lect 8-auctions

79
LECTURE 8: Cooperation in MAS (II): Negotiation via auctions Artificial Intelligence II – Multi-Agent Systems Introduction to Multi-Agent Systems URV, Winter-Spring 2010

Upload: antonio-moreno

Post on 15-Jan-2015

466 views

Category:

Technology


1 download

DESCRIPTION

MAS course at URV, lecture 8, coordination methods: auctions

TRANSCRIPT

Page 1: Lect 8-auctions

LECTURE 8: Cooperation in MAS (II): Negotiation via auctions

Artificial Intelligence II – Multi-Agent Systems

Introduction to Multi-Agent Systems

URV, Winter-Spring 2010

Page 2: Lect 8-auctions

Lecture overview

Negotiation protocolsAuctions

English/Dutch/FPSB/VickreyMulti-attributeCombinatorial

Page 3: Lect 8-auctions

MAS

Cooperative Benevolent

Independent Self-interested

EmergentDiscrete Without communication -

Implicit

With communication -

Explicit

NegotiatorsDeliberative

Cooperation hierarchy [last lecture]

Partial Global Planning

Contract Net

Reactive systems

Auctions

Page 4: Lect 8-auctions

Negotiation between multiple agents

Page 5: Lect 8-auctions

Need of negotiation in MASAgents may have incompatible goals, and resources to achieve these goals may be limited; in such cases competition and conflicts may ariseThe effects of the agents’ antagonistic behaviour needs to be limitedAgents must be able to reach compromises, resolve conflicts, allocate goods and resources by way of an agreementAgents’ interactions are governed by a set of rules: an interaction protocol

Page 6: Lect 8-auctions

Negotiation protocol elements (I)

Public elementsA negotiation set which represents the space of possible offers/proposals that the agents can makeThe protocol rules which govern the agents’interactions

Page 7: Lect 8-auctions

Negotiation protocol elements (II)

Private elementsThe set of strategies that the agents can use to participate in the negotiation process:

They are not dictated by the protocol itselfMay take into account the other agents’strategies

Learning mechanisms

Page 8: Lect 8-auctions

Protocol rules (I)Admission rules

When an agent can participate in a negotiation (e.g. eligibility criteria)

Interaction rulesSequence of admissible/valid actions (e.g. moments in which bids are allowed)

Validity rulesWhat constitutes a legal offer/proposal (e.g. a new bid must be higher than the last bid)

Outcome determination rulesWhen an agreement has been reached

Page 9: Lect 8-auctions

Protocol rules (II)

Withdrawal rulesWhen an agent can withdraw from the negotiation

Termination rulesWhen a negotiation ends unsuccesfully

Commitment rulesHow the commitments that agents make during the negotiation are managed

Page 10: Lect 8-auctions

Negotiation factorsNumber of attributes: one, many[Multi-attribute auctions]Number of agents:

One-to-oneOne-to-manyMany-to-many

Number of units: one, many[Multi-unit auction]Interrelated goods: one good or a number of goods that are substitutable or interdependent[Combinatorial auctions]

Page 11: Lect 8-auctions

Protocol evaluation criteria (I)

Social welfare – the sum of all agent’s payoffs or utilities in a given solutionPareto efficiency – a solution x is Pareto optimal if there’s no other solution x’ such that at least one agent is better off in x’ than in x, and no agent is worst off in x’ than in xIndividual rationality – an agent should not lose out by participating in a negotiation

Page 12: Lect 8-auctions

Protocol evaluation criteria (II)

Stability – mechanism should be designed to be non-manipulable: motivate each agent to behave in the desired mannerComputational efficiency – mechanisms should be designed so that when agents use them, as little computation is needed as possibleDistribution and communication efficiency –distributed protocol vs. minimum communication (time, money,…)

Page 13: Lect 8-auctions

Architecture for negotiating agents

Message interpretation

Message generation

Communication component

Decision making component

ProposalProposalhistory

Update

Query

Proposal

Message

Message

Knowledge base component

Environment model

Self-model

Opponent model

Negotiation history

Page 14: Lect 8-auctions

AuctionsA class of negotiation protocols which provide us with methods for allocating goods/resources based upon competition among self-interested parties

