On the Relations Between Search and Evolutionary Algorithms

Evolutionary algorithms are powerful techniques for optimisation whose operation principles are inspired by natural selection and genetics. In this paper we discuss the relation between evolutionary techniques, numerical and classical search methods and we show that all these methods are instances of a single more general search strategy, which we call the `evolutionary computation cookbook'. By combining the features of classical and evolutionary methods in different ways new instances of this general strategy can be generated, i.e. new evolutionary (or classical) algorithms can be designed. One such algorithm, GA , is described. 1 Introduction Evolutionary algorithms (EAs) are powerful optimisation techniques inspired by genetics and natural selection [6, 5, 2, 1, 9]. EAs are often referred to as global optimisation methods, to stress the fact that they can effectively explore very large solutions spaces without becoming trapped by local minima. Applying EAs to a new problem is s..