Presentation is loading. Please wait.

Presentation is loading. Please wait.

20-751 ECOMMERCE TECHNOLOGY SUMMER 2003 COPYRIGHT © 2003 MICHAEL I. SHAMOS eCommerce Technology 20-751 Agents and Auctions.

Similar presentations


Presentation on theme: "20-751 ECOMMERCE TECHNOLOGY SUMMER 2003 COPYRIGHT © 2003 MICHAEL I. SHAMOS eCommerce Technology 20-751 Agents and Auctions."— Presentation transcript:

1

2 20-751 ECOMMERCE TECHNOLOGY SUMMER 2003 COPYRIGHT © 2003 MICHAEL I. SHAMOS eCommerce Technology 20-751 Agents and Auctions

3 20-751 ECOMMERCE TECHNOLOGY SUMMER 2003 COPYRIGHT © 2003 MICHAEL I. SHAMOS What is an Agent? In real life, a person who acts on your behalf In eCommerce, a computer program that acts on your behalf Agents often perform tasks usually associated with humans But: an agent is just a computer program with certain properties Synonyms: –bot –daemon (a supernatural being of Greek mythology intermediate between gods and men)

4 20-751 ECOMMERCE TECHNOLOGY SUMMER 2003 COPYRIGHT © 2003 MICHAEL I. SHAMOS Agent Properties Autonomous –Acts by itself (independent of user) Reactive –Responds to its environment, initiates actions Communicative –Communicates with people and other agents Goal-driven –Acts until it accomplishes its purpose or learns that it can’t

5 Intelligent Agents: Key Features Proactive Don’t constantly need instructions Able to work unaided Learn Improve their actions with experience Adapt to user requirements Cooperate Share information with each other. Able to agree on subtasks E-commerce/ E-business Agents Information Management Agents Really smart Agents “Hidden” Agents SOURCE: BEN AZVINE

6 20-751 ECOMMERCE TECHNOLOGY SUMMER 2003 COPYRIGHT © 2003 MICHAEL I. SHAMOS Examples of Agents Search agents –Find web pages. FastSearch, GoogleFastSearchGoogle Metacrawlers –Search multiple indexes. Dogpile, MetaCrawlerDogpileMetaCrawler News agents –Locate relevant news stories. TotalNEWSTotalNEWS Monitors, update agents –Notify user when events occur, e.g. page is modified ChangeDetection, CyberAlert (company news), Enfish tracker (tracks email, web pages, files) MorningPaper ChangeDetectionCyberAlertEnfish trackerMorningPaper Instruction agents –How to do things. eHow (repair a roof)eHow

7 20-751 ECOMMERCE TECHNOLOGY SUMMER 2003 COPYRIGHT © 2003 MICHAEL I. SHAMOS Information Agents Addresses, phone numbers, reverse directories –AT&T AnyWho, BigYellow, InfoSpace (by address!)AT&T AnyWhoBigYellowInfoSpace Stock bots (financial information, charts, news) –StockPoint, Silicon Investor, biz.yahoo, 1jumpStockPointSilicon Investorbiz.yahoo1jump Filtering agents –Remove unwanted data not fitting profile

8 20-751 ECOMMERCE TECHNOLOGY SUMMER 2003 COPYRIGHT © 2003 MICHAEL I. SHAMOS Shopping Agents

9 20-751 ECOMMERCE TECHNOLOGY SUMMER 2003 COPYRIGHT © 2003 MICHAEL I. SHAMOS Shopping Agents Price bots –BestBookBuys, BottomDollar, Point&Shop, PillBot (medication)BestBookBuysBottomDollarPoint&ShopPillBot Sale locators –ShoppingList.com (brick & mortar), ValueFindShoppingList.comValueFind Auction notification –BidFindBidFind Recommenders –ActiveSalesAssistantActiveSalesAssistant

10 20-751 ECOMMERCE TECHNOLOGY SUMMER 2003 COPYRIGHT © 2003 MICHAEL I. SHAMOS Travel Agents Information about flights, trains, purchase tickets –USAirways, OrbitzUSAirwaysOrbitz Discount Hotels –hoteldiscount!comhoteldiscount!com Airplanes in flight –FlightView, FlyteComm, DFWFlightViewFlyteCommDFW –Chicago towerChicago tower

11 20-751 ECOMMERCE TECHNOLOGY SUMMER 2003 COPYRIGHT © 2003 MICHAEL I. SHAMOS Examples of Agents Negotiation agents Agent Builder ToolsTools CMU Bot List CMU Agent Page

12 20-751 ECOMMERCE TECHNOLOGY SUMMER 2003 COPYRIGHT © 2003 MICHAEL I. SHAMOS Avatars A simulated human being From Sanskrit: “Earthly incarnation of a Hindu god or goddess” Verbot

