FLORA: New Forensic Schedule Analysis Technique

Transcription

1 FLORA: New Forensic Schedule Analysis Technique Long D. Nguyen, Ph.D. 1 and William Ibbs 2 Abstract: While various factors such as float ownership, logic change, and resource allocation FLORA affect results of delay analysis, existing delay analysis techniques tend to ignore most if not all of them. To address this insufficiency this paper proposes a new schedule analysis technique called FLORA that simultaneously captures the dynamics of float, logic, and resource allocation in its analyses. FLORA analyzes not only the direct impact of a delay but also its secondary effect. The analysis process follows ten rules that are flexible and customizable. A case study is employed to illustrate its application. FLORA yields different and more reasonable outcomes compared to the window analysis technique. Each single analysis may also yield different or even conflicting results. By properly dealing with the current issues of schedule analysis, FLORA can be more reliable. Practitioners may readily accept its analyses and outcomes because they are able to specify, modify, and consent to the rules for schedule analysis to fit into a specific context in advance. Researchers may further evaluate the impacts of different factors on delay responsibility or apply FLORA to real projects to assess its strengths, weaknesses, and potential improvements. DOI: / ASCE :7 483 CE Database subject headings: Claims; Delay time; Litigation; Construction management; Forensic engineering. Introduction Time impact analysis in schedule delay situations is not simple. Various events caused by different parties occur during the course of contract work. These events may impact project schedules and costs, positively or negatively. They can delay, disrupt, or accelerate project completion. Thus a reliable forensic schedule analysis technique that helps evaluate the extent of project delay or acceleration of an event and its responsible party is essential. Proper and accurate analysis of delays is also requisite to allocate time-related costs to the responsible parties Hegazy and Zhang Unfortunately, today s preferred techniques, such as butfor and window analysis techniques, have substantial limitations and require improvement Mohan and Al-Gahtani In addition, industry practitioners do not agree which schedule analysis technique is preferable Arditi and Pattanakitchamroon 2006; Zack This paper presents a new schedule analysis technique float, logic, and resource allocation FLORA that simultaneously and comprehensively captures the dynamics of float, logic, and resource allocation during the course of work and thus analysis. An activity s total float TF in a project schedule may change over time. Critical paths/activities are therefore time dependent. Float 1 Construction Consultant, Jax Kneppers Associates, Inc., Walnut Creek, CA long@jaxkneppers.com 2 Professor of Construction Management, Dept. of Civil and Environmental Engineering, Univ. of California, Berkeley CA 94720; presently, President, The Ibbs Consulting Group, Inc., 5932 Contra Costa Rd., Oakland, CA corresponding author. drcwibbs@aol.com Note. Discussion open until December 1, Separate discussions must be submitted for individual papers. To extend the closing date by one month, a written request must be filed with the ASCE Managing Editor. The manuscript for this paper was submitted for review and possible publication on May 14, 2007; approved on January 4, This paper is part of the Journal of Construction Engineering and Management, Vol. 134, No. 7, July 1, ASCE, ISSN /2008/ /$ ownership is another issue that has increasingly concerned project participants Peterman 1979; Ponce de Leon 1986; Householder and Rutland 1990; Al-Gahtani and Mohan Some logical sequences between activities can also be changed to accommodate new progress and information. These are known as soft logic. Tamimi and Diekmann 1988 assert the need for reflecting the impact of logic change on project schedule. However, how logic change affects the results of schedule analysis is frequently ignored in current techniques. Ibbs and Nguyen 2007b identify possible extended effects of delays due to the disturbance of resource allocation in downstream work. FLORA solves these various problems in an integrated and interactive manner. Issues in Forensic Schedule Analysis A variety of schedule analysis techniques are available in the industry. Different techniques generally give different results for project parties Stumpf Thus extensive effort has been made to improve schedule analysis i.e., Alkass et al. 1996; Shi et al. 2001; Kim et al. 2005; Mbabazi et al. 2005; Al-Gahtani and Mohan Various issues have also been raised such as concurrent delays, pacing delays, fair treatment of noncritical activities, real time analysis, float ownership, scheduling options, and resource allocation Zack 2000; Arditi and Pattanakitchamroon 2006; Mohan and Al-Gahtani 2006; Ibbs and Nguyen 2007b. Current methods and their improvements can only solve one or some of these issues. In addition, the impact of logic change on delay responsibility has really not been addressed in these previous studies. The following sections will discuss critical issues and then show their relationship in forensic schedule analysis. Float and Float Ownership In the critical path method total float or slack is defined as the total amount of time that an activity can be delayed without delaying the project completion date. Since float is a critical asset JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT ASCE / JULY 2008 / 483

