Selective reduction of activity time uncertainty to reduce risk of unacceptable system performance

Abstract

One of the major problems in managing a new product development (NPD) project is the difficulty in accurately predicting the duration of project activities. Often only rough design information and preliminary product specifications are available at the early stages of project planning for NPD. Uncertainty concerning how much work must be performed to complete an activity and how productive assigned resources will be complicates the task of accurately estimating the expected time of activity duration. Minimizing the risk of unacceptable project performance (lost customers) is a major concern of decision makers in the NPD process. This research focuses on resource allocation in project environments with considerable activity time uncertainty, that is, new product development, and explores the implications of considering both the mean and variance of project duration in evaluating alternative approaches. We demonstrate that an objective of increasing the probability that a project completes by a target due date is a reasonable objective that requires the consideration of all activities in the project. In so doing, we demonstrate that resource allocation approaches considering both the mean and variance of activity time provide effective means for improving project performance and improvement approaches need not be confined exclusively to activities on the critical path.