Дисертації з теми "REAL CODED GENETIC ALGORITHM (RCGA)"

The Development scenario for renewable energy across the globe is changing rapidly in terms of capacity addition and grid interconnection. Penetration of renewable energy resources into grid is necessary to meet the elevated demand of electricity. In view of this penetration of solar and wind power growing enormously across the globe. Solar energy is widely escalating in terms of generation and capacity addition due its better predictability over wind energy. Electricity pricing is one of the important aspects for power system planning and it felicitates information for the electricity bidder for exact electricity generation and resource allocation. The important task is to forecast the electricity price accurately in grid interactive environment. This task is tedious in renewable integrated market due to intermittency issue. As renewable energy penetration into the grid is enhancing swiftly. An appropriate market model addressing the issues of related to renewable energy specially wind and solar is necessary. A novel solar energy-based market model is proposed for state level market along with the operating mechanism. The different component associated with grid and their functionality in the operation of grid is discussed. Challenges and possible solutions are addressed to implement the market model. Energy trading plays a crucial role in the economic growth of country. Renewable energy trading opens a new avenue for the economic growth. India is blessed with a rich solar energy resources, the solar power producers tapped the potential of solar up to appreciable extent, but due to lack of trading models and specific regulatory mechanism in context of renewable energy generation is main hurdle in competition among generators. Various market model developed for solar energy trading at state level electricity along with their trading mechanism is presented. Also features of the models are also addressed. xiii A Rigorous literature review on price forecasting is conducted with focus on impact of solar and wind energy on electricity price. The data of Australia electricity market is collected for price forecasting. The correlation among the inputs for price is calculated using correlation coefficient formula and selected the highly corelated input with price. Artificial Neural Network (ANN) is implemented to forecast the price by using historical data. The price is predicted for January to June month and weekly forecast of price for the same month is executed. The minimum MAPE is 1.94 for April month and 1.03 for third week of January. The research work is continued to investigate the impact of solar and wind energy on electricity price. The Long short-term memory (LSTM) is designed to forecast the electricity price considering the solar power penetration. The raw data of Austria market consists of actual day ahead load, forecasted day ahead load, actual day ahead price and actual solar generation is used. The reliability of forecasting model is analyzed by computation of confidence interval on MAPE. The research work is extended to investigate the impact of wind energy on electricity price. The Austria electricity market data is used for investigating the potential impact of wind energy on rice. The statistical analysis of the data is conducted for finding the suitability of the model. Decision tree model is designed and implemented and significant reduction in the forecasting accuracy of 5.802 is achieved for the data set using wind energy as input parameter. The future of solar energy in India is positive. The growth of solar energy in terms of capacity addition and grid interconnection programme is expanding day by day. To promote the solar energy trading in open market a suitable bidding mechanism must be designed for solar power producers. It becomes pertinent to design the bidding strategy for solar power producers to maximize their profit considering the uncertainty in the energy output. Hybrid Particle Swarm Optimization – Gravitational Search Algorithm (HPSO - GSA) is proposed for designing the optimal bidding strategy for solar PV power producer for designed solar energy xiv based Indian electricity market. The objective function is designed considering the constraint of uncertainty and energy imbalance in price. The proposed algorithm shows highest profit when compared with Real Coded Genetic Algorithm (RCGA), Particle Swarm Optimization (PSO) and Gravitational Search Algorithm (GSA). In the light of continual renewable energy growth and grid interconnection, a novel solar energy-based electricity market model addressing the issues of solar energy is proposed to make the system effective and reliable. This novel market may fill the promise of providing electricity at competitive cost for all in India. The various market models are proposed for trading the solar energy in competitive market for maximum utilization of untapped potential of solar energy. The various trading models may be implemented based on the application and suitability. The electricity price forecasting is an important aspects of power system planning and for renewable energy interactive grid price forecasting is crucial task due its intermittent nature. ANN model is proposed for price forecasting and significant improvement in MAPE is reported for Australia electricity market data. Further the investigation has been done on the impact of solar energy generation on electricity price using machine learning techniques (DT, RF, LASSO, XGBOOST and LSTM). The LSTM model accuracy is good in price forecasting with consideration of solar energy as input parameter. The investigation is extended for impact of wind energy on electricity price and Decision tree model accuracy is superior as compared to RF, LASSO, LR, SVR and DNN model. The bidding strategy for the designed solar based electricity model is proposed using HPSO-GSA method and profit calculation has been done for solar PV producers on real time data. The maximized profit has been obtained through HSPO-GSA method for two different sets of datasets.

