Minimalist Description: Simulation science (also known as "descriptive modeling") uses formal models of the world to support wise decisions. Simulation science and data science are like the right hand and the left hand of decision support. In fact, they both have a "Control" component. But simulation science is model-driven. Instead of using measurements of the world--which may be difficult to get or biased or noisy--to make decisions, simulation science supports decisions through representational models that capture cause-and-effect. There are three general kinds of simulation science methods, each of which transforms things at one level in the knowledge hierarchy into things a level higher up:
  • Modeling--These models describe, mathematically, the way we think the world works, including cause-and-effect: the rules of engagement, the best practices, the probabilities, the physics, the assumptions, etc., and particularly the interactions between the things we are modeling.
  • Simulation--Simulations bring a model to life by exercising it (usually) over time so that we can understand outcomes, and particularly the range of possible outcomes from best case to worst case.
  • Control--Finds efficient, effective strategies given the model, with all of its rules, constraints, and assumptions. One person's idea of effective may be to maximize the value of the most likely case while another person's idea of effective may be to minimize the likelihood of the worst case scenario. All of these dials and knobs fall within control.
Simulation science models (aka descriptive models) complement data science models. Often, both are necessary, for instance as checks and balances on each other, and we are proud to be one of the few small organizations we know of with capabilities in both areas. We think this gives us unique advantages in helping you generate value from your enterprise. We may, sometimes, wish to develop descriptive models before doing the data science because a descriptive model can be much less expensive, and in many cases what we learn from the descriptive model gives us a better understanding of the situation so we can do better data science faster, and at lower cost.  



S3 Data Science
S3 Data Science, copyright 2015.