Computational Science
Hi. This talk is about Computational Science Investigations. It was taken from a presentation given at the Super Computing Challenge at the Kickoff Conference in 2012. [Slide 1] Computational Science can be seen as the intersection of Computer Science, Mathematics and Science. Computational Science is the use of Mathematics in Computer Science to model real world problems and conduct a simulation experiments. [Slide 2] Computational Science is made possible by increases in Computational Power. This increase in power enables us to design and conduct experiments on models of systems that are too big, too expensive or too dangerous to experiment with in the real world.
Increased Computational Power also allows us to run multiple what if scenarios very quickly. We also collect and analyze large amounts of data produces by these models. Computational Science complements but does not replace field experimentation. Each approaches appropriate in different situations. [Slide 3] This is a diagram of the Computational Science Process also called the Computational Science ... We start with selecting a real world problem we're interested in studying. Next we need to simplify that problem. We use abstraction to simplify the real world into a more manageable reproduction of the real world, doing so produces a working model.
The next step is to go from the working model to a mathematical or behavioral model by representing the model in terms of formal mathematical or algorithmic terms. The next step is to translate the equations or algorithms into computer code, we see the role of modeling in the first three steps. The next step is to run simulations using the computational model as a virtual test bed. The simulation produces data, from that data we draw conclusions and then we interpret if our model has any basis in reality, does it reproduce data that was collected from the real world. If so perhaps it can be use to make predictions about the real world.
[Slide 4] Next, I'm going to talk a little about the Choice of Modeling Approach. We can describe our world in many ways, at one end of the spectrum, here the seen at the left, mathematical formulas can be used to describe behavior of the world as a hole or as an aggregate. Whereas on the right hand side we have logical rules or mathematical formulas that describe behavior of individuals in the world. You could say that the left hand side is a top down view or description whereas the right hand side here is a bottom up or individual view. Our choice of how we describe the real world problem will influence the kind of programming approach or programming language we use.
In our case using the agent based modeling paradigm we describe the behavior of individuals in a population and thus the analogous approach, agent based programming is used. [Slide 5] Different languages and programming environments are better suited for different types of technical approaches. This is a chart showing some of the different languages here on the right and the approaches that they are best suited for on the left. Well, I hope this is given you a broad introduction to Computational Science and you'll be learning more as we go along.
ComputationalScience-IreneLee / Page 1 of 2