Presentation is loading. Please wait.

Presentation is loading. Please wait.

IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 1 Agent-mediated electronic commerce Carles Sierra IIIA-CSIC Barcelona.

Similar presentations


Presentation on theme: "IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 1 Agent-mediated electronic commerce Carles Sierra IIIA-CSIC Barcelona."— Presentation transcript:

1 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 1 Agent-mediated electronic commerce Carles Sierra IIIA-CSIC Barcelona

2 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 2 Tutorial plan Context Agents and eCommerce Mechanisms Example: The fishmarket Example: Robot Navigation Negotiation Argumentation Electronic Institutions Future trends

3 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 3 Context

4 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 4 eCommerce evolution Electronic malls Portals: aggregate information and commerce resources, add services Auction-centred sites Vertical markets/portals Wholesale, consumer aggregation Differentiation: B2B, B2C, C2C Latest growth: B2B vertical markets

5 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 5 Internet growth 17 million users in million users world-wide in 1999 Disparity: Sweden 40.9%, Italy 8% of usage Users already sampled buying over the web (f.i. 40% in the UK) Many regular shoppers (f.i. 10% in the UK)

6 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 6 Expenses produced Disparity Shoppers in Europe Revenue 1999 Shoppers 2002 Revenue 2002 Finland 20 times more than Spain. 5.2 million EUR3,032 million 28.8 million EUR57,210 million

7 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 7 Examples C2C auctions: eBay –Consumer to consumer auctions –Services: auction engine, reputation management –Revenue source: advertising, auction commissions B2C: amazon –Catalog-based buying, auctions –Services: transactions, delivery, recommendation –Revenue source: advertising C2B: priceline –Reverse auction B2B: chendex, partMiner, metalsite –Catalog, auction –Revenue source: membership, per-transaction fees.

8 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 8 Current eMarkets The market is Internet! –Sellers. Usualy electronic, automated sites –Buyers. Not automated, humans. –Third parties. Few. Limited services (shopbots) Buying is based on buyers visiting sellers. eMarketplaces Protocols: –Single attribute auctions (several types) –Buying from catalog –Reverse auctions

9 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 9 Buyers/Sellers relationship One buyer, one seller (1:1) –A buyer interacts with a single seller at a single site –Trade has no regard to other buyers and sellers –Example: buying a book at amazon.com Many buyers, one seller (N:1) –Many buyers visit a single sales site –Trade depends on other buyers (auctions) One buyer, many sellers (1:M) –A buyer visits multiple sites simultaneously –Negotiation is possible. No regards to other buyers Many buyers, many sellers (N:M) –Many buers visit many sites –Coordination is possible –Many buyers and sellers visit a single site (exchange)

10 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 10 eMarketplace assumptions Single point: a market is a single meeting point for buyers and sellers Time spending: buyers are willing to spend time. Limited privacy: buyers are willing to surrender private information to sellers Price dominance: price is the main affector of buying decisions. No collusion: buyers/sellers do not collude. Familiarity: buyers can locate needed sellers. Interoperability: all sites understand each other

11 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 11 Do assumptions hold? Single point: electronic trade takes place at multiple sites, possibly inter-correlated; buyers may have myriad alternative markets. Time spending: buyers prefer to reduce time spent. Limited privacy: buyers may prefer not to reveal private information. Price dominance: other attributes are important too: delivery, quality recommendations, etc. No collusion: human players may not collude (but electronic ones may). Familiarity: buyers do not necessarily know sellers and how to find them across markets. Interoperability: each site is developed by a different company with possibly different ontologies.

12 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 12 Needs Buyers mechanisms for participating in multiple markets (1:M), selection of better ones Efficient mechanisms for locating markets, sellers, other buyers Interoperation standards: language, protocol, ontology Buyer tools for time-efficient buying Seller tools for dynamic pricing, promotion Buyer, seller negotiation protocols and strategies Enforceable, or self-enforceable contracts Trust mechanisms Means for payment and goods transaction Means for secure transactions Mechanisms for keeping players privacy Tools for analyzing market performance Protocols and tools for N:M interaction and trade.

13 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 13 Agents and eCommerce

14 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 14 Why Agents may help EC Autonomy. Agents work proactively, reactively and independently of human intervention. They can wait for good deals without diverting our attention. Personalisation. Agents can be equipped with a personal profile to reflect preferences. Social ability. The communication ability of agents can be used to negotiate over prices, services and transactions. Intelligence. Agents can learn and hence perform better over time. In EC scenarios this may equate to making more money. Many AI techniques can be applied.

15 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 15 Agents in ecommerce Product brokers: –Jango –PersonaLogic –Firefly Merchant brokering –Bargainfinder, Jango, Kasbah Buyer broker –Eyes. Negotiation through trusted third parties –Kasbah, Auctionbot, Fishmarket Basic limitation: Only price brokering. No product differentiation.

