反思前瞻规划优化施工流程Farook Hamzeh Glenn Ballard Iris D. Tommelein摘要研究的问题：如何改善前瞻规划在建设行业的做法来提高生产计划的可靠性？目的：为了评估前瞻规划的性能，寻找一个标准化的做法，使前瞻规划与活动执行有紧密的联系，来提高生产计划的可靠性。
研究设计/方法：本研究采用案例分析，行业访谈，和行业调查，以评估目前在北美、南美和欧洲的建设项目执行的前瞻规划。
研究结果：研究结果显示存在与去年规划系统规则的不符合，前瞻规划与标准化做法的不足，识别和清除限制的迟缓，而且没有对计划失败的分析。
关键词：前瞻规划，生产计划，生产控制，精益建设，最后的规划系统，规划建设。
简介建筑、工程与施工是受变化问题的困扰的，即破坏项目绩效和扰乱施工流程导致对项目时间、成本和质量造成的不利影响（Hamzeh等，2007年，霍普和Spearman2008年，萨利姆等。
2006年，克莱顿1966年）。
组织使用许多种不同的方法来维持生产流程的一致性和屏蔽产量内部业务流程以及外部环境的变化。
汤普森（1967）着重介绍了这些方法，其中包括：•预测•缓冲•平滑各种预测方法是用于预测在内部流程和生产原料中的变化。
然而，预测不能满足所有的变化，并且有许多限制：越详细的预测越不准确，越遥远的预测越容易出错。
（纳米亚斯2009年）。
缓冲用于减轻同时在输入侧和输出侧的工艺变化。
输入通常需要成功执行的任务包括：信息，先决条件工作，人力资源，空间，材料，设备，外部条件和资金（巴拉德＆Howell公司1994年，科斯基拉2000年）。
缓冲区可以采取的三种主要形式：时间，库存和产能。
时间缓冲是分配松弛的活动，利用额外的库存缓冲库存以应对供应的变化，以及用容量缓存，保留额外的容量，如加班或只在需要的时候维持机器工作，以适应激增的负荷。
平滑的供应和需求的变化是另一种方法，组织申请由于缓冲可能的不足，以满足所有的变化，是昂贵的，并可能导致满荷。
平滑需求的一个例子在丰田生产系统中平稳的工作负荷或平准化的倡导（莱克2004年）。
虽然变化破坏了项目的绩效，生产系统可以通过设计减少这种变化，并且可以通过上述提到的方法的组合来管理这种残差。
一个生产系统可以被定义为人员和资源的集合。
（例如，机械，设备，信息），被安排设计和制造产品（“货物”或“服务”）的价值给客户（Ballard等人，2007年）。
一个生产系统的基石是生产管理，例如最近的一个规划系统已成功实施建设项目（2004年巴拉德和Howell），以提高规划的可靠性，提高生产性能，在设计和施工作业中创建可以预测的工作流程。
在任何项目中，规划过程中可能遇到各种问题的困扰。
规划涉及到的越远的项目可能越不准确（纳米亚斯2009）。
当规划师把计划推向前线专家而不涉及他们的计划发展时是很难执行的工作日程。
在一厢情愿的想法的基础上开发的短期工作计划，贸易专家没有可靠的承诺，这种短期计划就会变得更短。
如果计划失败的原因并不在识别和处理不及时，进一步的失败是必然要发生的（哈姆泽2009年）。
此外，可靠的规划依赖于有效的约束分析与排除。
约束是一个活动启动前必须存在的先决条件（例如，以前的工作，信息，劳动力，材料，设备，工具，空间，天气等）。
管理约束条件可以通过识别资源冲突和提前解决这些问题帮助优化工作计划使工作开始。
如果没有约束，是难以管理和减少工作流程不确定性的，这往往导致过程的变化（蔡等，2003）。
考虑到上述挑战，最后规划系统主张项目规划中的以下步骤：•计划越详细情，执行工作越细致（科恩，2006年）•制定工作计划要执行工作的人制定工作计划•提前识别和消除工作的限制，作为一个团队，做好工作准备，提高工作计划的可靠性•在与项目参与方协调和积极谈判的基础上，做出可靠的承诺和工作执行•从失败的计划中学习经验，找出错误的根源，并采取预防措施（巴拉德，等。
2009年）尽管该系统的优点（阿拉尔孔和克鲁兹1997年，冈萨雷斯等人，2008），在许多建设项目目前的做法显示了执行不力前瞻规划，在较大差距长期规划（主阶段附表）和短期规划（承诺/每周的工作计划），降低了可靠性规划系统，并能够建立先见之明。
本文提出了一种前瞻规划实施的一种评估，为最新规划系统的一个过程，其重点介绍了统营规划系统一些不足之处，强调的前瞻规划每周工作规划成功的一个主要驱动力的作用，并建议进行前瞻规划的指导方针有关的活动故障，操作设计，及制约因素分析。
最新的规划系统最新规划系统开发者是格伦·巴拉德和格雷格·豪威尔，其是一生产计划和控制建设工作的系统，以协助平滑的变化流，发展规划先见之明，并减少施工作业中的不确定性。
系统最初是每周的工作计划水平，但很快就把处理工作流程中的变化扩大到覆盖整个计划和进度开发过程中，从主调度，通过前瞻规划，逐步调度达到每周工作规划中。
计划完成百分比（PPC）是一个度量，用于跟踪每周的工作计划水平的性能可靠性，通过测量相对于那些计划完成任务的百分比。
因此，它有助于评估的可靠性工作计划，并开始准备工作，以执行工作计划。
PPC并不是直接衡量项目的进展，而是承诺保持在何种程度上的措施，因此在何种程度上未来的工作负载可能是可预测的。
以前的研究发现PPC和劳动生产率之间的相关性（2008年刘和巴拉德）。
PPC上提高可能产生的二次冲击工作安全和质量需要进一步的研究（1998年，巴拉德和Howell Ballard等。
2007）。
有研究表明，尽管在LPS的优势，许多企业实施系统时都面临重大障碍。
（Ballard 等人，2007;哈姆泽，2009年维亚纳等）。
为建设项目的最新规划系统的成功实施提出了一个框架。
然而，当整个规划系统（主生产计划，阶段调度，前瞻的规划，每周工作规划）执行和更新设计，PPC 项目进展情况的指标，PPC和进步应随对方。
这可以表示为如权利要求一个复杂的假设，即：H1：如果前瞻任务是从一个阶段计划结构实现到项目的结束日期和中间里程碑，如果前瞻规划是准备应该怎样做，如果每周选择什么，可以做什么工作计划由应该做在临界的顺序没有游戏的系统，PPC将随项目的进展情况而不同。
如果我们接受这个假设，那么，如果不随PPC项目的进展情况，在虚拟链的某个地方就有一个破碎的链接。
图1显示了活动打破阶段的规划系统（砾石）的过程（岩石），然后在四个规划过程中的操作（卵石）不同的按时间跨度：主生产计划，阶段调度，前瞻规划和每周工作规划。
主生产计划是一个前端的规划过程中产生的时间表描述工作在进行整个项目的持续时间。
它涉及到项目级活动，并确定重要的里程碑日期主要集中在有关合同文件和拥有者的价值主张（Tommelein和Ballard，1997）。
相调度产生的时间表，覆盖项目的每个阶段，如地基，结构框架，或完成。
在协作规划设置项目团队：（1）定义项目阶段或里程碑，（2）将其分解成组成活动，（3）时间表向后的里程碑。
合并后从不同的项目方输入和在重要的阶段识别专家和团队执行之间逆相调度的平衡，从重要的阶段里程碑（哈姆泽2009年，巴拉德和豪威尔2004）。
前瞻规划是在生产控制（执行时间表）的第一步，通常包括一个为期6周的时间。
前瞻时段随正在执行的工作类型和上下文的不同而不同。
（例如，因为这种现象的出现，在概念设计，任务可以不在详细的预见水平很远的水平。
在工厂停工时，前瞻期延伸到年底关机。
在这项研究中，重点是正常的建设项目，并在这些4至6周的时间框架是常用的前瞻规划）。
在此阶段，活动被分解成水平的生产过程/操作，约束被识别，操作的设计，和准备作业就绪（巴拉德1997年，哈姆泽2009年）。
每周工作规划（WWP）也被称为承诺计划是最系统的详细计划，展现了工作的各专业组织之间的相互依存，和直接驱动的生产过程。
计划在这个级别的可靠性促进了质量分配和可靠的承诺，使生产单元从上游业务中的不确定性被屏蔽。
这个工作任务是一个详细的测量完成可的计划。
每个计划期结束时，作业被评论，评估它们是否是完整的，从测量规划中的可靠性。
对于不完整的任务，对计划失败的原因进行分析，并采取行动，这些原因是学习和持续改进的基础（巴拉德2000年）。
