please get one sentence the most take away and one-sentence criticism from this article Planning-stage estimation of hig

[answerhappygod]

please get one sentence the most takeaway and one-sentencecriticism from this article
Planning-stage estimation of highway project duration onthe basis of anticipated project cost, project type, and contracttype
Abstract
At the planning phase of project development, highway agenciesseek to estimate the time duration of project implementation forpurposes such as construction planning, contract administration,and workzone impact assessments. This paper investigates theestimation of highway project duration on the basis of variablesknown at the planning phase such as planned cost and project type,and contract type. Project types are pavement construction,rehabilitation, maintenance, traffic facility installation, andbridge construction. The contract types considered arefixed-duration and fixed-deadline contracts. Using a variety ofmodel specifications, the paper presents mathematical relationshipsbetween highway project duration and the magnitude of the plannedcost and project type, and contract type. The paper shows that allother factors remaining the same, the duration of fixed-datedeadline contracts generally exceed that of fixed-durationcontracts; and higher levels of planned cost translate non-linearlyinto greater project duration. The developed models can helpproject administrators provide improved estimates of projectduration and thus could help reduce project time delays. Also, themodels are useful in forecasting workzone durations, a key aspectof workzone user cost estimation.
1. Introduction
1.1. Background
Establishing realistic time duration at the planning phase ofproject development can help an agency meet public expectations ofproject duration and thus avoid the public relations problemsassociated with time overruns. Also, knowing the expected durationof a project can be useful in bid evaluation: in cases wherebidders are asked to specify construction periods, the engineer’sestimate of expected project duration can be used as one of the bidevaluation criteria. Furthermore, knowing the expected projectduration, the user costs of delays and safety associated with aproject implementation can be estimated for that project and canserve as an input for life-cycle cost analysis. The true durationof a project is known only after completion of the project, and isinfluenced by aberrations such as change orders, worker unrest, andmaterial shortages (Arditi et al., 1985, Kraiem, 1987, Majid andMcCaffer, 1998). A recent study in Zambia determined that theduration of road construction projects are influenced by delayedpayments, contract modification, economic problems, materialsprocurement, changes in specifications/drawings, poor supervisionand coordination on site, and construction mistakes (Kaliba et al.,2009). However, at the planning phase where such problems have notyet occurred, the variables associated with these root causes areunknown (Fig. 1). At the project planning phase of a project,agencies seek to predict the duration of a project implementationon the basis of data available at planning stage – project type,intended contract type, estimated cost at the planning phase or theproject award cost, etc. To provide agencies with such a model,this paper uses actual (final) durations and actual (final) projectcosts of past projects. To be consistent with standard modelingpractice, the use of the actual data on project duration and costare expected to yield more practical relationships compared tousing planning stage estimates of project duration and also ofproject cost. However, for purposes of applying the model topredict actual duration, actual cost is not available and henceplanned cost (that is, estimates of project cost at the planningphase) will need to be input into the model. Stoy et al. (2007),for example, provided estimates of building construction rate onthe basis of project size and climatic condition in Germany.
The estimation of highway project duration can therefore be madeon the basis of past experience or using historical data fromsimilar projects in similar contractual circumstances. Agenciesthat lack reliable models to predict time duration are likely toresort to the use of simple averages of time duration (thusimplicitly assuming a uniformly distributed duration function overvarious ranges of the project type and contract type variables)(Federal Highway Administration (FHWA), 1998, Lamptey et al.,2008). For example, using the averages, it may be estimated thatfive project duration days are needed for each lane-km or for each1000 dollars of pavement construction. However, for a 10 lane-km or$10,000 project, the actual project duration could be significantlyhigher or lower than 50 days.
The estimation of project duration is beneficial in variouscategories of construction management practices such asconstruction planning, contract administration and bid evaluation.Given the estimated project duration, the user cost of delays andsafety associated with a project implementation can be estimatedfor that project. Note that user cost, through detours orreduced-capacity workzones, is a direct function of projectduration. User cost is a key input for life-cycle cost analysis andmultiple criteria analysis of alternative schedules for assetrehabilitation or (re)construction. Better estimation of contractduration can facilitate agency engineers or contractors to specifymore reliable contract periods. This can help reduce time delay ofprojects and thus help avoid the negative publicity and politicalfallouts associated with project delays. Also, in cases wherebidders are asked to specify construction periods, the engineer’sestimate of expected project duration can be used as one of the bidevaluation criteria.
Project durations, like any other output of engineering systemmanagement, are not expected to be exact. This is because therewill always exist uncertainties in the system environment, such asmaterial imperfections, weather variations, economic uncertainties(inflation or depression), equipment malfunction, labor unrest,human error or incompetence. The introduction of stochasticelements into a duration model is therefore often necessary becausethey transform the model from an exact statement to a probabilisticdescription about expected outcomes. Thus, the probabilistic modelis less precise but more robust (Greene, 2000).
This paper reviews existing literature on the prediction ofproject duration and identifies the mathematical forms of modelsthat have been developed for this purpose. The paper then describesthe source data, presents preliminary plots of the trends, anddiscusses the methodology used for developing models to estimateproject duration as a function of project cost, project type, andcontract type. Using highway project and contract data from theState of Indiana, hazard-based duration models are developed. Inthis paper, the highway project types considered are: pavementconstruction, pavement rehabilitation (resurfacing), pavementmaintenance, traffic facility installation, and bridgeconstruction; the contract types are: fixed-duration contracts(where a fixed number of days or “available days” is specified) andfixed-deadline contracts (where a fixed completion date isspecified). At certain agencies, the common terminology for projectduration specifications is “calendar days” and “work days”. In theformer, weather delays, and/or holidays are included. To ensurethat the models perform satisfactorily in their predictivecapabilities, this paper carries out validation of the developedmodels using project data from Indiana that were not used in themodel development and also cross-validates the models usingcontract data from other states. Finally, the paper presents anexample to illustrate how an analyst could use the models topredict the project duration of a project on the basis of data thatare typically available at the planning phase.
