Professional Documents
Culture Documents
Lecture IS
Lecture IS
(Chapter 8)
BITS Pilani Gokul Kannan Sadasivam
Department of CS & IS
Hyderabad Campus
Outline
• Introduction
• Types of an Environment
• IS Models
• Knowledge Representations & other IS Models
Introduction
• Other Examples
• Intelligent Software Modules (Firewall, IDS/IPS)
• Fuzzy logic in Washing machines
Saturday, November 12, 202 BITS SS ZG531 (Pervasive Computing) 4 of 37
2
BITS, Pilani - Hyderabad Campus
What is an Agent?
• Regions: A & B
• Dirt or Clean Agent Program
• Actions
• Move right
• Move left
• Suck
• Nothing
Types of an Environment
• Fully Observable
• Able to get complete state of the environment.
• E.g., Vacuum Cleaner, Chess with a clock
• Partially Observable
• Not able to get complete environmental state.
• Some sensor readings not available.
• Some sensors become defective.
• E.g., Automated car, Part picking robot
• Unobservable
• No sensors
• Deterministic
• Able to predict the current state by knowing the previous state.
• E.g., Vacuum cleaner, Chess with a clock
• Stochastic
• Cannot predict the next state.
• E.g., Automated car, Part picking robot
• Episodic
• Current actions will not affect the next state of the environment.
• E.g., part picking robot
• Sequential
• Current actions will influence the next state of the environment.
• E.g., Vacuum cleaner, Automated car, Chess with a clock.
• Static
• The state doesn’t change until the action is completed.
• E.g., vacuum cleaner
• Dynamic
• The state changes quickly.
• E.g., Automated car.
• Semi dynamic
• The state changes after a duration.
• E.g., Chess with a clock
• Discrete
• E.g., Chess with a clock
• Continuous
• E.g., Automated car
• Unknown environment
• For many percepts, the system does not know the action(s).
• The system has to do some learning.
IS Models
• Environment:
– Fully observable
– Stochastic
– Episodic
– Static
• Planning-based IS.
• Environment same as
the Environmental-IS.
Node-Labelled Graph
• Propositional Logic
• Predicate Logic
• Textbook – Chapter 8
• Artificial Intelligence – Stuart Russell and Peter Norvig