Although there are many versions of evolutionary algorithms that are tailored to multi–criteria optimization, theoret ical results are apparently not yet available. Here, it is shown that results known from the theory of evolutionary algorithms in case of single criteri on optimization do not carry over to the multi–criterion case. At first, three different step size rules are investigated numerically for a selected prob lem with two conflicting objectives. The empirical results obtained by thes e experiments lead to the observation that only one of these step size rules may h ave the property to ensure convergence to the Pareto set. A theoretical a n lysis finally shows that a special version of an evolutionary algorithm wi th this step size rule converges with probability one to the Pareto set for thetest problem under consideration. Keywords—multi–criteriaoptimization, evolutionary algorithms, stochastic convergence to Pareto set
[1]
G. M.,et al.
Elliptic Integrals
,
1917,
Nature.
[2]
R. Bucy,et al.
Filtering for stochastic processes with applications to guidance
,
1968
.
[3]
Frank Kursawe,et al.
A Variant of Evolution Strategies for Vector Optimization
,
1990,
PPSN.
[4]
Peter J. Fleming,et al.
An Overview of Evolutionary Algorithms in Multiobjective Optimization
,
1995,
Evolutionary Computation.
[5]
Hajime Kita,et al.
Multi-objective optimization by genetic algorithms: a review
,
1996,
Proceedings of IEEE International Conference on Evolutionary Computation.
[6]
Jeffrey Horn,et al.
Multicriterion decision making
,
1997
.
[7]
Günter Rudolph,et al.
Evolutionary Search for Minimal Elements in Partially Ordered Finite Sets
,
1998,
Evolutionary Programming.