This paper analyzes the Self-Adaptation (SA) algorithm widely used to adapt strategy parameters of the Evolution Strategy (ES) in order to obtain maximal ES-performance. The investigations are concentrated on the adaptation of one general mutation strength (called SA) in (1;) ESs. The hypersphere serves as the tness model. Starting from an introduction into the basic concept of self-adaptation,...

A method is proposed to invert displacement and tilt measurements from a single monitoring location to constrain the nature of a volcanic pressure source. Evolutionary strategy is applied to a sequence of Mogi solutions and the program is shown capable of distinguishing between depth and volume increment of the magma source, in addition to its 3-dimensional spatial coordinates. The mechanism is...

Almost all approaches to multiobjective optimization are based on Genetic Algorithms, and implementations based on Evolution Strategies (ESs) are very rare. In this paper, a new approach to multiobjective optimization, based on ESs, is presented. The comparisons with other algorithms indicate a good performance of the Multiobjective Elitist

Learning in feedforward neural networks may fail due to several reasons. We give a model for the failure caused by premature saturation of hidden neurons. In order to avoid this type of failure, we suggest evolution strategies for the training of the networks. The properties of the chosen algorithm are demonstrated by some applications.

In this paper we present an Evolution Strategy (ES) approach towards the estimation of the location and strength of surface emissions of trace gases based on atmospheric concentration measurements and back-trajectory analyses. The details of the ES developed are outlined. The ES is tested using artificial emission maps at different grid 5 resolutions and the results compared to those obtained o...

The recently introduced family of natural evolution strategies (NES), a novel stochastic descent method employing the natural gradient, is providing a more principled alternative to the well-known covariance matrix adaptation evolution strategy (CMA-ES). Until now, NES could only be used for single-objective optimization. This paper extends the approach to the multi-objective case, by first der...

The family of natural evolution strategies (NES) offers a principled approach to real-valued evolutionary optimization by following the natural gradient of the expected fitness on the parameters of its search distribution. While general in its formulation, existing research has focused only on multivariate Gaussian search distributions. We address this shortcoming by exhibiting problem classes ...

Information geometric optimization (IGO) is a general framework for stochastic optimization problems aiming at limiting the influence of arbitrary parametrization choices: the initial problem is transformed into the optimization of a smooth function on a Riemannian manifold, defining a parametrization-invariant first order differential equation and, thus, yielding an approximately parametrizati...

Evolution strategies are a class of general optimisation algorithms which are applicable to functions that are multimodal, non-diierentiable, or even discontinuous. Although recombination operators have been introduced into evolution strategies, their primary search operator is still mutation. Classical evolution strategies rely on Gaussian mutations. A new mutation operator based on the Cauchy...

Differential evolution, termed DE, is a novel and rapidly developed evolution computation in recent years. There are some advantages of DE, including simple structure, easy use and rapid convergence speed. Besides, DE can be also applied on the complex optimization problem. However, there are some issues, such as premature convergence and stagnation, remaining in DE algorithm. To overcome those...