16 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 16 Two examples Jango & PersonaLogic

17 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 17 Personalized seller A personal seller is (usually) an animated character that uses conversational interactions to help a customer to use an electronic commerce site. It suggests products acording to a user profile and to his/her preferences. It helps as an answer to a user demand or proactively. It can be adapted through learning.

18 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 18 MAS for eCommerce Ad-hoc applications –Contract allocation. MAGNET. –Power load management as a computational market. –Control of shipment processes (EDI). Maquiladora. Generic applications –Auctions Auctionbot Fishmarket –Virtual markets Kasbah Bazar Metamall

19 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 19 Mechanisms Voting Auctions Bargaining

20 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 20 Introduction Agents inhabiting the same environment need to co- ordinate their activities to improve their individual or collective performance. The aim of DAI is to design intelligent systems that behave efficiently. A common assumption in many applications, specially in AMEC, is that agents are self-interested and utility maximisers. In others, agents are co-operative. DAI is divided in two big areas: Distributed problem solving, where the designer determines the protocol and the strategy (relation between state and action) of each agent, and Multi Agent Systems, where the agents are provided with an interaction protocol but chose the strategy to follow.

21 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 21 Protocol Avaluation Criteria Social Welfare: Is the addition of the utilities of all the agents for a given solution. It is a global measure a bit controversial given the difficulty in comparing the different utility functions. Pareto efficiency: A solution x is pareto efficient if there is no other x such that some agent improves without anyone else losing utility. It is also a global measure. Solutions that maximise the social welfare are a subset of those pareto efficients. Individual rationality: The participation of an agent in a negotiation is rational if the benefit it gets from the negotiated solution is not smaller that the benefit of not negotiating. A mechanism, or protocol, is individualoly rational if the participation is rational for all agents. Only such mechanisms are feasible.

22 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 22 Protocol Avaluation Criteria Stability: A protocol is stable if it is designed in such a way that motivates (selfish) agents to behave in a particular way. Those behaviours are called dominant strategies. When the preferred strategy for an agent depends on the strategies of others we have other criteria for stability: Nash equilibrium: Strategies S*(A)= are in equilibrium if for each agent i, S*(i) is the best strategy suposing that the others follow In many ocasions there is no equilibrium, in others, there are several. Moreover, agents can form coallitions to deviate from the behaviour in the equilibrium. Also, eficiency and stability may conflict. For example (prisoners dilema): C D C D 3,3 0,5 5,0 1,1

23 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 23 Protocol Avaluation Criteria Computational eficiency: Mechanisms should be designed in a way that gives the minimum computational cost to the agent. Distribution and communication: Distributed protocols are preferred to centralised ones because they are more robust. A trade-off with the amount of communication must be found. In this introduction well survey three coordination mechanisms: Votings Auctions Bargaining

24 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 24 Votings In a voting all agents give an input to a mechanism and the result of the mechanism is a solution for all agents. A, Agents, O, Outcomes, every agent has a preference relation over O, (> 1,..., > |A| ). We want a > * that represents the social preference, and that satisfies: 1) > * should exist for all sets of inputs. 2) > * Should be defined for all pairs o,o dO. 3) > * Asimetric and transitive. 4) The result should be pareto efficient: if for all i o > i o then o > * o. 5) The schema should be independent of irrelevant alternatives. 6) No dictators! That is, there must be no i such that o > i o implies o > * o independently of the others. Theorem [Arrow] No election rule satisfies all requirements.

25 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 25 Relax The different voting mechanisms relax some of the points. Unfortunately the sixth is very often relaxed. Usually the first one is relaxed. By relaxing the third we have the plural protocol (the usual in democratic systems). By introducing an irrelevant alternative we may get that a less prefered outcome wins. The binary protocol, on top of this problem it also gives different results depending on the order of the pairings.

26 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 26 More relax Another example is the Borda protocol, that assigns |O| points to the most preferred alternative, |O|-1 to the next, and so up to the last. Points are added up and the alternative with more points wins. With this protocol we can also have some paradoxes when we eliminate one alternative. The design of social mechanisms tries to define them in a way that no one cheats. For instance, random choice.

27 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 27 Auctions Auctions are mechanisms very frequent in MAS. They have been deeply analysed by economists. There are three types: 1) Of private value, e.g. a cake. 2) Of common value, e.g. treasure bonds. 3) Of correlated value, e.g. contracts. Protocols: English. If it is of private value, the strategy is to increase the bids until the reserve price. In those of correlated value the auctioneer may increase the price in predetermined amounts. Sealed bid. There is no dominant strategy. Dutch. Equivalent to sealed bid. They are very efficient. Vickrey. The dominant strategy is to bid for the reserve price.

