Computational evolutionary strategies

Methods based on fuzzy theory, neural nets, and evolutionary strategies are denoted as soft computing... 

An alternative approach for computational evolution, termed the Evolutionary Strategy (ES) was developed independently by Rechenberg and co-workers. The ES employs many of the same ideas as the GA, including mutation and selection. However, the ES always uses real-valued encoding. A particular evolution strategy is usually denoted as a i -I- k -ES. The parameter p refers to the constant population size, whereas the parameter X refers to the size of the pool out of which a new population is selected. Obviously, X must be at least as large as p. 

T. Back, F. Hoffmeister, and H.-P. Schwefel, in Proceedings of the Fourth International Conference on Genetic Algorithms, R. K. Belew and L. B. Booker, Eds., Morgan Kaufmann, San Mateo, CA, 1991, pp. 2-9. A Survey of Evolutionary Strategies. T. Back and H.-P. Schwefel, Evolutionary Computation, 1, 1 (1993). An Overview of Evolutionary Algorithms for Parameter Optimization. 

In class 1 type problems, the description of environment and specification is complete and the problem is completely described. However, in most cases, there are too many candidates of feasible solutions due to combinatorial explosion. Eor this type of problems, evolutionary computation methods, such as genetic algorithms, genetic programming, evolutionary strategies, and evolutionary programming, have been successfully applied. 

CAMD = computer-aided molecular design ES = evolutionary strategies GA = genetic algorithm GFA = genetic function approximation LOF = lack of fit LSE = least squares error MARS = multivariate adaptive regression spline PLS = partial least squares QSAR = quantitative structure-activity relationships RMSE = root mean squared error. 

Computational chemistry with its vast slew of optimization and related problems offers a wide scope for the application of genetic algorithms and evolutionary strategies. The above described applications are only two of several that have been investigated by researchers. The following briefly describes some related work in the application of evolutionary techniques to three areas relevant to computational chemistry and molecular design. 

The evolution of cooperation is frequently analysed in terms of the repeated Prisoner s Dilemma game. Computer simulations show that the emergence of cooperation is a robust phenomenon. However, the strategy which eventually gets adopted in the population seems to depend sensitively on fine details of the modelling process, so that it becomes difficult to predict the evolutionary outcome in real populations. 

The strategy of design, illustrated in Figure 8.1, consists of an evolutionary search of the feasible design space by means of a systematic combination of thermodynamic analysis, computer simulation and only limited experiments. The approach is generic for developing a RD process, at least for similar systems. The first element of similarity is the existence of an equilibrium reaction with water as product This raises the problem of possible aqueous-phase segregation. The second element is the similarity of thermodynamics properties over a class of substrates. However, while the fatty acids and fatty esters manifest a certain... 

Knowles, 1. D. and Come, D. W. (2000). Approximating the nondominated front using the Pareto archived evolution strategy. Evolutionary Computation, 8(2), pp. 149-172. 

J. Knowles and D. Come. The Pareto Archived Evolution Strategy A new baseline algorithm for Pareto multiobjective optimisation. In Proceedings of the 1999 Congress on Evolutionary Computation, pages 98-105, 1999. 

As we will see in a final chapter, dynamic Monte Carlo methods, genetic algorithms and evolutionary computational strategies help to determine the optimum structure of catalysts for maximum performance. 

A Computer-oriented algorithm for nonstationary random vibrations of polygonal Kirchhoff-plates has been developed. A fast and accurate BEM for calculation of undamped frequency response function has been used. Light hysteretic damping has been introduced subsequently, and the modified spectral Priestley-formulation has been applied in order to calculate evolutionary output power spectral density functions. It is hoped that this solution strategy will be of interest for workers in the field of probabilistic structural dynamics. 

---