Dissertations / Theses on the topic 'NATURE INSPIRED ALGORITHM'

This research focuses on nature-inspired optimisation algorithms, in particular, the Particle Swarm Optimisation (PSO) Algorithm and the Bees Algorithm. The PSO Algorithm is a population-based stochastic optimisation technique first invented in 1995. It was inspired by the social behaviour of birds flocking or a school of fish. The Bees Algorithm is a population-based search algorithm initially proposed in 2005. It mimics the food foraging behaviour of swarms of honey bees. The thesis presents three algorithms. The first algorithm called the PSO-Bees Algorithm is a cross between the PSO Algorithm and the Bees Algorithm. The PSO-Bees Algorithm enhanced the PSO Algorithm with techniques derived from the Bees Algorithm. The second algorithm called the improved Bees Algorithm is a version of the Bees Algorithm that incorporates techniques derived from the PSO Algorithm. The third algorithm called the SNTO-Bees Algorithm enhanced the Bees Algorithm using techniques derived from the Sequential Number-Theoretic Optimisation (SNTO) Algorithm. To demonstrate the capability of the proposed algorithms, they were applied to different optimisation problems. The PSO-Bees Algorithm is used to train neural networks for two problems, Control Chart Pattern Recognition and Wood Defect Classification. The results obtained and those from tests on well known benchmark functions provide an indication of the performance of the algorithm relative to that of other swarm-based stochastic optimisation algorithms. The improved Bees Algorithm was applied to mechanical design optimisation problems (design of welded beams and coil springs) and the mathematical benchmark problems used previously to test the PSO-Bees Algorithm. The algorithm incorporates cooperation and communication between different neighbourhoods. The results obtained show that the proposed cooperation and communication strategies adopted enhanced the performance and convergence of the algorithm. The SNTO-Bees Algorithm was applied to a set of mechanical design optimisation problems (design of welded beams, coil springs and pressure vessel) and mathematical benchmark functions used previously to test the PSO-Bees Algorithm and the improved Bees Algorithm. In addition, the algorithm was tested with a number of deceptive multi modal benchmark functions. The results obtained help to validate the SNTO-Bees Algorithm as an effective global optimiser capable of handling problems that are deceptive in nature with high dimensions.

A major factor that negatively impacts flight operations globally is adverse weather. To reduce the impact of adverse weather, avoidance procedures such as finding an alternative flight path can usually be carried out. However, such procedures usually introduce extra costs such as flight delay. Hence, there exists a need for alternative flight paths that efficiently avoid adverse weather regions while minimising costs. Existing weather avoidance methods used techniques, such as Dijkstra’s and artificial potential field algorithms that do not scale adequately and have poor real time performance. They do not adequately consider the impact of weather and its avoidance on passengers. The contributions of this work include a new development of an improved integrated model for weather avoidance, that addressed the impact of weather on passengers by defining a corresponding cost metric. The model simultaneously considered other costs such as flight delay and fuel burn costs. A genetic algorithm (GA)-based rerouting technique that generates optimised alternative flight paths was proposed. The technique used a modified mutation strategy to improve global search. A discrete firefly algorithm-based rerouting method was also developed to improve rerouting efficiency. A data framework and simulation platform that integrated aeronautical, weather and flight data into the avoidance process was developed. Results show that the developed algorithms and model produced flight paths that had lower total costs compared with existing techniques. The proposed algorithms had adequate rerouting performance in complex airspace scenarios. The developed system also adequately avoided the paths of multiple aircraft in the considered airspace.

