Professor of Decision Sciences
Project Management; Dynamic Programming; Uncertainty; Bayesian Updating; Optimal Stopping; Exponential Poisson Bandit
The authors investigate the cost of the opportunity delayed by working on one project with uncertain success rather than searching for a new project. The authors answer this question: how long should a firm work on a research project with uncertain success before abandoning it if the only alternative is to search for a new project to work on?Rather than treating the opportunity as an exogenous alternative, this approach endogenizes the opportunity value and the attendant cost of its delay. The authors consider cases with both a finite and an infinite number of potential projects.The authors derive the optimal stopping time and show that it increases with the discount rate, with the rate of arrival of project failure, and with the prior probability that the project can be successfully completed. The optimal stopping time decreases with the rates of new project arrival and project completion.