Where are auctions used?Telecommunication and TV licenses, mining rights, emission permits, airport gates and takeoff/landing slotsGoogle AdSenseExtremely popular as a means for conducting consumer-to-consumer (C2C) negotiations online (eBay, Yahoo)Collectibles (paintings, books, antiques)Agricultural products (flowers, fish, tobacco)

Page 15: Lect 8-auctions

Advantages of auctionsMarkets may not exist for what the seller wants to sell (phone spectrum)The seller does not know how much an item is worthCreate competition; enhances the seller’s bargaining powerFlexibilityLess time-consuming and expensive than negotiating a priceSimplicity in determining market-based prices

Page 16: Lect 8-auctions

Auction participantsThe auctioneer:

A seller who wants to sell goods at the highest possible price, orSomeone who wants to subcontract out contracts at the lowest possible price, orA buyer who wants to buy a good

The bidders:Buyers who want to acquire goods at the lowest possible price, orContractors who want to get a contract at the highest price, orSellers who want to sell their goods

Page 17: Lect 8-auctions

Properties of participantsAgents are:

Self-interested: they are seeking to maximise their own payoffsRational: they always prefer a larger payoff than a smaller one

Different attitudes towards risk:Risk-neutral: agents who bid conservatively and will never exceed their true valuationsRisk-prone: agents who are likely to raise their bids so that they are more likely to win

Page 18: Lect 8-auctions

Basic concepts (I)

Bids: offers by bidders to buy or sell the auctioned itemBuy bid: the price a bidder is willing to pay to own an itemSell bid: the price a bidder is willing to accept to sell an itemReservation price: Maximum/minimum price that a buyer/seller is willing to pay/accept for an item – usually private information

Page 19: Lect 8-auctions

Basic concepts (II)

Process a bid: the auctioneer checks the validity of a bid according to the rules of the auction protocol and updates its databaseClearance: matching between buyers and sellers and setting of the transaction priceClearing price: final transaction price that the winning bidder pays and the auctioneer receives

Page 20: Lect 8-auctions

Typical auction process (I)Buyers and sellers register with the auction houseThe auction event is set upThe auction is scheduled and advertised(local press, Internet, etc.)The actual bidding phase takes place according to the rules of the auction

Information may be revealed to participants to some extent during the bidding phase

Page 21: Lect 8-auctions

Typical auction process (II)

The bids are evaluated and the auction closes with a winner being determined if any of the bids is successfulThe transaction phase takes place: the buyer pays for the goods and the seller ships/delivers them

Page 22: Lect 8-auctions

Classification of auctions (I)

(1) Bidding rulesSingle good or combinatorialSingle attribute or multi-attributeOpen (outcry) or sealed-bidAscending or descendingSingle unit or multi-unit

Page 23: Lect 8-auctions

Classification of auctions (II)(2) Information revelation policy

When to reveal information: on each bid, at predetermined points in time, on inactivity, etc.What information

Bid: the price a seller would have to offer in order to tradeAsk: the price a buyer would have to offer in order to tradeAuction closure: known, unknown, after a period of inactivity

To whom: participants only, everyone

Page 24: Lect 8-auctions

Classification of auctions (III)

(3) Clearing policyWhen to clear: on each bid, on closure, periodically, after a period of inactivityWho gets what (allocation and winner determination problem)At what prices: first, second price or other

Page 25: Lect 8-auctions

Basic auction formats

Single sidedEnglishDutchFPSB-First price sealed bidVickrey

Multi-attributeCombinatorialOthers …

Page 26: Lect 8-auctions

English auction (I)Open-outcry and ascending-price auctionThe auctioneer announces an opening or the reserve priceBidders raise their bids and the auction proceeds to successively higher bids The winner of the auction is the bidder of the highest bid

Page 27: Lect 8-auctions
Page 28: Lect 8-auctions

English auction (II)

The best strategy is to bid a small amount above the previous high bid until one reaches its private value and then stopThe bidders gain information by observingthe others’ bids

Bid increase, doubts, time taken to bid, …

Page 29: Lect 8-auctions

English auction disadvantages (I)

The reserve price may not be met and the item may remain unsoldThe auctioneer may cheat by overstating a reserve price or present a reserve price that does not exist Susceptible to rings (collusions)In real life it can become complicated (voices, signals)

