A CASE STUDY OF LINE-OF-BALANCE BASEDSCHEDULE PLANNING AND CONTROL SYSTEM

Olli Seppänen1, Erno Aalto2

ABSTRACT

Line-of-Balance is a graphical technique which can be used to plan and manage work flow. It is suit-able for construction projects because of their large degree of repetition. Despite its strengths Line-of-Balance has not gained widespread use in construction industry internationally. However, it has beenused as the principal scheduling tool in Finland since 1980s. As a result of two decades of research anduse in industry, a comprehensive schedule planning and control system has been developed around lo-cation-based techniques. A computer software has facilitated implementation in constructioncompanies.

This paper describes a case study of 15,000 m2 office building project using location-basedmethods for schedule planning and control. Master schedule is based on Bill of Quantities wherequantities have been calculated based on the project’s Location Breakdown Structure. Beforeimplementation, different schedule alternatives were evaluated based on risk of interference, totalduration and cost. During construction, the master schedule sets constraints on lower level taskschedules, which were used to ensure the continuity of work for crews on a more detailed level. Themaster schedule was not updated even when there were deviations from the original plan. Instead thetask plans were updated to catch up with the original schedule. This prevented the problems fromaccumulating in downstream production.

In this case study the combination of PPC measurement (calculating the percentage of weeklyassignments complete) and task planning was piloted. Weekly plans were made by combiningassignments from all the task schedules. If the starting constraints had not been removed the task planwas updated to assess the effect on total production and to plan control actions. PPC measurement wasfound to improve task plan reliability.

Benefits of the approach included better schedule control and possibility to examine how deviationsfrom weekly plans affected the total schedule. Task planning provides information about how long amaster schedule task actually reserves a location. This information can be used in planning masterschedules of similar projects in future.

KEY WORDS

Line-of-Balance, Scheduling, Task planning, Production control

INTRODUCTION

BACKGROUND

The graphical line-of-balance scheduling methodis a planning method for continuous flow of pro-duction. Advantages of line-of-balance schedul-ing for the General Contractor include less

schedule risk because subcontractors can be kepton site, productivity benefits because the crewsare less likely to interfere with each other andmore realistic schedules as buffers can be easilyplanned and analysed (Kankainen & Seppänen2003). Despite its strengths, the method has notgained widespread use internationally. Accordingto the technical literature, lack of easy-to-use soft-

Flowline

271

1 PhD Student, Department of Civil and Environmental Engineering, Helsinki University of Technology,Technology Manager Dynamic System Solutions Ltd., Pohjoinen Rautatiekatu 29B, FIN-00100 Helsinki,Finland, E-Mail olli.seppanen@dss.fi, website www.dss.fi

2 MSc, Project Engineer, NCC Construction Ltd. Mannerheimintie 103A, FIN-00280 Helsinki, Finland,E-Mail: erno.aalto@ncc.fi, website: www.ncc.fi

ware solution implementing the method has beenone of the reasons for slow adoption. (Arditi,Tokdemir & Suh 2002).

In Finland, location-based planning methodshave been used widely in construction since1980s. The methods were brought to Finland andadapted to commercial construction by professorsKankainen and Kiiras from Helsinki University ofTechnology (e.g. Kiiras 1989, Kankainen &Sandvik 1993). In academic research tests it wasestablished that the use of modified flowline plan-ning increased productivity and decreased waitinghours for own workers and for subcontractors(e.g. Toikkanen 1989, Venermo 1992).

New research efforts to improve the schedulingskills of the Finnish industry were started in theend of 1990s by professor Kankainen’s researchgroup. The results included tools such as taskplanning (Junnonen 1998), project control charts,checklists to assess schedule’s feasibility(Kankainen & Kolhonen 2005) and new contractsto support location-based control. The researchresults were used in a software development pro-ject to design a new software able to be used as aplanning and control tool. The features of theresulting software DynaProject™ have beendescribed in Kankainen & Seppänen (2003).

