Environment I
Fully observable (accessible) vs. partially observable (inaccessible):
· Could be partially observable due to noisy, inaccurate or missing sensors, or inability to measure everything that is needed
· Model can keep track of what was sensed previously, cannot be sensed now, but is probably still true.
· Often, if other agents are involved, their intentions are not observable, but their actions are
· E.g
· chess – the board is fully observable, as are opponent’s moves.
· Driving – what is around the next bend is not observable (yet).
Environment II
Deterministic vs. stochastic (non-deterministic):
· Deterministic = the next state of the environment is completely predictable from the current state and the action executed by the agent
· Stochastic = the next state has some uncertainty associated with it
· Uncertainty could come from randomness, lack of a good environment model, or lack of complete sensor coverage
· Strategic environment if the environment is deterministic except for the actions of other agents
· Stochastic = the next state has some uncertainty associated with it
· Uncertainty could come from randomness, lack of a good environment model, or lack of complete sensor coverage
· Strategic environment if the environment is deterministic except for the actions of other agents
Examples:
Non-deterministic environment: physical world: Robot on Mars Deterministic environment: Tic Tac Toe game
Non-deterministic environment: physical world: Robot on Mars Deterministic environment: Tic Tac Toe game
Environment III
Episodic vs. sequential:
· The agent's experience is divided into atomic "episodes" (each episode consists of the agent perceiving and then performing a single action) and the choice of action in each episode depends only on the episode itself
· Sequential if current decisions affect future decisions, or rely on previous ones
· Examples of episodic are expert advice systems – an episode is a single question and answer
· Most environments (and agents) are sequential
· Many are both – a number of episodes containing a number of sequential steps to a conclusion
· Sequential if current decisions affect future decisions, or rely on previous ones
· Examples of episodic are expert advice systems – an episode is a single question and answer
· Most environments (and agents) are sequential
· Many are both – a number of episodes containing a number of sequential steps to a conclusion
Examples:
Episodic environment: mail sorting system Non-episodic environment: chess game
Environment IV
Discrete vs. continuous:
· Discrete = time moves in fixed steps, usually with one measurement per step (and perhaps one action, but could be no action). E.g. a game of chess
· Continuous = Signals constantly coming into sensors, actions continually changing. E.g. driving a car
· Continuous = Signals constantly coming into sensors, actions continually changing. E.g. driving a car
Environment V
Static vs. dynamic:
· Dynamic if the environment may change over time.
· Static if nothing (other than the agent) in the environment changes
· Other agents in an environment make it dynamic
· The goal might also change over time
· Not dynamic if the agent moves from one part of an environment to another, though it has a very similar effect
· E.g. – Playing football, other players make it dynamic, mowing a lawn is static (unless there is a cat...), expert systems usually static (unless knowledge changes)
· Static if nothing (other than the agent) in the environment changes
· Other agents in an environment make it dynamic
· The goal might also change over time
· Not dynamic if the agent moves from one part of an environment to another, though it has a very similar effect
· E.g. – Playing football, other players make it dynamic, mowing a lawn is static (unless there is a cat...), expert systems usually static (unless knowledge changes)
Environment VI
Single agent vs. multi agent:
· An agent operating by itself in an environment is single agent!
· Multi agent is when other agents are present!
· A strict definition of an other agent is anything that changes from step to step.
· A stronger definition is that it must sense and act
· Competitive or co-operative Multi-agent environments
· Human users are an example of another agent in a system
· E.g. Other players in a football team (or opposing team), wind and waves in a sailing agent, other cars in a taxi drive
· Multi agent is when other agents are present!
· A strict definition of an other agent is anything that changes from step to step.
· A stronger definition is that it must sense and act
· Competitive or co-operative Multi-agent environments
· Human users are an example of another agent in a system
· E.g. Other players in a football team (or opposing team), wind and waves in a sailing agent, other cars in a taxi drive
This nomination in really very helpful and motile to graps
ReplyDeleteThanks
ReplyDelete