In terms of technology, BIM is also part of the worldwide change Industry 4.0, which in essence is the trend toward automation and data exchange in manufacturing technologies and processes. Generative design is an iterative process that involves a program that will generate a certain number of outputs that meet certain constraints, so that a designer is able to fine tune the feasible project by changing minimal and maximal values of an interval in which a variable of the program meets the set of constraints, in order to reduce or augment the number of outputs to choose from. The initial idea of this thesis work was to manipulate few of the most basic geometric elements in order to get a complex parametric shape inspired from the honeycomb as the natures perfectly generated the element. This preliminary idea, together with the ambition to use this transformation for a façade system in a structural building led us to a series of decisions to try and connect two “worlds”, in the sense that we have a CAD environment that lets us create the geometry and a BIM environment where everything is represented by a specific level of information. This geometry is given a specific set of rules that drive and manipulate each of the elements it contains in a certain fashion. This methodology, as well as the communication and the interaction between the software adopted and their programming environments, is what makes the generative design possible. This result from the Grasshopper algorithm is then being created in the CAD environment in Rhinoceros3D, which then can be opened through Rhino.Inside.Revit and give us a direct real-time preview in the BIM environment in Revit. Through a long series of testing and experimenting with the geometry, we get to a point where we have a functional algorithm that creates and manipulates the geometry, in order to foster many design opportunities for structural and architectural designers.

This thesis investigates the core difficulties in the tracking field of computer vision. The aim is to develop a suitable tuning free optimisation strategy so that a real time tracking could be achieved. The population and multi-solution based approaches have been applied first to analyse the convergence behaviours in the evolutionary test cases. The aim is to identify the core misconceptions in the manner the search characteristics of particles are defined in the literature. A general perception in the scientific community is that the particle based methods are not suitable for the real time applications. This thesis improves the convergence properties of particles by a novel scale free correlation approach. By altering the fundamental definition of a particle and by avoiding the nostalgic operations the tracking was expedited to a rate of 250 FPS. There is a reasonable amount of similarity between the tracking landscapes and the ones generated by three dimensional evolutionary test cases. Several experimental studies are conducted that compares the performances of the novel optimisation to the ones observed with the swarming methods. It is therefore concluded that the modified particle behaviour outclassed the traditional approaches by huge margins in almost every test scenario.

As the number of nature-inspired algorithms increases so does the need to characterise these algorithms. A rigorous process to characterise algorithms helps practitioners decide which algorithms may offer a good fit for their given problem. One approach is to relate the characteristics of a problem's associated fitness landscape with the performance of an algorithm. The aim of this thesis is to capitalise on the notion of fitness landscape characteristics as a technique for analysing algorithm performance, and to provide a novel algorithm- and problem-independent methodology that can be used to present the strengths and weaknesses of an algorithm. The methodology was tested by developing a portfolio of six nature-inspired algorithms commonly used to solve continuous optimisation problems. This portfolio includes the performance of these algorithms with parameters both “out of the box" and after they have been tuned using an automated tuning technique. Each of the algorithms shows a different “resilience" profile to the landscape characteristics, and responds differently to the tuning process. In order to provide a more practical way to utilise the portfolio an automated “ranking" methodology based on two machine learning techniques was developed. Using estimates of the fitness landscape characteristics on benchmark problems, the best algorithm to use is estimated, and compared with the actual performance of each algorithm. While results show that predicting algorithm performance is difficult, the results are promising, and show that this is an area worth exploring further. This methodology has significant advantages over the current practice of demonstrating novel algorithm performance on benchmark problems, most importantly offering a practical, generalised overview of the algorithm to a potential practitioner. Choosing to use a technique such as the one demonstrated here when presenting a novel algorithm could greatly ease the problem of algorithm selection.

