A simple but effective evolutionary algorithm is proposed in this paper for solving complicated optimization problems. The new algorithm presents two hybridization operations incorporated with the conventional genetic algorithm. It takes only 4.1% ~ 4.7% number of function evaluations required by the conventional genetic algorithm to obtain global optima for the benchmark functions tested. Application example is also provided to demonstrate its effectiveness. / Singapore-MIT Alliance (SMA)

Cluster analysis is used to detect underlying group structure in data. Model-based clustering is the process of performing cluster analysis which involves the fitting of finite mixture models. However, parameter estimation in mixture model-based approaches to clustering is notoriously difficult. To this end, this thesis focuses on the development of evolutionary computation as an alternative technique for parameter estimation in mixture models. An evolutionary algorithm is proposed and illustrated on the well-established Gaussian mixture model with missing values. Next, the family of Gaussian parsimonious clustering models is considered, and an evolutionary algorithm is developed to estimate the parameters. Next, an evolutionary algorithm is developed for latent Gaussian mixture models and to facilitate the flexible clustering of high-dimensional data. For all models and families of models considered in this thesis, the proposed algorithms used for model-fitting and parameter estimation are presented and the performance illustrated using real and simulated data sets to assess the clustering ability of all models. This thesis concludes with a discussion and suggestions for future work. / Dissertation / Doctor of Philosophy (PhD)

This thesis studied a new variant of Evolutionary Algorithms called Gene Expression Programming. The evolution process of Gene Expression Programming was investigated from the practice to the theory. As a practice level, the original version of Gene Expression Programming was applied to a classification problem and an enhanced version of the algorithm was consequently developed. This allowed the development of a general understanding of each component of the genotype and phenotype separated representation system of the solution employed by the algorithm. Based on such an understanding, a version of the schema theory was developed for Gene Expression Programming. The genetic modifications provided by each genetic operator employed by this algorithm were analysed and a set of theorems predicting the propagation of the schema from one generation to another was developed. Also a set of experiments were performed to test the validity of the developed schema theory obtaining good agreement between the experimental results and the theoretical predictions.

A well-known paradigm for optimisation is the evolutionary algorithm (EA). An EA maintains a population of possible solutions to a problem which converges on a global optimum using biologically-inspired selection and reproduction operators. These algorithms have been shown to perform well on a variety of hard optimisation and search problems. A recent development in evolutionary computation is the Estimation of Distribution Algorithm (EDA) which replaces the traditional genetic reproduction operators (crossover and mutation) with the construction and sampling of a probabilistic model. While this can often represent a significant computational expense, the benefit is that the model contains explicit information about the fitness function. This thesis expands on recent work using a Markov network to model fitness in an EDA, resulting in what we call the Markov Fitness Model (MFM). The work has explored the theoretical foundations of the MFM approach which are grounded in Walsh analysis of fitness functions. This has allowed us to demonstrate a clear relationship between the fitness model and the underlying dynamics of the problem. A key achievement is that we have been able to show how the model can be used to predict fitness and have devised a measure of fitness modelling capability called the fitness prediction correlation (FPC). We have performed a series of experiments which use the FPC to investigate the effect of population size and selection operator on the fitness modelling capability. The results and analysis of these experiments are an important addition to other work on diversity and fitness distribution within populations. With this improved understanding of fitness modelling we have been able to extend the framework Distribution Estimation Using Markov networks (DEUM) to use a multivariate probabilistic model. We have proposed and demonstrated the performance of a number of algorithms based on this framework which lever the MFM for optimisation, which can now be added to the EA toolbox. As part of this we have investigated existing techniques for learning the structure of the MFM; a further contribution which results from this is the introduction of precision and recall as measures of structure quality. We have also proposed a number of possible directions that future work could take.