28 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 28 Auctions A centralized protocol: one auctioneer and many buyers The auctioneer puts a good for sale. Goods can sometimes be bundled or may have different attributes The buyers make offers The auctioneer determines who wins

29 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 29 Auctions: pros and cons Usually easier to prevent bidder lying Simple protocol Centralized: a single point of failure Allows collusion behind the scenes May favour the auctioneer

30 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 30 English Bidders free to raise their bid End: no more raises, winner: highest bidder at bid Strategy: a series of bids, based on private value, estimates of others valuations, their past bids Dominant strategy: bid a small amount more than current highest bid, stop when private value reached For correlated value, auctioneer increases price by constant or other rate

31 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 31 First-price sealed-bid auction Each bidder submits a bid not knowing others Highest wins, pays his bid Strategy: function of private value and beliefs about others valuations No dominant strategy. Best: bid less than true value How much less? Nash is computable if probability distribution of agents values is known Example: n agents, uniform value distribution, agent i has value v i, there is Nash if each agent i bids v i (n-1/n)

32 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 32 Dutch auction The auctioneer lowers the price until a bidder takes it The first bidder to speak takes it Strategy: equivalent to first-price sealed bid Advantage: auctioneer can do it fast!

33 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 33 Vickrey (second-price sealed-bid) Each bidder submits one bid, not knowing others The highest bid wins but pays the second price Strategy: base bid on private value and beliefs about others values Dominant strategy: bid true valuation –If bids more and this increment made him win, the agent may end up with a loss, since it may pay more than its true value –If it bids less, there is a smaller chance of winning, and the winner may end up paying less than his true value Therefore: bid true value regardless of others

34 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 34 Which auction to choose Computation criterion. Auctions with dominant strategies (Vickrey and English) are more efficient - no need to speculate regarding other bidders. Auctioneers revenue: –Second-price is less than the true price, however first price bidders under-bid. Which effect is stronger? –For risk-neutral bidders with private independent values, the effects are equivalent –For risk-averse bidders, dutch and first-price sealed-bid auctions maximize auctioneers revenue So, are revenues equivalent?

35 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 35 Real auctions In real auctions, values are not private As a result, for 3 or more bidders, English auctions provide auctioneer revenue higher than Vickrey does Explanation: when it observes other bidders increasing their bid, a bidder increases its own valuation Both English and Vickrey are better for the auctioneer than Dutch and first-price sealed-bid.

36 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 36 Collusion Bidders can coordinate their bids to lower them In English and Vickrey auctions collusion is dominant! Example. –Agents a, b and c values of the good are 10, 10, 12, respectively –They can agree to bid 5, 5, 6, respectively –If one defects, all observe that, and can increase to real value, so there is no benefit from defection

37 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 37 Avoiding collusion In the first-price sealed-bid and Dutch auctions, bidder collusion is not dominant, but possible: –In the previous example, after a, b, c decided on 5,5,6 it is beneficial for a, and b to bid more than 5. For any bid of c below 10 they can bid and win. In sealed bid, Dutch and Vickrey all bidders must identify each other and collude jointly. External bidder can win. In the English auction identifying is through bidding. Computerized anonymization can prevent identification and collusion.

38 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 38 Insincere Auctioneer Private value auctions –Vickrey: auctioneer can overstate the second highest bid to the winner –Solution: electronic signature Non-private value –English: auctioneer can use shills that bid in the auction to increase real bidders valuation –Any auction: auctioneer may bid, to guarantee a minimum price

39 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 39 Revenue In private or common value auctions, the four types of protocol are pareto eficient, they assign the resource to the buyers that value them most. English and Vickrey are more efficient because they have a dominant strategy. Buyers dont need to think on what the others are going to do. Theorem (revenue equivalence) The four protocols produce the same revenue to the auctioneer in auctions of private value with the values distributed independently and with risk- neutral buyers. The protocols are not completely protected against buyer coalitions, although sealed bid and Dutch do not favor collution. The electronic versions of the protocols go against collutions because they may avoid the mutual identification of buyers.

40 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 40 Bargaining In bargaining, agents may make deals that are mutually beneficial, but they are in conflict over which deal to chose. Negotiation mechanisms fall mainly on strategic bargaining. Axiomatic Theory. The desired solutions are not those found in a certain equilibrium, but those that satisfy a set of axioms. Classical axioms are those of Nash: outcome u*=(u 1 (o*), u 2 (o*)) must satisfy: Invariance: The numerical utilities of agents represent ordinal preferences, numerical values dont matter. Anonimity: Changing the labels of the players does not affect the outcome. Independence of irrelevat alternatives: if we eliminate some o, but not o*, o* is still the solution. Pareto eficiency: we cannot give more utility to both players over u*=(u 1 (o*), u 2 (o*)). Nash bargaining solution: o*=arg max o [ u 1 (o)- u 1 (o fallback) ] [ u 2 (o)- u 2 (o fallback) ]

41 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 41 Bargaining Strategic Theory: No axioms on the solution are given, the interaction is modelled as a game. The analysis consists on finding which strategies of the players are in equilibrium. It explains the behaviour of utility maximisers better than the axiomatic theory (where the notion of strategy does not make much sense). The theory of negotiation is basically here. Without assuming perfect rationality, the computational costs of the deliberation and the potential benefits of bargaining conflict. AI (and Agents) has many things to say on this task.