Page 30: Lect 8-auctions

English auction disadvantages (II)

Bidders can become carried away and overbidThe seller may not receive the maximum value for an itemVulnerable to shills

Person associated to the seller that pretends to be an enthusiastic customer

Phantom bidThe auctioneer calls a bid that no one has made

Page 31: Lect 8-auctions
Page 32: Lect 8-auctions

Dutch auction (I)

Open and descending-price auctionThe auctioneer announces a very high opening bidThen the auctioneer keeps lowering the price until a bidder accepts it – the winner pays the price of its bid

Page 33: Lect 8-auctions
Page 34: Lect 8-auctions
Page 35: Lect 8-auctions
Page 36: Lect 8-auctions

Dutch auction (II)

Bidders need to decide in advance the maximum amount that they are willing to bidA bidder must decide when to stop the auction (bid) based upon its own valuation of the commodity and its prior beliefs about the valuations of the other biddersNo relevant information on the valuation of the other bidders is disclosed during the process of the auction until it is too late

Page 37: Lect 8-auctions

Comments on Dutch auctions

Very efficient in real time: used in The Netherlands for selling fresh flowers, Ontario tobacco auction, fresh produce auctions, fish marketsThe effect of competition on the buyers is stronger in the Dutch auction than in the English one

Page 38: Lect 8-auctions
Page 39: Lect 8-auctions

First-price sealed-bid auction (I)Each bidder submits its own bid (usually in writing) without knowledge of the bids of othersTwo phases: (i) the bidding phase in which participants submit their bids(ii) the resolution phase in which the bids are opened and the winner is determinedThe highest bidder wins and pays the amount of its bid

Page 40: Lect 8-auctions
Page 41: Lect 8-auctions

First-price sealed-bid auction (II)

An agent’s strategy depends on its own valuation and prior beliefs about the other bidders’ valuationsA higher bid raises the probability of winning, but lowers the bidder’s profit if it wins Bidders are better off not bidding their true valuations but a small amount below it

Trade-off between the profit they can get and the probability of losing the auction

Page 42: Lect 8-auctions

Multi-unit auctions

In many kinds of auctions there is a certain quantity of units of the same item, which may be distributed between different buyersIn a multi-unit auction sealed bids are sorted from high to low, and the items awarded at highest bid price until the supply is exhausted

Page 43: Lect 8-auctions

Multi-unit auction example –

1000 units

200 units at 35 euros250 units at 30 euros300 units at 28 euros100 units at 27 euros150 units at 26 euros500 units at 25 euros225 units at 20 euros…

Awarded units (mean price: 29,5 euros/unit)

Page 44: Lect 8-auctions

Vickrey auction

A second-price sealed-bid auction, also known as uniform second-price sealed-bid or the philatelist auctionTwo distinct phases: the bidding phase and the resolution phaseThe highest bid wins but the bidder pays the amount of the second-highest bid

William Vickrey, Economics Nobel Prize 1996

Page 45: Lect 8-auctions

Why do Vickrey auctions work?

Although it seems that the auctioneer would make more money out of a first-price sealed auction, this is not the caseBidders adjust their bids upwards since they are not deterred by fear that they will have to pay too muchThe price that the winning bidder pays depends on the others’ bids alone and not on any action that the bidder undertakes

Page 46: Lect 8-auctions

Comments on Vickrey auctions

The best strategy is for the agent to bid its true valuationThe bidders do not waste time in counter-speculating what the other bidders will do

Page 47: Lect 8-auctions

Best strategy (I)

A bidder has a dominant strategy to bid an amount equal to their valuation, that is: bi = viConsider bidding v - x if your true value is vSuppose the highest bid other than yours is w

Case 1: If v - x > w you win the auction and pay w, just as if you had bid v

Page 48: Lect 8-auctions

Best strategy (II)

Case 2: If w > v you lose the auction and getnothing, just as if you had bid vCase 3: If v> w > v - x bidding v - x causes you to lose the auction, whereas if you had bid v, you would have won the auction and paid w for a net surplus of v - w