As a result of two decades of research, a com-plete schedule planning and controlling method-ology based on managing schedule risk has beendeveloped. Because of the shared aims with LeanConstruction, namely reduced waste and interfer-ence, Finnish results and tools have been pre-sented in IGLC conferences during the last fewyears (Kankainen & Seppänen 2003, Junnonen &Seppänen 2004, Soini & Leskelä & Seppänen2004, Seppänen & Kankainen 2004). This paperdescribes a case study using the risk-managementbased planning and control methodology of usinglocation-based tools to give a better idea abouthow the system works and how it can contribute toLean Construction knowledge. In addition tousing the location-based tools, the case studyexperiments with measuring reliability of taskschedules by use of PPC measure from Last Plan-ner™ system of production control (Ballard2000).

PREVIOUS RESEARCH AND PRACTICE

Earlier papers about Line-of-Balance describealgorithms and heuristics to optimize continuityof resource use, to achieve duration cuts and tomodel learning curves (e.g Kang et al 2001, Arditiet al 2001, Harris & Ioannou 1998). Researchmethods have been based mostly on simulation,not on real case studies or empirical research ofprojects. While this previous research gives good

theoretical foundation to the method, it doesn’tdescribe utilization of these concepts in real pro-jects.

However, experiences from using Line-of-Bal-ance in real projects indicate that most of the ben-efits are currently lost in implementation stage(Seppänen & Kankainen 2004). This is because ofvariability of production rates and inadequatecontrol mechanisms on site. Variability has beenwidely discussed in Lean Construction literature,for example by Tommelein, Riley and Howell(1998) in their Parade of Trades simulation. TheLast Planner System™ has been proposed toreduce variability by ensuring quality assign-ments, enabling learning during projects andmaking and committing to decisions near imple-mentation time and by the people responsible forthe work (e.g Ballard 2000). The Last PlannerSystem can be criticized because it doesn’t takeinto account production rates and it is easier tocontinue already begun work than begin workagain after interruption. Thus the Last Planner canachieve good results on one week but simulta-neously destroy prerequisites for continuing onsucceeding weeks. (Seppänen & Kankainen2004).

The objective of this paper is to show that thebenefits of the Line-of-Balance scheduling can beachieved in practice by using location-based pro-duction controlling tools. The research method iscase study with minimal researcher intervention(2 to 4 hours a week just to gather the data). To beable to compare the results with Last Planner casestudies also PPC (percentage of weekly assign-ments completed) was measured.

CASE STUDY

DESCRIPTION OF THE PROJECT

Opus business park is a 14 500 m2 office buildingin eastern Helsinki. It is composed of two sec-tions, which can be built independently of eachother and of parking hall below the main building.Both sections have six floors. The total schedule isfrom May 2004 to December 2005. Figure 1shows summary of the project. The project is thesecond part of larger three -part Opus BusinessPark development.

AVAILABLE STARTING DATA

NCC Construction has devoted a lot of resourcesto implement flowline based production controlsystems (Soini et al. 2004). The quantity take-offis done corresponding to the physical locations ofthe building. Labor consumption information hasalso been standardized within the company allow-

Proceedings IGLC-13, July 2005, Sydney, Australia

272 A Case study of line-of-balance based schedule planning and control system

ing for a very fast planning of first drafts of theschedule. Also the building services quantities areestimated based on project characteristics andsize. The productivity and quantity databasesinclude information about subcontracted work.The main principle is that subcontracted workshould be planned as if it were done with ownresources because otherwise effective control ofproduction flow is impossible.

In this case study, all the quantities had beendistributed to sections and floors so that theycould be directly utilized in flowline planning.

SCHEDULING PROCESS

Because of good starting data, it was possible tocreate many different alternative schedules in ashort period of time. Two main alternatives wereexamined: 1) completely continuous schedule and2) work continuous in sections but a breakbetween two sections.