42 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 42 Bargaining In bargaining, agents may make deals that are mutually beneficial, but they are in conflict over which deal to chose. Negotiation mechanisms fall mainly on strategic bargaining. Axiomatic Theory. The desired solutions are not those found in a certain equilibrium, but those that satisfy a set of axioms. Classical axioms are those of Nash: outcome u*=(u 1 (o*), u 2 (o*)) must satisfy: Invariance: The numerical utilities of agents represent ordinal preferences, numerical values dont matter. Thus, the utility functions must satisfy that for any f linear and increasing: u*(f(o), f(o fail ))=f(u*(o, o fail )) Anonimity: Changing the labels of the players does not affect the outcome. Independence of irrelevat alternatives: if we eliminate some o, but not o*, o* is still the solution. Pareto eficiency: we cannot give more utility to both players over u*=(u 1 (o*), u 2 (o*)).

43 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 43 Bargaining Strategic Theory: No axioms on the solution are given, the interaction is modelled as a game. The analysis consists on finding which strategies of the players are in equilibrium. It explains the behaviour of utility maximisers better than the axiomatic theory (where the notion of strategy does not make much sense). The theory of negotiation is basically here. Without assuming perfect rationality, the computational costs of the deliberation and the potential benefits of bargaining conflict. AI has many things to say on this task.

44 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 44 An application: The FishMarket

45 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 45 Fish Auction in Blanes Llotja GC GV S AC Buyers registration Fish show and auction Fishermen payments Fish and sellers registration Fish delivery and payment BUYERS ADMISSION SELLERS ADMISSION AUCTIONEER AV SELLERS SETTLEMENTS BUYERS SETTLEMENTS

46 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 46 Virtual fish auction

47 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 47 Auction boss Activates the FishMarket and controls all auctioning process. It may intervene talking to other agents. Closes the auction and shuts down the program. He customizes the program

48 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 48 Sets the auction parameters Auction boss

49 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 49 Controls the auction and closes the it whenall processes are dead. Auction boss

50 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 50 Fish admission The fish admissor interacts with the program through a browser i has the following functionalities Input the fish characteristics for its identification and packaging classification in boxes Initial price setting

51 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 51 Auctioneer In FM0.9 the auctioneer agent interacts with the system through a browser in which the actual information of the auction is displayed: which buyers and sellers participate, which round is the auction in, what product is being auctioned, initial pice, etc. The browser offers the following functionalities: Control the auctioning process Select the box to auction at any time Change the starting price Start the round Decide on multiple collisions Expell buyers due to insuficent credit Etc.

52 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 52 Auctioneer

53 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 53 Buyers Buyers can interact with the auction house (buyers admitter, buyers settlements and auctioneer) through a browser: Buyer identification Messages to and from the other agents in the auction house

54 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 54 Where in the auction house is the buyer Catalogue of the products to be auctioned Buyers in the auction house Buyer credit Round number Auctioned product Seller ID Seller name Initial price Buyers

55 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 55 Buyers Bidding price Winning price Credit update To bid

56 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 56 Buyers Information on the products sold To go to different places in the auction house To bidTo update the credit To leave the auction house Information in the evolution of the auction

57 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 57 Sellers Sellers interact with the auction house (sellers admitter, sellers settlements) through a browser : Seller ID Messages from the other agents in the market

58 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 58 Sold products Session earnings To go to different places in the auction house To include products in the market catalogue To leave the market Sellers

59 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 59 Modems Buyer agent Human buyer Seller agent Human sellet LAN Boss Buyers admitter Buyers settlements Auctioneer Sellers admitter Sellers settlements Fish admitter Accountant LLotja virtual Implementation Servidor

60 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 60 Monitoring

61 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 61 Tournements

62 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 62 eBuyer

63 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 63 Animation

64 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 64 Robot navigation

65 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 65 The problem Outdoor unknown environment navigation Legged robot No precise odometry (or very imprecise one) No location system (GPS) Visual feedback only No distance to objects estimation

66 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 66 Objectives Landmark based navigation (robust, animal-like) With the aim of leading the robot to an initially given visual target in an unknown environment Qualitative navigation (fuzzy distances) Map generation (topological, landmark based)

67 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 67 Robot Architecture Navigation System Pilot System Vision System RobotCamera Target information bids actions Look for target Identify landmarks Move to direction bids actions

68 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 68 Multiagent Navigation System MMTTRMREDE CO bids bids and illocutions information MM: Map Manager TT: Target Tracker RM: Risk Manager RE: REscuer DE: Distance Estimator CO: COmmunicator

69 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 69 Example Obstacle avoidance

70 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 70 Example Obstacle avoidance Topological map Landmark regions

71 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 71 Negotiation

72 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 72 Negotiation Commerce is about interaction –Between buyers and sellers at all stages: finding, purchasing, delivery. First generation –Passive web query –Simple interactions: auctions Second generation –Rich and flexible interactions Negotiation is the key type of interaction –Process by which groups of agents communicate with one another to try and come to a mutually acceptable agreement on same matter. –Many forms exist: auctions, contract net, argumentation. –It is key because agents are autonomous: an acquaintance needs to be convinced to be influenced. –Negotiation is achieved by making proposals, trading options, offering concessions.

