Evolutionary Computational Methods

In this section, we will begin with a short exposition of different evolutionary computational approaches. Then we will apply one of these methods to the design an alloy catalyst with optimal performance for a particular dissociation reaction. The performance of the catalyst is theoretically tested using the dynamic Monte Carlo method to predict the kinetics for a surface reaction. 

In evolution strategies, a method that is well suited for parameter optimization, crossover and mutation is used first to generate a large number of offspring. Selection is then used to reduce the offspring to a new population. The selection can include the old population or not, and there are many different methods to make the selection. 

The selection process is the place where the optimization in EC methods really takes place. This is where methods such as dynamic or kinetic Monte Carlo (DMC) simulations become important. They are used to compute the properties of a system or process. These properties are then converted to a fitness value. This fitness value is for satisfaction of a particular requirement of performance which is then operated on by the EC methods. The conversion is different for each system and property and also determines how effective the selection is. Dynamic Monte Carlo simulation, as we have already discussed, is a method to simulate elementary processes along with the actual rate. The method uses each individual reaction as an elementary event, which means that timescales comparable to actual experiments can be simulated. The reaction rate constants that it needs as input can be calculated using quantum chemical methods such as density functional theory, which results in what has been termed ab initio kinetics (see Chapter 3.10.4). 

The application of such methods to the modeling of overall kinetics has been described in detail in Chapter 3. EC and DFT can be combined to investigate the effect of replacing atoms or chemical groups by others. The simplest application would be to have the EC determine only the composition. This might be done with a straightforward GA. The coding can be done more or less as shown in Fig. 8.18. 

The genotype consists of a number of parts each of which correspond to an atom (or molecule). In principle, standard crossover and mutation can then be used e.g., this has been the procedure followed by the group of Nprskov to find new super-strong alloys Computer experiments have shown that adapting crossover and mutation to the problem can speed up that optimization. The combination EC plus DMC can also be used to carry out structural optimizations. 

---

J., Evolutionary computational methods to predict oral bioavailability QSPRs, Curr. Opin. Drug Disc. Dev. 2002, 5, 44-51. 

In this Volume, Evolutionary Computation methods are introduced and their application to a number of areas of current chemical interest is reviewed. Firstly, the differences between the various methods of Evolutionary Computation and the principles underlying them are outlined in the Chapter by Hugh Cartwright, which also discusses, in a generic way, the pseudo-evolutionary operators used in each case. Specific applications and detailed discussions of algorithms and methodologies are presented in subsequent Chapters. 

Bains, W., Gilbert, R., Sviridenko, L., Gascon, J.L., Scoffin, R., Birchall, K., Harvey, I. and Caldwell, J. (2002) Evolutionary computational methods to predict oral bioavailability QSPRs. Current Opinion in Drug Discovery cl Development, 5, 44-51. 

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. 

GAs or other methods from evolutionary computation are applied in various fields of chemistry Its tasks include the geometry optimization of conformations of small molecules, the elaboration of models for the prediction of properties or biological activities, the design of molecules de novo, the analysis of the interaction of proteins and their ligands, or the selection of descriptors [18]. The last application is explained briefly in Section 9.7.6. 

The rest of the chapter is mainly composed of five diverse examples of application of evolutionary developmental methods to problems in engineering and computer science. The chapter concludes with an attempt to foresee some future trends in EDS research and application, followed by a very short story that we hope will entertain and perhaps inspire. 

Genetic programming, a specific form of evolutionary computing, has recently been used for predicting oral bioavailability [23], The results show a slight improvement compared with the ORMUCS Yoshida-Topliss approach. This supervised learning method and other described methods demonstrate that at least qualitative (binned) predictions of oral bioavailability seem tractable directly from the structure. 

Coello Coello, C.A. (2002) Theoretical and numerical constraint handling techniques used with evolutionary algorithms A survey of the state of the art. Computer Methods in Applied Mechanics and Engineering, 191 (11—12), 1245-1287. 

Gulyaev Yu.V. Bukatova I.L. and Krapivin V.F. (1989a). Evolutionary computer technology. In E.P. Novitchikhin (ed.), Methods of Informatics in Radiophysical Investigations of the Environment. Science Publ., Moscow, pp. 25-43 [in Russian],... 

Evolutionary computation approaches are optimization methods. They are conveniently presented using the metaphor of natural evolution a randomly initialized population of individuals evolves following a crude parody of the Darwinian principle of the survival of the fittest. New individuals are generated using simulated evolutionary operations such as mutations. The probability of survival of the newly generated solutions depends on their fitness (how well they perform with respect to the optimization problem at hand) the best are kept with a high probability, the worst are rapidly discarded. 

The authors present some of the methods that we have developed and exploited in Aberystwyth for gathering highly multivariate data from bioprocesses, and some techniques of sound multivariate statistical analyses (and of related methods based on neural and evolutionary computing) which can ensure that the results will stand up to the most rigorous scrutiny. 

The boundaries to AI, within which evolutionary optimisers He, are woolly, and this may make it appear a mysterious subject to those who are unfamiliar with it even those whose research lies in AI find it hard to agree on the precise boundaries to the field. It encompasses a large family of algorithms, from neural networks and knowledge-based systems to those within evolutionary computing. Almost every major AI technique is now used within chemistry (Table 1), and in an increasing number of cases, AI is the method of choice. 

Although the use of evolutionary computing within science is still in its early stages, scientific applications are already notable for their diversity. Evolutionary methods have been used to optimise the geometry of molecules, the shape of propellers, the properties of polymers and the order in which chemicals are produced in industrial flow fines. They have been used in the study of oil extraction, the natural degradation of toxic chemicals in the environment, spectral deconvolution, the interpretation of microwave spectra and in a wide range of other areas. The chapters that follow provide further illustration of the potential of these intriguing and versatile techniques. 

Traditionally, the templates were chosen by trial and error or exhaustive enumeration. A computational method named ZEBEDDE (ZEolites By Evolutionary De novo DEsign) has been developed to try to introduce some rationale into the selection of templates [Lewis et al. 1996 Willock et al. 1997]. The templates are grown within the zeolite cavity by an iterative inside-out approach, starting from a seed molecule. At each iteration an action is randomly selected from a list that includes the addition of new atoms (from a library of fragments), random translation or rotation, random bond rotation, ring formation or energy minimisation of the template. A cost function based on the overlap of van der Waals spheres is used to control the growth of the template molecule ... 

Back, T., Fogel, D.B. and Michalewicz, Z., 1997. Handbook of Evolutionary Computation, Oxford University Press, New York, and Institute of Physics Pubhshing, Bristol. Becker, C. and Scholl, A., 2003. A survey on problems and methods in generalized assembly line balancing, European Journal of Operational Research, 168(3), 694-715. De Jong, K.A., 1975. An Analysis of the Behavior of a Class of Genetic Adaptive Systems, Doctoral dissertation, University of Michigan, Ann Arbor, MI (University Microfilms No. 76-9381). 

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. 

Genetic Algorithm (GA) was first proposed in 1975 by Holland [107]. As a strong and widely application method for random search and optimization, GA is one of the most influential methods of Evolutionary Computation [108]. During the past... 

---