2 project schedule and cost. Detailed discussion of these studies can be found elsewhere i.e., Prateapusanond 2003; Arditi and Pattanakitchamroon 2006; de la Garza et al While these studies recommend how float should be allocated and managed they do not in general provide a practical and systematic approach that can be used in forensic schedule analysis. A few approaches of total float management for schedule delay analysis have been proposed in recent years. Prateapusanond 2003 suggests that the owner and contractor each own half of total float available on any activity, namely allowable total float ATF. In addition, the number of delayed days that a party is held responsible for RDD ; if any, will equal the minimum value of: 1 the total delayed days of the entire project TDD ; and 2 the difference between the number of days that the party delays on the affected activity path PDD and its allowable total float, i.e. Fig. 1. Dynamics of float, logic, and resource allocation the question who owns float? has increasingly concerned contractual parties. The result of schedule delay analysis can be affected by the various views regarding who owns float Arditi and Pattanakitchamroon Consequently, float ownership and its use can be a major source of dispute when the project suffers from delay Prateapusanond For example, it is impossible to identify who is responsible for the 2-week project delay of the simple case in Fig. 1 a unless the parties have agreed on float ownership. Owner caused-, contractor-caused, and third partycaused, inexcusable, excusable/compensable, and excusable/ noncompensable delays are denoted as o, c, t, IE, EC, and EN, respectively, in this paper. Several studies have proposed different alternatives for total float ownership, sharing, and/or management. Householder and Rutland 1990 propose that the party who loses or gains as a result of fluctuation in the project cost should own and use float as a resource. de la Garza et al suggest that the contractor owns float but has to trade it on demand by the owner. Zack 1993 recommends the use of a joint-ownership-of-float provision and a systematic time-impact analysis of each delay event. Pasiphol and Popescu 1994 allocate total float to individual activities on the paths such that all activities are critical. Gong 1997 calculates safe float, which can be used without severely affecting the risk of project delay. Sakka and El-Sayegh 2007 propose a method that quantifies the impact of float loss on RDD = Min TDD,PDD ATF This concept of preallocation of total float is a workable and interesting idea. In the survey Prateapusanond 2003, the fact that most of participants generally agreed with this concept is evident. However, this concept alone is impractical if applied to delay analysis because it cannot capture the changing nature of activity paths during the course of work such as changes in critical paths and in logical sequences. In addition, six different examples used in Prateapusanond 2003 to illustrate the application of this delay analysis methodology are not representative. There is no activity that belongs to two or more paths a common situation in construction schedules. In such a situation the use of that proposed method can be impossible or problematic. Al-Gahtani and Mohan 2007 proposes a new total float management technique for delay analysis. It sets fairly reasonable rules for the entitlements of total float. If total float changes due to delay events the responsible party will be discredited total float for delays to the affected activity and will gain or lose total float of successor activities. However, the apportionment of concurrent delay in this method is arbitrary since it only considers the number of delays caused by each party rather than the degree of importance of different paths and/or activities on which these delays occur. Proper consideration of this degree in fairly apportioning concurrent delays is essential Ibbs and Nguyen 2007a. Calculation of owner- and contractor-caused delay days is also questionable. For instance, the fact that the sum of excusable/ compensable delays and inexcusable delays can be greater than total project delays is difficult to accept in the industry. Hard Logic versus Soft Logic Relationships involving both hard and soft logic are a key element in project scheduling. Four factors that govern the sequencing of activities are physical relationships among project components, trade interaction, path interference, and code regulations Echeverry et al In addition, sequencing constraints can be flexible or inflexible Echeverry et al Accordingly, hard or fixed logic is network logic requiring an only link definition due to inflexible constraints while soft, preferential, or discretionary logic is network logic configured with more flexible constraints. Soft logic draws extensive research which mostly focuses on schedule updates. Logic change is inevitable and complicated when a schedule contains some soft logic. Soft logic in network scheduling is unfortunately typical because it represents some of the more subtle connections between any two activities and thus 484 / JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT ASCE / JULY 2008

