Evolutionary Algorithms are search algorithms based on the Darwinian metaphor of “Natural Selection”. Typically these algorithms maintain a population of individual solutions, each of which has a fitness attached to it, which in some way reflects the quality of the solution. The search proceeds via the iterative generation, evaluation and possible incorporation of new individuals based on the current population, using a number of parameterised genetic operators. In this thesis the phenomenon of Self Adaptation of the genetic operators is investigated. A new framework for classifying adaptive algorithms is proposed, based on the scope of the adaptation, and on the nature of the transition function guiding the search through the space of possible configurations of the algorithm. Mechanisms are investigated for achieving the self adaptation of recombination and mutation operators within a genetic algorithm, and means of combining them are investigated. These are shown to produce significantly better results than any of the combinations of fixed operators tested, across a range of problem types. These new operators reduce the need for the designer of an algorithm to select

k.matthews, a.r.sibbald @ mluri.sari.ac.uk This paper explores the application of multiobjective Genetic Algorithms (mGAs) to rural land use planning, a spatial allocation problem. Two mGAs are proposed. Both share an underlying structure of: fitness assignment using Pareto-dominance ranking, niche induction and an individual replacement strategy. They are differentiated by their representations: a fixedlength genotype composed of genes that map directly to a land parcel's use and a variablelength, order-dependent representation making allocations indirectly via a greedy algorithm. The latter representation requires additional breeding operators to be defined and post-processing of the genotype structure to identify and remove duplicate genotypes. The two mGAs are compared on a real land use planning problem and the strengths and weaknesses of the underlying framework and each representation are identified. 1

Abstract. This paper explores the potential of applying Genetic Algorithms to land use planning, a spatial allocation problem. Two genotype representations are proposed: a fixed-length genotype composed of genes that map directly to a land parcel's use, and a variable-length, order-dependent representation making allocations indirectly via a greedy algorithm. The fixed-length genotype is used within a standard genetic algorithm framework but the variable-length genotype requires novel breeding operators to be defined and post-processing of the genotype structure to identify and remove duplicate genotypes. The two approaches are compared on a real land use planning problem and the strengths and weaknesses of each approach are identified. Key Words: representation, messy GAs, non-fitness information, land use planning