Goal-based Agents

Goal-based Agents

1. Problem and Actions represented as nodes

   • Cost & value of every action

2. Search Algorithm

   • Depth-first Search
     • Follow branches until they end
   • Breadth-first Search
     • Expand from the center of the graph exploring layers of nodes
   • A*
     • Determines cost with estimated distance to goal

3. State Machine

   • State transition tables (Rules)
     • New state = old state + input
   • Simple to code, easy to debug, hard-coded rules, flexible and intuitive

Bilder und Dateien der Mindmap