Hence you can only lose, and never gain, bybidding v - x

Page 49: Lect 8-auctions

English vs Dutch auctionsUnder some conditions, Dutch is equivalent to FPSB, and Vickrey is equivalent to EnglishDilemma: English (Vickrey) or sealed-bid (Dutch) auction?Ascending auction allows for collusion andentry deterrence, but is likely to achieve an efficient outcomeSealed-bid auction may prevent collusion, allows entry of weaker bidders, but leaves probability of inefficient outcome

Page 50: Lect 8-auctions

Solution: “Anglo-Dutch” auction

Auctioneer runs an English auction in which the price is raised continuously until all but two (or n) bidders have dropped outThe remaining bidders are then required to make a final sealed-bid offer that is no lower than the current asking price, and the winner pays the bid

Page 51: Lect 8-auctions

Advantages of Anglo-Dutch auction

The first stage of the auction (English) means less loss of efficiency than might result from pure sealed-bid auctionSealed-bid at the final stage induces some uncertainty about which of the finalists will win, and entrants are attracted by the chance to make it to the final

Page 52: Lect 8-auctions

Multi-attribute auctions

Multi-attribute or multi-dimensional auctions allow bidders to submit bids on more than one attribute or dimension of a goodVery common in procurement situationsThe attributes under negotiation are usually defined in advance and bidders can compete in open-cry or sealed-bid auctions

Page 53: Lect 8-auctions

Example: a procurement auction (I)

Consider a manufacturer which uses raw materials from a number of suppliers to produce finished goods Assume the manufacturer requires 1000 units of A by 31/03/2010The manufacturer sends a Request For Quotes (RFQ) to all potential suppliers of A

Page 54: Lect 8-auctions

Example: a procurement auction (II)

Suppliers reply with a bid bi

<quantity, delivery_date, price>The delivery_date and the quantity may deviate from those requested by the manufacturer The price offered by the various suppliers will differ tooThe manufacturer’s request could be covered by one or more bids in combination (sole or multiplesourcing)

Page 55: Lect 8-auctions

Winner determinationPreference elicitation through an appropriate scoring functionA multi-attribute offer has the form

vi =(vi1 ,…,vij ) where vij is the level of attribute j offered by i

Each bid is evaluated through a scoring function S(vi) which is a weighted averagefunction – often scaled from 0 to 1

Page 56: Lect 8-auctions

Example: 1000 units needed by 31/03/2010

S1: 500 units, 28/03, 30 euros/unitS2: 1000 units, 4/04, 38 euros/unitS3: 700 units, 2/04, 37 euros/unitS4: 300 units, 30/03, 42 euros/unit

If completeness is preferred => S3+S4 or S2S2 (1000 units, 38000 euros, by April 4th)S3+S4 (1000 units, 38500 euros, by April 2nd)

If sole-sourcing is preferred => S2If deadline has to be kept => S1+S4

S1+S4 (800 units, 27600 euros, by March 30th)

Page 57: Lect 8-auctions

Scoring functions and weightsAttributes may be interdependent: an additive scoring function does not suffice => could use multi-linear expression

Example with two attributes x,y

The scoring function, if revealed, serves as a guideline to the bidderThe winning bid is the one with the highest score

Page 58: Lect 8-auctions

Advantages of multi-attribute auctions

They allow more degrees of freedom for bidders: price may not be the only attribute of interestMore efficient information exchangeamong the market participants

Page 59: Lect 8-auctions

Combinatorial auctions

Auctioning multiple distinguishable items when bidders have preferences over combinations of items: complementarity & substitutabilityApplications: allocation of transportation tasks, bandwidth, manufacturing procurement, electricity markets

Example: auctioning a three-seater sofa, a two-seater one, two armchairs and a coffee table

What’s the best way to sell these goods? As a complete set? Each item separately? Allow combinations?

