See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/309384849

Removing Constraints to Make Tasks Ready in

Weekly Work Planning

Conference Paper · June 2016

DOI: 10.13140/RG.2.2.21789.10729

CITATIONS READS

0 68

3 authors:

Farook Hamzeh Emile Zankoul

American University of Beirut American University of Beirut
54 PUBLICATIONS 193 CITATIONS 11 PUBLICATIONS 9 CITATIONS

SEE PROFILE SEE PROFILE

Fatima El Sakka
American University of Beirut
3 PUBLICATIONS 0 CITATIONS

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Advances in Engineering Education View project

Organizational Issues and behavior View project

All content following this page was uploaded by Emile Zankoul on 23 October 2016.

The user has requested enhancement of the downloaded file. All in-text references underlined in blue are added to the original document
and are linked to publications on ResearchGate, letting you access and read them immediately.
 Creative Construction Conference 2016

Removing Constraints to Make Tasks Ready

in Weekly Work Planning
Farook Hamzeh*, Emile Zankoul, and Fatima El Sakka
Assistant Professor, Civil and Environmental Engineering Department, American University of Beirut, Beirut Riad El-Solh 1107 2020, Lebanon
Masters, Civil and Environmental Engineering Department, American University of Beirut, Beirut Riad El-Solh 1107 2020, Lebanon.
Masters Student, Civil and Environmental Engineering Department, American University of Beirut, Beirut Riad El-Solh 1107 2020, Lebanon

Abstract

The Last Planner System (LPS) have been used on construction projects to improve reliable work planning. Lookahead
planning is intermediate process that connects the master or phase schedule to the weekly work plan. Construction teams
employ lookahead planning to achieve various objectives including a breakdown down of activities into the level of operations,
operations design, and constraint removal to make tasks ready for execution. Tasks Made Ready (TMR) measures the
performance of lookahead planning in identifying and eliminating constraints to make activities ready for implementation. The
purpose of this paper is to study through computer simulation the impact of TMR on task execution, reliability of weekly work
planning, and project duration. Results show that TMR is a good predictor of project duration. These results advise planners
on the importance of the constraint removal process, how it influences the reliability of construction planning, and its impact
on project duration.

Keywords: Constraints, Last Planner System, Lookahead Planning, Tasks Made Ready, Workflow.

1. Introduction

Planning is a crucial determinate of the success of any project. It consists of directing a project to meet its
objectives in terms of time, cost, quality, and safety [1]. It follows then that any poor practice related to planning
could ultimately lead to the project’s failure. The traditional planning techniques, particularly the Critical Path
Method (CPM), were criticized and deemed insufficient and inadequate by many researchers [2].
One of the major deficiencies of CPM is that it focuses on the logical dependencies between activities with a
disregard for the continuity of workflow among different trades [3]. Another problem with CPM relates to
allocating time buffers to cover possible uncertainties in activities [4]. This exercise brings about many constraints
and especially prerequisite constraints that can disrupt the continuity of work. With these time buffers, it becomes
difficult to determine when a certain activity will finish. This adds to the complexity of making tasks ready when
needed resulting in a higher cost and longer project duration.
In the light of these and other inadequacies, the Last Planner System (LPS) has been developed for planning
and controlling construction projects. It has been used to ensure and improve the continuity of workflow while
increasing planning reliability by minimizing the difference between tasks that are planned and those that are
actually completed [5,6].
The LPS consists of four planning phases: (1) master scheduling which sets major project milestones; (2) phase
scheduling which identifies the activities that “should” be done as well as the handoffs among the different
specialty organizations involved in the project; (3) lookahead planning which breaks down the tasks from
processes to operations, identifies constraints, and assigns responsibilities to make tasks ready; and (4) weekly
work planning during which reliable promises and commitments to execute the work are made [7, 8, 9].