This thesis addresses orchestration of servers amongst multiple co-hosted internet services such as e-Banking, e-Auction and e-Retail in hosting centres. The hosting paradigm entails levying fees for hosting third party internet services on servers at guaranteed levels of service performance. The orchestration of server ensemble in hosting centres is considered in the context of maximising the hosting centre's revenue over a lengthy time horizon. The inspiration for the server orchestration approach proposed in this thesis is drawn from nature and generally classed as swarm intelligence, specifically, sophisticated collective behaviour of social insects borne out of primitive interactions amongst members of the group to solve problems beyond the capability of individual members. Consequently, the approach is self-organising, adaptive and robust. A new scheme for server ensemble orchestration is introduced in this thesis. This scheme exploits the many similarities between server orchestration in an internet hosting centre and forager allocation in a honeybee (Apis mellifera) colony. The scheme mimics the way a honeybee colony distributes foragers amongst flower patches to maximise nectar influx, to orchestrate servers amongst hosted internet services to maximise revenue. The scheme is extended by further exploiting inherent feedback loops within the colony to introduce self-tuning and energy-aware server ensemble orchestration. In order to evaluate the new server ensemble orchestration scheme, a collection of server ensemble orchestration methods is developed, including a classical technique that relies on past history to make time varying orchestration decisions and two theoretical techniques that omnisciently make optimal time varying orchestration decisions or an optimal static orchestration decision based on complete knowledge of the future. The efficacy of the new biomimetic scheme is assessed in terms of adaptiveness and versatility. The performance study uses representative classes of internet traffic stream behaviour, service user's behaviour, demand intensity, multiple services co-hosting as well as differentiated hosting fee schedule. The biomimetic orchestration scheme is compared with the classical and the theoretical optimal orchestration techniques in terms of revenue stream. This study reveals that the new server ensemble orchestration approach is adaptive in a widely varying external internet environments. The study also highlights the versatility of the biomimetic approach over the classical technique. The self-tuning scheme improves on the original performance. The energy-aware scheme is able to conserve significant energy with minimal revenue performance degradation. The simulation results also indicate that the new scheme is competitive or better than classical and static methods.

Financial contagion refers to a scenario in which small shocks, which initially affect only a few financial institutions or a particular region of the economy, spread to the rest of the financial sector and other countries whose economies were previously healthy. This resembles the “transmission” of a medical disease. Financial contagion happens both at domestic level and international level. At domestic level, usually the failure of a domestic bank or financial intermediary triggers transmission by defaulting on inter-bank liabilities, selling assets in a fire sale, and undermining confidence in similar banks. An example of this phenomenon is the failure of Lehman Brothers and the subsequent turmoil in the US financial markets. International financial contagion happens in both advanced economies and developing economies, and is the transmission of financial crises across financial markets. Within the current globalise financial system, with large volumes of cash flow and cross-regional operations of large banks and hedge funds, financial contagion usually happens simultaneously among both domestic institutions and across countries. There is no conclusive definition of financial contagion, most research papers study contagion by analyzing the change in the variance-covariance matrix during the period of market turmoil. King and Wadhwani (1990) first test the correlations between the US, UK and Japan, during the US stock market crash of 1987. Boyer (1997) finds significant increases in correlation during financial crises, and reinforces a definition of financial contagion as a correlation changing during the crash period. Forbes and Rigobon (2002) give a definition of financial contagion. In their work, the term interdependence is used as the alternative to contagion. They claim that for the period they study, there is no contagion but only interdependence. Interdependence leads to common price movements during periods both of stability and turmoil. In the past two decades, many studies (e.g. Kaminsky et at., 1998; Kaminsky 1999) have developed early warning systems focused on the origins of financial crises rather than on financial contagion. Further authors (e.g. Forbes and Rigobon, 2002; Caporale et al, 2005), on the other hand, have focused on studying contagion or interdependence. In this thesis, an overall mechanism is proposed that simulates characteristics of propagating crisis through contagion. Within that scope, a new co-evolutionary market model is developed, where some of the technical traders change their behaviour during crisis to transform into herd traders making their decisions based on market sentiment rather than underlying strategies or factors. The thesis focuses on the transformation of market interdependence into contagion and on the contagion effects. The author first build a multi-national platform to allow different type of players to trade implementing their own rules and considering information from the domestic and a foreign market. Traders’ strategies and the performance of the simulated domestic market are trained using historical prices on both markets, and optimizing artificial market’s parameters through immune - particle swarm optimization techniques (I-PSO). The author also introduces a mechanism contributing to the transformation of technical into herd traders. A generalized auto-regressive conditional heteroscedasticity - copula (GARCH-copula) is further applied to calculate the tail dependence between the affected market and the origin of the crisis, and that parameter is used in the fitness function for selecting the best solutions within the evolving population of possible model parameters, and therefore in the optimization criteria for contagion simulation. The overall model is also applied in predictive mode, where the author optimize in the pre-crisis period using data from the domestic market and the crisis-origin foreign market, and predict in the crisis period using data from the foreign market and predicting the affected domestic market.

