Transcript Artificial Intelligence in an Agent

Artificial Intelligence in an AgentBased Model
By John Walsh
Agent-Based Model

Equation based modeling

Diverse field of research

Models scenarios very high detail

Impossible in equation based modeling
Goal

Investigate detailing agent behavior beyond
simple rules

Simple predator-prey interaction

Non-grid-based setup
Steps

Basic artificial intelligence for each agent

Basic environment and agents

Define interaction between predator and prey

Seeding

Graph populations over time

Reproduction among agents

Output results to a file
Steps

Advanced collision detection

Food for prey

Stable prey-food scenario

Stable with predators added

Food-seeking prey

Fleeing
Current
Equation-Based
• Lotka-Volterra two species model
Agent-Based