Agent-based simulations are used in a wide variety of fields, from biology to economics. An agent-based simulation gives individual actors rules at the actor level but then we look for emergent patterns and behavior at a macro level . Frequently these actors can move in 2D and 3D space, so these models tend to have a strong spatial component

Over the last 30 years we have seen the rise of several different spatially aware agent-based simulation frameworks, such as Mesa (Python) or Mason/GeoMason (Java). During that same time we have seen the emergence of vertically and horizontally scaling of distributed applications. There have also been many advances in rapidly handling large streams of data as well as loosely coupling pieces of an application.

The talk will cover a brief introduction to spatially aware agent-based modeling. The rest of the talk will cover building an agent based simulation for finding optimal animal corridor locations. We will cover the architecture, then show the agent running, and then visualize the results. The application will take advantage of data streaming, messaging, and horizontal scaling to create a robust and efficient framework.

The goal of this talk is for you to leave understanding spatially aware agent-based simulations and how they can leverage modern cloud capabilities. It doesn’t matter if you are in finance, social media, or pharmaceuticals - agent-based modeling can help your understanding of your vertical. With this knowledge in hand you will sure to be a hit at your next cocktail party. Come for the knowledge and stay for the witty banter.