3 makes the schedule more subjective and less apparent to parties other than the schedule creator. Several models have been proposed to handle the soft logic in schedule updating i.e., Tamimi and Diekmann 1988; El-Sersy 1992; Hanks 1999; Fan et al. 2002; Fan and Tserng The impact of soft logic on the project duration and critical paths is also significant Wang No schedule analysis technique easily addresses the impact of logic change on delay responsibility, and some ignore it completely. Such an impact may be considerable since the logic change is caused by different sources that can be ultimately traced to the contractual parties. Before applying the delay events of the next time period, the isolated delay type IDT technique Alkass et al only requires incorporating any changes to the as-planned schedule logic that occurred beyond the previous time period. This can be insufficient and inappropriate, as discussed next. Fig. 1 b illustrates the effect of logic change on delay responsibility. The as-planned schedule is 10 weeks. The project is delayed 2 weeks. At Week 1 there is a 2-week owner-caused delay on Activity A. Until this point the 2-week project delay is excusable and compensable. At Week 8 the contractor causes a 3-week delay on the same activity. This would delay the project for another 3 weeks. However, the contractor changes the soft logic of Activities B and C from finish-start FS to start-start SS. The result is no additional 3-week project delay. In this situation it would be unfair to conclude that the 2-week project delay is excusable and compensable given that the contractor delays Activity A more than the owner does. As this small example illustrates, logic change should be considered when assigning delay responsibility. Resource Allocation Resource allocation can also affect delay responsibility. The need for incorporating resource allocation in schedule delay analysis has been known for years. Pinnell 1992 suggests that the work plan in the form of a bar chart or network diagram should be resource loaded. An ideal delay analysis method should consider real resource allocation profiles Mohan and Al-Gahtani Until recently though schedule analysis explicitly and separately incorporated resource allocation. Ibbs and Nguyen 2007b propose steps to enhance window analysis by incorporating resource analysis inherently in the delay calculation. Among other things, they include the possible extended effect of delays due to changes in resource allocation and the positive/negative effect of resource allocation on delay responsibility. Dynamics of Float, Logic, and Resource Allocation The previous sections demonstrate that float and its ownership, logical sequences, and resource allocation really affect delay responsibility. These three issues are discussed separately. To improve the reliability of schedule analysis, they clearly should be considered. Whether they should be treated discretely or jointly in schedule analysis needs to be further considered. Our premise is that float, logic, and resource allocation actually have interrelationships that require them to be considered in an integrated fashion in any schedule analysis. Resource leveling is traditionally neglected in the calculation of float Householder and Rutland Nevertheless a noncritical activity may be resource critical because it will extend project duration if it does not release resources on time Fondahl In addition the use of soft or preferential logics, artificial activity durations, or constraints can sequester total float Prateapusanond Thus fair float ownership specification also requires nonsequestering of float. Schedule analysis should therefore address the dynamics of float, logic, and resource allocation in an integrated manner. Fig. 1 c depicts the dynamics of float, logic, and resource allocation in schedule analysis. The as-planned project duration is 9 weeks with four activities A, B, C, and D. The maximum allowable number of workers on this site is ten. At Week 5 the owner issues a change order that extends Activity D 3 weeks. The project would not be delayed since the change order only consumes float of Activity D. However the required number of workers during weeks 6 8 would be 12, which exceed the allowable allocation of workers. To accommodate this problem the contractor has to reschedule Activity C by removing the FS logic between A and C and adding FS logic between D and C. This logic change delays the project 2 weeks. Consequently, the change order does not simply consume time float but alters the schedule s downstream logic and resource allocation and delays the project. Forensic schedule analysis should capture this dynamics properly to provide a more reasonable result. FLORA attempts to fulfill this need. New Forensic Schedule Analysis Technique As a new time impact analysis technique FLORA addresses the dynamics of float, logic, and resource allocation in its analyses. It considers ownership and use of float in apportioning delay responsibility. Float is shared based upon a prior agreed-to basis between the owner and contractor. For instance, the owner and the contractor may mutually agree that each owns half of the total float available on any activity Prateapusanond FLORA evaluates not only the direct impact of a delay on project schedule but also its secondary effect. The secondary effect of a delay may be a mandatory logic change and/or the disturbance of resource allocation in downstream activities caused by the corresponding delay. FLORA uses a set