Recent innovations and progress in machine learning algorithms from the Natural Language Processing (NLP) community have motivated efforts to apply these models and concepts to proteins. The representations generated by trained NLP models have been shown to capture important semantic and structural understanding of proteins encompassing biochemical and biophysical properties, among other key concepts. In turn, these representations have demonstrated application to protein engineering tasks including mutation analysis and design of novel proteins. Here we use this NLP paradigm in a protein engineering effort to further red shift the emission wavelength of the red fluorescent protein mNeptune684 using only a small number of functional training variants ('Low-N' scenario). The collaborative nature of this thesis with the Department of Chemistry and Biomolecular Sciences explores using these tools and methods in the rational design process.

Submitted by Renata Lopes (renatasil82@gmail.com) on 2016-02-05T17:32:25Z No. of bitstreams: 1 diegonascimentoarcanjo.pdf: 1706072 bytes, checksum: 2329ddd810b5aca8da733c7793937d65 (MD5)
Approved for entry into archive by Adriana Oliveira (adriana.oliveira@ufjf.edu.br) on 2016-02-26T11:52:47Z (GMT) No. of bitstreams: 1 diegonascimentoarcanjo.pdf: 1706072 bytes, checksum: 2329ddd810b5aca8da733c7793937d65 (MD5)
Made available in DSpace on 2016-02-26T11:52:47Z (GMT). No. of bitstreams: 1 diegonascimentoarcanjo.pdf: 1706072 bytes, checksum: 2329ddd810b5aca8da733c7793937d65 (MD5) Previous issue date: 2014-07-24
CAPES - Coordenação de Aperfeiçoamento de Pessoal de Nível Superior
Neste trabalho é proposto uma metodologia multi-estágio utilizando algoritmos bio-inspirados para a resolução do processo de Restabelecimento de Sistemas Elétricos de Distribuição. O primeiro estágio consiste na solução de uma função multi-objetivo visando a determinação da configuração final das chaves do sistema após isolados os ramos defeituosos (configuração de pós-contingência). Neste estágio, a modelagem da função multi-objetivo busca uma configuração adequada de chaves para minimizar a carga não suprida, as perdas do sistema, o número de chaveamentos, penalizando as violações aos limites operativos do sistema e considerando a presença de consumidores prioritários. Adicionalmente, a restrição de radialidade é assegurada em cada configuração utilizando, caso necessário, uma técnica de abertura de laço. A partir da configuração final obtida no primeiro estágio, são identificadas as chaves que foram manobradas. O segundo estágio da metodologia busca a determinação da sequência de chaveamento levando em conta a minimização da energia não suprida. Essa formulação permite que o tempo de manobra das chaves possa ser considerado. Sendo necessário, é realizado, ainda neste estágio, cortes mínimos discretos de carga para cada manobra executada. Em ambos os estágios foram utilizadas algoritmos bio-inspirados como métodos de solução dos respectivos problemas de otimização não-lineares inteiros mistos. As técnicas utilizadas são: Algoritmos Genéticos, Método da Eco Localização de Morcegos (Bat Algorithm) e Método da Reprodução dos Pássaros Cuco (Cuckoo Search). Os desenvolvimentos do algoritmo proposto foi implementado no ambiente MatLab®. Os resultados obtidos foram comparados com outras metodologias conhecidas da literatura comprovando a eficiência e robustez da técnica proposta.
This dissertation proposes a methodology for solving multi-stage process of Restoration on Power Distribution Systems using Nature-Inspired Algorithms. The first stage consists in solving a fitness multi-objective function in order to determine the final configuration of the switches after the faulted branches were isolated (post-contingency configuration). In this stage the multi-objective function seeks through the suitable configuration to minimize the undelivered power, the power losses, the number of switching, penalizing for violation in the system operational limits and taking in consideration the presence of priority load in the system. Additionally the radiality constraint is improved using an open loop technique. After the final configuration is obtained, for the first stage, the switches which were maneuvered are identified. The second stage of the methodology is to determine the sequence of switching taking into account the minimization of energy not supplied. This formulation allows to consider the switching operation time. If necessary, the minimum discrete load shedding procedure is made for each maneuvered switch. In both stages Nature-Inspired Algorithms to solve mixed integer nonlinear programming problems were used. The techniques used are: Genetic Algorithms, Bat Algorithm and Cuckoo Search. The developments of the proposed algorithm were implemented in MatLab ® environment. The results obtained were compared with other well-known methodologies showing the efficiency and robustness of the proposed technique.