73 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 73 Negotiation components Negotiation objects. Issues of the agreements. Number of them, types of operations on them. Negotiation protocols. Rules that govern the interaction: permissible participants, valid actions, negotiation states. Agents reasoning model. Decision making apparatus. From simple bidding to complex argumentation. Challenges –Trust –Protocol engineering –Reasoning models

74 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 74 Negotiation object example Real State Agency. Seller b and buyer a. Issues={Address,Surface,Rooms,Brightness,Price,Garage} Negotiation thread:

75 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 75 Negotiation protocol

76 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 76 Negotiation reasoning model Each agent a negotiates over a number of issues that have a: 1) Delimited range [min j, max j ] 2) Monotonic scoring function V j a : [min j, max j ]-> [0,1] 3) Relative importance, w j a The utility function for an agent a has the following form: The negotiation protocol consists of an iterative process of offers and counteroffers until a deal is reached.

77 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 77 Tactic: Concession

78 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 78 Tactic: Imitative

79 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 79 Tactic: trade-offs Price:2 Quality:5 Price:9.9 Quality:1.1 ? A B

80 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 80 Trade-off Mechanism (I) Trade-off is lowering of utility on some issues and simultaneously demanding more on others. Steps: given x (as offer) and y (bs offer) –(1) Generate all / subset of contracts with the same utility ( ) iso a ( ) = {x | V a (x) = } –(2) selection of a contract (x´) that agent a believes is most preferable by b. B a (U b (x´) > U b (x)) U a (x´) + U b (x´) > U a (x) + U b (x) (maximization of joint utility) U a (x) = U b (x´) Step (2) is an uncertain evaluation: must model B a

81 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 81 Fuzzy Similarity Select a contract from iso a ( ) = {x | V a (x) = } that is closest or most similar to y. Implications of this choice: –not the probable choice of the other, but rather, the closeness of two contracts Not modeling of others but the domain –need a logic of degrees of truth (Zadeh) as opposed to binary truth values of true or false

82 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 82 Definition of Similarity Sim( ) defined as: Sim(x,y) = j J w j Sim j (x j,y j ) Sim j (x j,y j ) = 1 i m (h i (x j ) h i (y j )) where w j is the agent´s belief about the importance the other places on each issue in negotiation h i ( ) is ith comparison criteria function (e.g warmth) is the conjunction operator (e.g minimum) is the equivalence operator (e.g 1-| h i (x j )-h i (y j )|)

83 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 83 An Example of Similarity D colours {yellow,orange,green,cyan,red,...} Similarity of colours according to different perceptive criteria: Temperature (warm v.s cold colours) Luminosity Visibility Memory dynamicity h t = {(yellow, 0.9), (violet, 0.1), (magenta, 0.1), (green, 0.3), (cyan, 0.2), (red, 0.7),...} h l = {(yellow, 0.9), (violet, 0.3), (magenta, 0.6), (green, 0.6), (cyan, 0.4), (red, 0.8),...} h v = {(yellow, 1), (violet, 0.5), (magenta, 0.4), (green, 0.1), (cyan, 1), (red, 0.2),...}

84 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 84 Similarity of Colours Sim colour (yellow, green) = min( 1- |h t (yellow)- h t (green)|, 1-| h l (yellow)- h l (green)|, 1- |h v (yellow)- h v (green)|)= min(0.4,0.7,0.1) = 0.1 Sim colour (yellow, red) = min( 1- |h t (yellow)- h t (red)|, 1-| h l (yellow)- h l (red)|, 1- |h v (yellow)- h v (red)|)= min(0.8,0.9,0.2) = 0.2 yellow is more similar to red than to green on these criteria sim(yellow,green) and sim(yellow,red) sim colour (colour,colour) = 1 i m (h i (x colour ) h i (y colour )) i={temperature,luminosity,visibility}

85 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 85 The Trade-off Algorithm y x x y ? X´ complexity kn To be beneficial to the other the preference of the other must match the similarity function trade-off a (x,y) = arg max z iso a ( ) {Sim(z,y)}

86 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 86 Tactic: Issue-set manipulation

87 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 87 Agent Architectures

88 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 88 Case-based negotiating agent

89 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 89 Fuzzy Agent

90 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 90 GA populations

91 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 91 GA on negotiating agents

92 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 92 Argumentation

93 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 93 Argumentation Autonomy leads to negotiation and to argumentation. Many problems cannot be solved by a simple offer/counter offer negotiation protocol. When arguing, agent offers may include knowledge, information, explanations. The dialogue includes critiques on each others proposals. Agents must be able to generate arguments as well as rebutting and undercutting other agents arguments. Which argument to prefer may depend on logical criteria or on social considerations. A logically-based approach to building agents seems natural.

