Better Construction Project Management through Better Scheduling

2018Final ReportPDFTech ReportFHWA-SC-10-03SPR No. 721Best practicesConstruction managementCritical path methodProject managementSchedulingUncertaintyCritical Path MethodProjectDelayUnited StatesFederal Highway Administration (U.S.)Mullen, Robert LHuynh, Nathan N.Ahmed, FahimPadmanabhan, BhavyaUniversity of South CarolinaSouth Carolina. Dept. of Transportation. Office of Materials and ResearchFederal Highway Administration Resource Center (U.S.)US Transportation CollectionThis project investigated the effect of using Critical Path Method (CPM) schedules on South Carolina transportation projects. The analysis used 2,097 projects let after February 2007 and substantially completed by August 2015. Among these projects, 55.22% had CPM schedules. Among the projects without CPM schedules, 47.49% were delayed beyond the original contract completion date, and 16.08% were delayed beyond the date adjusted by change orders; the average delay beyond the original contract completion date was 28.5 days, and the average delay beyond the adjusted completion date was \u201314.6 days (i.e., they were completed 14.6 days before the adjusted completion date). Among the projects with CPM schedules, 54.32% were delayed beyond the original contract completion date, and 14.50% were delayed beyond adjusted completion date; the average delay beyond the original contract completion date was 54.8 days, and the average delay beyond the adjusted completion date was \u20139.8 days. Chi-square test results indicated that projects with larger bid amounts and longer durations were more likely to have a CPM schedule. The t-test results indicated that the Bridge, General, and Hot-mix Asphalt Paving projects with CPM schedules had a statistically significant longer average delay beyond the original contract completion date than projects without CPM schedules. Conversely, the t-test results indicated that the short duration (less than six months) projects without CPM schedules had longer average delay beyond the adjusted completion date than projects with CPM schedules. It should be noted that for the SCDOT, projects with CPM schedules are those with a significant amount of risk and complex scope whereas projects without CPM schedules are those with low risk and limited scope. A binary logistic regression model was developed to predict project delay and this model indicated that the variable \u201ctotal bid amount\u201d was statistically significant at the 95% confidence level. A neural network model was also developed using change order remarks and texts from daily work reports as inputs. This model correctly predicted a project will be delayed 53.7% of the time. Best-fit distributions were developed for delayed and early completion projects. Using these distributions and the Total Law of Probability, the probability that a project will be delayed can be estimated. Based on the literature view and SCDOT current practice, it is recommended that the SCDOT considers using a more methodical approach that examines multiple criteria for identifying those projects that need CPM schedules. Additionally, the SCDOT may consider incorporating the probabilistic information about project delay in the schedules, such as providing a range for the activity duration in the CPM schedule instead of simply using the expected duration. Lastly, it is recommended that the SCDOT consider including a provision in future contracts to ensure that all personnel (SCDOT and contractors) are using the same schedule for project management.85