of general rules, called FLORA s rules, for time impact analysis Table 1. These ten rules are flexible and enable contractual parties to customize them to fit into a specific context. Most are straightforward. For instance, Rule 4 follows the principle of the total float entitlement in Al-Gahtani and Mohan That is, the responsible party will be discredited any change of total float on the affected activity and gain or lose in the total float of successor activities Al-Gahtani and Mohan Rule 5 is codification of the current general practice that the owner will grant the contractor a time extension if there is a third party-caused delay and the owner will gain or lose total float for excusable and noncompensable delays. Due to the flexibility of the FLORA s rules, however, the project parties may agree to assign any change in total float for excusable and noncompensable delays to the contractor. Fig. 2 illustrates the decision logic of FLORA for forensic schedule analysis. FLORA can apply to either real-time or after-the-fact analysis. A real-time analysis activates when a delay event or a logic change occurs while an after-the-fact analysis is performed for each delay event or logic change in chronological order. FLORA first defines the baseline schedule by following Rule 1. It then allocates total float of all activities in the baseline schedule to the owner and contractor based on the second rule. If a delay event or logic change occurs a primary analysis and secondary analysis, if necessary, will start. Rules 3 8 will be applied in these analyses. Assigning project delay days to the JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT ASCE / JULY 2008 / 485

4 Table 1. FLORA s Rules for Time Impact Analysis Rule Description 1 Real-time analysis chooses the as-planned schedule as a baseline schedule. After-the-fact analysis develops a baseline schedule based on the as-planned and as-built schedules after changing errors found in the as-planned schedule. 2 Total float of each activity is shared between the owner and contractor, namely owner s total float TF o and contractor s total float TF c, based on the agreed basis e.g., 0 100, 50 50, Total float of new activities, which are added later to project schedules, will also be shared in the same manner. 3 An analysis may cover the whole time span of a delay event or logic change. If two or more delays occur in the same timeframe, the analysis in this overlapping timeframe will include all of these delays. 4 If the owner or contractor causes a delay or acceleration event, any increase decrease in the total float of an activity will add to deduct from the responsible party s total float of the corresponding activity. 5 If the third party causes a delay event such as force majeure, any increase decrease in the total float of an activity will add to deduct from the owner s total float of the corresponding activity. 6 Any increase decrease in the total float of an activity due to an approved logic change will be shared based on the agreed basis. 7 Float is an expiring resource. A party may freely use the other party s total float if the other does not use it and this free ride does not subsequently cause project delay. Otherwise, the party has to hold delay responsibility for total float he/she has overused. 8 Any increase decrease in the total float of an activity due to the secondary effect of a delay or acceleration will add to deduct from the responsible party s total float of the corresponding activity. 9 Total float of an activity for a certain party will be increased accordingly if lately the consumption of this float contributes to project delays. 10 Any project delay or acceleration due to an approved logic change will be shared between the parties based on the agreed basis. responsible party and updating owner-owned and contractorowned total floats can be carried out simultaneously. If there is a delay event or logic change that has not been addressed, the analysis will continue. Otherwise final delay responsibility of each project party will be determined by summing all delay days he/she has caused in the above analyses. The following case study demonstrates the application of FLORA. Case Study The small project has eight activities and is planned to finish in 12 days. Fig. 3 depicts the as-planned schedule in the form of the linked bar charts. The as-planned schedule, which is also the baseline schedule Rule 1, does not present resource constraints explicitly for simplicity. Its owner and contractor agree to the FLORA s rules Table 1 without any modification. Rule 2 is also specified by equally shared total float of all activities. This equally shared float ownership is considered fair by many authorities Prateapusanond Accordingly, total float is distributed to owner s total float TF o and contractor s total float TF c Fig. 3. Finally, total float and project schedule changes due to an approved logic change are equally shared between the owner and the contractor. Several delay events occur during construction. FLORA with its rules Table 1 and process Fig. 2 helps apportion delay responsibility between the owner and the contractor. Real-time forensic schedule analysis under FLORA will apply to this case. Results of the window analysis technique are also given for com- Fig. 2. FLORA process flowchart for real-time analysis Fig. 3. As-planned schedule 486 / JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT ASCE / JULY 2008