Page 60: Lect 8-auctions

Sequential vs parallel auctionsSequential auctions: run individual auctions one after the other

Impossible to determine best strategy because game tree is hugeInefficiencies can result from future uncertainties

Parallel auctions: run individual auctions in parallelInefficiencies can still result from future uncertaintiesDifficult to keep track of several simultaneous auctions with substitutable and interdependent goods (at least for human agents)

Page 61: Lect 8-auctions

Bundles of items

Combinatorial auctions: bids can be submitted on combinations (bundles) of itemsExample

Bundle1: the two-seater sofa and the tableBundle2: the armchairs and the tableBundle3: the two-seater sofa and the armchairs

Page 62: Lect 8-auctions

300 $

400 $

350 $

275 $

325 $

Page 63: Lect 8-auctions

Interest of combinatorial auctions

Bidder’s perspectiveAvoids the need for lookahead

Auctioneer’s perspectiveOptimal bundling of itemsWinner determination problem: Decide on the winning bids so as to maximize the sum of bid prices (i.e. revenue)

Page 64: Lect 8-auctions

Issues on combinatorial auctionsShould the combinations of items on which bids are allowed be restricted?

Limits number of possibilities to be consideredSingle-shot or iterative combinatorial auctions?

In an iterative auction, what information should be made available to bidders from round to round?

What is the objective of the auctioneer? Sell all itemsMaximise revenueMinimise number of buyers

Page 65: Lect 8-auctions

Advantages of auctions

Flexibility, as protocols can be tailor-madeLess time-consuming and expensivethan negotiating a priceSimplicity in determining the market prices

Page 66: Lect 8-auctions

Disadvantages of auctions

Winner’s curseLying auctioneerSnippingOthers: rings, revelation of private information, …

Page 67: Lect 8-auctions

Winner’s curse

What bidders suffer when they win an auction byoverestimating how much something is worth andtherefore bidding too much – the most optimistic bidder winsMore severe in common value auctions as one’s valuation depends on the others’A win may actually mean a loss …

Page 68: Lect 8-auctions

Example of winner’s curse

Suppose several petroleum firms are bidding on the drilling rights to a piece of tundra. No firm is sure how much oil is underneath the property, so they hire a team of engineers to poke at the surface rocks and make a guess. The guesses will likely range from too low to too high. Some firm's engineer will probably guess right, but that firm won't win the auction. The winner will be the firm whose engineer was the most overoptimistic. The winning firm won't ultimately get as much oil as their engineers promised, meaning the firm paid too much. In short, the auction "winner" is ultimately a loser.

Page 69: Lect 8-auctions

Facing the winner’s curse

Anticipate the effects of the winner’s curse before bidding:

Assume that you have the highest estimateRational bidding: correct downwards

Any piece of information is useful to the bidder – beware of rings !!!

How aggressively others bidHow many remain in the biddingWhen others apparently drop off the bidding

Page 70: Lect 8-auctions

Lying auctioneer (I)

In the Vickrey auction, the auctioneer may overstate the second-highest bidSolution: using cryptographic electronic signaturesCannot happen in the other 3 protocols

Page 71: Lect 8-auctions

Lying auctioneer (II)

In English auctions: shills[Read NYTimes article]In sealed-bid auctions: Auctioneer may place a bid himself (reservation price)In Vickrey auctions: the auctioneer may bid more than its true reservation price

Page 72: Lect 8-auctions

SnippingSnipping: bidding very late in the auction in the hope

that other bidders do not have time to respond and you can snatch a bargain. This is an issue, in particular in online auctions

Bidders may bid late in order to avoid:increasing the price of the item early on in the auctionrevealing their preferences early in the auction (especially experts)bidding wars with other like-minded agents

Page 73: Lect 8-auctions
Page 74: Lect 8-auctions
Page 75: Lect 8-auctions

Online auction systems

Online auctions are increasingly popular: last minute package holidays, C2C negotiations (eBay), B2B negotiationsPopularity and success of online auctions can be attributed to:

Simple mechanisms usedAllow millions of people with no specialist knowledge of negotiation protocols and strategies to exchange goods

Page 76: Lect 8-auctions
Page 77: Lect 8-auctions

Pitfalls in auction design

Collusive behaviourEntry deterrence Setting reservation prices at the right levelCommunication and computational efficiencyAdvanced auction formats may be cognitively complex and difficult to understand by participants

Page 78: Lect 8-auctions

Readings for this weekChapter 8 in M.Fasli’s book (Agent technology for e-commerce)New York Times article (June 2000): In online auctions, rings of biddersChapter 14 in M.Wooldridge’s book (An Introduction to MultiAgent Systems), 2nd edition

Page 79: Lect 8-auctions