We propose a framework for understanding the complex activities of Sim
ulation Experiment Design and Analysis (SEDA), a framework that provid
es a basis for developing SEDA software. This paper presents the frame
work and illustrates it with a brief example. The definitions of six S
EDA components form the core of the framework. The framework itself in
cludes a dynamic model and a static model. The dynamic model is a gene
ric description of the sequential nature of SEDA. The static model, in
contrast is a very specific description of the SEDA computer system s
tructure for our chosen problem domain: comparison of queueing network
s in terms of their expected performance. We argue that the proposed S
EDA framework is a good one on the basis of its properties and by comp
aring it with related frameworks. We close by briefly summarizing our
prototype SEDA software, which was used to evaluate our framework.