_____________________________
* Tel.: +961 1 350000 Ext: 3616
E-mail address: fh35@aub.edu.lb

450
 Of the LPS different phases, lookahead planning plays a critical role as it bridges between master and phase
scheduling from one side and weekly work planning from the other side. It serves to break down tasks from the
process level employed in the master and phase schedules into the operation level of detail that is needed for
commitment. Moreover, as previously mentioned, tasks are made ready during this phase after identification and
removal of different constraints [7, 9]. These constraints could subsume design or construction related information,
equipment and machinery, labor, space, predecessor activities, weather conditions, and others [10, 11]. Thus, the
reliability of the construction workflow can increase when tasks are made ready for execution just on time.
In the LPS, the reliability of the construction workflow is generally measured using the metric “Percent Plan
Complete” denoted as PPC. PPC measures the percentage of tasks completed at the end of a fixed time interval
relative to those that were planned at the beginning of this interval [7, 8]. As for measuring the performance of the
lookahead planning phase, two metrics “tasks Anticipated” (TA) and “tasks made ready” (TMR) are used. TA
measures the effectiveness of the lookahead planning in anticipating the tasks that are going to be executed in the
future. On the other hand, TMR assess the performance of the lookahead planning in identifying and removing the
different types of constraints, previously mentioned, for tasks to be ready for execution. TMR denoted as TMR (i,
j) refers to the percentage of tasks anticipated on the lookahead plan i weeks ahead of the execution week j [12,
13].
Several studies have targeted the make-ready process [14, 15, 16]. In fact, one of the studies has explored the
impact of the make-ready process on PPC. The percent of constraints removed relative to all constraints identified
on the lookahead plan one week before execution was measured using a metric called “Percent of Constraint
Removal” denoted as PCR1. A strong correlation between PCR1 and PPC was found in three case study projects
[14, 15]. Other studies have shown that productivity increases with increased reliability of workflow which is
reflected as a higher PPC [6]. However, there are no detailed studies addressing: (1) the planning actions followed;
(2) the parameters that affect the make-ready process; and (3) the effect of modifying these parameters on the
reliability of planning, measured in PPC and schedule performance, measured in project duration [13].
In acknowledgement of the need to explore such aspects, this paper aims at studying the relation between TMR,
PPC, and project duration on construction projects by simulating real project conditions. The behaviours of TMR,
PPC, and project duration are studied under various conditions enhancing the understanding of the effect of making
tasks ready on planning reliability and project duration. For these purposes, a theoretical framework is developed
for the make-ready process within lookahead planning. This is accomplished through building a simulation model
for the make-ready process and testing different scenarios by varying different parameters. This will be further
expanded in the methodology section below.

2. Methodology

The research method includes 4 main stages: (1) developing a conceptual model for weekly work planning
starting form a lookahead schedule, (2) building the corresponding mathematical depiction of system into a
computer simulation model; (3) designing an experiment to understand the impact of TMR on project duration
while varying input parameters and measuring the output variables as shown in table 1; and (4) analyzing the
results. Once the model was set and verified, it was used as a laboratory for testing various scenarios and project
conditions.

Weekly work planning process

The lookahead process as it progresses from six weeks ahead of execution until the week of execution was
modeled as follows.
(1) 6 to 3 weeks ahead of execution: Tasks obtained from the phase or master schedule are added on the six
weeks lookahead plan. During this period, gross constraints, which relate to phases and processes, are identified
and removed. These constraints subsume, for example, design and material information which necessitate long
lead time to be removed [5, 9]. Tasks are then broken down into processes and operations.
(2) 2 weeks ahead of execution: Tasks are continuously broken down until they match the level of detail
necessary for production at the weekly work plan level. During this week, specific constraints, relating to specific
tasks or operations, are identified and then removed in order to make tasks ready. These constraints include
prerequisite tasks, material, space, information, and other resources.
(3) 1 week ahead of execution: At this stage, the process of pulling and screening of tasks is conducted. Pulling
helps in identifying the tasks that should be made ready based on the actual site demand. As for screening, it helps
determine the actions needed in order to remove the different types of constraints, previously discussed. This
prevents commitment to tasks that cannot be made ready [7, 9]. At the end of this process, tasks are divided into
two categories: tasks that are made Ready (R % of all) and those that are Not-Ready (1-R % of all). Out of the

