition,other supporting model structures for resources,scopes, and quality have been also developed.Detailed model descriptions are found in Park(2001).

4. CASE STUDIES

The developed dynamic project model is beingapplied to the construction of 27 bridges inorder to help effectively manage changes andprepare a robust construction plan. Theconstruction is a part of a $400 millionDesign/Build/Operate/Transfer project forroadway improvements along State Route 3from its intersection with State Route 128 inBurlington, MA north to its terminus at theNew Hampshire border. The developmentprocess is expected to span 42 months with theproject completion achieved in February, 2004.The project scope includes widening the 21-mile of the state roadway and the existing 15underpass bridges, and renovating 12 overpassbridges. This paper presents a case study of theTreble Cove Road Bridge Construction, one ofbridge renovation projects, demonstrating howthe case project has suffered from changes andproviding construction policies to minimizechange impact on the project performanceincluding labor policies and schedule buffering.

4.1 Simulating the Actual Performance

The simulated actual duration is 559 workingdays. This is 168 days longer than the CPM-based duration of the base case, which is 391working days. The difference in thecompletion time is mainly caused by a lot ofnon-value adding iterations among design andconstruction activities. Actually, theconstruction team is working to address suchissues that the design development of the

Treble Cove Road Bridge project was alreadyshown significant delay and construction hasnot been yet started. Some of these issues aredue to the fact that this project was awarded tothe contractor before the detailed scope of theproject has been established. As a result,changes on the design work were frequentlyrequested from the owner side during sketchplan, final plan, and shop drawing submittal,which resulted in a lot of design iterations. In

addition, this case project was the firstdesign/build contract for the members ofdevelopment team in the owner side, expectedlevel of coordination among the owner,designer and constructor has not been met todate and design iterations encountered weredifficult to handle. Based on interviews with

the design and construction team, thesechallenges in the design development wererepresented as Highly Unreliable in theproject model and the simulated actualdurations for those activities show how muchnon-value adding iterations caused by changescan affect the project progress in a quantitativemanner.

4.2 Policy Implications

In order to examine the effectiveness of

different construction policies, simulationswere done adapting the actual case withdifferent scenarios for managerial decisions onchange or rework, labor control, buffering, andsome important time variables. As a result ofthe simulations, the following policyimplications were obtained (refer to Figure 4 tosee the model simulation).

First, a higher managerial change ratio tendedto reduce costs but lengthen the projectduration. However, it is hard to generalize thisresult, since the tradeoff of change and rework

is highly dependent on construction systemconditions at the time when a decision is made.This implies that effective change managementrequires an operational level approach ratherthan a long term policy, and it should beaccompanied with well preparation ofrelatively long-term policies such as laborcontrol policies, schedule buffering anddelivery methods.

In connection with labor policies, flexiblelabor control was found to be effective for the

case project in terms of schedule and costreduction. In contrast, overtime contributed tofacilitating the project schedule to some extentbut its effectiveness is questioned, onceincreased project costs are considered.Overtime applied for the case project loweredproductivity and increased change rate, asworkers fatigue was accumulated. In fact, theeffectiveness of labor control policies can varydepending on the nature of a project. However,


5/9

5

many success stories of concurrentconstruction projects like our case projectconfirm the above policy implications,demonstrating that having flexibility in laborcontrol contributes to reducing the projectduration and costs by assigning workforce in atimely manner.

In addition, the case project has been simulatedwith various buffering scenarios; not havingbuffer, having uniform buffer, and havingbuffer based on activities characteristics. Thesimulation results showed that applied bufferscontributes to reducing the upstream changeimpact and non-value adding iterations. As aresult, the resource idle time and waste werereduced, which made it possible to moreeffectively utilize given workforce. Inparticular, buffering based on activitiescharacteristics turned out to have mosteffectively enhanced the schedule and costperformance.

Lastly, the simulations done with differenttime variable scenarios demonstrate thatshortening a required time for labor hiring andRFI reply contributes to enhancing the projectschedule and cost performance. In particular,RFI reply time greatly affected the projectperformance. Shortening RFI reply time byhalf could facilitate the project progress by12% and reduce the project costs by 10%. Incontrast, when RFI reply time was doubled,duration and costs were increased by 29% and24% respectively. These simulation resultsimply that for this case project, coordinationamong the project functions is crucial to thesuccess of the project. Consequently, thedecision-making process in design andconstruction should be shortened andinformation flow among project functionsshould be streamlined to assist in reducing thedecision-making time.

In conclusion, although the obtainedsimulation results can vary depending onproject settings, they well demonstrate how thedynamic project model can contribute toenhancing the project performance in a realworld setting by providing effective changemanagement plans and policy guidelines.Additionally, the simulation results also imply

that model-based construction policies can bemore effective, when combined with othermanagerial efforts such as reducing a processtime and increasing the level of coordinationamong project functions.

5. CONCLUSIONS

Construction involves a lot of non value-adding change iterations due to its structuralproblems, in particular when construction isperformed concurrently. This has necessitatedthe development of a tool that can effectivelymanage construction changes. This paperaddressed the challenging issue by introducingthe concept of dynamic change management toconstruction planning and management.Although the research results discussed thusfar need to be further refined and developed,

they demonstrated that the dynamic changemanagement approach and the developedproject model would help prepare a morerobust construction plan against uncertaintiesand provide policy guidelines, by taking intoconsideration the context in which aconstruction project is being developed.

6. REFERENCES

Fazio, P., Moselhi, O., Theberge, P. andRevay, S. (1988), Design Impact ofConstruction Fast-Track, ConstructionManagement and Economics, Vol. 6, No. 2,pp 195-208

Ford, D and Sterman, J. (1997), DynamicModeling of Product Development Processes,

Sloan School of Management, Working Paper3943-97, MIT, Cambridge, MA

Huovila, P., Koskela, L., and Lautanala, M.(1994), Fast or Concurrent: The Art ofGetting Construction Improved, Proceedings,pp 143-158, The second workshop on lean

construction, Santiago

Kwak, S. (1995), Policy Analysis of HanfordTank Farm Operations with System Dynamics

Approach, Doctoral Thesis, Dept. of NuclearEngineering, MIT, Cambridge, MA

Lyneis, J. (1999), Dynamics of ProjectPerformance, Course Material, Dept. of Civiland Envr. Eng. at MIT


6/9

Park, M. (2001), "Dynamic Planning and

Control Methodology for Large-Scale

Concurrent Construction Projects",

Doctoral Thesis, Dept. of Civil

Engineering, MIT, Cambridge, MA

Pena-Mora, F and Park, M. (2001),

"Robust Control of Cost Impact on Fast-

tracking Building Construction Projects',

Journal of Construction Engineering and

Management, ASCE

Tighe, J (1991), Benefits of Fast Tracking

are a Myth, International Journal of

Project Management, Vol. 9, No. 1, pp

49-51

Williams, G.V. (1995), Fast-Track Prosand Cons: Considerations for Industrial

Projects, Journal of Management in

Engineering, Vol. 11 No 5, pp 24-32,

Sep/Oct. 1995

Figure 1: Changes as Iteration Trigger


7/9

Figure 2: Behaviors of Change and Rework

Figure 3a: Change Option Loop


8/9

Figure 3b: Quality ManagementThoroughness Loop

Figure 3c: Downstream

Reprocess Iteration Loop


9/9

Documents

Dynamic Change Management for Fast-Tracking Construction Projects