ABSTRACT
As described in Chapter 5, there are many ways to implement a particular design of a discrete event simulation (DES). If the implementation is correctly done, the variations should produce equivalent results in the sense that they convey the same information to the decision maker (they may not be numerically identical because of the stochastic behavior of the simulations). The variants may not be equivalent, however, in terms of their efficiency. Efficiency here refers to both the number of calculations required to obtain a set of results, which determines the time required to run the analyses of the model, and the effort needed to build the model. Although with the vast speed of current computers, this may seem like an unimportant aspect of simulation, it can have a major impact on the quality of the analyses and the usability of the model. If the simulation takes a long time to produce results, then analysts may be reluctant to carry out the necessary range of analyses required to validate the model and to represent structural and parameter uncertainty. In addition, unsound simplifications may be implemented, all in the pursuit of faster execution times. Among any series of implementations that yield equivalent results, the most efficient one should always be preferred.