451
Not-Ready tasks, a certain ratio P has a probability of having the constraints removed. These tasks, which are not
Ready but can be made Ready (denoted as CMR) are advanced to the weekly work plan (WWP). The remaining
portion (1-P) cannot be made Ready (CNMR) and, thus, is not advanced to the WWP. However, this portion will
undergo the same analysis in the following lookahead planning cycle.
It should be noted that any task undergoes a process of shielding before joining the WWP. In this process, tasks
are tested against five quality criteria including definition, soundness, sequence, size, and learning. Any task must
be well defined, constraint-free, in a proper sequence, and adequately sized to match capacity and must allow
learning for continuous improvement. By ensuring that only such tasks are moved to the WWP, the shielding
process protects downstream tasks from uncertainties of upstream tasks and, thus, ameliorates plan reliability [12,
17].
Excess tasks that are ready but not critical are added to the workable backlog and are executed in case of extra
available capacity. Some tasks other than those pebbles that were broken at the beginning of this cycle are added
at the last moment to the WWP. These tasks (denoted as New) add more complexity to the planning process and
might be hard to be made ready just in time for completion.
(4) During the execution week: Tasks that are actually unconstrained (denoted as Ready Ready-RR) and
constrained tasks that were made ready during execution are successfully executed. On the other hand, some tasks
that were perceived as ready before and they were actually not ready or got constrained during execution fail to be
executed. These tasks are called “not quite ready” and denoted as 1-RR. As for the CMR tasks, coordinated effort
must be made in order to remove their constraints and execute them. These tasks are denoted as NR while those
that cannot be made ready are denoted as 1-NR. Regarding the New tasks, a portion of them denoted as N will be
made ready and executed while the second portion denoted as 1-N cannot be made ready or executed.
At the end of this week, Percent Plan Complete (PPC) is calculated by dividing the number of tasks executed
by the number of those planned. TMR is calculated by dividing the Made Ready task at WK0 (execution week)
that have survived from WK2 (two weeks ahead of execution) over the total number of tasks on WK2. This
lookahead planning process is repeated until all the tasks are executed denoting the end of the project.
Table 16 below summarizes the parameters used as well as their corresponding values.

Table 16. Test Parameters and Values

Parameters Description Values

Rocks The number of tasks at the level of detail of processes in the project. 2,080
Ratio of the number of tasks at the lookahead level that an activity gets Randomly selected from {2,
Rocks to pebbles broken into. 3, 4}
R Percentage of tasks that are made ready during WK1. {0.1, 0.3, 0.5, 0.7, 0.9, 1}
RR Percentage of Ready tasks that are actually ready or unconstrained. {0.1, 0.3, 0.5, 0.7, 0.9, 1}
Percentage of tasks that are Not-Ready but have a chance to become ready
P some time prior to their scheduled execution. {0.1, 0.3, 0.5, 0.7, 0.9, 1}

Percentage of Not-Ready tasks that will be made ready during the

NR execution week. {0.1, 0.3, 0.5, 0.7, 0.9, 1}

N Percentage of New tasks made ready during the execution week. {0.1, 0.3, 0.5, 0.7, 0.9, 1}
NBR Ratio of New tasks at WWP to tasks at the end of WK2. {0.1, 0.3, 0.5, 0.7, 0.9, 1}
Number of tasks added to the WWP at the last moment without undergoing
New lookahead planning process. NBR*Broken Pebbles

It is equivalent to all the possible combinations of values that the

Total number of parameters can take which is equal to 66, given that each of the 6 parameter 46,656
runs can take 6 different values {0.1, 0.3, 0.5, 0.7, 0.9, 1}.

Simulation Model and Experiment

The conceptual model was then built into a discrete event simulation model in Any Logic 7 (university edition,
educational version). To test the relationship between TMR, PPC, and project duration, an experiment was
designed to find the possible impacts on PPC and project duration when varying TMR. The experiment tests the
impact of various combinations of parameters. To vary TMR, the governing parameters R, RR, P, NR, New, and

452
NBR are varied in almost all combinations. The resulting TMR, PPC, and project duration are recorded and
analyzed as per the parameters shown in table 1. This experiment can resemble almost any type of project. Having
six parameters taking on six different possible values, there are 46, 656 possible combinations studied. Results are
shown in the next section.

