Workflows

Depending on the type of study that you’re doing there are different step-wise approaches or workflows to follow. The first important distinction should be made between a data science and a simulation study.

Data science workflows

CRISP-DMOSEMNGeron (2017)
1Business understandingGather dataFrame the problem and look at the big picture
2Data understandingClean dataGet the data
3Data preparationExploreExplore the data to gain insights
4ModelingModelPrepare the data to better expose the underlying data patterns with ML algorithms
5EvaluationInterpretExplore many different models and short-list the best ones
6DeploymentFine-tune your models and combine them into a great solution
7Present your solution
8Launch, monitor, and maintain your system

Simulation study workflows

The first question to ask is what simulation paradigm are you following? If you haven’t decided yet, take a look at the guides to select between System Dynamics, Discrete-Event Simulation and Agent-Based Modeling (see, e.g., [ 1]). With a modeling paradigm different approaches exist. Let’s take a closer look at them one by one.

System Dynamics

Figure 1. Overview of the SD modeling approach according to Richardson and Pugh \[\]

Discrete-event simulation

Agent-based modeling

References

  1. Behdani B. Evaluation of paradigms for modeling supply chains as complex socio-technical systems. InProceedings of the 2012 Winter Simulation Conference (WSC) 2012 Dec 9 (pp. 1-15). IEEE.
  2. Richardson GP, Pugh III AI. Introduction to system dynamics modeling with DYNAMO. Productivity Press Inc.; 1981 Jan 1.