94 A B ++ Hang Mirror ++ Hang Picture Hang Mirror ++ Hang Mirror S S

95 A B ++ Hang Mirror ++ Hang Picture ++ Hang Mirror I know agent B has a nail S S

96 A B ++ Hang Mirror ++ Hang Picture ++ Hang Mirror ? S S

97 A B ++ Hang Mirror ++ Hang Picture ++ Hang Mirror ++ Hang Mirror S S

98 A B ++ Hang Mirror ++ Hang Picture ++ Hang Mirror ++ Hang Mirror S S S S ++ Hang Mirror

99 A B ++ Hang Mirror ++ Hang Picture ++ Hang Mirror ++ Hang Mirror ? S S S

100 A B ++ Hang Mirror ++ Hang Picture ++ Hang Mirror ++ Hang Mirror S S S

101 A B ++ Hang Mirror ++ Hang Picture ++ Hang Mirror ++ Hang Mirror ? S S S

102 A B ++ Hang Mirror ++ Hang Picture ++ Hang Mirror ++ Hang Mirror S S S

103 A B ++ Hang Mirror ++ Hang Picture ++ Hang Mirror ++ Hang Mirror ? S S S S

104 A B ++ Hang Mirror ++ Hang Picture ++ Hang Mirror ++ Hang Mirror S S S

105 A B ++ Hang Mirror ++ Hang Picture ++ Hang Mirror ++ Hang Mirror OK!!! S S S

106 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 106 Multi-context agents Units: Structural entities representing the main components of the architecture. Logics: Declarative languages, each with a set of axioms and a number of rules of inference. Each unit has a single logic associated with it. Theories: Sets of formulae written in the logic associated with a unit. Bridge Rules: Rules of inference which relate formulae in different units.

107 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 107 planner undercutting module rebutting module resource manager social manager goal manager An argumentative agent

108 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 108 GOAL MANAGER A module

109 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 109 DONE:G:goal(X),R:X ==> G:done(X) ASK:G:goal(X),G:not(done(ask(X))),G:not(done(X)),R:not(X),P:not(plan(X,Z)) ==> CU:ask(self/G,self/All,goal(X),[]),G:done(ask(X)) RESOURCE: CU>answer(self/RM,self/G,have(X,Z),[])==> R:X PLAN: CU>answer(self/_,self/G,goal(Z),P)==> P:plan(Z,P) MONITOR: G:goal(X),R:not(X),P:plan(X,P) ==> G:monitor(X,Z) NEW_GOAL: CU>inform(self/_,self/_,newGoal(X),_) ==> G:goal(X) FREE: R:X,GM:not(goal(X,_)) ==> R:free(X) FREE2: R>free(X),R>X ==> CU:free(X) FAILURE_R: R>done(ask(X,Y))FAILURE_P: P>done(ask(X,Y)) [t1] [t2] ==> GM:fail_R(X,Y) ==> GM:fail_P(X,Y) Bridge rules

110 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 110 Electronic Institutions

111 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 111 Electronic Institutions Institutions are the rules of the game in a society or, more formaly, are the humanly devised constraints that shape human interaction The major role of institutions in a society is to reduce uncertainty by establishing a stable (but not necessarily efficient) structure for human interaction D.C.North: Institutions, Institutional Change and Economic Performance. Cambridge (1990)

112 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 112 Agent-Mediated Institutions (fundamental elements) Role. Standardized patterns of behaviour required of all agents playing a part in a given functional relationship. Agent. The players of the institution. Each agent may take on several roles. Dialogic Framework. Ontologic elements and communication language (ACL) employed during an agent interaction. Scene. Agent meetings whose interaction is shaped by a well-defined protocol. Each scene models a particular activity. Performative Structure. Complex activities composed of multiple scenes specified as connections among scenes. Normative Rules. Determine both subsequent commitments and constraints on (dialogic) agent actions.

113 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 113 Performative structure (rationale) Complex activities can be specified by establishing relationships among scenes that: capture causal dependency among scenes; define synchronisation mechanisms involving scenes; establish paralellism mechanisms involving scenes; define choice points that allow roles leaving a scene to choose which activity to engage in next; and establish the role flow policy among scenes.

