Evolution%20Strategies

1
Evolution Strategies

• Originally developed in Germany in the early 60s
  with Rechenberg and Schwefel being the main
  contributors. Main ideas include
• floating point representation with standard
  derivation (x,?) where x is a vector in
  n-dimensional space and ? is the standard
  derivation that influeces how solutions are
  changed x x N(0, ?).
• mutation is the sole operator.
• in most approaches ? is changed dynamically (e.g.
  using Rechenbergs 1/5 Rule).
• employs a two-member evolution strategy.
• Was later generalized to support multi-membered
  evolution strategies
• employs uniform crossover and averaging
  crossover.
• each member of the population has the same chance
  to reproduce (selection doesnt consider
  fitness).
• weakest individual is eliminated to keep a
  constant population size.

2
?????-ES and ?????-ESs.

• developed by H.P. Schwefel 2-member population
  ((11)-ES) is generalized to multi-membered
  populations. 2 approaches are supported
• (???)-ES
• ? individuals produce???offsprings
• the population consisting of?????individuals (the
  old generation and the parents) is reduced to ?
  using selection.
• relies on replacement
• (???)-ESs
• lifetime of individuals is limited to one
  generation.
• ? individual produce ? offsprings (? gt ?) with
  the best ? surviving.
• generates the new generation from the scratch
• Moreover, the standard deviation undergoes
  evolution.

3
Evolutionsstrategie and GAs

• Differences ES and trad. GAs
• real-code (ES) vs. binary string representation
  (GA)
• selection is performed implicitly by removing
  unfit individual deterministically (ES) GAs
  employ a stoachastic selection process, and does
  not rely on removal..
• selection after recombination (ES) selection
  before recombination (GA).
• different handling of constraints ES supports
  un-equalities as a part of the problem
  specification, and disqualifies illegal
  offspring moreover, ES adjusts control
  parameters if illegal offspring occur too
  frequently. GAs, on the other hand, employ
  penalty functions.
• mutation is less important for traditional GAs
  crossover is less important for ESs.
• Some scientists, e.g. Fogel, claim that ES and
  GAs are not fundamentally different (see also
  Hoffmeisters paper 141).