Soar
Architecture
Philosophy and Methodological Assumptions
Description of the Architecture
Problem Spaces
Long-Term Production Memory
Short-Term Working Memory
Preference Memory
Decison Cycle
Goal-Directed Behavior
Percetual-Motor Components
Chunking
Agent Properties
Style of Control
Impasse-Driven Control
Serial Processing
Parallel Processing
Interruptible Processing
Hierarchical Organization
Symbolic World Model
Size of the Knowledge Base
Black Box Approach
Global Representation and Uniform Access to Knowledge
Homogeneous (Uniform) Knowledge Representation
Associative Memory
Reflexive Learning
Generalization
Minimum Commitment Strategy
Reflexive Response to Stimuli (Reactive Response)
Match-Based Attention
Deliberation/Operation Speed
Goal-Directed
Default Knowledge
Movement to Routine Behavior
Distractibility
Adaptive Default Behavior
List of Agent Properties for All Architectures
Capabilities
Single Learning Mechanism
Multi-Method Learning
Learning by Instruction
Simple Concept Acquisition
Abstraction
Learning by Analogy
Transfer of Learning
Universal Weak Methods
Planning and Situated Action
Problem Solving (Inductive and Deductive Reasoning)
Replanning
Meta-Reasoning
Expert-Systems Reasoning
Natural Language Understanding
Perception
Robotic Tasks
Focused Behavior and Processing/Selective Attention
Responding Intelligently to Interrupts and Failures
Human-like Math Capability
Soar and Simlulated Tactical Air Flight
Soar as a Model of Human Cognition
List of Capabilities for All Architectures
Environmental Considerations
Static Environments
Dynamic Environments
Environmental Consistency
Real World Environments
Simulated Environments
Complex Environments
Knowledge-Rich
Unpredictable
Search-Intensive Environments
List of Environmental Considerations for All Architectures
Issues
References
Table of Contents
Current Location: Soar - Top