Transcript Slide 1

AN AGENT-BASED ARCHITECTURE FOR ONLINE DISPUTE RESOLUTION SERVICES
DAVIDE RUA CARNEIRO *
Supervisors: Paulo Novais, José Neves
University of Minho
School of Engineering
Computer Science and Technology Center
* [email protected]
Alternative Dispute Resolution
In the past, contracts were
signed in the physical
presence of all the parties.
Now, contracts are signed electronically and
sometimes even automatically, with the
parties possibly anywhere in the world. This
poses new challenges to the legal systems.
islab.di.uminho.pt
Retrieval of Information
Generation of Solutions
In order for parties to make good decisions they should have access
to a set of important information. Using a Case-based Reasoning
(CBR) approach, the system is able to generate information about:
- BATNA – Best Alternative to a Negotiated Agreement;
- WATNA – Worst Alternative to a Negotiated Agreement;
- MLATNA – Most Likely Alternative to a Negotiated Agreement;
- Similar cases – Past cases that are similar to the current one and
may be of interest;
- ZOPA – Zone of Potential Agreement;
Sometimes, case-based approaches can have drawbacks, mainly
when there is a poor database. In that sense, we developed a module
based on a Genetic Algorithm (GA) to generate solutions for the
resolution of the conflict.
Several approaches exist that aim to
solve conflicts out of courts, namely
Arbitration, Negotiation, Mediation or
Facilitation. All of them aim to be a
cheaper and faster alternative to
litigation, which is nowadays slow and
inefficient.
However, these already traditional
approaches can be improved with the
support of technology and, specifically
through
Artificial
Intelligence
techniques.
Environments for Conflict Resolution
Solving the Conflict
In the search for more efficient methods we are developing the
concept of Environments for Conflict Resolution. In this approach,
Ambient Intelligence is used to collect context information from the
context of interaction that can support the lifecycle of the conflict
resolution platform, allowing the software agents to make better and
more informed decisions.
When parties have enough information and the system has solutions
to suggest (obtained either through CBR or GA) the conflict resolution
can start. It consists of a negotiation in which the system or the
parties successively propose solutions using mobile or web interfaces.
Under this approach, each
solution is seen as an
individual
of
the
population,
which
is
evolved
through
the
successive application of
three genetic operators:
Crossover, Heredity, and
Mutation.
These operators enable
the natural emergence of
species: lines of evolution
that generate solutions
that are better for a given
party. This is denoted by
the
colors
of
the
individuals.
During the negotiation the solutions generated can then be
proposed to the parties according to the changes in the context of
interaction (e.g. levels of stress, personal conflict style).
This work is included in TIARAC - Telematics and Artificial Intelligence in Alternative
Conflict Resolution PTDC/JUR/71354/2006, a research project supported by FCT. The
work of Davide Carneiro is also supported by a doctoral grant (SFRH/BD/64890/2009).
Uma Escola a Reinventar o Futuro – Semana da Escola de Engenharia - 24 a 27 de Outubro de 2011