Hereby Recommend That the Dissertation Prepared under Our Su- Pervision by Soraya Rana Entitled Examining the Role of Local Optima and Schema Processing in Genetic Search Be Accepted as Fulfilling in Part Requirements for the Degree of Doctor

OF DISSERTATION EXAMINING THE ROLE OF LOCAL OPTIMA AND SCHEMA PROCESSING IN GENETIC SEARCH Several factors contribute to making search problems easy or di cult. One of these factors is the modality of the tness landscape. Quite often, when local search algorithms fail to locate the global optimum, it is because the algorithm converged to a local optimum. The majority of local search methods in use today maneuver through the search space using local neighborhood information around a single point to guide the search. In that search paradigm, the number of local optima that occur in the search space has a tremendous e ect on search performance. Genetic algorithms are a population based search algorithm that use an ever changing neighborhood structure, based on the population mixture and genetic operators, to sample points in the search space. Genetic algorithms are believed to process schemata, where a schema is a subpartition of the search space, rather than individual points. When genetic algorithms fail to locate the global optimum, the typical analysis is that the schema information in the optimization problem was misleading. Consequently, it is unclear how local optima can a ect such algorithms. While misleading schema information may be one source of problem di culty for genetic algorithms, the existence of multiple high quality local optima may also be responsible for misleading the genetic algorithm. This research explores the relationship between local optima, schema processing and genetic algorithm behavior from a variety of perspectives. Soraya Rana Department of Computer Science Colorado State University Fort Collins, Colorado 80523 Summer 1999 iii TABLE OF CONTENTS