Farook Hamzeh, Glenn Ballard & Iris D. Tommelein(2012) Rethinking Lookahead Planning to OptimizeConstruction Workflow. Lean Construction Journal 2012pp 15-34 Lean Construction Journal 2011/licenses/by-nc-nd/3.0/15 Rethinking Lookahead Planning to Optimize Construction WorkflowFarook Hamzeh1; Glenn Ballard2; Iris D. Tommelein3AbstractResearch Question: How to improve lookahead planning practices in the constructionindustry to increase the reliability of production planning?Purpose: To assess the performance of lookahead planning, advise a standardized practice tosupport a strong linkage between Lookahead planning and activity execution, and improve the reliability of production planning.Research Design/Method: This study employs case study analysis, industry interviews, and anindustry survey to assess the current implementation of lookahead planning on construction projects in North America, South America, and Europe.Findings: The study findings indicate the existence of non-compliance with Last Planner®System rules, inadequate lookahead planning and standardized practices, sluggish identification and removal of constraints, and absence of analysis for plan failures. Limitations: The authors’ active role on the projects used as case studies may constitute alimitation to the research methods and tools used. The industry survey may have not covered all companies applying the Last Planner System. The suggested framework should be custom tailored to different projects to cater for size, culture, etc. Implications: This research provides a framework for applying the Last Planner System rulesduring lookahead planning. It aims at increasing the success of the making activities ready, designing operations, and ultimately improving PPC.Value for practitioners: The study presents to industry practitioners applying the LastPlanner System a standardized framework for implementing lookahead planning on construction projects. The paper also highlights the use of two metrics to assess the performance of lookahead planning at a given point in time and to monitor performanceover a period of time or between projects.Keywords: Lookahead planning, production planning, production control, lean construction,the Last Planner System, construction planning.1 Corresponding Author- Assistant Professor, Department of Civil and Environmental Engineering, 406E Bechtel,American University of Beirut, Riad El Solh, Beirut 1107 2020, Lebanon, Farook.Hamzeh@.lb2 Research Director, Project Production Systems Laboratory andAssociate-AdjunctProfessor, Civil and Environmental Engineering Department, 215 McLaughlin Hall, University. of California,Berkeley, CA 94720-1712,USA, ballard@3 Professor, Dept. of Civil and Environmental Engineering, and Director, Project Production Systems Laboratory, 215-A McLaughlin Hall, University of California, Berkeley, CA 94720-1712, USA ,tommelein@Hamzeh, Ballard, & Tommelein: Rethinking Lookahead Planning to Optimize Construction WorkflowLean Construction Journal 2011/licenses/by-nc-nd/3.0/page 16 Paper type: Full PaperIntroductionArchitecture, Engineering, and Construction (AEC) processes are plagued with problemsassociated with variations that undermine project performance and disrupt workflow leadingto detrimental impacts on project’s duration, cost, and quality (Hamzeh et al. 2007, Hoppand Spearman 2008, Salem et al. 2006, and Crichton 1966). Organizations use a number ofdifferent methods to maintain consistency in production flow and to shield production fromvariations in internal business processes as well as the external environment. Thompson(1967) highlighted some of these methods including:•Forecasting•Buffering•SmoothingVarious forecasting methods are used to anticipate variations in internal processes andin inputs to production. However, forecasts cannot cater for all variations and have manylimitations: the more detailed a forecast is the more off it will be, the farther a forecastlooks into the future the less accurate it becomes, and forecasts are always wrong (Nahmias2009).Buffering is used to mitigate process variations on both the input and output sides. Inputs typically needed for successful execution of tasks include: information, prerequisitework, human resources, space, material, equipment, external conditions, and funds (Ballard& Howell 1994, Koskela 2000).Buffers can take on one of three main forms: time, inventory and capacity. Time buffersallocate slack