2012-RSK16

Utilizing Optimization Technique to Enhance Cost and Schedule Risk Analysis

Risk Track

RSK16_Presentation_UtilizingOptimizationTechniquetoEnhanceCostandSchedRiskAnalysis_Smith

Abstract:

Advancements in Monte Carlo simulations have enabled cost and schedule uncertainty analysis to be conducted on large, complex programs which could never be analyzed using conventional methods. Analysts are now able to predict with increasing confidence the budgetary requirements, schedule forecast, and risk mitigation measures on projects at all points during their lifecycle. These advancements in Monte Carlo technologies and methodologies have further lowered the barrier to entry for conducting uncertainty analysis; have reduced simulation times to mere fractions of a second; and have allowed for a comprehensive, integrated cost, schedule and risk analysis. However, despite the improvements in Monte Carlo analysis, the methodology is still limited in scope because stochastic optimization has been largely ignored. We propose that the next stage of evolution for integrated cost and schedule Monte Carlo analysis investigate optimization of the project plan and scope given uncertain inputs and constraints in the budget, resource availability, and capability requirements.

Deterministic optimization methods have become widespread, but we wish to go a step further and draw from the fields of stochastic optimization, network analysis, and control theory to account for uncertainty in cost and schedule optimization models. Through a presentation of our proposed methods for achieving the next stage in evolution of Monte Carlo analysis and a frank discussion with the Cost Estimating community, we will refine our vision for the future. Monte Carlo analysis has enabled the community to achieve what was out of reach merely a decade ago and, with the community’s help, we will push the limits of Monte Carlo simulations yet again, and equip estimators and analysts with the tools and methodologies to provide results never before possible.

Author(s):

Colin Smith
Booz Allen Hamilton
Colin Smith is an Associate for Booz Allen Hamilton’s Decision Analytics Team. He has a background in software engineering, artificial intelligence and cost modeling. He has performed the role of lead software engineer on past projects including a resource transition model for BRAC moves and cost modeling and data visualization effort for the US Navy Design for Affordability project, for which the team was recognized with a company-wide Booz Allen Professional Excellence Award for Innovation. He is currently the Technical Director of Real Time Analytics and has been heavily involved in that technology since its inception. His other research interests include artificial intelligence in general and more specifically evolutionary algorithms and swarm intelligence. Colin received a B.S. in Computer Science, Summa Cum Laude, and a M.S. in Computer Science from the Georgia Institute of Technology.

Brandon Herzog
Booz Allen Hamilton
Brandon Herzog is a Senior Consultant with Booz Allen Hamilton’s Decision Analytics Team. In addition to his experience in software engineering, he has a background in financial modeling and cost and schedule risk analysis. Mr. Herzog started his career building sales forecasting models for a Fortune 500 retailer, and since joining Booz Allen Hamilton has acted as a lead software engineer on various analytic tools including a program health management tool for the Navy and a cost and schedule risk analysis model for NASA’s Cost Analysis Division, where he also served as an analyst on the Joint Confidence Level (JCL) Cost Assessment for NASA’s James Webb Space Telescope. Currently he leads development of Dice, Booz Allens integrated cost and schedule risk analysis tool. He holds a bachelors degree in Computer Science from Stanford University.