7. Summary and conclusion
The paper developed models to estimate the expected duration ofa highway project on the basis of estimated project cost at theplanning phase, project type, and contract type. However, inreality, actual cost could also be a function of planned or actualtime duration, and future work could examine possible endogenous orsimultaneous equation relationships between the planned or actualcost on one hand and the planned or actual duration on the otherhand. The project types are pavement construction, pavementrehabilitation (resurfacing), pavement maintenance, trafficfacility installation, and bridge construction; contract types arethose based on specification of a fixed number of days for theproject and those based on a fixed-date deadline. The models inthis study were developed using data from the state of Indianaspanning the years 1996–2001.
In its introductory sections, the paper duly recognizes that theactual duration of projects is typically a function of “root”causes that occur during the construction process (such as designerrors, site conditions, and labor troubles) and planning-phasevariables that are known at the planning phase (such as projecttype and estimated cost). At the project planning phase, agenciesseek to predict the duration of a project implementation on thebasis of data available at planning stage – planned project cost,project type, intended contract type, etc. To provide agencies withsuch a model, this paper uses actual (final) durations and actual(final) project costs of past projects. To be consistent withstandard modeling practice, the use of the actual data on projectduration and cost are expected to yield more practicalrelationships compared to using planned values of project durationand cost. However, at the planning phase, the actual cost is notavailable and hence planned cost will need to be input into themodel to predict actual duration.
Furthermore, there is the issue of interconnectivity betweenproject duration and cost that will need to be further investigatedin future research. The hypothesis that duration and cost areinterdependent can be studied by showing that their estimationsimultaneously yields a superior model than when they are estimatedseparately. This interdependency, if proven, could be attributed tothe fact that duration and cost are influenced by similarexplanatory factors such as bid price, level of bid competition,project size, and project type. In future research in this area,the first step could be to duly recognize that most existingduration or cost model types such as single-equation regressionequations, simply ignore the fact that many explanatory variablesare not truly exogenous, independent variables. Econometric theorysuggests that ignoring the other equations is not good modelingpractice because the single-equation OLS estimator will be biased.For example, it could be shown that the output of the durationequation is correlated with error term of the cost equation.Fortunately, there exist techniques to address simultaneity biasand thus to consistently estimate the coefficients in such a systemof simultaneous equations. The most common approach is called themethod of instrumental variables (IV). When several instrumentalvariables are available, they are combined via regression (thefirst stage) and then used in a second regression. In standardstatistical texts, this is referred to this as two-stage leastsquares (2SLS or TSLS).
Cross-validation of the models were carried out to ascertainthat they can be used to predict project duration for projectsoutside of Indiana. The modeling results suggest that for all theproject types, the time duration of a project increases withproject cost. This relationship is not only non-linear but alsoshows differences across the project types. It was also determinedthat for the domain of project costs and types given in this paper,projects with a fixed-date deadline generally have greater timeduration compared with those with a specified number of days, allother factors remaining the same.
The paper investigated the appropriateness of probabilisticmodel specifications by conducting Weibull analysis, a relativelyflexible modeling technique that yielded survivor curves and hazardfunctions for project duration. The probabilistic models enhancethe prediction of project durations by introducing stochasticelements into the duration model, thus transforming the predictionprocess from an exact deterministic statement to a robuststochastic description.
The marginal effects estimated in the paper can also help theagency to predict the effect of a change in contract type, projectcost, and other factors on project duration. In predicting theproject duration, the developed model (Eq. (1)) and not theelasticities, should be used.
The models developed in this study were validated using datafrom other states and the results were satisfactory. Also exampleapplications were provided to show how an agency could use thedeveloped models to predict the duration of their projects on thebasis of the project type, contract type, and project cost. Asevidenced from the satisfactory goodness-of-fit values obtained(i.e., χ2 and R2),the developed models explain a reasonably good amount of thevariance in the data.
The developed models can help in several management functionsincluding construction planning and management, projectadministration, and facility preservation planning. From thecontractors’ perspective, the models could serve as an empiricaltool at the letting stage for specifying an appropriate contractperiod as part of their bid while at the award phase. Also,contractors may use the models to derive more reliable projectduration estimates so that they can make appropriate plans andarrangements for equipment and labor utilization, and otherresource plans that are influenced by project duration. From theagency perspective, the models can be used by contract or projectadministrators to establish more reliable project implementationdurations and thus to specify more reliable construction periods.This would not only lead to reduced likelihood and magnitude oftime delays but would also lower the risk of disputes regardingdelay responsibility and liquidated damages charges. Anotherimportant application of the project duration models is theprediction of highway workzone duration and concomitant user costsassociated with the workzone. These user costs are an essentialcomponent of life-cycle costing for highway preservationevaluation; as such, the project duration models developed in thisstudy can help in assessing workzone impacts associated withalternative highway preservation treatments and life-cyclepreservation strategies. Furthermore, because workzones translateinto user costs of delay, safety, and traveler inconvenience, anyextended workzone durations due to unexpected contract prolongationcan lead to increased traffic delay and safety problems, road-userdissatisfaction, and political ramifications. At the current timewhen agencies seek to burnish their image as responsible stewardsof taxpayer-funded public infrastructure, they can hardly affordthe negative publicity associated with project time delays. Assuch, agencies are expected to embrace efforts geared towardreliable prediction of project duration at the planning phase ofproject development.