to an activity, inventory buffers utilize extra stock to account for supplyvariations, and capacity buffers reserve extra capacity such as using overtime or maintainingmachinery used only when needed to accommodate surges in load.Smoothing variations in supply and demand is another method that organizations applysince buffering may not be enough to cater for all variations, is costly to apply, and may leadto complacency. An example of smoothing demand is leveling the work load or heijunka asadvocated in the Toyota Production System (Liker 2004).Although variation undermines project performance, production systems can be designed to reduce them and to manage residuals utilizing a combination of the abovementioned methods.A production system can be defined as a collection of people and resources (e.g., machinery, equipment, information) arranged to design and make a product (“goods” or“services”) of value to customers (Ballard et al. 2007). A cornerstone of a production systemis production management such as the Last Planner System, which has been successfullyimplemented on construction projects (Ballard and Howell 2004) to increase the reliability ofplanning, improve production performance, and create predictable workflow indesign andconstruction operations.On any project, the planning process can be plagued by various problems. Planning involves forecasts that can be inaccurate the further they project into the future (NahmiasHamzeh, Ballard, & Tommelein: Rethinking Lookahead Planning to Optimize Construction WorkflowLean Construction Journal 2011/licenses/by-nc-nd/3.0/page 17 2009). It is hard to execute work schedules when Planners push plans to frontline specialistswithout involving them in plan development. Short-term work plans developed on the basis ofwishful thinking and in absence of reliable promises from trade experts are more likely to fallshort during execution. And if causes of plan failures are not identified and dealt with in atimely fashion, further failures are bound to happen (Hamzeh 2009). Moreover, reliableplanning depends on effective constraint analysis and removal. Constraints are those prerequisites required to be present before an activity can start (e.g., previous work,information, labor, material, equipment, tools, space, weather, etc.). Managing constraintscan help optimize work plans by identifying resource conflicts and resolving them prior towork start. Without constraint removal, it is hard to manage and reduce work flow uncertainties that often cause process variations (Chua et al. 2003).Taking into account the challenges mentioned above, the Last Planner System advocatesthe following steps in project planning:•Plan in greater detail as you get closer to performing the work (Cohn 2006) •Develop the work plan with those who are going to perform the work•Identify and remove work constraints ahead of time as a team to make work ready andincrease reliability of work plans•Make reliable promises and drive work execution based on coordination and active negotiation with project participants•Learn from plan failures by finding root causes and taking preventive actions (Ballard,et al. 2009)Despite the advantages of this system (Alarcón and Cruz 1997, Gonzalez et al. 2008), the current practice on many construction projects shows a poor implementation oflookahead planning resulting in a wide gap between long-term planning (master and phaseschedules) and short-term planning (commitment/weekly work plans) reducing the reliabilityof the planning system and the ability to establish foresight.This paper presents an assessment of lookahead planning implementation as one processin the Last Planner System, highlights some inadequacies in operating the planning system,emphasizes the role of lookahead planning as a prime driver to the success of weekly workplanning, and suggests guidelines for performing lookahead planning pertaining to