114 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 114 Specification tool

115 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 115 Future trends

116 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 116 Challenges Relate ideal and actual agent behaviour Relate roles of agents and their inherent qualities Generate ontologies, interaction standards and social conventions Generate new products, services and practices

117 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 117 Interaction Dramatic change of supply chains Decrease of customers prices Old markets will change New markets for commodities soon –Power, telephone, bandwidth, … Auctions intermediated with agents

118 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 118 Negotiation Soon: Applications where –Interactions are very fast –Interactions are repeated –Each trade is of relative small value –The process is repeated over long times Need for a significant value Preference elicitation complex. Need for learning –The product is relatively easy to specify

119 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 119 Challenges for negotiation Trust –Confidentiality, integrity, authentication and non-repudiation. –Safe payment and delivery –Supervised interaction first –Reputation (e.g. via Chambers of Commerce) Protocol standards

120 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 120 Challenges for negotiation Preference modelling –Dynamics of preferences –Different ontologies –Fuzziness –Learning Argumentation Protocols –From fixed to dynamic –Negotiation of protocols

121 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 121 Applications Travel agencies Retail industry:Sainsbury and Otto Versand Auction house for general trade Procurement applications (Bangemans challenge winners)

122 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 122 Engineering Adaptability Mobility Trust modelling Legal issues Open vs closed markets Electronic Institutions: Heterogeneity, trust and scalability, exception handling and societal change

123 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 123 Other challenges New products: market places and agent servers New standards: FIPA, W3C, P3P. FIPA and OMG signed agreement Security: preferences, public key servers and signature management by agent servers Privacy protection according to EU 95/46 directive

124 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 124 Vision Mobile devices Context perception Deregulation Disappearing computer That is: To put the customer at the heart of the business

125 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra

126 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 126 A tool: Islander

127 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 127 Basic intuitions There are situations where individuals interact in ways that involve Commitment, delegation, repetition, liability and risk. These situations involve participants that are: Autonomous, heterogeneous, independent, not-benevolent, not-reliable, liable. These situations are not uncommon: Markets, medical services, armies and many more. It is usual to resort to trusted third parties whose aim is to make those interactions effective by establishing and enforcing conventions that standardize interactions, allocate risks, establish safeguards and guarantee that certain intended actions actually take place and unwanted situations are prevented. These functions have been the basis for the development of many human institutions. They are even more necessary when interaction is among agents.

128 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 128 Concerns of an (electronic) institution Manage the identity of participants Define and validate requirements on participant capabilities Establish interaction conventions Facilitate effective interactions Enforce satisfaction of commitments

129 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 129 Three Fundamental Elements Dialogic Framework: Linguistic and ontological conventions to make efficient communication among agents. Performative Structure: Complex activities specified as connections among scenes (agents meetings whose interaction is shaped by a well-defined protocol). Norms: Consequences of agent actions within scenes.

130 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 130 Dialogic Framework We define a dialogical framework as a tuple DF = where: –O stands for the ontology; –I is the set of illocutionary particles; –L stands for a representation language; –R I is the set of internal roles; –R E is the set of external roles; and –R S is the list of relationships over roles;

131 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 131 Communication Language CL expresions are formulae of the form (i ( i r i ) ) where: –i is an illocutionary particle in I; – i can be either an agent variable or an agent identifier; – r i can be either a role variable or a role identifier in R I R E ; – represents the adressee(s) of the message and can be: ( k r k ) the message is addressed to a single agent. r k the message is addressed to all the agents playing the role r k. all the message is addressed to all the agents of the scene. – is an expression in the content language. – can be either a time variable or a value time stamping the illocution

132 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 132 FM dialogic framework

133 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 133 FM dialogical framework

134 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 134 Scene Scene is a pattern of multi-agent conversation. Scene is specified by a finite state oriented graph where the nodes represent the different states and oriented arcs are labelled with illocution schemes or timeouts. During the execution new agents can join the scene or some of the participants can leave the scene at definite states depending on their role. An scene can be multiple-instantiated and played by different groups of agents.

135 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 135 Scene The graph has a single initial state non reachable once left. Set of final states not connected to other states. For each role is defined a set of access and exit states. Minimum and maximum number of agents per role Final states must be exit states for each role. Initial state must be an access state for each role which minimum is greater than zero.

136 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 136 Auction room scene

137 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra (inform (?x auctioneer) (buyer) open_auction(?n)) 2(inform (!x auctioneer) (buyer) (open-round(?r)) 3(inform (!x auctioneer) (buyer) to-sell(?good-id)) 4(inform (!x auctioneer) (buyer) buyers(?buyers_list)) 5(inform (!x auctioneer) (buyer) offer(!good-id,?price)) 6(inform (!x auctioneer) (buyer) offer(!good-id,?price)) 7,8(commit (?y buyer) (!x auctioneer) bid(!good_id, ?price)) 9(inform (!x auctioneer) (buyer) withdrawn(!good-id)) 10(inform (!x auctioneer) (buyer) collision(?price)) 11(inform (!x auctioneer) (buyer) sanction(?buyer-id)) 12(inform (!x auctioneer) (buyer) expulsion(?buyer-id)) 13(inform (!x auctioneer) (buyer) sold(!good-id,?price, ?buyer-id)) 14,15(inform (!x auctioneer) (buyer) end-round(!r)) 16(inform (!x auctioneer) (buyer) end-auction(!n) ) Auction room scene

138 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 138 Performative Structure Complex activities can be specified by establishing relationships among scenes that: –capture causal dependency. –define synchronisation mechanisms. –establish parallelism mechanisms. –define choice points that allow roles leaving a scene to choose which activity to engage in next. –establish the role flow policy.

