MEET OUR FINALISTS:

This project involved a decision analysis application of portfolio management for the $400 annual Internal Capital Allocation Program at Dow Chemicals. The total amount requested across all projects exceeded $600 million, which led Dow to make the project allocation decisions based on cost estimates provided by managers for individual projects. However, this motivated managers to systematically provide low estimates for costs. There was also no incentive for managers to return surplus amount if projects finished with a surplus. We developed a data-driven near optimal mechanism that addressed these issues. It (i) provided a quantification for the variability in actual project costs, (ii) This variability provided a basis to determine a strategic reserve for potential cost overages, (iii) encouraged managers to truthfully report cost estimates. The mechanism was implemented with success at the firm.  

Asteroids and other near-earth objects (NEOs) pose a significant threat to our planet. Advance detection is essential to respond to any object on a collision course, but detection and tracking technologies require substantial investments. We provide a multi-attribute utility framework to analyze which NEO detection technologies offer the best decision alternatives using a stylized model of the uncertainties, objectives, and tradeoffs inherent to decisions involving low-probability, high-consequence events.

To address the challenge of effectively engaging many stakeholders and experts in real-world decision analysis processes, both for knowledge construction and for stakeholder engagement, we have developed the Collaborative Value Modeling (CVM) framework. The CVM combines large-scale participatory Web-Delphi processes with smaller-scale decision conferencing or workshop processes to promote agreement in different modeling stages of multicriteria decision analysis (MCDA).