activitybreakdown, operation design, and constraint analysis.The Last Planner SystemThe Last Planner System as developed by Glenn Ballard and Greg Howell is a system forproduction planning and control used to assist in smoothing variations in construction workflow, developing planning foresight, and reducing uncertainty in construction operations. Thesystem originally tackled variations in workflow at the weekly work plan level but soonexpanded to cover the full planning and schedule development process from master scheduling to phase scheduling through lookahead planning to reach weekly work planning.Percent Plan Complete (PPC) is a metric used to track the performance of reliable promising at the weekly work plan level by measuring the percentage of tasks completedrelative to those planned. It thus helps assess the reliability of work plans and initiatesHamzeh, Ballard, & Tommelein: Rethinking Lookahead Planning to Optimize Construction WorkflowLean Construction Journal 2011/licenses/by-nc-nd/3.0/page 18 preparations to perform work as planned. PPC is not a direct measure of project progress, butrather a measure of the extent to which promises are kept, and hence the extent to whichfuture work load may be predictable. Previous research has found a correlation between PPCand labor productivity (Liu and Ballard 2008). Possible secondary impacts of PPC onimprovingwork safety and quality require further research (Ballard and Howell 1998, Ballard et al.2007). Despite the advantages of the LPS, research has shown that many organizations facesignificant hurdles when implementing the system (Ballard et al. 2007; Hamzeh, 2009; Vianaet al. 2010). Hamzeh (2011) presented a framework for successful implementation of the LastPlanner System on construction projects.However, when the entire Last Planner System (master scheduling, phase scheduling,lookahead planning, and weekly work planning) is executed and updated as designed, PPCshould be an indicator of project progress; i.e., PPC and progress should vary with each other.This claim can be expressed as a complex hypothesis; namely:H1: If lookahead tasks are drawn from a phase schedule structured toachieve the project end date and intermediate milestones, and iflookahead planning makes ready what SHOULD be done, and if weeklywork plans are formed from what CAN be done selected from whatSHOULD be done in the order of criticality without gaming the system, PPCwill vary with project progress.If we accept this hypothesis as an assumption, it follows that if PPC does not vary withproject progress, there is a broken link somewhere in the hypothesized chain. Figure 1 shows the Last Planner System where activities are broken down from phases(boulders) to processes (rocks) then to operations (pebbles) across four planning processeswith different chronological spans: master scheduling, phase scheduling, lookahead planning,and weekly work planning.Master scheduling is a front-end planning process that produces a schedule describingwork to be carried out over the entire duration of a project. It involves project-level activities and identifies major milestone dates mostly in relation to contract documents andthe owner’s value proposition (Tommelein and Ballard 1997).Phase scheduling generates a schedule covering each project phase such as foundations, structural frame, or finishing. In a collaborative planning setup the projectteam: (1) defines a project phase or milestone, (2) breaks it down into constituent activities,and (3) schedules activities backward from the milestone. After incorporating input fromdifferent project parties and identifying hand-offs between specialists, the team performsreverse phase scheduling back from important phase milestones (Hamzeh 2009, Ballard andHowell 2004).Lookahead planning is the first step in production control (executing schedules) and usually covers a six week time frame. Lookahead time periods vary