Please enable JavaScript.
Coggle requires JavaScript to display documents.
Reflex and Goal-Based Agents: Decision Making - Coggle Diagram
Reflex and Goal-Based Agents: Decision Making
Introduction
A goal-based agent may look beyond the present to choose the best course of action to accomplish its objective, in contrast to a basic reflex agent that just considers the environment at hand while making decisions.
Model-Based Reflex Agents
These agents maintain an internal world model that combines
Memory of past experiences
Current perceptual inputs
- The model continuously updates as the agent operates, influencing its behavior. For example, a robot vacuum cleaner uses this approach to navigate around furniture while remembering which areas it has already cleaned.
4 types of agents
Simple reflex
Model-based reflex
Goal-based
Utility-based
Rational decision making
Expected utility calculation
Risk assessment
Useful
Relatively simple to code
Easy to debug
Follows hard
Intuitive to model
Flexible
Embed in classes
Types
Deterministic
Non-deterministic
Learning agents
Improve performance based on experience
Can adapt to changing environments
Hybrid agents
Combine multiple agent approaches
Leverage benefits of different architectures
Implementation Methods
Planning algorithms
A* search
Dijkstra's algorithm
Monte Carlo Tree Search
Machine learning approaches
Reinforcement learning
Supervised learning
Neural networks
Search techniques
Breadth-first search
Depth-first search
Heuristic search
Web Links
https://www.open.edu/openlearn/science-maths-technology/systems-modelling/content-section-3?active-tab=description-tab
https://study.com/academy/lesson/goal-based-agents-definition-examples.html
https://www.spiceworks.com/tech/tech-general/articles/what-is-fsm/
Modelling a system
Example - Miner Bob
Representation of knowledge
System / agents have to make decisions
Environment types
Fully vs partially observable
Deterministic vs stochastic
Single-agent vs multi-agent
Episodic vs sequential
Performance measures
Success rate
Efficiency metrics
Resource utilization
Example
Waiting at home
Chasing
Evading
Pathfinding agent
Chess-playing AI
Finite State Machine
State defines behaviour and produces action
Transitions between states
Rules and conditions that lead to transition
Input events = triggers
Hierarchical FSMs
Use cases in game AI