3. Results and Discussion

The experiment’s objective is try all different combinations of the parameters shown in Table 1 and then analyse
the resulting TMR, PPC and Project Duration. These results are shown in Figure 1. For each graph, the x-axis
corresponds to the run number where each run consists of a different scenario defined by a certain combination of
the project parameters R, RR, P, NR, N and NBR. It is important to note that although N and NBR have no major
effect on TMR, they were included as part of the experiment’s parameters due to their effect on PPC and project
duration.
In order to get a clearer view of the results, the different scenarios are sorted according to increasing project
duration. Linear trend lines are also added to both TMR and PPC to have a better idea of these results’ overall
trend. Moreover, the project duration’s y-axis is fixed at 100 weeks even though some runs reached durations over
100. Adding these relatively few high values would have negatively altered the graph’s scale, resulting in an
insignificant representation of results.
The results show that as project duration decreases both PPC and TMR tend to increase as a general trend. It is
also worth mentioning that the project duration’s decreasing pattern is not linear but rather logarithmic which
means that when PPC and TMR are low, working on increasing their values can greatly improve project
performance. However, after a certain point, improving project duration becomes more challenging since the rate
of improvement becomes less significant. In fact, a jump of TMR from 0.035 to 0.25 resulted in a total duration
reduction of over 40 weeks whereas a jump of TMR from 0.25 to values close to 1 resulted in a reduction slightly
greater than 10 weeks. This raises questions on the economic sense of increasing TMR beyond a certain threshold.
When the results are evaluated on a scenario by scenario basis, more interesting inferences can be made,
especially regarding PPC. The PPC graph reveals that very high values are reached (values close to 1) for project
durations as high as 77 weeks which is 20 weeks greater than the lowest possible duration. The same cannot be
said about TMR since there are no instances where a high project duration is coupled with a high TMR. This is
the case because PPC does not take into account cannot be made ready (CNMR) tasks unlike TMR. So, a high
PPC coupled with a long duration simply means that the project is under-committing by having a high number of
cannot be made ready tasks even though all the ready and can be made ready activities are being executed at a
high rate.
Although the TMR trend is closer to that of project duration than PPC, it is still not completely linear, having a
range of value for given project durations. This range is due to the presence of New tasks which are not part of the
original weekly work plan. A high number of New tasks coupled with a low N (making them ready) will negatively
affect project duration without affecting TMR since the latter does not take New into account. Similarly, a high
number of New activities and a high N would positively impact project duration without affecting TMR. A high N
is a sign of an agile project team who can work fast on removing constraints for tasks that emerge at the last
moment. Despite this range of TMR resulting from variations in New and N, TMR still proves to be a better and
more reliable indicator of project duration than PPC.
1
0,8
0,6
PPC

0,4
0,2
0
0 5000 10000 15000 20000 25000 30000 35000 40000 45000
Simulation Run

453
 1
0,8
0,6
TMR

0,4
0,2
0
0 5000 10000 15000 20000 25000 30000 35000 40000 45000
Simulation Run

100
Project Duration (weeks)

90
80
70
60
50
0 5000 10000 15000 20000 25000 30000 35000 40000 45000
Simulation Run

Figure 17. Results for TMR, PPC, and Project Duration

4. Conclusions

This study presents results from simulation experiments conducted to investigate the relationship between
TMR, PPC, and project duration. The results confirm the importance of the make-ready process and how constraint
removal can influence project duration. The practical impact of this finding suggests that improving the team’s
ability to perform sound lookahead planning, design operations to match load and capacity, and remove constraints
before and during execution can have a profound impact on improving plan reliability measured by PPC and
overall project duration.
As TMR increases, the project duration decreases in a clear fashion. However, the decrease in project duration
flattens and becomes asymptotic for higher values of TMR. When New is high, the impact of TMR on project
duration is shared with the impact of N. If TMR and N are high then project duration is at a minimum. A high N
is a sign of an agile project team who can work fast on removing constraints for tasks that emerge at the last
moment.
The results show that a high TMR can result in a reduced project duration but the same cannot be always said
about PPC. A schedule with a high number of ‘cannot be made ready’ (CNMR) tasks can result in a high PPC but
would most probably result in a long project duration. PPC can also be gamed by a project team on a real-life
project to appear high by under committing.
In conclusion, TMR serves as a better indicator for project duration than PPC. The practical implications of this
finding invites practitioners to put more focus on sound look ahead planning where future weekly work plans are
studied early, operations are designed to match load and capacity, and constraints are successfully removed before
and during execution. The fruits of this focus would show through improved plan reliability measured by PPC, a
higher rate of constraint removal measured by TMR, and a reduced overall project duration resulting from
executing more tasks on-time and enabling a timely start for more downstream tasks.