Nowadays modern electric power systems consist of large-scale and highly complex interconnected transmission systems, thus transmission expansion planning (TEP) is now a significant power system optimisation problem. The TEP problem is a large-scale, complex and nonlinear combinatorial problem of mixed integer nature where the number of candidate solutions to be evaluated increases exponentially with system size. The accurate solution of the TEP problem is essential in order to plan power systems in both an economic and efficient manner. Therefore, applied optimisation methods should be sufficiently efficient when solving such problems. In recent years a number of computational techniques have been proposed to solve this efficiency issue. Such methods include algorithms inspired by observations of natural phenomena for solving complex combinatorial optimisation problems. These algorithms have been successfully applied to a wide variety of electrical power system optimisation problems. In recent years differential evolution algorithm (DEA) procedures have been attracting significant attention from the researchers as such procedures have been found to be extremely effective in solving power system optimisation problems. The aim of this research is to develop and apply a novel DEA procedure directly to a DC power flow based model in order to efficiently solve the TEP problem. In this thesis, the TEP problem has been investigated in both static and dynamic form. In addition, two cases of the static TEP problem, with and without generation resizing, have also been investigated. The proposed method has achieved solutions with good accuracy, stable convergence characteristics, simple implementation and satisfactory computation time. The analyses have been performed within the mathematical programming environment of MATLAB using both DEA and conventional genetic algorithm (CGA) procedures and a detailed comparison has also been presented. Finally, the sensitivity of DEA control parameters has also been investigated.

This thesis describes the use of a Real-Time Evolutionary Algorithm (RTEA) to optimise an Artificial Neural Network (ANN) on-line (in this context “on-line” means while it is in use). Traditionally, Evolutionary Algorithms (Genetic Algorithms, Evolutionary Strategies and Evolutionary Programming) have been used to train networks before use - that is “off-line,” as have other learning systems like Back-Propagation and Simulated Annealing. However, this means that the network cannot react to new situations (which were not in its original training set). The system outlined here uses a Simulated Legged Robot as a test-bed and allows it to adapt to a changing Fitness function. An example of this in reality would be a robot walking from a solid surface onto an unknown surface (which might be, for example, rock or sand) while optimising its controlling network in real-time, to adjust its locomotive gait, accordingly. The project initially developed a Central Pattern Generator (CPG) for a Bipedal Robot and used this to explore the basic characteristics of RTEA. The system was then developed to operate on a Quadruped Robot and a test regime set up which provided thousands of real-environment like situations to test the RTEA’s ability to control the robot. The programming for the system was done using Borland C++ Builder and no commercial simulation software was used. Through this means, the Evolutionary Operators of the RTEA were examined and their real-time performance evaluated. The results demonstrate that a RTEA can be used successfully to optimise an ANN in real-time. They also show the importance of Neural Functionality and Network Topology in such systems and new models of both neurons and networks were developed as part of the project. Finally, recommendations for a working system are given and other applications reviewed.

In solving noise-free and noisy optimization problems, candidate initialization and sampling play a key role, but are not deeply investigated. It is of interest to know if the entire search space has the same quality for candidate-solutions during solving different type of optimization problems. In this thesis, a comprehensive investigation is conducted in order to clear those doubts, and to examine the effects of variant sampling methods on solving challenging optimization problems, such as large-scale, noisy, and multi-modal problems. As a result, the search space is segmented by using seven segmentation schemes, namely: Center-Point, Center-Based, Modula-Opposite, Quasi-Opposite, Quasi-Reflection, Supper- Opposite, and Opposite-Random. The introduced schemes are studied using Monte-Carlo simulation, on various types of noise-free optimization problems, and ultimately ranked based on their performance in terms of probability of closeness, average distance to unknown solution, number of solutions found, and diversity. Based on the results of the experiments, high-ranked schemes are selected and utilized on well-known metaheuristic algorithms, as case studies. Two categories of case studies are targeted; one for a singlesolution- based metaheuristic (S-metaheuristic) and another one for a population based metaheuristic (P-metaheuristic). A high-ranked single-solution-based scheme is utilized to accelerate Simulated Annealing (SA) algorithm, as a noise-free S-metaheuristic case study. Similarly, for noise-free P-metaheuristic case study, an effective population-based algorithm, Differential Evolution (DE), has been utilized. The experiments confirm that the new algorithms outperform the parent algorithm (DE) on large-scale problems. In the same direction, with regards to solving noisy problems more efficiently, a Shaking-based sampling method is introduced, in which the original noise is tackled by adding an additional noise into the search process. As a case study, the Shaking-based sampling is utilized on the DE algorithm, from which two variant algorithms have been developed and showed impressive performance in comparison to the classical DE, in tackling noisy largescale problems. This thesis has created an opportunity for a comprehensive investigation on search space segmentation schemes and proposed new sampling methods. The current study has provided a guide to use appropriate sampling schemes for a given types of problems such as noisy, large-scale and multi-modal optimization problems. Furthermore, this thesis questions the effectiveness of uniform-random sampling method, which is widely used in of S-Metaheuristic and P-Metaheuristic algorithms. / UOIT

