A hybrid graphical-simulation analysis of a health systems application

Computer simulation is often used to optimize practical network queueing models. In many situations the number of possible alternatives that could be simulated is large, and individual runs may be expensive. A simple approximation technique can be used to reduce the number of options actually simulated by providing a pictorial representation of the service process in a network of queues under simplifying assumptions. A simulation experiment can then be performed by detailed analysis of the selected subset of alternatives. This hybrid graphical-simulation approach was used in the design of a United States Air Force health improvement program.