Completely continuous schedule would havehad the same end date as partially continuousschedule but both sections would have been fin-ished at approximately the same time. Partiallycontinuous schedule achieved much of the samebenefit but enabled the first section to be finishedearlier thus reducing the risk of exceeding thetotal duration. The project team decided to imple-ment the partially continuous alternative and takethe break between sections into account incontracts with subcontractors.

It was not possible to change the sequence ofsections because the parking hall had to be handedover before the second section could be started.This was because the second section was used as atemporary parking lot for customers of the neigh-

boring supermarket. If the second section couldhave been built first, the project duration wouldhave decreased by one month.

In the final schedule the production rates havebeen synchronized and buffers have been plannedbetween the most important activities. All taskdurations are based on quantities, resources andproductivity data from earlier projects or fromFinnish productivity database, which has beencreated as a joint effort of the industry (Olenius etal. 2000). The final master schedule is shown infigure 2. The project has been divided into threesections. Parking hall deck has to be completedbefore earthworks of module lines 27-20 couldbegin. Otherwise the sections are independent ofeach other so they could be built in any sequenceor simultaneously. The schedule has been plannedso that the structure is built from down up to sec-tion in module lines 30-27 first and then the sameresources continue to the larger section 27-20.Finishes have been planned to be continuouswithin the section. When the same subcontractorwas used in the both sections the break of onemonth before the work could continue in the nextsection was taken into account in the call for ten-ders and the actual contract. There was a contrac-tual milestone after the finish date of the firstsection for all major subcontractors.

Risk analysis of the schedule indicated that thestructure was the riskiest part of the project. As aconsequence, buffers were added between struc-ture and finishes. For example, there is a lag ofthree floors before concrete floor finishing worksstarts in the building.

Flowline

Olli Seppänen, Erno Aalto 273

Figure 1: Summary information about the Opus Business Park development

WEEKLY CONTROLLING PROCESS

Task scheduling

The weekly control process used throughout theproject was based on task planning method(Junnonen & Seppänen 2004). The schedule ofupcoming master schedule tasks was explodedinto more accurate level and the quantities wereupdated. This process started in the beginning ofthe project so by planning just one task accuratelyeach week it was possible to always be well aheadof production. While the master schedule looks atproduction flow and required production rates at“macro” level, the task schedules are used to plancontinuous work for each crew. Task schedulesare constrained by the master schedule so that thetask schedule must finish all subactivities in alocation before the next master schedule taskbegins in that location. Task schedules areupdated weekly to always correspond with thecurrent situation but the master schedule is neverupdated. This is because updating the masterschedule may lead the site management into falsesense of safety. This is because updating theschedule shifts the problems towards the end ofproject and leads to hurry in the end of project(Kankainen & Kolhonen 2005). Updating the taskschedule may have the same effect but becausethe master schedule sets the boundaries in which itcan be updated, the effect is more localized and itis possible to react earlier.

Example of task scheduling: plasterboardwalls

“Plasterboard walls” task illustrates the use oftask scheduling and how the buffers were actuallyutilized in the project. Figure 3 shows part of themaster schedule of the first section. The originalintention was to build plasterboard walls onlyafter the roof was completely waterproof so thatthe risk of walls getting wet because of rain wouldbe removed. However, because of delay of thestructure and the roof, the walls would havestarted at least 4 weeks late with the originalconstraint.

The problem was solved in task scheduling byexploding the master schedule activity to threesubtasks: wall frames and installing board on oneside, electrical piping inside walls and insulationand installing board on the other side (figure 4).The sequence was changed so that the first floorwhich was more complex (auditorium, dentist andother special spaces) than the other floors wasdone last. Contract was made with one subcon-tractor to install both the bulkheads and the plas-terboard walls. One crew went through all thelocations installing bulkheads. The next crewstarted a bit later and followed in the samesequence installing wall frames and the board onone side. When the bulkheads crew had gonethroughout the building, they started to install theinsulation wool and second board. Buffers were

Proceedings IGLC-13, July 2005, Sydney, Australia

274 A Case study of line-of-balance based schedule planning and control system

Figure 2: Master schedule of the Opus project—For clarity, only the construction and building services tasks whichreserve the whole location are shown

planned between the first board and the secondbecause different subcontractor was involved ininstalling the electrical piping. Even though theplasterboard walls task ends two weeks later thanin the master schedule, the succeeding tradesincrementally catch up the delay. Tiling work and

vinyl floor covering work can start according tothe original master schedule.