Face Recognition is one of the crucial areas of research because of its widespread application. It is also a real world problem that can be solved with the help of computational intelligence techniques . Now, days area of computational intelligence is gaining a lot of interest. Many of the nature –inspired methodologies and approaches comes under this area are used to solve the real-world problem to which traditional approaches are infeasible, ineffective and less efficient. Computational intelligence primarily includes artificial neural networks, evolutionary computation, swarm intelligence and fuzzy logic. Face Recognition is a two step process i.e. features extraction and recognition process. In the feature extraction phase Gabor kernel is used to smoothen the images and PCA is used for feature extraction. Later on an evolutionary algorithm is used to find optimal features and an evolutionary algorithm is used to recognize an input image. We worked on both the phases to increase efficiency, by applying various evolutionary techniques. There are so many techniques available to solve the problem that the application developer gets into dilemma that which technique is most appropriate to use. In our thesis we have tried to solve the problem of face recognition using some of the evolutionary techniques like Ant Colony Optimization(ACO), Particle Swarm Optimization(PSO) ,Hybrid ACO/PSO, Biogeographical Based Optimization(BBO), Extended BBO and holistic technique like Principal Component Analysis(PCA). We analyzed and compared the results of all of these techniques to make it clear that which one is appropriate. Performance analysis is performed using Olivetti research Laboratory (ORL) face database and Cohn-Kanade database. We have shown the performance on the basis of time taken for recognition and accuracy in recognition. We found that different technique are best on both the parameters but if we are trying to find the appropriate one, Extended BBO serves the purpose.

In this dissertation a new architecture of antenna array is proposed in which antenna elements are placed on a Hexagonal Prism in order to obtain split beam radiation pattern.The Hexagonal Prism Array(HPA) is synthesized for amplitude excitation ,complex excitation and relative distance using a novel optimisation technique.As a performance criteria the split beam radiation pattern is evaluated for side lobe level(SLL) and half power beam width(HPBW).The optimisation technique is experimented on conventional arrays like Linear and Planar and a comparative study with the published literature is also performed.An attempt to model the HPA on a commercial software in which radiator is a patch elements.

