PhD Thesis, School of Electrical, Electronic and Computer Engineering,University of Western
May 2008, 281 p
The Genetic Programming paradigm, which applies the Darwinian principle of evolution to hierarchical computer programs, has produced promising breakthroughs in various scientific and engineering applications. However, one of the main drawbacks of Genetic Programming has been the often large amount of computational effort required to solve complex problems. There have been various amounts of research conducted to devise innovative methods to improve the efficiency of Genetic Programming. This thesis has three main contributions. It firstly provides a comprehensive overview of the related work to improve the performance of Genetic Programming and classifies these various proposed approaches into categories. Secondly, a new static population variation scheme (PV) is proposed, whereby the size of the population is varied according to a predetermined schedule during the execution of the Genetic Programming system with the aim of reducing the computational effort with respect to that of Standard Genetic Programming. Within this new static scheme the initial population size is made to be different from the initial size of the Standard Genetic Programming such that the worst case computational effort is never greater than that of the Standard Genetic Programming. Various static schemes for altering population size under this proposal are investigated using a comprehensive range of standard problems to determine whether the nature of the "population variation", i.e. the way the population is varied during the search, has any significant impact on Genetic Programming performance. It is shown that these population variation schemes do have the capacity to provide solutions at a lower computational cost compared with the Standard Genetic Programming. Thirdly, three innovations for dynamically varying the population size during the run of the Genetic Programming system are proposed. These are related to what is called Dynamic Population Variation (DPV), where the size of the population is dynamically varied
using a heuristic feedback mechanism during the execution of the Genetic Programming with the aim of reducing the computational effort. The efficacy of these innovations is examined
using the same comprehensive range of standard representative problems. It is shown that these new ideas do have the capacity to provide solutions at a lower computational cost
compared with standard genetic programming and previously reported algorithms. Finally, further interesting research potentials for population variation are identified together with some of the open areas of research within the Genetic Programming and also possible future trends in this discipline.