The example above also illustrates the fact thatbuffers are needed if the master schedule isplanned on rough level of detail. They give flexi-bility in task scheduling phase. In this example,the subtasks flowed through all the buildinginstead of finishing all the subtasks first in a loca-

Flowline

Olli Seppänen, Erno Aalto 275

Figure 3: Original master schedule plan and actuals for plasterboard walls task and its predecessors and successors.Solid lines indicate the original plan, dotted lines the actual

Figure 4: The task schedule for plasterboard walls task and its predecessors and successors

tion before moving on to the next location. Thelocation is reserved for each trade longer than inthe master schedule.

Collecting actual data from the site by usingcontrol charts

Every week on Tuesday, the actuals from the lastweek were compared with the weekly assign-ments derived from all of the task schedules. Thedata was collected from site by using the controlchart, a matrix of locations and tasks which showswith color codes the status of each location(Kankainen & Seppänen 2003). The control chartwas prepared for each task schedule of the project.It was used to communicate the schedule status tosubcontractors and the management team in sub-contractor meetings every Wednesday.

Figure 5 shows the control chart of the taskschedule presented above. Each cell has fournumbers and a color code. The top left numbershows the planned calendar week when the taskshould begin in a location. The top right numberindicates the planned finish week. The bottom leftnumber indicates the actual start week and thebottom right number actual finish week or thecompletion rate if the task has begun in the loca-tion but hasn’t been finished yet. Green colormeans that the task has been completed, yellowcolor that the task has begun but is running late,red color that the task hasn’t been begun and isrunning late and the blue color that the task hasbegun and is on time. In this example, the subcon-tractor used the same men to install bulkheads and

wall frames which slowed down the progress.They also had problems in completely finishing alocation before moving on to the next location.This can be seen immediately from the controlchart because the second and third floors of bulk-heads task are yellow. The impact on the electricaltrade is also readily apparent by the red square onthe second floor of the electrical piping task.

Measuring reliability of task schedules byusing PPC

The reliability of task plans was measured by cal-culating the percentage of planned assignmentscompleted (PPC) during the week. This measureis the same as in the Last Planner™ system of pro-duction control (e.g Ballard 2000) but the assign-ments are the result of Line of Balance -based taskscheduling process. PPC was used in this researchin order to calculate a baseline reliability for justusing location-based techniques without actuallyusing the Last Planner™ system. This wouldallow comparison between the two systems andshow the benefits of possible combination.

Calculating reliability of task schedules is anew concept in Finnish projects. Earlier just thecontrol chart and plotting actuals to Line-of-Bal-ance charts or status line to Gantt chart have beenused to measure success of the control system.Some evidence was found that the PPC measure-ment improved the results. The PPC increasedfrom the base 50 % to about 70 %. However, thePPC couldn’t be consistently held at the 70 %level and the average PPC for the whole projectwas 63 %. This is probably because the subcon-

Proceedings IGLC-13, July 2005, Sydney, Australia

276 A Case study of line-of-balance based schedule planning and control system

Task status / Task scheduling exampleNCC Rakennus Oy OPUS 3version 9.4.2005 19:22Foreperson: Mäkitalo Jukka Planner: Aalto Erno

