A novel schedule forecasting methodology (named Nostradamus) was envisioned/tested for low-volume, highly-complex, new product development (NPD) projects to contribute towards managing risk of a unique, large-scale program.  The goal was to offer objective real-data based product-delivery schedule forecasts with acceptably high precision.  Nostradamus is designed to combine “past (most recent) manufacturing performance” of a  manufacturer (or project management firm) with “current information” from the products that are presently under manufacturing,  to provide  an objective delivery date (or project completion date) forecasts for not-yet-delivered products.  The past information used, conforms with the “reference class concept” described by Nobel Prize laureate Daniel Kahneman and coworkers, and possesses the highest similarity to products under manufacturing, for which schedule forecasting is required.  Additionally, the Nostradamus software program ranks a list of components of the products that are heavily affecting the delivery dates, or delivery delays. Hence, targeted measures can be taken to favorably affect the product delivery dates and reduce the schedule risk.  The product’s major components (or tasks), along with their expected completion dates (ECDs), both defined/determined by  the manufacturer,  constitute the “current information” and are provided by the manufacturer at each Program Management Review event. This set of product components, along with their ECDs and manufacturer’s estimated product delivery date, collectively define the Line-of-Balance (LOB) information.  The “Accuracy Level” probability distribution function (PDF) of the ECDs is defined  and calculated for a most recently delivered product and then utilized  in a subsequent  Monte Carlo simulation that used the “current information,”  generating a product delivery failure probability (DFP) graph.   This DFP was used to find delivery date forecast at any probability level.  Results of the tests indicated that, over the project duration,  a time-averaged forecast imprecision value of -5%  was achieved for Nostradamus, as compared to 140% by the manufacturer.

Keywords
ANALYTICS, BEST PRACTICES, DECISION, DESIGN, DISCRETE EVENT SIMULATION, DISRUPTIVE INNOVATION, EMERGING TECHNOLOGIES, PROBABILITY, RISK ASSESSMENT, SIMULATIONS, STATISTICS

Additional Keywords
Schedule forecasting, Manufacturing, New Product Development, Line of Balance, Accuracy Level, Expected Completion Date, Product Delivery Failure Probability, Monte Carlo Simulation, Reference Class Forecast, Project Management, Nostradamus, Complexity