OVERVIEW

Input data are key ingredients of simulation modeling. Such data are used to initialize simulation parameters and variables, or construct models of the random components of the system under study. For example, consider a group of milling machines on the shop floor, whose number is to be supplied as a parameter in an input file. You can define a parameter called, say, No_of_Milling_Machines in the simulation program, and set it to the group size as supplied by the input file. Other examples are target inventory levels, reorder points, and order quantities. On the other hand, arrival streams, service times, times to failure and repair times, and the like are random in nature (see Chapter 3), and are specified via their distributions or probability laws (e.g., Markovian transition probabilities). Such random components must be first modeled as one or more variates, and their values are generated via RNGs as described in Chapter 4.

The activity of modeling random components is called input analysis. From a methodological viewpoint, it is convenient to temporally decompose input analysis into a sequence of stages, each of which involves a particular modeling activity:

Stage 1. Data collection

Stage 2. Data analysis

Stage 3. Time series data modeling

Stage 4. Goodness-of-fit testing

The reader is reminded, however, that as in any modeling enterprise, this sequence does not necessarily unfold in a strict sequential order; in practice, it may involve multiple backtracking and loops of activities.

Data are the grist to the input analysis mill,...