MV4025 Cognitive and Behavioral Modeling for Simulations

This course focuses on the primary technologies used to model cognition and behavior in order to create agents that represent human beings in simulations. Topics include the dominant technologies in use, the tools used to support them, and their application to the various capabilities required of an agent. The modeling technologies covered include the production-system approaches common in artificial intelligence/cognitive science/psychology, as well as the finite-state, automata-inspired approaches that are part of engineering practice in computer-generated force simulations and the computer entertainment industry. The full scope of the modeling problem will be addressed, from sensation and perception through situation awareness and action selection, to action execution. Approaches to modeling communication and behavior moderators (e.g. experience, emotion, fatigue) will also be discussed.

Prerequisite

CS3310

Lecture Hours

3

Lab Hours

2

Course Learning Outcomes

On completing this course, students will:

  • Learn and gain in-class experience with the traditional architectural concepts and tools used to create behaviors for simulations, such as finite state machines and hierarchical models.
  • Learn about and create models that create behavior based on utility functions (incremental or “whole COA”) produced by engineers, including A* search for tactical path finding.
  • Apply reinforcement learning to create behavior models and gain a view of the wide range of experimental techniques used to improve its performance.
  • Understand game solving approaches to behavior creation.
  • Understand the issues around partial knowledge of the simulated environment (fog of war) and some tools that are helpful in dealing with it.
  • Improve their performance in efficiently reading complex technical literature and conducting careful computational experiments.