FLEXIBILITY IN ENGINEERING DESIGN by Richard de Neufville and Stefan Scholtes
Reviewed by Todd Schenk, Massachusetts Institute of Technology
For all our engineering marvels, the world has seen its share of white elephants and design failures – bridges to nowhere, communications networks out-of-date by the time they are deployed, and dikes overtopped leading to flooding. Perhaps these problems are to be expected; designers and engineers are always confronted with significant biophysical, social and economic uncertainty, and no model is perfect. Furthermore, the world is constantly changing, making different designs optimal at different points in time. The conventional approach to design is to use average or worst-case forecasts, add in safety factors and proceed as though we know what the future holds. Unfortunately, this leads to wasted resources when projects are overdesigned and failure when they are underdesigned.
de Neufville and Scholtes make a compelling case for another approach. Instead of preparing a single forecast of a possible future, they argue for building in flexibility so that engineered projects can be adapted easily as conditions change. One example they present is the bridge over the Tagus River in Lisbon, Portugal. When it was constructed in the 1960s, its designers had the foresight to make it strong enough to hold a second deck at some point in the future, if that became necessary. They also included a railroad station underneath the toll plaza in case rail connections were ever added. Thirty years later, their foresight paid off when a second deck carrying commuter rail lines was constructed at modest cost and little disruption to the transportation system.
While it may sound simple, the challenges associated with shifting the way project design proceeds, from the traditional predict, plan and build approach to something more iterative should not be underestimated. Designers and engineers need to rethink the way they use forecasting models. Budgeting and capital planning need to allow for longer-term project adaptation. de Neufville and Scholtes introduce a variety of tools to support this shift, including phasing projects and investments, Monte Carlo simulation for exploring scenarios, and dynamic forecasting to highlight uncertainties.
Strategies for making planning and design nimbler can not come soon enough, given that climate change is increasing uncertainty and budgets are regularly stretched. Flexibility should become the new design norm. This book represents both a manifesto for this important shift as well as an early guidebook instructing us how to get there.