Title: MS Ecommerce course 20-853 Electronic Negotiation Summer 2004
1MS Ecommerce course 20-853 Electronic
NegotiationSummer 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/gilpin/ec
20-853/ec20-853.htm
2Course 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
3Motivation
4Automated 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
5Automated 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
6Systems 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
7A bold vision How automated negotiation
techniques could play a role in different stages
of an ecommerce transaction
8Automated 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
9Automated 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
10Automated 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
11Example 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
12Basics
- Agenthood,
- utility function,
- evaluation criteria of multiagent systems
13Agenthood
- 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
14Utility 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.
15Criteria 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