Evolutionary Algorithms - Self-Adaptive Heuristics for Evolutionary Computation - page 25

upper and lower bounds for the expected runtime of an EA and the success prob-
ability. Various proof techniques have been developed to bound the runtime of
EAs, e.g. for the runtime of the ( μ + 1)-EA on simple pseudo-boolean functions
[160]. Other papers in this field concentrate on parameter control techniques,
e.g. Jagerskupper's analysis [63], [64] of an EA with Rechenberg's 1/5th rule
[114].
2.5 Summary
1. EAs are the attempt to adapt the powerful Darwinian evolution to computer
science for search and optimization problems. They simulate the principles of
inheritance, variation and natural selection with genetic operators. But this
bloomy description means no more than EAs perform purely randomized
search in the domain space.
2. Various operators and algorithmic schemes have been proposed, a few of
them more, many less meaningful. The famous EA variants GAs, ES and
EP have more and more grown together and cannot be distinguished any-
more. This topic concentrates on ES, which are appropriate for numerical
optimization problems.
3. PSOAs exhibit similar elements like EAs, in particular ES. Operators like
mutation, crossover and selection can be found in the PSOA equations.
4. Theoretical research comprises the fields no free lunch, Holland's schema the-
orem, the building block hypothesis, various dynamical systems approaches
and statistical mechanics analysis. Very successful is the analysis of EAs
from the point of view of Markov chains. In the last years, a classical run-
time analysis of EAs got more and more attention.
5. The contribution of research in EA can be seen as old-fashioned AI research:
reducing exponentially large search spaces by appropriate representations,
operators and heuristics. The success of the search is often a result of the
engineer's knowledge, also an old-fashioned AI wisdom and a symptom of
the no free lunch theorem.
6. EAs have successfully been applied to domains like control, pattern recogni-
tion or automatic design. Sometimes the practitioner should be aware that
classic search techniques may be faster or offer other advantages. But nev-
ertheless, randomized search has grown to a strong, practicable and well-
understood search technique, which is used all over the world in research
and engineering.