Control modeRegistered: Olli Seppänen / DSS Ltd.DYNAProject v2.0.07

Key The task has not begunThe task is in progress

Task completedThere is no task for this position

The task is in progress but running lateThe task has not begun, running late

On time and pausedLate and paused

1 2 3 4 5 6 7 8 9 10 11

51 522 350 5152 248 5051 5247 4848 5045 4746 4843 4543 46

4 463 42 62 31 452 251 351 5248 250 5148 51

52 42 64.3%

5 66 25%4 55 45%3 43 90%2 32 66 7

6 7

5 66 0%4 54 63 42 57 8

7 8

7 7

6 7

5 6

8 9

9 9

8 9

7 8

7 7

9 10

10 10

9 10

8 9

8 8

10 11

12 13

11 12

10 11

9 10

13 15

15 15

14 15

13 14

11 13

10 11

18 19

17 18

16 17

15 16

14 15

1 2 3 4 5 6 7 8 9 10 11

+S

truc

ture

mod

30-2

7

+C

oncr

ete

floor

finis

hing

wor

k1

+R

oofin

g

Bui

ldin

gse

rvic

esco

rrid

orbu

lkhe

ads

Wal

l fra

mes

and

boar

don

one

side

Ele

ctric

alpi

ping

ofw

alls

Insu

latio

nw

ool a

ndbo

ard

on2n

dsi

de

Pla

ster

wor

k

Pai

ntin

g

Tili

ngw

ork

1

Vin

ylflo

orco

verin

g

1 2 3 4 5 6 7 8 9 10 11

Projekti Section Floor

OPUS 3 MOD30-27

Roof

5

4

3

2

1

Figure 5: Control chart shows graphically the schedule status of each task and location

tractors committed to the task schedules but not toweekly targets based on the task schedules. ThePPC measurement wasn’t exposed to other peoplethan the authors. Figure 6 shows the PPC asfunction of time.

Bad PPC weeks usually had one or two criticaltasks going with lower production rate thanplanned or in wrong sequence, which alsoimpacted all the following trades causing every-one to fail in their assignments. In addition theimpact of some holiday seasons (Christmas week53-1, Easter weeks 12-13) was inadequatelyreflected in weekly plans (Estonian subcontrac-tors had more days off than anticipated). Devia-tions were mainly concentrated on a few trades(bulkheads / plasterboard walls and electricaltrade). Bulkheads / plasterboard walls subcontrac-tor was an Estonian company with another largejob in Finland which caused excessive resourcedemands. The subcontractor wasn’t familiar with

the Line-of-Balance technique and tended to workin wrong location or in multiple locations at thesame time. Most of the deviations of electricaltrade were caused by long lead times of switch-boards and resulting material problems. The devi-ations in the end of the observation period (weeks21-24) were caused by resource problems becausethe required production rate in the second sectionwas higher than in the first section and the subcon-tractors continued with the same amount ofresources.

Updating task schedules

All existing task schedules were updated next totake changed circumstances and actual productionrates into account. Actual production rates wereused in task schedule updates to make schedulesmore accurate and to show if more resources wereneeded. The master schedule sets the boundaries

Flowline

Olli Seppänen, Erno Aalto 277

Figure 6: Percentage of Plan completed as function of time (calendar week)

Figure 7: Actual situation (dotted lines) compared to the master schedule (solid lines). Forecasts based on the currenttask schedules are shown with dashed line.

for planning so the problems couldn’t be pushedfarther than the end of master schedule task in alocation. The aim was to make the best possibleforecasts for the rest of the task while preservingcontinuous production for as many workers aspossible. Most effort was expended on updatingthe next week’s plan because the reliability of thenext week’s task schedules was measured by PPC.

If there was sufficient information on a masterschedule task which was about to begin in the nextfew weeks, a new task schedule was planned. Thefirst draft of the task schedule used accurate quan-tities taken from current drawings and estimatedproduction rates. It was planned by the authors.Before beginning of production, the task schedulewent through multiple rounds of comments by thesubcontractor during contract negotiations, thesuperintendent and procurement people. Thecomments were used to refine the task scheduleuntil everyone was ready to commit to it. The fin-ished task schedule was included as part of thesubcontract.

