is a collection of multiple autonomousagents, each acting towards its objectives while all interacting in a shared environment, being able to communicate and possibly coordinate their actions.
Agent
[Russell & Norvig] An agent is anything that can perceive its environment (through its sensors) and act upon that environment (through its effectors). [Wooldridge & Jennings] An agent is a computer system that is situated in some environment, and that is capable of autonomous action in this environment in order to meet its design objectives/delegated goals.
Autonomous Agent Properties
autonomous
the agent is self goal-directed and acts without requiring user initiation and guidance; it can choose its own goal and the way to achieve it; its behavior is determined by its experience; we have no direct control over it
reactive
the agent maintains an ongoing interaction with its environment, and responds to changes that occur in it
proactive
the agent generates and attempts to achieve goals; it is not driven solely by events but takes the initiative
sociable
the agent interacts with other agents (and possibly humans) via cooperation, coordination, and negotiation; it is aware and able to reason about other agents and how they can help it achieve its own goals
coordination
is managing the interdependencies between actions of multiple agents (not necessarily cooperative)
cooperation
is working together as a team to achieve a shared goal
negotiation
is the ability to reach agreements on matters of common interest
Agent’s behavior is described by the agent function that maps percept sequences to actions.
Agent program
runs on a physical architecture to produce agent function.
Rational agent
chooses whichever action maximizes the expected value of the performance measure given the percept sequence to date and whatever bulit-in knowledge the agent has.
is a function that maps a state onto a real number; it captures “quality” of a state. If an agent prefers one world state to another state then the former state has higher utility for the agent.
There are different agent architectures with different capabilities and complexity.
Auctions
Auction
mechanism for allocating resource among self-interested agents [Shoham & Leyton-Brown 2009] An auction is a protocol that allows agents (=bidders) to indicate their interests in one or more resources and that uses these indications of interest to determine both an allocation of resources and a set of payments by the agents.
How offers are made: by whom, when, what their content is
Clearing rules
Who gets which goods (allocation) and what money changes hands (payment).
Information rules
What information about the state of the negotiation is revealed to whom and when.
Valuation Models
Common value
the good has the same value to all agents example: a 100 dollar note
Private value
an agent A’s valuation of the good is independent from other agent’s valuation of the good example: a painting, John Lennon’s last dollar bill
Correlated value
valuations of the good are related, i.e. the more other agents are prepared to pay, the more agent A prepared to pay. example: purchase of items for later resale
Agent’s payoff from participating in an auction
if winner: payoff = item’s valuation – price paid for the item; if not winner: payoff = zero
multiple units of the same item are available for auction
Multi-Item Auctions
bidding for multiple items grouped together
Reverse Auctions
The buyer issues a request for bids to his providers.
Multi-Attribute Auctions
Negotiation over further attributes beyond price, e.g. color, weight, or delivery time
Single-Item Auctions - Basic Auction Mechanisms
English
Japanese
Dutch
First-Price
Second-Price
English Auction
Auctioneer starts the bidding at some reservation price. Bidders then shout out ascending prices (minimum increments). Once bidders stop shouting, the high bidder gets the good at that price.
Japanese Auction
Same as an English auction except that the auctioneer calls out the prices All bidders start out standing. When the price reaches a level that a bidder is not willing to pay, that bidder sits down. Once a bidder sits down, they can't get back up the last person standing gets the good.
Dutch Auction
The auctioneer starts ahigh value; it descends clock at some. At some point, a bidder shouts “mine!" and gets the good at the price shown on the clock. Good when items need to be sold quickly (similar to Japanese auction) No information is given away during auction.
First-price sealed bid auction
Bidders write down bids on pieces of paper. Suctioneer awards the good to the bidder with the highest bid. That bidder pays the amount of his bid.
Other consideration: communication complexity, private information revelation, ...
Dutch and First-price Sealed Bid
Strategically equivalent: an agent bids without knowing about the other agents’ bids - a bidder must decide on the amount he's willing to pay, conditional on having placed the highest bid