Simulation Theory

One of the methods of analyzing a system is creating its model and simulating its performance. In this case you carry out all the experiments with the model, not the real system, but get the results that describe the real system.

Simulation is used when lots of various scenarios are available and there are many relationships between objects; thus, analytical calculation models become very complex.

In Business Studio you create an activity model. Assuming that the model accurately describes reality, it is assumed that property values obtained during the simulation will somehow coincide with the real values.

To simulate activity performance, first, you describe the activity as a model of tasks, processes, events, and resources. Then based on these created models, simulation engine generates objects that simulate reality: simulation task instances, simulation event instances, etc. In addition, indirect interaction between task instances is also simulated, for example, task instances compete for available resource instances. For each object instance that is simulated, properties are saved. As a result, simulation provides the analyst with the statistical data on process time, cost, etc.

Thus, if the analyst possesses information on statistical distribution of real task times and the cost of resources, then by means of simulation method he gets the following data with some accuracy:

• time and cost properties of process instances,
• process cost distributions,
• load of time resources (this data makes it possible to identify bottlenecks and underloaded resource instances),
• cost of resources needed to perform a process instance,
• etc.

Please note that it’s impossible to create a model that is an exact copy of reality; thus, the analyst creates a model with the acceptable accuracy.

1. The analyst formulates the tasks that should be solved. Examples of tasks include:
   • To validate the created model. To do this, you need to run simulation of the model that consists of several process diagrams. The property values you’ll get as a result of simulation, you need to compare to the real data. If the difference is significant, it’s recommended to adjust the model.
   • To optimize the current activity. This is the case when the model is created to find the critical properties and optimize their values.
   • To create a new model.
2. The activity is presented as a model of processes and tasks.
3. Events instantiation models are created.
4. For each task model the processing time and the technological wait time are defined, as well as resources needed to perform each task instance. Resources can be either work or material. Also, fixed costs may be also included into the task cost.
5. For each resource model cost properties are defined.
6. The analyst specifies process models that should be simulated and runs the simulation. The simulation engine creates instances of start events and activates performance of each task instance according to process diagrams. All the time and cost properties are calculated.
7. The results are then analyzed.