13 20-751 ECOMMERCE TECHNOLOGY SUMMER 2003 COPYRIGHT © 2003 MICHAEL I. SHAMOS Agent Technologies Table-driven (data lookup) Rule-based Goal-directed Utility-based inputs “ ”

14 20-751 ECOMMERCE TECHNOLOGY SUMMER 2003 COPYRIGHT © 2003 MICHAEL I. SHAMOS Rule-Based Agents Condition-action rule: if car-in-front-is-braking then start-braking SOURCE: ANDREAS GEYER-SCHULZ

15 20-751 ECOMMERCE TECHNOLOGY SUMMER 2003 COPYRIGHT © 2003 MICHAEL I. SHAMOS Business Rules Grocery store example IF inBasket(french_fries) AND NOT asked(ketchup) THEN ask(ketchup) ; ask “Would you care for ketchup to go ;with your french fries?” Rules that learn IF inBasket(french_fries) THEN prob(want_ketchup) = SQL( ) ; query might involve customer data and ; demographics IF prob(want_ketchup) > 0.3 AND NOT asked(ketchup) THEN ask(ketchup)

16 20-751 ECOMMERCE TECHNOLOGY SUMMER 2003 COPYRIGHT © 2003 MICHAEL I. SHAMOS Fuzzy Logic Traditional set theory: Set membership. If b i = 1 then e i  T else e i  T S = { a, b, c, d, e, f, g } b = 0, 0, 1, 1, 0, 1, 1 then T = { c, d, f, g } b is the membership function Fuzzy set theory The membership function can be any value in [0, 1] Often interpreted as a probability S = { a, b, c, d, e, f, g } b = ½, 0, 1, ¾, 0, 1, ¼ Now what is T? (g is 25% in T, 75% not in T)

17 20-751 ECOMMERCE TECHNOLOGY SUMMER 2003 COPYRIGHT © 2003 MICHAEL I. SHAMOS Fuzzy Logic IF temperature IS hot AND humidity IS sticky THEN air_conditioning = high What is hot? Is 60° hot? 70°? 80°? 90°? –60° is definitely not hot; 90° is definitely hot; everything else is “in between” hot(60) = 0; hot(90) = 1; hot(75) = 0.4 etc. What is sticky”? 80%, 90%, 100%? Fuzzy logic has fuzzy inference rules, e. g. A  B = A*B HOT COLD TEMPERATURE, °F

18 20-751 ECOMMERCE TECHNOLOGY SUMMER 2003 COPYRIGHT © 2003 MICHAEL I. SHAMOS Goal-Directed Agents Actions are evaluated with respect to goals Will this action get me closer to the goal state? SOURCE: ANDREAS GEYER-SCHULZ

19 20-751 ECOMMERCE TECHNOLOGY SUMMER 2003 COPYRIGHT © 2003 MICHAEL I. SHAMOS Static versus Mobile Agents Static Agent System Mobile Agent System SOURCE: MITSUBISHI

20 20-751 ECOMMERCE TECHNOLOGY SUMMER 2003 COPYRIGHT © 2003 MICHAEL I. SHAMOS Contract Formation Execution Framework Business Logic Formation Fulfilment Execution Framework Business Logic Formation Fulfilment Buyer Seller Contract SOURCE: CHRIS PREIST, HP

21 20-751 ECOMMERCE TECHNOLOGY SUMMER 2003 COPYRIGHT © 2003 MICHAEL I. SHAMOS Event Scheduling Execution Framework Business Logic Formation Fulfilment Execution Framework Business Logic Formation Fulfilment Buyer Seller contract and schedule execution SOURCE: CHRIS PREIST, HP

22 20-751 ECOMMERCE TECHNOLOGY SUMMER 2003 COPYRIGHT © 2003 MICHAEL I. SHAMOS Obligation Pending Execution Framework Business Logic Formation Fulfilment Execution Framework Business Logic Formation Fulfilment Buyer Seller An obligation is pending. Fulfilment is triggered. SOURCE: CHRIS PREIST, HP

23 20-751 ECOMMERCE TECHNOLOGY SUMMER 2003 COPYRIGHT © 2003 MICHAEL I. SHAMOS Invoke Processes Execution Framework Business Logic Formation Fulfilment Execution Framework Business Logic Formation Fulfilment Buyer Seller Processes to fulfill the obligation are retrieved and invoked SOURCE: CHRIS PREIST, HP

24 20-751 ECOMMERCE TECHNOLOGY SUMMER 2003 COPYRIGHT © 2003 MICHAEL I. SHAMOS Execute Processes Execution Framework Business Logic Formation Fulfilment Execution Framework Business Logic Formation Fulfilment Buyer Seller Processes are executed SOURCE: CHRIS PREIST, HP

