The Deep Dive System: Reimagining Global Climate Change with Artificial Intelligence and Virtual Reality

Sarah Saad, Thomas Palazzolo, Chencheng Zhang, Robert Reynolds, Ashley Lemke, John O'Shea, Calien O'Shea

Keywords: tipping points, Machine Learning, Evolutionary Algorithms, Cultural Algorithms, Virtual Reality

Abstract: The Deep Dive system was designed originally to predict ancient paleolithic site locations using AI and Virtual Reality. The focus of the project was a Land Bridge that connects what is now Michigan to Canada. During the recent Ice Age it was above water from 10,000 to 8000 B.P. During that time, it was a migration route for caribou. The assumption is that where caribou go, hunters will follow. AI path planning approaches were used to predict caribou migration pathways over time on the Land Bridge. This work investigates the impact that herd size has on the migration routes taken. Three models of herd behavior are used by an evolutionary algorithm, Cultural Algorithm, to produce optimal values for each of the herd models for different sized herds. The question of interest is whether there was a “tipping point” for one or more of the models that lead to a bifurcation of the optimal path over time? The results suggest what models are most appropriate to answer this question but the reasons why the bifurcation took place.