A Pedagogy-Oriented Modeling-and-Simulation Environment for AI Scenarios

Programming assignments are an effective way for students to investigate many important and fun aspects of AI. However, for any task of reasonable breadth and depth, especially involving complex, compositional agents, actions, and graphics, most of the programming effort goes into mundane, tedious, time-consuming, and error-prone administrative aspects. Moreover, time constraints usually result in designs that are not extensible or reusable for subsequent assignments, which repeat this overhead throughout the semester. This pedagogy-oriented modeling-and-simulation framework provides convenient support capabilities to get students quickly playing with a wealth of agent-based AI content. It contains extensive, highly configurable, yet user-friendly, engineering, physics, and communication models for arbitrary components and task environments. These components are managed automatically in a stochastic Monte Carlo simulation that allows students to define, test, and evaluate their quantitative performance for a wide range of controlled experiments.