with the type of workbeing performed and the context. (For example, in conceptual design, tasks cannot beforeseen at a detailed level very far in advance because of the phenomenon of emergence. Inplant shutdowns, the lookahead period extends to the end of the shutdown. In this research,the focus is on normal construction projects, and on those 4 to 6 week time frames arecommonly used in lookahead planning). At this stage, activities are broken down into theHamzeh, Ballard, & Tommelein: Rethinking Lookahead Planning to Optimize Construction WorkflowLean Construction Journal 2011/licenses/by-nc-nd/3.0/page 19 level of production processes/operations, constraints are identified, operations are designed,and assignments are made ready (Ballard 1997, Hamzeh 2009).Weekly work planning (WWP) also known as commitment planning represents the mostdetailed plan in the system, shows interdependence between the works of various specialistorganizations, and directly drives the production process. Plan reliability at this level ispromoted by making quality assignments and reliable promises so that the production unitwill be shielded from uncertainty in upstream operations. The work assignment is a detailedmeasurable commitment of completion. At the end of each plan period, assignments arereviewed to assess whether they are complete or not, thus measuring the reliability of theplanning. For incomplete assignments, analyzing the reasons for plan failures and acting onthese reasons is the basis of learning and continuous improvement (Ballard 2000). Figure 1: Planning processes in the Last Planner System.The Last Planner System relates to deliberative and situated action planning as described by Senior (2007) combining aspects of both worlds. On one hand, deliberativeplanning takes place at the master and phase scheduling level where a premeditated courseof action is specified in setting milestones and identifying handoffs. On the other hand, thelookahead and weekly work plans are closer to the situated planning model where plans takeHamzeh, Ballard, & Tommelein: Rethinking Lookahead Planning to Optimize Construction WorkflowLean Construction Journal 2011/licenses/by-nc-nd/3.0/page 20 into account changes in the environment affecting inputs and outputs of constructionactivities.However, a question remains unanswered: how can the AEC industry advance the implementation of the lookahead planning within the Last Planner System to improveconstruction workflow and the reliability of planning?Accordingly, this paper reports an assessment of the current implementation of the LastPlanner System in construction, presents analytical data, highlights concerns with the currentpractice, and lays out recommended procedures to perform lookahead planning aiming atproducing more reliable production plans.MethodologyThis paper summarizes research conducted to study the role of lookahead planning within the Last Planner System in improving construction workflow and increasing thereliability of planning. Research involves results from two construction projects and preliminary results from a survey addressing Last Planner implementation (Hamzeh 2009).Case study research was the methodology adopted in this study because:1. It is appropriate for answering questions pertaining to ‘how’ and ‘why’ whe n no control for behavioral events is required and when research focuses on contemporaryaffairs.2. It uses both quantitative and qualitative methods to explain phenomena.3. It utilizes multiple sources of evidence in a natural setting that encompasses temporaland contextual facets of the variables monitored.4. It uncovers the dynamics of events explaining the phenomenon under study.5. It provides qualitative understanding when arriving at conclusions and analyzing results (Meredith 1998, Stuart et al. 2002, Yin 2003).The case studies involve two health care projects in the United States. Both projects employed