The resulting set of task schedules were up-to-date, took into account the availability ofresources and were based on actual circum-stances. From these task schedules the productionobjectives for the next week were established.These objectives were communicated to the sub-contractors and foremen. Their success was eval-uated in the next Tuesday’s schedule update bycalculating the PPC.

Comparing actuals to the original masterschedule

After updating the task schedules the status ofmaster schedule was evaluated based on com-puter-calculated forecasts (Seppänen &Kankainen 2004). If the delay of a task endan-gered the continuous flow of another task, controlactions were planned by updating the task sched-ules to minimize the risk of interference. Actualsituation and resource availability of the subcon-tractors as well as the cost effects of accelerationwere evaluated to arrive at the best solution. Ifinterference couldn’t be avoided, task planningwas used to estimate the optimal time to continueproduction for the disturbed trade. Figure 7 showsa few activities with the status of calendar week11/2005.

It can be seen from the figure that the scheduleof the first section has been caught up. The firstfloor being late doesn’t matter because there is alot of float in the first section. Each task has hadcontinuous work and the second section can becaught up because the same subcontractors will beused. An example of learning effects is the struc-ture which had problems in the first section but

had the planned production rate for the entireduration of the second section.

The project management was very satisfiedwith the schedule performance of the project. Intheir earlier projects the deviations in the structurewould have affected all the trades until the end ofthe project. Now the original schedule was mostlycaught up before summer holiday season.

CONCLUSIONS

Line-of-Balance based schedule planning andcontrol system in combination with PPC measure-ment worked well. The system allowed the projectteam to see the total effects of deviations com-pared to the original master schedule and simulta-neously measure the reliability of short-term taskschedules. Because of measurement the projectteam was motivated to plan better task schedulesand to implement them better. By combining thesystems, the site can set realistic objectives,commit to them and maintain control of the over-all schedule. The 63 % mean PPC achieved is amoderately good result because optimizing itwasn’t the main purpose of the research and it wasnot exposed to subcontractors or workers.

The main result of the case study was thatweekly task scheduling ensures effectively theimplementation of the original master scheduleand prevents the problems from shifting to the endof project. Just monitoring and controlling themaster schedule isn’t enough because the subcon-tractors should be involved in the planning pro-cess to get their commitment. However, the startdates and production rates planned in the Line-of-Balance master schedule were found to be mostlyrelevant throughout the project so they can beused as basis of procurement. Evidence for impor-tance of continuous production with constant pro-duction rate was found because there weredeviations in the second section which requiredhigher production rate. The actual production ratefor all the trades was equal to the production ratein the first section.

Largest problems in the case study arose fromthe lack of subcontractor resources and overlap-ping production in many locations, instead of fin-ishing one before moving to another. If theproduction overlaps in many locations, the nexttrade can’t begin on time and problems with costeffects occur. Prerequisites of starting (especiallydesign) caused some problems which could havebeen prevented by doing detailed constraint listsand controlling them before starting of produc-tion. However, most of the other constraints arealready explicitly planned in task scheduling (forexample resource constraints, space constraints,procurement needs).

Proceedings IGLC-13, July 2005, Sydney, Australia

278 A Case study of line-of-balance based schedule planning and control system

Because of the clear benefits in terms of pro-duction continuity and predictability of the masterschedule, location-based planning and control canbe recommended. Planning a location-basedschedule isn’t enough without effective controltools. To ensure good implementation, mutualunderstanding of the schedule and effects ofresource decisions by all parties is critical. Forthis purpose, the control chart proved to be themost powerful communication tool of the currentstatus because it was easily understood byeveryone.

In future research, the cost benefits of location-based control should be shown. Other projecttypes should be studied to find out whether theresults generalize to less repetitive project types.

