Decision-support software based on discrete-event simulation (DES) has been undergoing development for use in planning and scheduling work in the Orbiter Processing Facility (OPF) at Kennedy Space Center. The software is also suitable for adaptation to similar use in other facilities characterized by low-volume processing and by hierarchical processing environments with multiple subfacilities and with dependencies of tasks on resources and on the prior completion of other tasks.

DES is traditionally applied to mass-production manufacturing environments, where processing volumes are large. DES can enhance decision-support systems by providing the means for mathematical modeling of the effects of uncertainties in parameters. In DES, parameters are treated as being variable with probability distributions; this makes it possible to perform computational simulations to quantify the effects of changes in parameters upon critical outputs, before changing the parameters in a real process or system. In mass-production environments, parameters such as queue lengths and throughput rates are relevant, but in the dynamic processing environment of the OPF, different parameters become relevant.

The developmental software system includes subsystems that implement mathematical models for planning the flow of work at two levels: (1) the OPF task level and (2) the interfacility level. These models have been built by use of commercially available DES software.

The model for the OPF task level is a generic, reusable one that simulates a network of tasks. As each simulated task begins, an attempt to obtain the resources necessary to complete the task is made. If the resources cannot be obtained, the task waits. The output information provided by this model includes the amount of time each task waited for resources, amounts of resources utilized, and the total duration (and variance of duration) of a project.

The model for the interfacility level can also be characterized as a launch-processing-simulation model. This model evaluates the number of launches per year, based on expected process durations and numbers of OPF shifts available. Inasmuch as fixed "milestone" dates are available as inputs, a deviation from a planned launch date can be evaluated in the near term. The model can be used to evaluate workers' overtime needed to meet a planned launch date in the near term and the expected number of launches per year in the long term.

This work was done by JoAnn Leotta of Kennedy Space Centerand Robert R. Safford and Matt Archer of the University of Central Florida. For further information, access the Technical Support Package (TSP) free on-line at under the Mathematics and Information Sciences category, or circle no. 160on the TSP Order Card in this issue to receive a copy by mail ($5 charge).