the Last Planner System for production control investing heavily in employeetraining and in different aspects of lean construction. The owner on both projects is a strongadvocate of lean and integrated project delivery systems. The authors also conducted someinterviews with industry practitioners who worked on these projects.The first author spent fourteen months on project one working for ten months as a researcher and four months as an intern. While he was more of an observer in the researcherposition, he had more input into the process while working as an intern helping the development of a planning process. The first author also spent around eight months onproject two helping in the implementation of the Last Planner System and applying a newsoftware application to manage the planning process.Prior to case study research, an industry-wide survey was conducted among Last PlannerSystem industry users in the US, south America, and Europe. The survey addressed engineersand managers working for owners, architects/engineers, and contractors. The companieswere identified with help of the Lean Construction Institute (LCI). The survey aimed atassessing the implementation of the system, informing research on obstacles faced in theHamzeh, Ballard, & Tommelein: Rethinking Lookahead Planning to Optimize Construction WorkflowLean Construction Journal 2011/licenses/by-nc-nd/3.0/page 21 current practice, and providing feedback required in shaping the formation of guidelines forimprovement.The research process follows an inductive reasoning scheme adjusted to the specific situation. Accordingly the research process comprised multiple steps of evaluating andassessing the current practice, developing guidelines for improvements, and testing thesesuggested guidelines. In collecting data (e.g., PPC historical project data, master scheduledata) several methods were employed such as: conducting short interviews, attending weeklyor pull/phase scheduling sessions, attending value stream planning sessions, and performingexercises to assess the performance of the lookahead process and weekly work planning.Case Study OneThe first case study is a 555-bed hospital and medical campus in San Francisco, California. The $1.7 billion project comprises a 16-story hospital including two below-gradefloors. This is a project to study because of:•implementing integrated project delivery (IPD) and integrated form of agreement (IFOA)•engaging project partners who are interested in experimenting with lean practices •applying Last Planner System for production planning and control,•utilizing target value design (TVD) to steer design towards meeting the owner’s valueproposition•using building information modeling (BIM) extensivelyAs explained above, we assume for the moment that if the Last Planner System is implemented correctly, PPC and project progress should vary together. To determine if theyin fact do vary together, and hence if implementation is correct, an investigation wasperformed to compare slips or gains on the master schedule with each update to weekly workplanning performance expressed in percent plan complete (PPC). This investigation comprisedan in-depth study of the master schedule built in Primavera P6 scheduling software including:•Monitoring project milestones with each update.•Tracking changes to the schedule in terms of rescheduling, adding activities, addingdetail into current activities, changing sequence/logic, altering durations, and addingmodules suggested during phase planning sessions.