Using Probability Management to Predict Performance on Navy Military Construction Projects
Abstract
Naval Facilities Engineering Systems Command (NAVFAC) currently has no predictive tool to assess the cost and schedule growth of Military Construction (MILCON) projects. The current NAVFAC assessment of project performance is accomplished by tracking metrics to determine if project performance is degrading. This is a reactive process with lagging indicators and does not allow NAVFAC to proactively manage the performance of projects before degradation occurs. Utilizing historical data of NAVFAC executed MILCON projects, the project team applied statistical analysis techniques developed by Dr. Sam Savage to create a proof-of-concept data analytic tool. Using unique forms of Monte Carlo analysis, the project team developed a model that takes project parameter inputs of total cost and duration and predicts final cost, duration, and probability of meeting original targets. The project team used over 800 clean project data points from NAVFAC's ieFACMAN (Interoperable Enterprise Facilities Management) database to generate probability distributions. These probability distributions were then converted into Stochastic Information Packets (SIP) as inputs into Dr. Savages Microsoft Excel add-on, ChanceCalc, to generate the project teams data analytic model. The model indicates that for all NAVFAC MILCON, a new project has a 90 likelihood to be over budget and 88 likelihood to be beyond schedule. It is suggested that further analysis and modeling include project parameters with intact relationships (e.g., the relationship between cost and schedule) and expanding available data from other enterprise systems to improve predictive factors and analytic tool capabilities.
Document Details
- Document Type
- Technical Report
- Publication Date
- Apr 01, 2022
- Accession Number
- AD1170401
Entities
People
- Karl Coulson
- Scott Sobieralski
- Tim Dahms
Organizations
- Naval Facilities Engineering Systems Command