5 Table 2. Delay Events and Their Secondary Effects Day Description 2 The contractor delays 1 day on Activity A. To bring the schedule back as planned, the contractor changes some logical relationships by altering the relationship between Activities E and F from finish-start to start-start and adding a new finish-start relationship between Activity E and H. 4 The owner fails to allow Activity B to proceed on time. Activity B now takes 4 days. 5 The contractor fails to mobilize resources to start Activity B until Day 6. 5 and 6 The owner does not respond to the request for information RFI on Activity C timely. This inaction delays activity C 2 days. 6 The contractor stops the work on Activity D without any reasonable excuse. 7 and 8 Activity D continues being stopped due to inclement weather. 10 and 11 The owner makes a change order which extend Activities E and G 2 more days. Activity F requires a lot of workers to finish on its last day Day 12. This means that Activities E, F, and H cannot be performed concurrently on Day 12 since the contractor is unable to send adequate workers in such a fast change and notice. As a result, the contractor has to temporarily stop Activity G on Day 12 and restarts it as soon as Activities E and F finish. parisons. Table 2 summarizes delay events during the course of work. Day 2: 1-Day Contractor-Caused Delay on Activity A Activity A is delayed 1 day by the contractor on Day 2. This delay would extend the project 1 day, from 12 to 13 days. To recover this 1-day delay the contractor changes the soft logic between Activities E and F from finish-start to start-start and adds a new logic finish-start between E and H. These changes ensure the project completion in 12 days. Thus by using a delay analysis technique like window analysis the contractor has no responsibility for his/her delay on Activity A. Fig. 4 illustrates the analyses for this delay and the corresponding logic changes using FLORA and Rules The TF column shows any difference in total float that an activity has after and before the occurrence of the corresponding event and analysis. For instance, the TF in Fig. 4 a is determined by subtracting the total float of an activity after the delay on Activity A occurs the schedule in Fig. 4 a from that of the same activity before the delay on Activity A occurs the baseline schedule on Fig. 3. Fig. 4 shows two analyses. The first analysis is the direct impact of the delay on project schedule Fig. 4 a. The project is delayed 1 day for which the contractor is responsible. This delay does not change total float of any activity. Next, the contractor has to revise some construction sequencing as a result of this delay. Fig. 4 b portrays the secondary analysis. These changes of the relationships help accelerate the project 1 day. Total floats of several activities are changed as well. Specifically, the total floats of Activities B and E are increased 1 day while those of Activities C, D, and G are decreased 1 day. FLORA s Rule 6 ensures these changes in total float are shared between the owner and the contractor. Similarly, the 1-day project acceleration due to the logic changes is equally shared between the two parties Rule 10. That is, each party accelerates 0.5 day or delays 0.5 day. In sum, with the 1-day delay on Activity A at Day 2 and its secondary effect, the contractor is responsible for 0.5 delay day while the owner is responsible for 0.5 delay day. This result is different from the one derived from the window analysis previously mentioned. Importantly, the secondary analysis can also be considered an independent analysis without affecting or changing the results of delay responsibility. Day 4: 1-Day Owner-Caused Delay on Activity B The owner delays 1 day on Activity B at Day 4 Fig. 5. This delay does not delay the project since Activity B is a noncritical activity. Instead it consumes the whole 1-day total float of this activity and causes Activities B and E to become critical. Activity E is now critical because the early start of Activity E cannot delay unless F and, hence the project, are delayed. Following FLORA s Fig. 4. Analyses for contractor-caused delay on Activity A at Day 2 Fig. 5. Analysis for owner-caused delay on Activity B at Day 4 JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT ASCE / JULY 2008 / 487

6 Fig. 6. Analysis for concurrent delays on B and C at Day 5 Fig. 7. Analysis for concurrent delays on C and D at Day 6 Rule 4 the owner will be responsible for this decrease in total float of Activities B and E. For that reason, TF o of these activities will be deducted, from 0.5 to 0.5 day. The owner has no responsibility at the moment since this delay event does not cause any project delay. In terms of delay responsibility, a window analysis would derive the same conclusion. Day 5: 1-Day Concurrent Delays, Contractor, and Owner Caused, on Activities B and C Day 5 experiences concurrent delays. The contractor causes a delay to Activity B while the owner delays Activity C. Although the owner delays Activity C by 2 days Days 5 and 6, Rule 3 dictates that these 2 days be analyzed separately. The project is extended 1 day from 12 to 13 days for the events occurring until Day 5. Fig. 6 shows the analysis of the concurrent delays. The project is delayed 1 day. Notably, total float of Activities B and C is zero in the updated schedules on Day 4 Fig. 5 and Day 5 Fig. 6. In other words, both Activities B and C are critical before