This thesis successfully addresses the issues faced by particle filters in high-dimensional state-spaces by comparing them with genetic algorithms and then using genetic algorithm theory to address these issues. Sequential Monte Carlo methods are a class of online posterior density estimation algorithms that are suitable for non-Gaussian and nonlinear environments, however they are known to suffer from particle degeneracy; where the sample of particles becomes too sparse to approximate the posterior accurately. Various techniques have been proposed to address this issue but these techniques fail in high-dimensions. In this thesis, after a careful comparison between genetic algorithms and particle filters, we posit that genetic algorithm theoretic arguments can be used to explain the working of particle filters. Analysing the working of a particle filter, we note that it is designed similar to a genetic algorithm but does not include recombination. We argue based on the building-block hypothesis that the addition of a recombination operator would be able to address the sample impoverishment phenomenon in higher dimensions. We propose a novel real-coded genetic algorithm particle filter (RGAPF) based on these observations and test our hypothesis on the stochastic volatility estimation of financial stocks. The RGAPF successfully scales to higher-dimensions. To further strengthen our argument that whether building-block-hypothesis-like effects are due to the recombination operator, we compare the RGAPF with a mutation-only particle filter with an adjustable mutation rate that is set to equal the population-to-population variance of the RGAPF. The latter significantly and consistently performs better, indicating that recombination is having a subtle and significant effect that may be theoretically explained by genetic algorithm theory. After two successful attempts at validating our hypothesis we compare the performance of the RGAPF using different real-recombination operators. Observing the behaviour of the RGAPF under these recombination operators we propose a mean-centric recombination operator specifically for high-dimensional particle filtering. This recombination operator is successfully tested and compared with benchmark particle filters and a hybrid CMA-ES particle filter using simulated data and finally on real end-of-day data of the securities making up the FTSE-100 index. Each experiment is discussed in detail and we conclude with a brief description of the future direction of research.

This thesis is concerned with the design optimization of single-storey steel portal frame buildings. In the UK, such buildings account for 90% of all single-storey buildings and 50% of all constructional steelwork used. Two different types of steel portal frames are considered: conventional hot-rolled steel portal frames, which can achieve spans of up to 50 m, and cold-formed steel frames, which while less popular in the UK, may be more efficient for spans around 12 m. A real-coded niching genetic algorithm is used for the purposes of the design optimization. By adopting a niching strategy, the diversity of the population is effectively maintained and so increases the probability in searching for the optimum solution in the design space. The efficiency of the real-coded niching genetic algorithm is demonstrated through design examples of both hot-rolled steel and cold-formed steel portal frames. For the design optimization of hot-rolled steel portal frames, the optimization algorithm is used to minimize the material cost of the portal frame, per square m on plan, taking into account both the hot-rolled steel cross-sections and the eaves haunch size. In all cases, a frame spacing of 6 m is adopted. Both ultimate and serviceability limit states are considered, with deflection limits recommended by the Steel Construction Institute. It is shown that serviceability deflections govern for the design of a 50 m span portal frame, where material costs increase by 60% compared to an ultimate limit state design only. For small span frame, i.e., span of 10m, material cost only increases by 19%. For the design optimization of the cold-formed steel portal frame, the same algorithm is applied to minimize the material cost of the main frame members. In addition, frame spacings of both 4 m and 6 m are considered. For the case of a 12 m span frame, with rigid joints, it is shown that the frame design is not sensitive to serviceability deflections and that the frame is 24% cheaper (in terms of material costs per square m) than using hot-rolled steel. When the effects of semi-rigid joints and stressed-skin action are included, it is shown that the cost of members is further reduced by 32%.

Simultaneous lot-sizing and scheduling problem is the problem to decide what products to be produced on which machine and in which order, as well as the quantity of each product. Problems of this type are hard to solve. Therefore, they were studied for years, and a considerable number of papers is published to solve different lotsizing and scheduling problems, specifically real-case problems. This work proposes a Real-Coded Genetic Algorithm (RCGA) with a new chromosome representation to solve a non-identical parallel machine capacitated lot-sizing and scheduling problem with sequence dependent setup times and costs, machine cost and backlogging. Such a problem can be found in real world production line at furniture manufacturer in Sweden. Backlogging is an important concept in this problem, and it is often ignored in the literature. This study implements three different types of crossover; one of them has been chosen based on numerical experiments. Four mutation operators have been combined together to allow the genetic algorithm to scan the search area and maintain genetic diversity. Other steps like initializing of the population and a reinitializing process have been designed carefully to achieve the best performance and to prevent the algorithm from trapped into the local optimum. The proposed algorithm is implemented and coded in MATLAB and tested for a set of standard medium to large-size problems taken from the literature. A variety of problems were solved to measure the impact of different characteristics of problems such as the number of periods, machines, and products on the quality of the solution provided by the proposed RCGA. To evaluate the performance of the proposed algorithm, the average deviation from the lower bound and runtime for the proposed RCGA are compared with three other algorithms from the literature. The results show that, in addition to its high computational speed, the proposed RCGA outperforms the other algorithms for non-identical parallel machine problems, while it is outperformed by the other algorithms for problems with the more identical parallel machine. The results show that the different characteristics of problem instances, like increasing setup cost, and size of the problem influence the quality of the solutions provided by the proposed RCGA negatively.