abstract: This dissertation presents a portable methodology for holistic planning and optimization of right of way infrastructure rehabilitation that was designed to generate monetary savings when compared to planning that only considers single infrastructure components. Holistic right of way infrastructure planning requires simultaneous consideration of the three right of way infrastructure components that are typically owned and operated under the same municipal umbrella: roads, sewer, and water. The traditional paradigm for the planning of right way asset management involves operating in silos where there is little collaboration amongst different utility departments in the planning of maintenance, rehabilitation, and renewal projects. By collaborating across utilities during the planning phase, savings can be achieved when collocated rehabilitation projects from different right of way infrastructure components are synchronized to occur at the same time. These savings are in the form of shared overhead and mobilization costs, and roadway projects providing open space for subsurface utilities. Individual component models and a holistic model that utilize evolutionary algorithms to optimize five year maintenance, rehabilitation, and renewal plans for the road, sewer, and water components were created and compared. The models were designed to be portable so that they could be used with any infrastructure condition rating, deterioration modeling, and criticality assessment systems that might already be in place with a municipality. The models attempt to minimize the overall component score, which is a function of the criticality and condition of the segments within each network, by prescribing asset management activities to different segments within a component network while subject to a constraining budget. The individual models were designed to represent the traditional decision making paradigm and were compared to the holistic model. In testing at three different budget levels, the holistic model outperformed the individual models in the ability to generate five year plans that optimized prescribed maintenance, rehabilitation and renewal for various segments in order to achieve the goal of improving the component score. The methodology also achieved the goal of being portable, in that it is compatible with any condition rating, deterioration, and criticality system. / Dissertation/Thesis / Ph.D. Construction 2012

This thesis deals with the problem of TV commercials planning. This problem is getting more difficult as the number of different TV stations grows and naive algorithms become unable to scale with this number. In this work, we deal with the possibility to use evolutionary algorithms to solve this problem. The work also contains an implementation of a software, which is capable of testing the various versions of the algorithm and comparing them to a naive one. The results indicate that evolutionary algorithms are a suitable technique to solve the problem at hand. Powered by TCPDF (www.tcpdf.org)

Abstract In industry there is a high demand for algorithms that can efficiently solve search problems. Evolutionary Computing (EC) belonging to a class of heuristics are proven to be well suited to solve search problems, especially optimization tasks. They arrived at that location because of their flexibility, scalability and robustness. However, despite their advantages and increasing popularity, there are numerous opened questions in this research area, many of them related to the design and tuning of the algorithms. A neutral technique called Pseudo Redundancy and related concepts such as Updated Objective Grid (UOG) is proposed to tackle the mentioned problem making an evolutionary approach more suitable for ''real world'' applications while increasing its robustness and efficiency. The proposed UOG technique achieves neutral search by objective function transformation(s) resulting several advantageous features. (a) Simplifies the design of an evolutionary solver by giving population sizing principles and directions to choose the right selection operator. (b) The technique of updated objective grid is adaptive without introducing additional parameters, therefore no parameter tuning required for UOG to adjust it for different environments, introducing robustness. (c) The algorithm of UOG is simple and computationally cheap. (d) It boosts the performance of an evolutionary algorithm on high dimensional (constrained and unconstrained) problems. The theoretical and experimental results from artificial test problems included in this thesis clearly show the potential of the proposed technique. In order to demonstrate the power of the introduced methods under "real" circumstances, the author additionally designed EAs and performed experiments on two industrial optimization tasks. Although, only one project is detailed in this thesis while the other is referred. As the main outcome of this thesis, the author provided an evolutionary method to compute (optimal) daily water pump schedules for the water distribution network of Sopron, Hungary. The algorithm is currently working in industry.