•Analyzing the incremental slip or gain on schedule with each update. Assessing the weekly work planning performance was performed by monitoring PPC witheach weekly update. To adequately compare weekly PPC and incremental schedule changeson the master schedule, an aggregate PPC figure was calculated over 3 or 4 weeks, which isthe average cycle for master schedule updates.Figure 2 shows the relationship between percent plan complete (PPC) and the incremental schedule difference in days (+ is schedule delay and – is schedule gain). If systemimplementation was correct (drawing tasks into the weekly work plan that are critical withoutgaming the system), the higher the PPC the lower the schedule slip should be. However,results show a weak correlation of 0.28 and a covariance of 1.01 between PPC and negativeHamzeh, Ballard, & Tommelein: Rethinking Lookahead Planning to Optimize Construction WorkflowLean Construction Journal 2011/licenses/by-nc-nd/3.0/page 22 incremental schedule difference. Therefore, a clear relationship between weekly performance and overall schedule performance was not found.Moreover, the highlighted area in Figure 2 shows an increase in project delays (+ increment) for a higher PPC when more work was actually completed on the weekly schedule3. These results suggest that in Case 1 project, there is no evidence that variation inthe incremental schedule difference can be explained by PPC. This may indicate a poorlinkage between weekly work plans and the master schedule.Since earned value analysis is used to measure progress it may sometimes tempt teamsto work out of sequence to maximize progress gains in a certain period. While this may havean impact on the relationship between PPC and incremental schedule difference, it is still inviolation of the caveats stipulated in hypothesis H1 and a result of an incorrect developmentof the master schedule into a lookahead plan and an incorrect development of the lookaheadplan into a weekly work plan. That is why both working out of sequence and gaming thesystem are indications of poor implementation of the LPS and a poor linkage between weeklywork plans and the master schedule.‐80‐60‐40‐20204060800%10%20%30%40%50%60%70%80%90%100%25‐Oct‐2007 8‐Nov‐2007 22‐Nov‐2007 6‐Dec‐2007 20‐Dec‐2007 3‐Jan‐2008 17‐Jan‐2008 31‐Jan‐2008 14‐Feb‐2008 28‐Feb‐2008 13‐Mar‐2008 27‐Mar‐2008 10‐Apr‐2008 24‐Apr‐2008 8‐May‐2008 22‐May‐2008 5‐Jun‐2008 19‐Jun‐2008 3‐Jul‐2008 17‐Jul‐2008 31‐Jul‐2008 14‐Aug‐2008 28‐Aug‐2008 11‐Sep‐2008 25‐Sep‐20089‐Oct‐200823‐Oct‐2008 PPC Incremental Schedule Difference (days)Figure 2: The relationship between Percent Plan Complete (PPC) and incrementalschedule difference for case study 1Case Study TwoThe second case study project involved rehabilitation of 23 buildings, new constructionof 14 buildings, and landscape works for Fort Baker Retreat Group LLC at Fort Baker, Sausalito, California (Quakenbush 2008). The project budget was $103 million. This project3 While the two negative increments may be due to optimistic updating at the user level, the positive and negativeincrements do not correlate with PPC figures.Hamzeh, Ballard, & Tommelein: Rethinking Lookahead Planning to Optimize Construction WorkflowLean Construction Journal 2011/licenses/by-nc-nd/3.0/page 23 was chosen because the schedule was broken down into separate phases that were monitoredand recorded.This study focuses on the production management practices performed on the rehabilitation section of the project where the general contractor (GC) used the Last PlannerSystem to manage production. The GC developed the master schedule for the project usingphase scheduling sessions in collaboration with trade contractors and subcontractors. Theteam focused on the project handover and everything in between project start and projecthandover was decided using pull sessions. The GC continuously incorporated results fromthese sessions into the master schedule that was built in Microsoft Project and updated on aweekly basis. This schedule was developed to allow for a flow of construction trades from onebuilding to another. The project team used the Last Planner System to develop and manageweekly work planning based on the master schedule. Although the team did not applylookahead planning, they conducted weekly meetings for last planners who。