碩士
國立高雄第一科技大學
電機工程研究所碩士班
102
In this paper, the main objective is to identify the parameters of motors, which includes a BLDC motor and an induction motor. The real-coded genetic algorithm (RGA) is adopted to identify all parameters of motors, and the standard genetic algorithm (SRGA) and adaptive genetic algorithm (ARGA) are compared in the rotational angular speeds and fitness values, which are the inverse of square difference of angular speeds. From numerical simulations and experimental results, it is found that the SRGA and ARGA are feasible, the ARGA can effectively solve the problems of slow convergent speed and premature phenomenon, and the ARGA is more accurate in identifying system’s parameters than the SRGA. From the comparisons of ARGAs in identifying parameters of motors, the best ARGA method is obtained and could be applied to any other areas of expertise.

The management of Water Distribution Systems (WDSs) involves making decisions about various operations in the network, including the scheduling of pump operations and setting of disinfectant dosing rates. There are often conflicting objectives in making these operational decisions, such as minimising costs while maximising the quality of the water supplied. Hence, the operation of WDSs can be very difficult, and there is generally considerable scope to improve the operational efficiency of these systems by improving the associated decision making process. In order to achieve this goal, optimisation methods known as Genetic Algorithms (GAs) have been successfully adopted to assist in determining the best possible solutions to WDS optimisation problems for a number of years. Even though there has been extensive research demonstrating the potential of GAs for improving the design and operation of WDSs, the method has not been widely adopted in practice. There are a number of reasons that may contribute to this lack of uptake, including the following difficulties: (a) developing an appropriate fitness function that is a suitable description of the objective of the optimisation including all constraints, (b) making decisions that are required to select the most appropriate variant of the algorithm, (c) determining the most appropriate parameter settings for the algorithm, and (d) a reluctance of WDS operators to accept new methods and approaches. While these are all important considerations, the correct selection of GA parameter values is addressed in this thesis. Common parameters include population size, probability of crossover, and probability of mutation. Generally, the most suitable GA parameters must be found for each individual optimisation problem, and therefore it might be expected that the best parameter values would be related to the characteristics of the associated fitness function. The result from the work undertaken in this thesis is a complete GA calibration methodology, based on the characteristics of the optimisation problem. The only input required by the user is the time available before a solution is required, which is beneficial in the WDS operation optimisation application considered, as well as many others where computationally demanding model simulations are required. Two methodologies are proposed and evaluated in this thesis, one that considers the selection pressure based on the characteristics of the fitness function, and another that is derived from the time to convergence based on genetic drift, and therefore does not require any information about the fitness function characteristics. The proposed methodologies have been compared against other GA calibration methodologies that have been proposed, as well as typical parameter values to determine the most suitable method to determine the GA parameter values. A suite of test functions has been used for the comparison, including 20 complex mathematical optimisation problems with different characteristics, as well as realistic WDS applications. Two WDS applications have been considered: one that has previously been optimised in the literature, the Cherry Hills-Brushy Plains network; and a real case study located in Sydney, Australia. The optimisation problem for the latter case study is to minimise the pumping costs involved in operating the WDS, subject to constraints on the system, including minimum disinfectant concentrations. Of the GA calibration methods compared, the proposed calibration methodology that considered selection pressure determined the best solution to the problem, producing a 30% reduction in the electricity costs for the water utility operating the WDS. The comparison of the different calibration approaches demonstrates three main results: 1. that the proposed methodology produced the best results out of the different GA calibration methods compared; 2. that the proposed methodology can be applied in practice; and 3. that a correctly calibrated GA is very beneficial when solutions are required in a limited timeframe.
http://proxy.library.adelaide.edu.au/login?url= http://library.adelaide.edu.au/cgi-bin/Pwebrecon.cgi?BBID=1325448
Thesis (Ph.D.) - University of Adelaide, School of Civil, Environmental and Mining Engineering, 2008

