A Cost Estimation Risk Assessment Process Implementing Lurie-Goldberg’s Algorithm for Generating Correlated Vectors of Random Numbers
A true cost estimate is not a single value. A true cost estimate is a probability distribution of anticipated cost. There are many high-quality cost estimating tools in the marketplace that provide “point solution” estimates, but there are few tools that perform a rigorous risk assessment sufficient to develop an accurate probability distribution of estimated cost.
This paper describes a generic process and supporting tool used to compute an accurate probability distribution of anticipated cost. The risk assessment techniques described in this paper are generic. They can be applied to any estimating application, as long as that application provides an automated interface which allows the tool to manipulate the application’s elementary cost drivers and harvest the point solution estimates. For purposes of this paper, the SEER-SEM software estimation tool was used for the cost estimation engine.
The basis of the process is a series of Monte-Carlo simulations: the first simulation iterating over the elementary cost drivers for the model of choice, the second, recursively iterating over the computations of the first simulation’s effort and schedule. The simulation inputs are correlated by applying the 1996 Lurie-Goldberg algorithm for generating correlated vectors of random numbers using Cholesky decomposition.
This paper details the issues related to the implementation of the Lurie-Goldberg algorithm when applied to the risk assessment of a cost estimate. The paper also describes the implementation of the autoRisk process into an autoRisk tool, and the tool’s current use in the preparation of credible, defensible acquisition cost estimates for leadership, decision-makers, and external oversight partners.
Paul is the Chief Technology officer at OPS Consulting LLC. He is responsible for overseeing strategy and development of the Company’s technical offerings.
Paul has 27 years experience in systems, software, and mechanical engineering, the last 12 providing on-site engineering consulting in the Intelligence arena. Paul is currently working as a senior cost analyst, performing cost realism analyses for the National Security Agency. Prior to joining OPS, Paul was president of Syntactic Solutions LLC, a small engineering consulting firm. Paul also worked for 20 years at BAE SYSTEMS performing systems and software development for several DOD programs including the Peacekeeper in Minute Man Silos, the Terrier and Tartar Weapons Direction Systems and the Joint Services Unmanned Arial Vehicles program. Paul also worked as an instrument design engineer and analyst for the DOE’s Solar Data Network.
Paul holds a Masters of Mechanical Engineering from Catholic University of America and a BS in Physics from East Stroudsburg University.
He lives in Silver Spring, Maryland, with his wife and six children. He is an avid sailor and an active triathlete.