MS Ecommerce course 20-853 Electronic Negotiation Spring 2004

1
MS Ecommerce course 20-853 Electronic
NegotiationSpring 2004

• Professor Tuomas Sandholm
• School of Computer Science
• Carnegie Mellon University

Instructors web page www.cs.cmu.edu/sandholm
Course web page http//www.cs.cmu.edu/sandholm/
ec20-853/ec20-853.htm
2
Course content at a high level

• Covers the state-of-the-art
• Covers
• game-theoretic aspects
• computational aspects
• Additional readings (and proofs of claims) are
  available on the web site of my PhD-level course
  Foundations of Electronic Marketplaces

3
Motivation
4
Automated negotiation systems

• Agents search make contracts
• Through peer-to-peer negotiation or a mediated
  marketplace
• Agents can be real-world parties or software
  agents that work on behalf of real-world parties
• Increasingly important from a practical
  perspective
• Developing communication infrastructure
  (Internet, WWW, NII, EDI, KQML, FIPA, Concordia,
  Voyager, Odyssey, Aglets, AgentTCL, Java Applets,
  ...)
• Electronic commerce on the Internet Goods,
  services, information, bandwidth, computation,
  storage...
• Industrial trend toward virtual enterprises
  outsourcing
• Automated negotiation allows dynamically formed
  alliances on a per order basis in order to
  capitalize on economies of scale, and allow the
  parties to stay separate when there are
  diseconomies of scale

5
Automated negotiation systems

• Fertile, timely area
• Deep theories from game-theory computer science
  merge
• Started together in the 1940s Morgenstern von
  Neumann
• There were a few decades of little interplay
• Upswing of interplay in the last few years
• It is in this setting that the prescriptive
  (normative) power of game theory really comes
  into play
• Market rules need to be explicitly specified
• Software agents designed so as to act optimally
• unlike humans ("As far as the laws of mathematics
  refer to reality, they are not certain and as
  far as they are certain, they do not refer to
  reality. - Albert Einstein)
• Computational capabilities can be quantitatively
  characterized, and prescriptions can be made
  about how the agents should use their computation
  optimally

6
Systems with self-interested agents
(computational or human)

• Mechanism (e.g., rules of an auction) specifies
  legal actions for each agent how the outcome is
  determined as a function of the agents
  strategies
• Strategy (e.g., bidding strategy) Agents
  mapping from known history to action
• Rational self-interested agent chooses its
  strategy to maximize its own expected utility
  given the mechanism
  gt strategic analysis required for
  robustness
  
  
  gt noncooperative game theory
• But computational complexity
• In executing the mechanism
• In determining the optimal strategy
• In executing the optimal strategy
• Has significant impact on prescriptions
• Has received little attention in game theory

7
A bold vision How automated negotiation
techniques could play a role in different stages
of an ecommerce transaction
8
Automated negotiation techniques in different
ecommerce stages

• 1. Interest generation
• Funded adlets that coordinate
• Avatars for choosing which ads to read
• Customer models for choosing who to send ads and
  how much to offer
• 2. Finding
• Simple early systems BargainFinder, Jango
• Meta-data, XML
• Standardized feature lists on goods to allow
  comparison
• How do these get (re)negotiated
• Different vendors prefer different feature lists
• Shopper agents need to understand the new lists
• How do machine learning algorithms cope with new
  features?
• Want to get a bundle gt need to find many vendors

9
Automated negotiation techniques in different
ecommerce stages...

• 3. Negotiating
• Advantages of dynamic pricing
• Right things sold to (and bought from) right
  parties at right time
• World becomes a better place (social welfare
  increases)
• Further advantages from discriminatory pricing
• Can increase social welfare
• Fixed-menu take-it-or-leave-it offers -gt
  negotiation
• Cost of generating disseminating catalogs?
• Other customers see the price?
• Negotiation overhead?
• Personalized menus (check customers web page,
  links to from it, what other similar customers
  did, customer profiles)
• Generating/printing the menu may be intractable,
  e.g. mortgages 530
• Negotiation will focus the generation, but vendor
  may bias prices offerings based on path
• Preferences over bundles
• Coalition formation

10
Automated negotiation techniques in different
ecommerce stages...

• 4. Contract execution
• Digital payment schemes
• Safe exchange
• Third party escrow companies
• Tradesafe Inc.
• Tradenable Inc.
• i-Escrow Inc.
• Sometimes an exchange can be carried out without
  enforcement by dividing it into chunks
  SandholmLesser IJCAI-95, Sandholm96,97,
  SandholmFerrandon ICMAS-00, SandholmWang
  AAAI-02
• 5. After sales

11
Example applications

• Application classes
• B2B (business-to-business), e.g. procurement
  RFPs/RFQs, buying consortia (e.g. Covisint),
• B2C (business-to-consumer), e.g. goods, debt
• C2C (consumer-to-consumer), e.g. eBay
• Task and resource allocation in computer systems
  (networks, computational grids, storage systems)
• Just a few example application areas
• Electricity markets
• Manufacturing subcontracting
• Transportation exchanges
• Stock markets
• Collaborative filtering

12
Basics

• Agenthood,
• utility function,
• evaluation criteria of multiagent systems

13
Agenthood

• We use economic definition of agent as locus of
  self-interest
• Could be implemented e.g. as several mobile
  agents
• Agent attempts to maximize its expected utility
• Utility function ui of agent i is a mapping from
  outcomes to reals
• Can be over a multi-dimensional outcome space
• Incorporates agents risk attitude (allows
  quantitative tradeoffs)
• E.g. outcomes over money

Lottery 1 0.5M w.p. 1 Lottery 2 1M w.p.
0.5 0 w.p. 0.5 Agents strategy is
the choice of lottery
Risk aversion gt insurance companies
14
Utility functions are scale-invariant

• Agent i chooses a strategy that maximizes
  expected utility
• maxstrategy Soutcome p(outcome strategy)
  ui(outcome)
• If ui() a ui() b for a gt 0 then the agent
  will choose the same strategy under utility
  function ui as it would under ui
• Note that ui has to be finite for each possible
  outcome
• Otherwise expected utility could be infinite for
  several strategies, so the strategies could not
  be compared.

15
Criteria for evaluating multiagent systems

• Computational efficiency
• Distribution of computation
• Communication efficiency
• Social welfare maxoutcome ?i ui(outcome)
• Requires cardinal utility comparison
• but we just said that utility functions are
  arbitrary in terms of scale!
• Surplus social welfare of outcome social
  welfare of status quo
• Constant sum games have 0 surplus. Markets are
  not constant sum
• Pareto efficiency An outcome o is Pareto
  efficient if there exists no other outcome o
  s.t. some agent has higher utility in o than in
  o and no agent has lower
• Implied by social welfare maximization
• Individual rationality Participating in the
  negotiation (or individual deal) is no worse than
  not participating
• Stability No agents can increase their utility
  by changing their strategies
• Symmetry No agent should be inherently
  preferred, e.g. dictator