Optimisation is a branch of mathematics that was developed to find the optimal solutions, among all the possible ones, for a given problem. Applications of optimisation techniques are currently employed in engineering, computing, and industrial problems. Therefore, optimisation is a very active research area, leading to the publication of a large number of methods to solve specific problems to its optimality. This dissertation focuses on the adaptation of two nature inspired algorithms that, based on optimisation techniques, are able to compute approximations for zeros of polynomials and roots of non-linear equations and systems of non-linear equations. Although many iterative methods for finding all the roots of a given function already exist, they usually require: (a) repeated deflations, that can lead to very inaccurate results due to the problem of accumulating rounding errors, (b) good initial approximations to the roots for the algorithm converge, or (c) the computation of first or second order derivatives, which besides being computationally intensive, it is not always possible. The drawbacks previously mentioned served as motivation for the use of Particle Swarm Optimisation (PSO) and Artificial Neural Networks (ANNs) for root-finding, since they are known, respectively, for their ability to explore high-dimensional spaces (not requiring good initial approximations) and for their capability to model complex problems. Besides that, both methods do not need repeated deflations, nor derivative information. The algorithms were described throughout this document and tested using a test suite of hard numerical problems in science and engineering. Results, in turn, were compared with several results available on the literature and with the well-known Durand–Kerner method, depicting that both algorithms are effective to solve the numerical problems considered.
A Optimização é um ramo da matemática desenvolvido para encontrar as soluções óptimas, de entre todas as possíveis, para um determinado problema. Actualmente, são várias as técnicas de optimização aplicadas a problemas de engenharia, de informática e da indústria. Dada a grande panóplia de aplicações, existem inúmeros trabalhos publicados que propõem métodos para resolver, de forma óptima, problemas específicos. Esta dissertação foca-se na adaptação de dois algoritmos inspirados na natureza que, tendo como base técnicas de optimização, são capazes de calcular aproximações para zeros de polinómios e raízes de equações não lineares e sistemas de equações não lineares. Embora já existam muitos métodos iterativos para encontrar todas as raízes ou zeros de uma função, eles usualmente exigem: (a) deflações repetidas, que podem levar a resultados muito inexactos, devido ao problema da acumulação de erros de arredondamento a cada iteração; (b) boas aproximações iniciais para as raízes para o algoritmo convergir, ou (c) o cálculo de derivadas de primeira ou de segunda ordem que, além de ser computacionalmente intensivo, para muitas funções é impossível de se calcular. Estas desvantagens motivaram o uso da Optimização por Enxame de Partículas (PSO) e de Redes Neurais Artificiais (RNAs) para o cálculo de raízes. Estas técnicas são conhecidas, respectivamente, pela sua capacidade de explorar espaços de dimensão superior (não exigindo boas aproximações iniciais) e pela sua capacidade de modelar problemas complexos. Além disto, tais técnicas não necessitam de deflações repetidas, nem do cálculo de derivadas. Ao longo deste documento, os algoritmos são descritos e testados, usando um conjunto de problemas numéricos com aplicações nas ciências e na engenharia. Os resultados foram comparados com outros disponíveis na literatura e com o método de Durand–Kerner, e sugerem que ambos os algoritmos são capazes de resolver os problemas numéricos considerados.

During the past 10 years Computational Statistics techniques' importance has been always increasing and it is actually still doing. This is due to more than one factor. On one side, the technological growth of the last decades has made computational power and storage capacity incredibly affordable and accessible. In particular, the storage capacity has improved much more than the computational resources. This resulted in an unbalanced situation in which there are huge masses of data but there is not (and, probably, there won't be for a long time) enough computing power to exhaustively process it. On one side, it is always more widely felt the need of extracting knowledge from databases, being such phase considered as a critical activity in many decision making processes. Such databases are actually considered as great potential knowledge repositories in a dormant state. Computational Statistics and, more properly, Data Mining techniques are, probably, the unique way to extract such knowledge from the forementioned raw data sources and this explains the always growing interest around such methodologies. This has led Data Mining to meet other research fields like machine learning, operation research and, more in general artificial intelligence. One of the actually biggest problem that many Data Mining techniques have to deal with is combinatorial optimization that, in the past, has led many techniques to be taken apart and, now, makes them applicable only within certain bounds. Since other research fields like Artificial Intelligence have being (and still are) dealing with such problems, their contribute to statistics (and viceversa) has been very significative. An indicator of this phenomenon could be the introduction of Neural Networks in Statistics. This thesis tries to go in this direction, in particularly about the use of Nature inspired optimization algorithms, which have been proven to be powerful instruments for attacking many combinatorial optimization problems, when exhaustive enumeration and evaluation of all possible candidate solution to a problem is not computationally affordable. In particular, Nature inspired algorithms have been designed for attacking the combinatorial optimization problems that affects Classification and Regression Trees algorithms, which are considered powerful instruments for knowledge extraction and decision making support. A Java software for quickly building Classification and Regression Trees (by using a very fast procedure) has also been written to fulfill the need for a flexible framework to support the research for developing the nature inspired algorithms. A Forward Search-based methodology has also been proposed to improve the stability of the Trees.