25 20-751 ECOMMERCE TECHNOLOGY SUMMER 2003 COPYRIGHT © 2003 MICHAEL I. SHAMOS Mark Obligation Fulfilled Execution Framework Business Logic Formation Fulfilment Execution Framework Business Logic Formation Fulfilment Buyer Seller SOURCE: CHRIS PREIST, HP The obligation is fulfilled

26 20-751 ECOMMERCE TECHNOLOGY SUMMER 2003 COPYRIGHT © 2003 MICHAEL I. SHAMOS Signal Fulfilment Execution Framework Business Logic Formation Fulfilment Execution Framework Business Logic Formation Fulfilment Buyer Seller SOURCE: CHRIS PREIST, HP Notify seller

27 20-751 ECOMMERCE TECHNOLOGY SUMMER 2003 COPYRIGHT © 2003 MICHAEL I. SHAMOS Acknowledge Fulfilment Execution Framework Business Logic Formation Fulfilment Execution Framework Business Logic Formation Fulfilment Buyer Seller SOURCE: CHRIS PREIST, HP Acknowledge fulfilment

28 20-751 ECOMMERCE TECHNOLOGY SUMMER 2003 COPYRIGHT © 2003 MICHAEL I. SHAMOS Cooperating Agents SOURCE: PETER FINGARPETER FINGAR

29 Networks of Agents Local Objects: Financial agent Purchasing agent Inventory management agent Customer Services agent Other objects Beverage Intermediate Agent 3 4 1 2 5 DB2 DB1 Soup Intermediate Agent Intermediate Agents: Interfacing Networking and searching Optimal Matching Gourmet-to-Go COCA-COLA CAMPBELL SOURCE: MOHSEN JAFARI, RUTGERS

30 20-751 ECOMMERCE TECHNOLOGY SUMMER 2003 COPYRIGHT © 2003 MICHAEL I. SHAMOS Major Ideas Agents are the wave of the future –laziness + information overload = agents Agent systems are object-oriented and distributed Agents are mobile Agents negotiate with and talk to other agents

31 20-751 ECOMMERCE TECHNOLOGY SUMMER 2003 COPYRIGHT © 2003 MICHAEL I. SHAMOS Auctions Process for matching buyers and sellers and setting prices. Allows supply and demand to operate “a market institution with an explicit set of rules determining resource allocation and prices on the basis of bids from the market participants” -- McAfee & Macmillan (1987) A protocol for exchanging bids and determining a winner Thousands of different protocols possible, depending on auction rules Auctions are complicated

32 20-751 ECOMMERCE TECHNOLOGY SUMMER 2003 COPYRIGHT © 2003 MICHAEL I. SHAMOS Auctions Foreign exchange ($2T per day) Securities ($100B per day) Government debt –treasury bills > $1T per year, municipal bonds Public works (roads, bridges) Private construction Frequency spectrum Oil drilling rights Fishing quotas Cars (Japan Aucnet) Sheep and cattle (Australia)

33 20-751 ECOMMERCE TECHNOLOGY SUMMER 2003 COPYRIGHT © 2003 MICHAEL I. SHAMOS Price Discovery In a marketplace, how are prices set? Slow movement toward equilibrium, called tatonnement (from French tâtonner, “to feel one’s way”) Exchange of price information in many steps Why? No one wants to reveal his price at the start Could be instantaneous if prices were revealed to a neutral agent

34 20-751 ECOMMERCE TECHNOLOGY SUMMER 2003 COPYRIGHT © 2003 MICHAEL I. SHAMOS Valuations Private value : value of the good depends only on the bidder’s own preferences –Refrigerator to be used at home Common value : bidder’s value of an item determined entirely by independent valuation –Treasury bills Correlated value : bidder’s value depends partly on own preferences & partly on others’ values –Manufacturing contract whose tasks can be subcontracted out –Goods for resale (art purchased by a dealer)

35 20-751 ECOMMERCE TECHNOLOGY SUMMER 2003 COPYRIGHT © 2003 MICHAEL I. SHAMOS Auction Rules When does the auction start and end? Who can participate? (public, registered users, trading partners) What type of auction? Is the number of bidders (& Identities) known? How are bids submitted? (timing, increments) How many bids? Can bids be withdrawn? Who can see the bids? (open, sealed) How is winner determined? How is price determined? Payment terms AUCTION RULES AFFECT THE FINAL PRICE

36 20-751 ECOMMERCE TECHNOLOGY SUMMER 2003 COPYRIGHT © 2003 MICHAEL I. SHAMOS English Auctions One item, one seller, multiple bidders Bidders call out increasing bids Auction ends when bidding stops What price will win this auction? –Private values: $10, $17, $18, $20, $23 Answer: the smallest bidding increment over the second-highest bid NOT the maximum anyone is willing to bid

