In this paper, we address some issues related to evaluating and testing evolutionary algorithms. A landscape generator based on Gaussian functions is proposed for generating a variety of continuous landscapes as fitness functions. Through some initial experiments, we illustrate the usefulness of this landscape generator in testing evolutionary algorithms.

In XCS classifier fitness is based on the relative accuracy of the classifier prediction [3]. A classifier is more fit if its prediction of the expected payoff is more accurate than the prediction given by the other classifiers that are applied in the same situations. The use of relative accuracy has two major implications. First, because the evaluation of fitness is based on the relevance that...

We present a new mechanism by which prior employment can influence transitions to other firms. We propose that some employees divert effort toward unproductive activities to learn about their own fitness for alternative employment. Based on the results of this costly learning experience, or “experiment,” some employees will transition into other firms or launch their own ventures, while others ...

Computational evolution provides a mechanism for searching a space of potential solutions according to a specified fitness function. In design, the search for potential solutions is often interleaved with changes in the requirements for the solution. We propose a model of design that uses a co-evolutionary computational process. Specifically, we present design as a parallel search for both desi...

The phenotypic view of selection assumes that genetic responses can be predicted from selective forces and heritability or in the classical quantitative genetic equation: R = h2S. However, data on selection in bird populations show that often no selection response is found, despite consistent selective forces on phenotypes and significant heritable variation. Such discrepancies may arise due to...

In the paper it is demonstrated how evolutionary techniques can be used for the process of generating benchmark circuits covering a wide scale of testability properties. To calculate the value of fitness function the approach based on analytical evaluation of testability parameters is used. The solutions which cannot be synthesized by a design system are avoided from the process of developing a...

It is widely acknowledged that the system functionality captured in a system model has to match organisational requirements available in the business model. However, fitness measures are rarely integrated in design methodologies. The paper proposes a framework to ease the generation of fitness measures adapted to a given methodology in order to quantify to which extent there is fit between the ...

We sketch a novel and improved version of Boorse’s biostatistical theory of functions. Roughly, our theory maintains that (i) functions are non-negligible contributions to survival or inclusive fitness (when a trait contributes to survival or inclusive fitness); (ii) situations appropriate for the performance of a function are typical situations in which a trait contributes to survival or inclu...

We present a model that demonstrates the evolution of groups composed of cooperative individuals performing specialised functions. Specialisation and cooperation results from an evolutionary process in which selection and reproduction is based on individual fitness. Specialists come to help (through the donation of resources) their non-kin group members, optimising their behaviour as a team and...

This work introduces a general mathematical framework for non-stationary fitness functions which enables the exact definition of certain problem properties. The properties’ influence on the severity of the dynamics is analyzed and discussed. Various different classes of dynamic problems are identified based on the properties. Eventually, for an exemplary model search space and a (1, λ)-strategy...