The management of Water Distribution Systems (WDSs) involves making decisions about various operations in the network, including the scheduling of pump operations and setting of disinfectant dosing rates. There are often conflicting objectives in making these operational decisions, such as minimising costs while maximising the quality of the water supplied. Hence, the operation of WDSs can be very difficult, and there is generally considerable scope to improve the operational efficiency of these systems by improving the associated decision making process. In order to achieve this goal, optimisation methods known as Genetic Algorithms (GAs) have been successfully adopted to assist in determining the best possible solutions to WDS optimisation problems for a number of years. Even though there has been extensive research demonstrating the potential of GAs for improving the design and operation of WDSs, the method has not been widely adopted in practice. There are a number of reasons that may contribute to this lack of uptake, including the following difficulties: (a) developing an appropriate fitness function that is a suitable description of the objective of the optimisation including all constraints, (b) making decisions that are required to select the most appropriate variant of the algorithm, (c) determining the most appropriate parameter settings for the algorithm, and (d) a reluctance of WDS operators to accept new methods and approaches. While these are all important considerations, the correct selection of GA parameter values is addressed in this thesis. Common parameters include population size, probability of crossover, and probability of mutation. Generally, the most suitable GA parameters must be found for each individual optimisation problem, and therefore it might be expected that the best parameter values would be related to the characteristics of the associated fitness function. The result from the work undertaken in this thesis is a complete GA calibration methodology, based on the characteristics of the optimisation problem. The only input required by the user is the time available before a solution is required, which is beneficial in the WDS operation optimisation application considered, as well as many others where computationally demanding model simulations are required. Two methodologies are proposed and evaluated in this thesis, one that considers the selection pressure based on the characteristics of the fitness function, and another that is derived from the time to convergence based on genetic drift, and therefore does not require any information about the fitness function characteristics. The proposed methodologies have been compared against other GA calibration methodologies that have been proposed, as well as typical parameter values to determine the most suitable method to determine the GA parameter values. A suite of test functions has been used for the comparison, including 20 complex mathematical optimisation problems with different characteristics, as well as realistic WDS applications. Two WDS applications have been considered: one that has previously been optimised in the literature, the Cherry Hills-Brushy Plains network; and a real case study located in Sydney, Australia. The optimisation problem for the latter case study is to minimise the pumping costs involved in operating the WDS, subject to constraints on the system, including minimum disinfectant concentrations. Of the GA calibration methods compared, the proposed calibration methodology that considered selection pressure determined the best solution to the problem, producing a 30% reduction in the electricity costs for the water utility operating the WDS. The comparison of the different calibration approaches demonstrates three main results: 1. that the proposed methodology produced the best results out of the different GA calibration methods compared; 2. that the proposed methodology can be applied in practice; and 3. that a correctly calibrated GA is very beneficial when solutions are required in a limited timeframe.
Thesis (Ph.D.) - University of Adelaide, School of Civil, Environmental and Mining Engineering, 2008

碩士
國立臺北科技大學
自動化科技研究所
100
Motion planning of redundant robot manipulators is a difficult problems encountered in the field of robotics. In this paper, we use Forward Kinematic and Optimal Control to avoid singularity problem in motion planning that use Inverse Kinematic to compute. It is a large quantity of computing in searching optimal result process, due to this reason, we use Genetic Algorithm and Particle Swarm Optimization Algorithm to reduce complexity and use parallel computing architecture of GPU(CUDA) to improve computing speed.

碩士
逢甲大學
化學工程學所
99
In this thesis, we consider the issue of robust controller design problem for interval time delay systems. The design is based on optimization of load disturbance rejection and weighting of set point response, a two degree of freedom control system architecture is introduced. We use real coded genetic algorithms (RCGA) to design the optimal two degree of freedom PI controller that not only stabilizes all of the interval time delay plant family but also minimizes the integral absolute error (IAE) of the worst-case plant. The controller design problem is a min-max optimization problem. Then, we can extend the technique proposed for the single loop control system to cascade control system. Numerical experiments indicate that proposed scheme is efficient. As an application, we applied the proposed design scheme to the problems of designing robust two degree of freedom PI controller for bioreactor and CSTR. The result of this proposed simulation was satisfactory.

