. The theory of evolutionary computation has been enhanced rapidly during the last decade. This survey is the attempt to summarize the results regarding the limit and finite time behavior of evolutionary algorithms with finite search spaces and discrete time scale. Results on evolutionary algorithms beyond finite space and discrete time are also presented but with reduced elaboration. Keywords: evolutionary algorithms, limit behavior, finite time behavior 1. Introduction The field of evolutionary computation is mainly engaged in the development of optimization algorithms which design is inspired by principles of natural evolution. In most cases, the optimization task is of the following type: Find an element x 2 X such that f(x ) f(x) for all x 2 X , where f : X ! IR is the objective function to be maximized and X the search set. In the terminology of evolutionary computation, an individual is represented by an element of the Cartesian product X \Theta A, where A is a possibly...

The fundamental distinction between ordinary evolutionary algorithms (EA) and co-evolutionary algorithms lies in the interaction between coevolving entities. We believe that this property is essentially game-theoretic in nature. Using game theory, we describe extensions that allow familiar mixing-matrix and Markov-chain models of EAs to address coevolutionary algorithm dynamics. We then employ concepts from evolutionary game theory to examine design aspects of conventional coevolutionary algorithms that are poorly understood.

Abstract not found

Recent literature shows that learning in oligopoly games might in the long run result in the Cournot or in the Walras equilibrium. Which outcome is achieved seems to depend on the underlying learning dynamics. This paper analyzes the forces behind the learning mechanisms determining the long run outcome. Apart from the fact that there is a difference between social and individual learning, the key parameter is shown to be the degree of rationality of the learning agents: Learning the Cournot strategy requires the agents to acquire a large amount of information and to use sophisticated computational techniques, while the Walras strategy can be shown to be a particular ‘low rationality result‘.

sponding groups of loci are independent each other in the k--th generation (that is, p (x 1 ; : : : ; xm ) = j=1 p j (x i j1 ; : : : ; x i jd j ) and p j is the pdf of i j1 , : : :, i jd j --th loci in the k--th generation). Then, it is easily shown that the same groups of loci in the (k + 1)--th generation are independent each other if recombination maintains this independence. Second, we consider a special case of the above assumptions on the fitness function and loci. If f is completely separable and all the loci in the k--th generation are independent each other (that is, L = m), it can be shown that all the loci in the (k + 1)--th generation are independent each other if recombination is any of one--point, multi--point, and uniform crossovers. This is shown by the fact that the pdf after these recombinations is described as the sum of the products of the pdf of some loci and that of the other loci before the recombinations (Nomura and Shimohara, 2001). Finally, as a mo

In this paper, as one approach for mathematical analysis of evolutionary algorithms with real number chromosomes, we focus our attention on crossovers, give a general framework of the description for the change of the distribution of the population through them, and verify the properties of crossovers based on the framework. This framework includes various crossover which have been proposed and we apply our result to these crossover methods. 1

We study a simple overlapping generations economy as an adaptive learning system. The learning is via a so-called genetic algorithm process. We first investigate performances of Holland’s standard GA (SGA), Arifovic’s augmented GA (AGA), and Birchenhall’s selective transfer GA (STGA), Bullard and Duffy (BDGA) as a model of population learning. In addition, we also investigate these learning algorithms variant. Second, compared to population learning, we also implement the GAs as a model of individual learning. An “ecological ” approach showing “inter-generation ” aspect of the GA to learning problems is therefore modelled. Finally, We visit a further approach called “open learning” model, about endogenising learning in which agents learn how to learn. The results we obtain confirm previous statement that the stability of the Pareto superior equilibrium of the model, i.e. the low inflation equilibrium, is robust independent of precise learning variant. Furthermore, we show that individual agents with heterogeneous learning schemes eventually coordinate on the equilibrium. We offer the interpretation of convergence to the equilibrium. 1 I.

the Social Contrivance of Money ” as an exemplary, Neo-classically misconstructed model of nowadays society and its economic activities. The failure to notice the existence of unpaid labor in the framework of the model, and the implicit dependence of the outcomes on unpaid (omitted) reproduction work has apparently gone unchallenged so far. The logical inconsistencies of the model and their androcentric backgrounds are discussed in this paper. Furthermore, we aim to debate (but not explicitly formulate) an alternative modeling approach based on the use of Genetic Algorithms to include – at least some – crucial features of modern society in its whole heterogeneity.