Acknowledgements

Simulation models and experiments presented in this paper were supported by the Civil Engineering
Department, and University Research Board (URB) grant ‘Short-term Planning’ at the American University of
Beirut. All support is gratefully acknowledged. Any opinions, findings, conclusions, or recommendations
expressed in this paper are those of the authors and do not necessarily reflect those of the contributors.

454
 References

[1] Laufer, A., Shapira, A., Cohenca-Zall, D., & Howell, G. A. (1993). Prebid and preconstruction planning process. Journal of construction
engineering and management, 119(3), 426-44
[2] Koskela, L., Howell, G., Pikas, E., & Dave, B. (2014). If CPM is so bad, why have we been using it so long?
[3] Seppänen, O. (2009). "Empirical Research on the Success of Production Control in Building Construction Projects." PhD, Helsinki
University of Technology Espoo, Finland
[4] Goldratt, E. M. (1997). Critical Chain, North River Press
[5] Hamzeh, F.R. (2009) Improving construction workflow – the role of production planning and control, PhD Dissertation, Civil and
Environmental Engineering Department., University of California, Berkeley, CA, p. 273.
[6] Liu, M. and Ballard, G. (2011) Work flow variation and labor productivity: Case study. Journal of Construction Engineering and
Management, ASCE, 27(4), 236–42
[7] Ballard, G. (2000) The last planner system of production control, Ph.D. Dissertation, Faculty of Eng., School of Civil Eng., The University
of Birmingham, UK, 192 pp.
[8] Ballard, G. and Howell, G. (2003) An update on last planner Conference of the International Group for Lean Construction, Blacksburg,
Virginia, USA Julio Martinez (ed.) 11, 13 pp.
[9] Hamzeh, F.R., Ballard, G. and Tommelein, I.D. (2012) Rethinking lookahead planning to optimize construction workflow. Lean
Construction Journal, 15–34
[10] Koskela, L. (1999) Management of production in construction: A theoretical view, in Tommelein, I.D. and Ballard, G. (eds.), Annual
Conference of the International Group for Lean Construction, University of California, Berkeley, CA, USA, 7, pp. 241–52.
[11] Lindhard, S. and Wandhal, S. (2012) Improving the making ready process - exploring the pre-conditions to work tasks in construction.
Annual Conference of the International Group for Lean Construction, San Diego, USA, 20, pp.451
[12] Ballard, G. (1997) Lookahead planning: The missing link in production control. Annual Conference of the International Group for Lean
Construction, Gold Coast, Australia, Luis Alarcon (ed.), 5, 13–26.
[13] Hamzeh, F. R., Saab, I., Tommelein, I. D. and Ballard, G. (2015) Understanding the role of “Tasks anticipated” in lookahead planning
through simulation, Automation in Construction, 49(A), January, 18–26
[14] Jang, J.W. and Kim, Y.W. (2007) Use of percent constraint removal to measure the make ready process. Conference of the International
Group for Lean Construction, Michigan, USA Christine Pasquire and Patricia Tzortzopoulos (eds), 15, 529–38.
[15] Jang, J. and Kim, Y. (2008). The relationship between the make ready process and project schedule performance. Annual Conference of
the International Group for Lean Construction, Manchester, UK, 16, 647–56
[16] Hamzeh, F., Aridi, O. (2013) Modeling the last planner system metrics: A case study of an AEC company, Proceedings of the 21st Annual
Conference of the International Group for Lean Construction IGLC, Fortaleza, Brazil, pp. 599–608.
[17] Ballard, G., Hamzeh, F.R. and Tommelein, I.D. (2007) The Last Planner Production Workbook- Improving Reliability in Planning and
Workflow, 2nd edn, Lean Construction Institute, San Francisco, California, USA, p. 81 .

455

View publication stats