A vast majority of the Internet of Things (IoT) devices will be connected in a topology where the edge-devices push data to a local gateway, which forwards the data to a cloud for further processing. In sizeable outdoor deployment regions, the edge-devices may experience poor connectivity due to their distant locations and limited transmission power. Repeaters or relays must be placed at a few locations to ensure reliable connectivity to either a gateway or another node in the network. A big challenge in achieving reliable connectivity and coverage is the outdoor propagation environment being heterogeneous. Engineers often deploy networks based on resource-intensive field visits, detailed surveys, measurements, initial test deployments, followed by fine-tuning. For scalability to large scale IoT deployments, automated network planning tools are essential. Such tools should predict connectivity based on the edge-device locations using available Geographical Information System (GIS) data, identify the need for relays/repeaters, and, if needed, suggest the number of relays needed with their locations. Furthermore, such tools should also be extended to suggest the minimum number and locations of base stations to maximise coverage. In this thesis, we discuss coverage estimation procedure for the deployment of out- door Internet of Things (IoT). In the first part of the thesis, a data-driven coverage estimation technique is proposed. The estimation technique combines multiple machine- learning-based regression ideas. The proposed technique achieves two purposes. The first purpose is to reduce the bias in the estimated received signal strength arising from estimations performed only on the successfully received packets. The second purpose is to exploit commonality of physical parameters, e.g. antenna-gain, in measurements that are made across multiple propagation environments. It also provides the correct link function for performing a nonlinear regression in our communication systems context. Next, a method to use readily available geographic information system (GIS) data (for classifying geographic areas into various propagation environments) followed by an algorithm for estimating received signal strength (which is motivated by the initial section of the thesis) is proposed. Together they enable quick and automated estimation of coverage in outdoor environments. In the second part of the thesis, we propose an automated network deployment frame- work. Our proposed methodology uses either Ant Colony Optimisation (ACO) or Differential Evolution (DE) to identify the number and locations of relays for meeting specified quality of service constraints using a black box, received signal strength estimation oracle, that provides signal strength estimates between candidate pairs of transceiver locations in a heterogeneous deployment region. We discuss adaptations of our techniques to handle scenarios with multiple gateways. Further, we show the effectiveness of these algorithms to find suitable candidate base station locations to provide coverage in a heterogeneous propagation environment that meets the specified quality of service constraints. We then demonstrate the effectiveness of our algorithms in two deployment regions. It is anticipated that these will lead to faster and more efficient deployment of outdoor Internet of Things.

A dissertation submitted in fulfilment of the requirements for the degree of Master of Science to the School of Electrical and Information Engineering, 2020
Inertial stabilization systems are essential in ensuring that the optical system tracks the target of interest and rejects any disturbance. The work presented in this dissertation focuses on optimizing the Proportional-Integral-Derivative (PID) controller in order to ensure that the gimbal used in the chosen inertial stabilization system follows the Line-of-Sight (LOS) rate command input (which represents the target velocity) and rejects disturbance. The main objective of this research was to compare which optimization methods for tuning the PID controller work best for the one-axis gimbal stabilization system. The methods compared are three nature-inspired meta-heuristic algorithms; the Teaching Learning Based Optimization (TLBO) algorithm, the Flower Pollination Algorithm (FPA) and the Genetic Algorithm (GA). This research also involved tuning the parameters of the algorithms themselves in order for the algorithms to optimize the controller. This work also encompasses tuning the common algorithm parameters including the population size and search space bounds, tuning algorithm-specific parameters for each algorithm that requires this, and comparing whether dynamic or static parameters are better suited for the problem instances presented. These parameters were optimized for three different problem instances, which represent different target motions and additional disturbances in the system. It was found that different parameters work best for different problem instances and that this research favoured the TLBO when comparing the algorithm performances overall
CK2021