and after the concurrent delays on Day 5 occur. Each single delay event would have caused project delay if the other had not occurred. As such, both contractor and owner are responsible for this 1-day project delay. Window analysis would yield 1-day concurrent delays, where the contractor is typically granted a time extension but not delay damages. FLORA goes an extra step in this scenario. Activity B is delayed by 2 days, 1 day by the owner at Day 4 and the other day by the contractor at Day 5. While the delay event at Day 4 does not directly result in the project delay as analyzed above, it contributes to the 1-day project delay in the present analysis. The contractor-caused delay on Activity B at Day 5 would not have made Activity B critical if the owner-caused delay on Activity B at Day 4 had not existed. It is therefore unfair to neglect this owner-caused delay in the current analysis. Float ownership plays a role in this apportionment. The contractor has owned 0.5 day of total float of Activity B until Day 5 while the owner overuses 0.5 day of that due to his/her 1-day delay at Day 4 Fig. 5. Thus the contractor only overuses 0.5 day of total float of this activity due to her 1-day delay at Day 5. Following Rule 7, the owner has to be held responsible for the total float she has overused. That is, the 1-day delays of the owner on Day 4 and the contractor on Day 5 on Activity B each include a 0.5 day consuming total float and a 0.5 day causing the project delay. Therefore, together with the owner-caused delay on Activity C at Day 5 FLORA divides the 1-day project delay in the current analysis at Day 5 into a 0.5 day of the concurrent delays and 0.5 day of the excusable and compensable delay. This result differs from the window analysis result, which treats the whole 1-day project extension as a concurrent delay. In addition, TF c and TF o of Activity B will be increased 0.5 day from 0.5 to 0 based on Rule 9 Fig. 6. Day 6: 1-Day Concurrent Delays, Owner, and Contractor Caused, on Activities C and D Concurrent delays also occur at Day 6. The owner continues delaying Activity C. The contractor delays Activity D on the same day. Consequently, the project is delayed 1 day Fig. 7. The delay on Activity D however does not cause the project delay. This contractor-caused delay only consumes total float since the contractor owns 1.5 of total float of Activity D before the current analysis at Day 6. TF c of Activity D is deducted from 1.5 to 0.5, which is still positive. In contrast, the owner-caused delay on Activity B solely extends the project 1 day. That is, the 1-day project delay is an excusable and compensable delay. Window analysis would provide the same result. Total floats TF, TF c, and TF o need updating. The ownercaused delay on Activity C increases the total float of Activities B and E. Rule 4 allocates the increase to TF o for the same activities. TF o of B and E gains 1 day from 0 to 1 and from 0.5 to 0.5, respectively Fig. 7. Although TF of Activity D does not change, TF c and TF o of this activity are changed. This can be explained by the fact that the contractor-caused delay on Activity D consumes total float while the owner-caused delay on Activity C adds to total float of Activity D. This cancels out the change in total float of Activity D. Based on Rule 4, however, 1 day is shifted from TF c to TF o Days 7 and 8: 2-Day Third Party-Caused Delay on Activity D Unexpected inclement weather delays Activity D at Days 7 and 8. The project completion date is not affected by this delay Fig. 8. That is, the delay only consumes the total float of Activity D. Fig. 8. Analysis for third party-caused delay on D at Days 7 and / JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT ASCE / JULY 2008

7 Fig. 9. Analysis for owner-caused delays on E and G at Days 10 and 11 Following Rule 5, TF o of Activity D is deducted 2 days due to this float consumption. Window analysis would also yield no critical delay in the period of Days 7 and 8. Days 10 and 11: 2-Day Owner-Caused Delays on Activities E and G The owner issues a change order that extends Activities E and G 2 more days Table 2. Fig. 9 a illustrates the direct impact of these delays on the project schedule. The project is delayed 1 day, from 14 to 15 days. This is an excusable and compensable delay. Window analysis for the same time period would give the same result. The delays also cause changes in total float of Activities D, F, G, and H. The owner s total float of Activity F and H will gain 1 day while that of Activity G will lose 1 day Rule 4. It should be noted that TF c and TF o of Activities B and D become zero because these activities completely finish at the current analysis. This follows Rule 7 that treats total float as an expiring resource. The secondary effect of these delays is the infeasibility of the planned resource allocation. As described in Table 2, Activity F requires an excessive number of workers to finish on its last day Day 12. The contractor has to temporarily stop Activity G for 1 day and restart at Day 13 when E and F are completed. As a result the project is delayed 1 more day. If the dynamics of logic and resource allocation are not considered, which is the case in traditional window analysis, this 1-day project delay would be an inexcusable delay. FLORA provides the secondary analysis for this infeasible resource allocation and changed logic situation Fig. 9 b. Two new relationships are added as