REFERENCES

Arditi, D., Tokdemir, O.B. & Suh K. (2001).“Effect of Learning on Line-of-BalanceScheduling” International Journal of ProjectManagement, Elsevier Science Ltd andIPMA, No. 19. 265–277

Arditi, D., Tokdemir, O.B. and Suh K. (2002).“Challenges in Line-of-Balance Scheduling”Journal of Construction Engineering andManagement, ASCE, New York, NY, Vol.128(6) 545–556

Ballard, G. (2000). The Last Planner™ System ofProduction Control. PhD Dissertation. Schoolof Civil Engineering. University of Birming-ham, U. K.

Harris, R. B. & Ioannou P. G. (1998). “SchedulingProjects with Repeating Activities”. Journalof Construction Engineering and Manage-ment. ASCE, New York, NY, 124(4), 269-278

Junnonen, J-M. (1998). Tehtäväsuunnittelu jalaatupiiriohjattu tuotannonohjaus. (Taskplanning and quality circles in productioncontrol). Licentiate Thesis, Construction Eco-nomics and Management, Department ofCivil and Environmental Engineering, Hel-sinki University of Technology. Espoo,Finland.

Junnonen, J-M. & Seppänen, O. (2004). “Taskplanning as a part of production control”. Pro-ceedings of the 12th international conferenceof Lean Construction (IGLC12). Elsinore,Denmark

Kang, L. S., Park, I. C., Lee B. H. (2001). ”Opti-mal Schedule Planning for Multiple, Repeti-tive Construction Process”. Journal ofConstruction Engineering and Management,ASCE, New York, NY, 127(5). 382–390

Kankainen, J. & Sandvik, T. (1993).Rakennushankkeen ohjaus. (Controlling aconstruction project). Confederation of Finn-

ish Construction Industries, RakennustietoOy, Helsinki. Finland.

Kankainen, J. and Kolhonen, R. (2005).Rakennushankkeen aikataulusuunnittelu javalvonta. (The schedule planning and controlof a construction project). Rakennustieto Oy.Helsinki, Finland.

Kankainen, J. & Seppänen, O. (2003). “A Line-of-Balance based schedule planning and con-trol system”. Proceedings of the 11th interna-tional conference of Lean Construction(IGLC11). Blacksburg, Virginia.

Kiiras, J. (1989). OPAS ja TURVA,Erityiskohteiden työnaikaista ohjaustapalveleva aikataulu- ja resurssisuunnittelu.(A schedule and resource planning system forthe implementation phase of special projects).Helsinki University of Technology. Construc-tion Economics and ManagementPublications.

Olenius, A., Koskenvesa, A. and Mäki, T. (edi-tors) (2000). Ratu aikataulukirja 2001.Conferederation of Finnish ConstructionIndustries RT and Rakennustietosäätiö.Tampere, Finland.

Seppänen, O. & Kankainen, J. (2004). “An empir-ical research on deviations in production andcurrent state of project control.” Proceedingsof the 12th international conference of LeanConstruction (IGLC12). Elsinore, Denmark.

Soini, M., Leskelä, I. & Seppänen, O. (2004).“Implementation of Line-of-Balance basedscheduling and project control system in alarge construction company”. Proceedings ofthe 12th international conference of Lean Con-struction (IGLC12). Elsinore, Denmark

Toikkanen, A. (1989). Erityiskohteen aikataulunlaadinta ja työnaikainen ohjaus. (Schedulingand control of a Special Construction Project).Master’s Thesis, Construction Economics andManagement, Department of Civil and Envi-ronmental Engineering, Helsinki Universityof Technology, Espoo, Finland. 109 pp.

Tommelein, I.D., Riley, D., and Howell, H.(1998). “Parade game: Impact of work flowvariability on succeeding trade performance.”Proceedings of the 6th international confer-ence of Lean Construction (IGLC6), Guarujá,São Paulo, Brazil

Venermo, T. (1992). Pienen erityiskohteentuotannon suunnittelu ja ohjaus. (Control andplanning of a small special constructionobject). Master’s Thesis, Construction Eco-nomics and Management, Department ofCivil and Environmental Engineering, Hel-sinki University of Technology, Espoo,Finland. 119 pp.

Flowline

Olli Seppänen, Erno Aalto 279