This work explores the technical challenges that emerge when applying bio-inspired optimisation methods to real-world engineering problems. A number of new heuristic algorithms were proposed and tested to deal with these challenges. The work is divided into three main dimensions: i) One of the most significant industrial optimisation problems is optimising renewable energy systems. Ocean wave energy is a promising technology for helping to meet future growth in global energy demand. However, the current technologies of wave energy converters (WECs) are not fully developed because of technical engineering and design challenges. This work proposes new hybrid heuristics consisting of cooperative coevolutionary frameworks and neuro-surrogate optimisation methods for optimising WECs problem in three domains, including position, control parameters, and geometric parameters. Our problem-specific algorithms perform better than existing approaches in terms of higher quality results and the speed of convergence. ii) The second part applies search methods to the optimization of energy output in wind farms. Wind energy has key advantages in terms of technological maturity, cost, and life-cycle greenhouse gas emissions. However, designing an accurate local wind speed and power prediction is challenging. We propose two models for wind speed and power forecasting for two wind farms located in Sweden and the Baltic Sea by a combination of recurrent neural networks and evolutionary search algorithms. The proposed models are superior to other applied machine learning methods. iii) Finally, we investigate the design of water distribution systems (WDS) as another challenging real-world optimisation problem. WDS optimisation is demanding because it has a high-dimensional discrete search space and complex constraints. A hybrid evolutionary algorithm is suggested for minimising the cost of various water distribution networks and for speeding up the convergence rate of search.
Thesis (Ph.D.) -- University of Adelaide, School of Computer Science, 2020

碩士
國立金門大學
電子工程學系碩士班
106
In this study, two techniques "software algorithm" and "hardware reconfiguration circuit" were separately proposed to improve the power output of high concentration photovoltaic (HCPV) systems under partial shading conditions (PSC). The maximum power point (MPP) of an HCPV system varies rapidly under changing environmental conditions. Therefore, a "software algorithm" maximum power point tracking (MPPT) with rapid response is needed. However, the power-voltage (P-V) curve of a solar system under PSC exhibits multi-peaks of each interval. As a result, although the tracking speed of conventional MPPT algorithms is rapid, they usually couldn’t accurately track the global maximum power point (GMPP) under PSC. To overcome this issue, researchers have proposed using various natural-inspired algorithms for MPPT, such as genetic algorithm (GA) etc. Although GA could accurately track GMPP under PSC, it has complicated calculation and slow tracking speed. In order to further improve the tracking speed of GA with non-decreasing accuracy, a novel modified genetic algorithm (MGA) is proposed in this study. The proposed MGA integrates the calculation processes of conventional firefly algorithm (FA) and difference evolution (DE) algorithm for MPPT of HCPV systems under PSC. In the second part of this thesis, the strategy of using "hardware reconfiguration circuit" to improve total output power of an HCPV system is studied. Conventional reconfiguration circuit methods rearrange serial and parallel connections of a solar array under PSC which could increase the total output power of the system. However, they need complicated switching control algorithms and large number of switches and sensors. In order to improve this strategy and overcome its disadvantages, a novel modified circuit reconfiguration (MCR) method is proposed in this research. The total-cross-tied (TCT) topology is adopted in the MCR to simplify its switching control algorithm and reduce the number of switches and sensors. The performance of the two proposed techniques were first simulated and then evaluated by practical hardware. In the first phase, an HCPV circuit model and various PSC patterns with different solar radiance were established using the M-file and Simulink toolbox of Matlab software for preliminary simulation and parameters fine tuning. In this second phase, the hardware evaluation experiments of the proposed MGA were conducted using a solar I-V curve simulator, while actual LED light sources and HCPV modules were adopted for MCR evaluation. Evaluation results demonstrate that when compared with conventional GA, the execution time and tracking accuracy of the proposed MGA could improve around 69.4% and 4.16%, respectively. In addition, when compared with conventional FA, the execution time and tracking accuracy could improve around 42.9% and 1.85%, respectively. On the other hand, when compared with conventional series-connection and TCT methods, the total output power of an HCPV system using the proposed MCR method could increase around 29.12% and 40.11%, respectively. The advantage of the proposed MGA is fast tracking speed with high accuracy, while the MCR method can successfully effectively increase output power with simplified control algorithm and reduced hardware cost. Both methods can be implemented not only to HCPV systems but also different solar systems. In addition, the proposed prototype architecture is capable of being extended to larger scale solar arrays.