37 20-751 ECOMMERCE TECHNOLOGY SUMMER 2003 COPYRIGHT © 2003 MICHAEL I. SHAMOS Dutch Flower Auction Many items, one seller, multiple bidders Price declines with time according to a clock First bidder to accept a price wins the auction, buys the quantity he wants Auctioneer resets the price clock, restarts Price tends to go up as supply decreases Very fast, often used for perishables: flowers, fish

38 20-751 ECOMMERCE TECHNOLOGY SUMMER 2003 COPYRIGHT © 2003 MICHAEL I. SHAMOS Dutch Flower Auction Clock SOURCE: BADM.SC.EDUBADM.SC.EDU SIMULATION ONTARIO AUCTION WEBCAM

39 20-751 ECOMMERCE TECHNOLOGY SUMMER 2003 COPYRIGHT © 2003 MICHAEL I. SHAMOS Fish Auction SOURCE: ALLOT.COMALLOT.COM FISH AUCTION

40 20-751 ECOMMERCE TECHNOLOGY SUMMER 2003 COPYRIGHT © 2003 MICHAEL I. SHAMOS Single Sealed-Bid Auctions One item, one seller, multiple bidders Bidders each submit ONE bid, which they cannot withdraw The highest bid wins If a bidder values the item at $v, how much should he bid? ANSWER: less than $v, depending on the number of bidders and his estimate of the behavior of other bidders

41 20-751 ECOMMERCE TECHNOLOGY SUMMER 2003 COPYRIGHT © 2003 MICHAEL I. SHAMOS Sealed-Bid Second-Price (Vickrey) Auctions One item, one seller, multiple bidders Bidders each submit ONE bid, which they cannot withdraw The highest bid wins, but the price paid is the amount of the second-highest bid If a bidder values the item at $v, how much should he bid? ANSWER: exactly $v

42 20-751 ECOMMERCE TECHNOLOGY SUMMER 2003 COPYRIGHT © 2003 MICHAEL I. SHAMOS Revenue Equivalence Theorem William Vickrey (1961) The following auction rules produce the same theoretical revenue for the seller if buyers are risk- neutral in a private-value auction: –English –Dutch (one item) –single sealed-bid –second-price sealed-bid (Vickrey) Nobel Prize, 1996

43 20-751 ECOMMERCE TECHNOLOGY SUMMER 2003 COPYRIGHT © 2003 MICHAEL I. SHAMOS Trading Rules of a Double Auction BUYERSSELLERS lowest highest Trading price = (buy price + sell price) / 2 Highest buyer matched to the lower seller, 2nd highest buyer is matched to the 2nd lowest seller, etc. Condition: buy price  sell price 5432154321 5.1 4.1 2.8 2.4 1.6 PRICE: 3.3 PRICE: 3.2 PRICE: 2.9 SOURCE: JUNLING HU

44 20-751 ECOMMERCE TECHNOLOGY SUMMER 2003 COPYRIGHT © 2003 MICHAEL I. SHAMOS NASDAQ National Association of Securities Dealers Automated Quotations 61,000 computers 6500 stocks Continuous double auctions (CDA) ARCA Integrated BookBook

45 Some Auction Types Auction Double-sided Single-sided Sealed-bid Outcry Sealed-bid Dutch English First Price or Vickrey Call Market Descending Ascending CDA Clearing House Asynchronous Synchronous SOURCE: JUNLING HU CDA = CONTINUOUS DOUBLE AUCTION, E.G. NEW YORK STOCK EXCHANGE

46 20-751 ECOMMERCE TECHNOLOGY SUMMER 2003 COPYRIGHT © 2003 MICHAEL I. SHAMOS eAuction Types Used in Practice SOURCE: BEAM & SEGEV

47 Auction Types SOURCE: LAUDON & TRAVER

48 20-751 ECOMMERCE TECHNOLOGY SUMMER 2003 COPYRIGHT © 2003 MICHAEL I. SHAMOS Buyer Seller register Pref. alertsearch select displaybid Registration Pref.& Target Product Auction Rules Short list auction target define update register notify Setup auction cancel update START bids offer select END auction evaluate registrationProduct description & auction setup bidding Auction close & evaluating bids Delivery & Payment SOURCE: JERRY GAO Online Auction Structure

49 20-751 ECOMMERCE TECHNOLOGY SUMMER 2003 COPYRIGHT © 2003 MICHAEL I. SHAMOS Major Ideas Auctions are the most important market mechanism Primarily an information exchange protocol Very complicated: small changes to rules cause large changes in behavior Agents can be used in auctions

50 20-751 ECOMMERCE TECHNOLOGY SUMMER 2003 COPYRIGHT © 2003 MICHAEL I. SHAMOS Q A &


Download ppt "20-751 ECOMMERCE TECHNOLOGY SUMMER 2003 COPYRIGHT © 2003 MICHAEL I. SHAMOS eCommerce Technology 20-751 Agents and Auctions."

Similar presentations


Ads by Google