139 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 139 Performative Structure A performative structure can be seen as a network of scenes. We introduce transitions to mediate between connections of scenes. Arcs connecting scenes and transitions labelled with constraints. The specification allows to express that an scene can be running simultaneously multiple-times at execution time. Determines wether agents moving between scenes join current executions of the target scene(s) or whether new executions are started.

140 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 140 FM performative structure

141 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 141 Norms The normative rules defines the consequences of agents actions within the institution. Such consequences are: –some actions can impose obligations to agents. –can vary the paths that agents can follow. We set out the predicate Obl for the notion of obligation. –Obl(x,,s): meaning that agent x is obliged to do in scene s. Norms are specified by three elements: –Antecedent: the actions that provoke the activation of the norm and restrictions over illocution scheme variables. –Defeasible antecedent: the actions that agents must carry out in order to fulfill the obligations. –Consequent: the set of obligations For instance, a buyer winning a bidding round is obliged to go later on to the buyers settlement scene to pay for the good.

142 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 142 Norms

143 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 143 Electronic Institution An electronic institution is defined as a tuple EI = where: –DF stands for a dialogic framework. –PS stands for a performative structure. –N stands for a set of norms.

144 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 144 ISLANDER Textual specification language for electronic institutions. ISLANDER editor: specification and verification tool for electronic institutions. Combines textual and graphical specifications. Verification –Integrity –Liveness –Protocol Correctness –Norm Correctness

145 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 145 Islander Load Import Files Edition Verification GraphEditor Edition Panels Save Export Files Islander Textual Spec. Textual Spec. XML Spec. errors correct? User ISLANDER modules

146 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 146 Infrastructure The infrastructure is in charge of: –allowing agents interaction –giving agents the information they need for participating in the institution –checking that agents do not violate the institution rules. Social infrastructure composed of: –Institution Manager –Governor Creator –Scene Managers –Governors

147 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 147 Governor Agent developed to mediate between the institution and participating agents. Each participating agent is connected to a governor that controls that it behaves according to the institution specification. Facilitates to the agent some information about the state of the institution. Coordinates with other agents of the infrastructure for the correct execution of the institution.

148 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 148 Institution Execution

149 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 149 Institution Execution

150 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 150 Institution Execution

151 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 151 Institution Execution

152 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 152 Institution Execution

153 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 153 Institution Execution

154 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 154 Scene Execution

155 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 155 Institutions for AMEC: FM+

156 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 156 FM+ MAS User 1 Buyer agents creation Virtual auctions federation Virtual auction 1 Buyers coordination User 2 User k Real auction 1 Real auctions Users... Global DB external data B B B B B Virtual auction 2 Virtual Auction N Real auction 2 Real auction N B B B

157 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 157 Institutional Agents Good Register (GR) Provides the virtual environment with all the events related to the product entry from the real auction place. One agent for each auction place. Remote Control (SRC) Performs all the tasks to participate in an specific auction under the management of an agent with buyer role. It is an institution object, one for each buyer. Auction Broker (AB) Transmits the events related to the auction of the real world to the virtual agents who take part in the auction and to the DB. One agent for each auction place. DB Manager (DBM) Manages the database that contains information about all the auction places. Auction Admitter (aad) Controls the access of the buyers to the auction place. One agent for each auction place. Buyer Admitter (bad) Controls the access of the buyers to the auction place federation. One agent for the whole federation.

158 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 158 MASFIT Institution (Islander)

159 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 159 Scene example 1.Auction-Broker: Inform(StartAuction) 2.Auction-Broker -> RemoteControl: Inform(StartRound) 3.Auction-Broker ->RemoteControl: Inform(Good to Sell,Price,Boxes) 4.RemoteControl -> Auction-Broker: Propose(Bid) 5.RemoteControl -> Auction-Broker: Propose(Bid) 6.Auction-Broker -> RemoteControl: Inform(Result) 7.Auction-Broker->RemoteControl Failure(Collision/Sanction/Expulsion) 8.Auction-Broker -> RemoteControl: Inform(Result) 9.Auction-Broker -> RemoteControl: Failure(Withdrawn) 10.Auction-Broker -> RemoteControl: Inform(EndAuction)

160 IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 160 Real world FM+ real world interaction Inst. Agent Auction Broker Inst. Agent Good Register Specific Protocol Auction Soft (AUTEC) Human buyers Virtual world Agent buyers Electronic Institution Governor


Download ppt "IIIA-CSIC SBIA02-AMEC Tutorial.. Carles Sierra. 1 Agent-mediated electronic commerce Carles Sierra IIIA-CSIC Barcelona."

Similar presentations


Ads by Google