dotted arrows. This secondary analysis demonstrates that the owner is responsible for this additional 1-day project delay. In other words, this delay is excusable and compensable instead of inexcusable, as computed by the traditional window analysis. Table 3 summarizes the results of FLORA and the window analysis technique. From the 4-day project delay, window analysis would show 1-day inexcusable, 1-day concurrent, and 2-day excusable, and compensable delays. Half-day inexcusable, 0.5- day concurrent, and 3-day excusable and compensable delays are indicated by this application of FLORA. Each single analysis may also yield different or even conflicting outcomes. This confirms that project progress factors play a significant role in forensic schedule analysis. The differences between the two results derived from FLORA and the window analysis technique are significant. The 1-day difference of excusable and compensable delays is really considerable given that the project is only delayed 4 days. This difference leads to the change in damages paid recovered by the owner the contractor. Other differences in the results of inexcusable delays and concurrent delays also affect resultant damages. In a single analysis, at Day 5 for instance, FLORA yields 0.5-day excusable and compensable and 0.5-day concurrent delays while the window analysis technique does 1-day concurrent delays. A shift in a 0.5 day from concurrent delay to excusable and compensable delay apparently changes the associated damages. This is because the contractor is typically granted time extension only for concurrent delays, whereas he/she receives delay damages for excusable and compensable delays. As such, the outcomes of delay claims and disputes are impacted by the project progress factors. Discussion FLORA solves various issues in forensic schedule analysis. By capturing the dynamics of float, logic, and resource allocation, it also helps solve other schedule analysis dilemmas such as pacing delays, acceleration, concurrent delays, sequestering of float, fair treatment of noncritical activities, and real-time analysis. A pacing delay is the deceleration of the work by the contractor owner Table 3. Summary of Forensic Schedule Analysis Analysis/window date Type of delays day Project duration day Excusable/compensable Excusable/noncompensable Inexcusable Concurrent a Total a Results of FLORA window analysis technique. JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT ASCE / JULY 2008 / 489

8 due to a delay to the end date of the project caused by the owner contractor to maintain balanced progress with the updated project schedule Zack Zack 2000 notes that pacing delays relieve the owner contractor of some delay damages it otherwise may have owed to the contractor owner since they can cause concurrent delays and/or float consumption. FLORA indirectly considers this issue in analyses because its rules clearly address float ownership and consumption. Additionally, the rules weigh acceleration as equally important as delay. That is why we call FLORA a new time impact analysis or forensic schedule analysis technique rather than a delay analysis technique per se. The proper treatment of logic change in such analyses enables FLORA to deal with any sequestering of total float in project schedules. Prior approved float sharing and lucid rules for float consumption helps treat noncritical activities fairly. Finally, FLORA can work for both real-time and after-the-fact schedule analyses. The rules of FLORA are flexible. A flexible and more accurate delay analysis technique is valuable Alkass et al Project parties may follow a certain view of float ownership and allocation in their specific project as long as the parties agree to that view. That is, total float can be owned by the owner or the contractor or shared between the two Rule 2. Instead of being assigned to the owner, changes in total float due to third-partycaused delays can alternatively be assigned to the contractor or shared between the parties Rule 5. In addition, total float changes and delay or acceleration due to approved logic changes can be shared based on any ratio rather than equally shared between the project parties Rules 6 and 10. In addition to inheriting strengths of available delay analysis techniques FLORA is by and large more advantageous. The window snapshot and traditional time impact analysis techniques do not directly capture the impacts of float ownership, logic changes, and resource allocation on delay responsibility. The comparison in the above case study is evident. The isolated delay type technique Alkass et al does not deal with logic changes properly. The fact that the IDT technique incorporates delays in one shot in each window period is not practical Mohan and Al- Gahtani As previously discussed, the preallocation of total float Prateapusanond 2003 is unrealistic since the critical path s and, hence the total float of each activity, can change during the course of work. A total float management technique Al-Gahtani and Mohan 2007 only addresses float ownership and not the other problems discussed above. FLORA has several weaknesses. It is somewhat more complicated than window analysis. Window analysis however becomes arduous if the window sizes are set to small time periods to gain more accuracy. By incorporating the secondary effect of delays, FLORA requires project records about logic changes and resource allocation together with delay