Description
Set-based
Evolution (SBE) concerns the evolution of sets. The relation between
solutions and individuals in nature might be interpreted in the case of
SBE to a relation between sets and species. Sets might be related to
conceptual solutions in design, portfolios in market's analysis,
strategies in games etc.
The
following distinguish SBE from common EC:
-
A set is
commonly associated with a cluster of performances in the objective
space. Each point in the objective space is associated with a different
solution, which belongs to the set.
-
Therefore,
comparing between sets for both the enhancement of the evolution and for
the selection stage is based on comparing between clusters of
representatives and not between single
representatives.
-
In traditional
evolution, individual solutions are compared and optimized. In contrast,
in SBE, the purpose is to search and select a preferred set a-priory to
selecting the final solution out of that winning set. It is important to
note that although just one solution is selected out of the set, it is
possible, that later it could be substituted by another member of the
set. This might be as a result of changes in market demands/situation or
as a contra to opponent actions.
-
A set, such as
dealt with here, is commonly associated with an abstract description,
which distinguishes it from the other sets. Such an abstract description
opens the way for incorporating human preferences within the
evolutionary search.
Workshop
Goals
The goals of the
workshop are to: 1. Present the state-of-the-art on SBE. 2. Highlight the
importance of SBE, 3. Search for new implications and implementations of
SBE, 4. Suggest new research directions that will improve and extend its
use.
Papers
Prospective
participants are invited to submit abstracts/short papers (not more than 4
pages) on SBE. Please send your
contributions to the workshop organizer by 25.6.08.
More
details to appear in this website.