CPSP 134 Update of Mathematical and Statistical Distributions Modelling of Estimates

Highways projects are highly complex and are delivered in a dynamic environment that is constantly exposed to different risks and uncertainties that need to be understood by the various stakeholders involved. There are a plethora of risks and uncertainty due to which the actual cost exceeds the budgeted / estimated amount. If the RIS funding is less than the actual cost of the RIS, Highways England are either unable to meet their efficiency targets or have to borrow more from the government. This results in loss of confidence in Highways England's ability and puts future RIS funding at risk. Hence, it is very important to have robust cost estimates taking into account comprehensive foreseeable risks and uncertainties. The uptake of mathematical models in risk/uncertainty and contingency analysis of project costs in the Highways construction industry has been slow as compared to other industries. Highways England since 2007 have been using mathematical models and Monte Carlo simulations to produce probabilistic range estimates. The estimating process has not seen any fundamental change since 2007. There is a need to review the current mathematical model deployed in delivering estimates to identify any gaps and update it to bring it abreast of latest developments in mathematical modelling application to estimating.  

2.2 Scope The main aim of this project is to incorporate use of explicit, empirically mathematical and statistical distributions or models of cost elements to produce more realistic forecasts of individual project estimates by reducing what has been termed project optimism bias i.e. the demonstrated systematic tendency of project appraisals and estimates to be overly optimistic. Highways England wishes to review existing approach and update the process of simulation based cost estimation using Monte Carlo techniques and mathematical modelling in line with current best practice used in cost estimation.

The scope will include but not limited to

- Review of current mathematical and statistical distribution models used in Highways England estimates - Production of a report on what improvements are required to align Highways England estimating with current mathematical & statistical analysis best practice in cost estimation - Identify a model and associated process / procedural documents to implement the updated and revised approach for the review and approval of Highways England. This may include: o Assessment of the potential use of a modified RET to produce and probabilistic output and EMV for each individual cost element o Review and assessment of the appropriate correlation between series / cost elements and update of RET Determine the effect of correlations or interdependencies between cost elements on the estimate of the total project costs o Review of the levels at which simulation is best to take place o Identification of any gaps and proposals to address these areas - produce report for issue to EM / HoCP o The actions requi