Discrete event simulation (DES) modeling and analysis for strategic program planning and management

Management of technology programs of the type undertaken by NASA is difficult for several reasons, including the large amounts of uncertainty and inherent risks associated with the costs, milestone completion dates, safety, and other performance measures. Unfortunately, mitigation of risk attributed to technology development has been frequently overlooked in scholarly research. This dissertation illustrates a methodology based on discrete event simulation and optimization to aid in the planning and of long range, strategic programs. The methodology is built around a framework that allows a program manager to assess and optimize resource allocation for a particular program objective and set of constraints. It provides the capability to analyze a program comprised of a portfolio of technology development projects and determine how best to ensure that at least one program will mature sufficiently.;The technology maturation life cycle is modeled in terms of sector-specific discrete development stages. Work is completed in a series of successive stages and funding is allocated separately for each stage. The duration of each stage has an associated underlying development time assumed to take the form of a Beta distribution. Shape parameters for the Beta distributions are functions of the baseline development time, the perceived level of development complexity, and funding allocation.;The simulation model execution begins by identifying the portfolio of desired technologies. The model samples the appropriate probability density function for each stage, yielding individual development times. Total development time for each technology is the sum of the corresponding transition times. Total development time for a portfolio of technologies is simply the longest individual technology development time. The output from the simulation model is a joint probability distribution over various performance measures of interest, including cost, completion time, and probability of success.;OptQuest and Arena software provided the platform for the simulation modeling effort. Arena software proved to be very flexible and eliminated some of the constraints exhibited by traditional program management constructs such as PERT or CPM. Furthermore, the combined functionality of these software packages provided a powerful optimization modeling tool and is demonstrated in the three illustrative examples contained in this document.