Cost overruns in projects have historically been a problem spread over different sectors. Generating accurate estimates early in the project life cycle is difficult. However, uncertainty in cost estimates can be assessed. This report sought to analyze the consequences of neglecting correlation between cost estimates in the uncertainty analysis performed at WSP Management Stockholm. The uncertainty analysis consisted of a model using Monte Carlo simulation with PERT-distributed three-point estimates from expert judgement as input. 7 Swedish projects, 4 in infrastructure and 3 in construction, with a total expected cost of 7.1 billion SEK were analyzed; all in early stages of the project life cycle. Estimated costs were correlated successively in 3 different correlation scenarios with correlation coefficients 0.3, 0.7 and 1; later compared with the uncorrelated scenario. Results showed that: Each scenario led to an increase in the result’s standard deviation of 68%, 129% and 168% in average for correlation coefficients 0.3, 0.7 and 1 respectively. Furthermore, if the 85th percentile in the resulting cumulative distribution was to be used to determine the cost frame of the project, assuming no correlation in a scenario with the 10 most uncertain cost estimates correlated would lead to an underestimation of the final cost with 3%, 6% and 8% for correlation coefficients 0.3, 0.7 and 1 respectively. Lastly, if standard deviation in the results is low there is a reason to question the results. Estimating correlation in the model accounts for a wider range of outcomes and may be more realistic in an uncertain reality. Furthermore, correlation can compensate for the uncertainty that is lost in calculations as estimates are broken down into more detail.
| Identifer | oai:union.ndltd.org:UPSALLA1/oai:DiVA.org:uu-355223 |
| Date | January 2018 |
| Creators | Lauritzen, Artur |
| Publisher | Uppsala universitet, Institutionen för fysik och astronomi |
| Source Sets | DiVA Archive at Upsalla University |
| Language | Swedish |
| Detected Language | English |
| Type | Student thesis, info:eu-repo/semantics/bachelorThesis, text |
| Format | application/pdf |
| Rights | info:eu-repo/semantics/openAccess |
| Relation | UPTEC ES, 1650-8300 ; 18 025 |
Page generated in 0.0022 seconds