Steps in the simulation-based planning projects

A conceptual model is a formal description of the simulation model to be developed. I.e. the relevant elements of the real system as well as their relations are represented at an appropriate level of detail. It defines the general structure of the simulation model and determines the requirements for data analysis. 

In the data analysis step historical data of the real system to be modelled are analysed to extract relevant parameters for the simulation model and the involved stochastic processes. The simulation model is the implementation of the conceptual model accompanied by the required parameters obtained from the data analysis step as some sort of a computer program. 

To derive recommendations for improving the system performance, simulation experiments have to be conducted. I.e. control parameters of the simulation model are altered in a structured way to investigate their effects on the performance measures. A specific setting of the control parameters is called a configuration of the system. 

As a simulation model typically contains stochastic elements, a simulation run is a stochastic process. Hence, the outcome of simulation runs are stochastic variables. Therefore, a certain number of simulation runs has to be conducted to handle the stochasticity 

In the case of single-objective problems, the aim is to find the optimal configuration of the system under study. Once found, this configuration is the sole recommendation for the management. The objective typically represents an expected performance measure. However, since stochastic events may influence the performance measure, it is advisable to incorporate measures of variability (like the variance of a performance measure) as a further decision criterion. This immediately leads to a multi-objective problem.  

---

Figure 4.6 Steps in the simulation-based planning projects...

---