碩士
國立成功大學
電機工程學系碩博士班
90
With the significant discovery of photosensitivity in optical fibers, a new class of in-fiber component has been developed, called the fiber Bragg grating (FBG). The fiber Bragg gratings have many advantages, such as low loss transmission, immunity to electromagnetic interference, easy fabrication, make the intro-core grating an ideal candidate for use in telecommunications and sensory field. A method of extracting the strain profile along a fiber Bragg grating for the known reflection spectrum is described. By combining the T-matrix analysis method for calculating the reflection spectrum together with a real-coded genetic algorithm, we obtain a promising method for the spectrum synthesis. The synthesis procedure is based on a real-coded genetic algorithm that relates to the non-uniform grating pitch associated with the loading strain field. The strain-optic effect in an optical fiber, therefore, is considered. Several examples of the synthesis strain profile in fiber Bragg gratings for the band-pass, power discriminator filters, and EDFA gain flatten filter are presented. Including the design variables in length of grating and difference in refractive indices, the accuracy of the matching spectrum could be improved.

碩士
國立高雄第一科技大學
電機工程研究所碩士班
102
The main objective of this paper is to identify the parameters of an induction motor, which is dynamically formulated by various mathematical models of the electromechanical system. In dynamic modelling, we make some assumptions to simplify the dynamic models. In system identification, we adopt the real-coded genetic algorithm (RGA) to find the parameters of various mathematical models of the induction motor. It can be successfully identified from the convergent fitness, and the identified parameters are compared and discussed. The RGA not only simplifies the measurement steps, but also quickly obtains the parameters.

碩士
國立成功大學
光電科學與工程研究所
92
In this thesis we will propose that the methods of arbitrary strain distribution sensing with real-code genetic algorithm to analyze the reflection spectrums of Fabry-Perot-like of fiber Bragg gratings. The arbitrary strain distribution along the fiber Bragg gratings is recovered inversely from the Fabry-Perot-like reflective spectrum using the genetic algorithm optimization process. 　　The proposed methods permit accurate strain reconstruction with no restrictions on the applied strain profile or on the grating length. We demonstrate the validity and accuracy of these techniques by reconstructing the strain distribution along the grating with nonmonotonic strain profile, which includes: a constant strain distribution, linearly gradient strain, and strain discontinuities.

博士
逢甲大學
化學工程學系
101
In this dissertation, a novel and efficient real-coded genetic algorithm (RGA) termed as RGA-RDD is developed for solving single-objective optimization problems. The RGA-RDD is equipped with three effective and distinct operators; they are the ranking selection (RS), direction-based crossover (DBX) and dynamic random mutation (DRM). The RS operator is used to eliminate the bad solutions and reproduce good solutions, making the whole population to achieve a better average fitness. For exploring better solutions, the DBX operator applies an adaptive scheme to direct the crossover toward a direction that facilitates a significant increase in fitness. The DRM operator can effectively prevent the premature convergence and at the same time increases the precision of the searched solution. Unlike conventional RGAs that are usually operated in a series manner, the proposed RGA-RDD constitutes a parallel loop for the crossover (DBX) and mutation (DRM) operations, which therefore greatly enhances the possibility of locating a global optimum. The effectiveness and applicability of the proposed RGA-RDD are demonstrated through a variety of benchmarked problems included the type of unconstrained and constrained real-parameter optimization. The performance of the RGA-RDD is further examined by a comparative study with several state-of-the-art evolutionary algorithms (EAs). Extensive simulation results reveal that the proposed RGA-RDD is effective in solving varied real-parameter optimization problems and that it provides much better performance than most comparative methods. As a specific application, we further apply the developed RGA-RDD to the mathematical modeling and optimal design of a horizontal metal-organic chemical vapor deposition (MOCVD) reactor. A detailed 3D model of the MOCVD reactor is developed, which includes a reaction kinetic mechanism to describe the gas-phase and surface reactions occurring in the reactor, and a comprehensive heat transfer scheme to express the heat transferring between the inner reactor walls and the outer tube cooling gas. To estimate the model parameter, a data-driven optimization technique which incorporates a uniform design technique, a neural network auxiliary model and the RGA-RDD is proposed. The presented 3D MOCVD model was verified to be in a good agreement with experimental data and has been shown to present an excellent ability to predict the GaAs film growth rate and uniformity. Besides, the comprehensive model enables the systematic investigation of the microscopic transfer phenomena and reaction dynamics that relate to the dimensionless groups such as the Reynolds number (Re), Prandtl number, Peclet number (Pe) and Grashof number (Gr). We found that the “cold finger” and buoyancy-driven transverse rolls can occur in the AIX 200/4 horizontal MOCVD reactor when Pe 20 and Gr/Re2 4000. To improve the film growth performance and satisfy quality demands, the proposed data-driven optimization technique is then applied to search for a set of operating conditions that optimizes the film growth rate distribution on substrate. The optimized results indicate that a significantly better GaAs film growth performance can be achieved for the horizontal MOCVD reactor based on using the comprehensive 3D mathematical model and the proposed data-driven optimization scheme.