and acceleration events. Fortunately, these records can be readily obtained if the project team updates and documents project progress well. Finally, its analysis takes more effort since schedule delay or acceleration and the impacts on different types of total floats TF, TF, TF c, and TF o must be computed. Conclusions Various factors affect the results of delay analysis. Different views of float, float ownership, logic change, concurrent delays, resource allocation, etc. may lead to different results. They should therefore be considered in schedule analyses to ensure more reliable outcomes. Current delay analysis techniques tend to overlook most if not all of them. While achieving modest success, recently proposed techniques try to incorporate some of these factors. They mainly deal with concurrent delays and float ownership. FLORA addresses various issues that remain unsolved and/or neglected in forensic schedule analysis. Its analyses effectively capture the dynamics of float, logic, and resource allocation. It can be used for either real-time or after-the-fact analysis. The analysis processes are based on ten rules, which are flexible and customizable. A case study is used to illustrate its application. FLORA may yield different results compared to other available schedule analysis techniques like window analysis because it is more inclusive of project progress factors. By properly dealing with the issues of schedule analysis, FLORA can be more reliable. Finally, the parties may find the results of its analysis less contentious since they specified and agreed to FLORA s rules for schedule analysis in advance. References Al-Gahtani, K. S., and Mohan, S. B Total float management for delay analysis. Cost Eng., 49 2, Alkass, S., Mazerolle, M., and Harris, F Construction delay analysis techniques. Constr. Manage. Econom., 14 5, Arditi, D., and Pattanakitchamroon, T Selecting a delay analysis method in resolving construction claims. Int. J. Proj. Manage., 24 2, de la Garza, J. M., Prateapusanond, A., and Ambani, N Preallocation of total float in the application of a critical path method based construction contract. J. Constr. Eng. Manage., , de la Garza, J. M., Vorster, M. C., and Parvin, C. M Total float traded as commodity. J. Constr. Eng. Manage., 117 4, Echeverry, D., Ibbs, C. W., and Kim, S Sequencing knowledge for construction scheduling. J. Constr. Eng. Manage., 117 1, El-Sersy, A. H An intelligent data model for schedule updating. Ph.D. dissertation, Univ. of California, Berkeley, Calif. Fan, S. L., and Tserng, H. P Object-oriented scheduling for repetitive projects with soft logics. J. Constr. Eng. Manage., 132 1, Fan, S. L., Tserng, H. P., and Wang, M. T Development of an object-oriented scheduling model for construction projects. Autom. Constr., 12 3, Fondahl, J. W The development of the construction engineer: past progress and future problems. J. Constr. Eng. Manage., 117 3, Gong, D Optimization of float use in risk analysis-based network scheduling. Int. J. Proj. Manage., 15 3, Hanks, D. R Soft logic An overview. Cost Eng., 41 2, Hegazy, T., and Zhang, K Daily windows delay analysis. J. Constr. Eng. Manage., 131 5, Householder, J. L., and Rutland, H. E Who owns float? J. Constr. Eng. Manage., 116 1, Ibbs, W., and Nguyen, L. D. 2007a. Alternative for quantifying fieldoverhead damages. J. Constr. Eng. Manage., , Ibbs, W., and Nguyen, L. D. 2007b. Schedule analysis under the effect of resource allocation. J. Constr. Eng. Manage., 133 2, Kim, Y., Kim, K., and Shin, D Delay analysis method using delay section. J. Constr. Eng. Manage., , Mbabazi, A., Hegazy, T., and Saccomanno, F Modified but-for method for delay analysis. J. Constr. Eng. Manage., , Mohan, S. B., and Al-Gahtani, K. S Current delay analysis techniques and improvements. Cost Eng., 48 9, Pasiphol, S., and Popescu, C Qualitative criteria combination 490 / JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT ASCE / JULY 2008

9 for total float distribution. AACE Int. Trans. of the Annu. Meet., The Association for the Advancement of Cost Engineering International AACEI, DCL Peterman, G. G Who owns float? Waerme Stoffuebertrag., 21 2, Pinnell, S. S Construction scheduling disputes: Proving entitlement. Constr. Lawyer, 12 1, pp Ponce de Leon, G Float ownership: specs treatment. Cost Eng., 28 10, Prateapusanond, A A comprehensive practice of total float preallocation and management for the application of a CPM-based construction contract. Ph.D. Dissertation, Virginia Polytechnic Institute and State Univ., Blacksburg, Va. Sakka, Z. I., and El-Sayegh, S. M Float consumption impact on cost and schedule in the construction industry. J. Constr. Eng. Manage., 133 2, Shi, J. J., Cheung, S. O., and Arditi, D Construction delay computation method. J. Constr. Eng. Manage., 127 1, Stumpf, G. R Schedule delay analysis. Cost Eng., 42 7, Tamimi, S., and Diekmann, J Soft logic in network analysis. J. Comput. Civ. Eng., 2 3, Wang, W. C Impact of soft logic on the probabilistic duration of construction projects. Int. J. Proj. Manage., 23 2, Zack, J. G Claimsmanship: Current perspective. J. Constr. Eng. Manage., 119 3, Zack, J. G Pacing delays The practical effect. Cost Eng., 42 7, Zack, J. G Delay and delay analysis: Isn t it simple? Proc., 1st ICEC and IPMA Global Congress on Project Management, International Cost Engineering Council, Ljubljana, Slovenia. JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT ASCE / JULY 2008 / 491