An analysis of the dynamic behavior of Evolution Strategies applied to Trav-eling Salesman Problems is presented. For a special class of Traveling Salesman Problems a stochastic model of the optimization process is introduced. Based on this model diierent features determining the optimization process of Evolution Strategies are analyzed. In addition the stochastic model is extended to explain s...

We present a formal framework for the implementation of a concep tual data model in terms of an internal data model within a two level architecture taking into account schema consistency and instance com patibility Within that framework the consequences of schema changes are investigated leading to a homogeneous formalization of existing ap proaches to schema evolution On that basis we introduc...

Traditional convolutional neural networks (CNN) are stationary and feedforward. They neither change their parameters during evaluation nor use feedback from higher to lower layers. Real brains, however, do. So does our Deep Attention Selective Network (dasNet) architecture. DasNet’s feedback structure can dynamically alter its convolutional filter sensitivities during classification. It harness...

Let P = fP0; :::; Png be a partition of a nite abelian group A and L(P ) the complex vector space generated by the indicator functions of the sets P0; :::; Pn. In [1] it was proved that P de nes an abelian association scheme on A if L(P ) equals the space L(P ) of all Fourier transforms of functions in L(P ). In this paper it is shown that P de nes an abelian association scheme on A if and only...

ABSTRACT Recent advances in the theory of evolutionary algorithms have indicated that a hybrid method known as the evolutionary-gradient-search procedure yields superior performance in comparison to contemporary evolution strategies. But the theoretical analysis also indicates a noticeable performance loss in the presence of noise (i.e., noisy fitness evaluations). This paper aims at understand...

Behavior planning of a vehicle in real-world traffic is a difficult problem. Complex systems have to be build to accomplish the projection of tasks, environmental constraints, and purposes of the driver to the dynamics of two controlled variables: steering angle and velocity. This paper comprises two parts. First, the behavior planning for the task of intelligent cruise control is proposed. The...

|A new formulation for coordinate system independent adaptation of arbitrary normal mutation distributions with zero mean is presented. This enables the evolution strategy (ES) to adapt the correct scaling of a given problem and also ensures invariance with respect to any rotation of the tness function (or the coordinate system). Especially rotation invariance, here resulting directly from the ...

A new adaptation scheme for adapting arbitrary normal mutation distributions in evolution strategies is introduced. It can adapt correct scaling and correlations between object parameters. Furthermore, it is independent of any rotation of the objective function and reliably adapts mutation distributions corresponding to hyperellipsoids with high axis ratio. In simulations, the generating set ad...

The usual approach to p laus ib le reasoning is to associate a v a l i d i t y measure wi th each fac t or r u l e , and to compute from these a v a l i d i t y measure for any deduction tha t is made. This approach is shown to be inappropr iate for some classes of problems, p a r t i c u l a r l y those in which the evidence is not i n t e r n a l l y cons is tent . Two current p laus ib le re...

Motivated by parallel optimization, we experiment EDA-like adaptation-rules in the case of λ large. The rule we use, essentially based on estimation of multivariate normal algorithm, is (i) compliant with all families of distributions for which a density estimation algorithm exists (ii) simple (iii) parameter-free (iv) better than current rules in this framework of λ large. The speed-up as a fu...