The success of fuzzy application to solve the control problems depends on a number of parameters, such as fuzzy rules and membership functions. The problem of generating desirable fuzzy rules is very important in the ...
详细信息
ISBN:
(纸本)9781424465880
The success of fuzzy application to solve the control problems depends on a number of parameters, such as fuzzy rules and membership functions. The problem of generating desirable fuzzy rules is very important in the development of fuzzy systems, which are usually decided upon subjectively. This paper describes a very simple and straightforward fuzzy rule generation and optimization technique by using the particle swarm optimization algorithm (PSO). The proposed algorithm can obtain a set of fuzzy rules which cover the examples set in iterative process. The proposed method is tested with promising results.
In order to solve the problem of linearization, complexity and poor accuracy for parameter estimate of Muskingum Routing Model at present, this paper introduces three modern intelligent algorithms - Genetic algorithm ...
详细信息
ISBN:
(纸本)9781424469284
In order to solve the problem of linearization, complexity and poor accuracy for parameter estimate of Muskingum Routing Model at present, this paper introduces three modern intelligent algorithms - Genetic algorithm (GA), Simulated Annealing algorithm (SA) and particle swarm optimization algorithm (PSO) for the parameter calibration of Muskingum model. Through specific simulation, the results of five methods are produced. Then according to the calculation, comparison and analysis of five methods comprehensively, it is found that the results of three modern intelligent algorithms are fit significantly and better than traditional methods.
In this paper, an artificial neural network (ANN) based on hybrid algorithm combining particleswarmoptimization (PSO) with back-propagation (BP) is proposed to forecast the daily streamflows in a catchment located i...
详细信息
In this paper, an artificial neural network (ANN) based on hybrid algorithm combining particleswarmoptimization (PSO) with back-propagation (BP) is proposed to forecast the daily streamflows in a catchment located in a semi-arid region in Morocco. The PSO algorithm has a rapid convergence during the initial stages of a global search, while the BP algorithm can achieve faster convergent speed around the global optimum. By combining the PSO with the BP, the hybrid algorithm referred to as BP-PSO algorithm is presented in this paper. To evaluate the performance of the hybrid algorithm, BP neural network is also involved for a comparison purposes. The results show that the neural network model evolved by PSO-BP algorithm has a good predictions and better convergence performances.
In order to overcome the inherent deficiency in particle swarm optimization algorithm such as premature convergence, this paper presents crossover operator and mutation operator to improve particleswarmoptimization ...
详细信息
ISBN:
(纸本)9783642163876
In order to overcome the inherent deficiency in particle swarm optimization algorithm such as premature convergence, this paper presents crossover operator and mutation operator to improve particle swarm optimization algorithm, which is called IPSO. At the meantime, a new hybrid algorithm model is presented, which combine improved PSO algorithm and simulated annealing (SA) algorithm. The experimental results show that the proposed algorithm can reach the goal completely and the speed of convergence was greatly fast for optimization of Sphere, Griewank and Rastrigrin functions. The stability and robustness of proposed algorithm have been enhanced greatly. Its performance is superior to the standard PSO obviously.
The job shop scheduling problem is a well-known NI) hard problem, on which genetic algorithm is widely used However. due to the lack of the major evolution direction. the effectiveness of the regular genetic algorithm...
详细信息
ISBN:
(数字)9783642134951
ISBN:
(纸本)9783642134944
The job shop scheduling problem is a well-known NI) hard problem, on which genetic algorithm is widely used However. due to the lack of the major evolution direction. the effectiveness of the regular genetic algorithm is restricted In this paper. we propose a new hybrid genetic algorithm to solve the job shop scheduling problem The particle swarm optimization algorithm is introduced to get the initial population, and evolutionary genetic operations are proposed We validate the new method on seven benchmark datasets. and the comparisons with some existing methods verify as effectiveness
Fault diagnosis of electronic circuit is important for safety of the device and relevant power system. In the study, support vector regression (SVR) classifiers combined with the particle swarm optimization algorithm ...
详细信息
ISBN:
(纸本)9781424451944
Fault diagnosis of electronic circuit is important for safety of the device and relevant power system. In the study, support vector regression (SVR) classifiers combined with the particle swarm optimization algorithm (POSA) are applied to construct diagnostic model of electronic circuit, and the diagnostic system structure of electronic circuit is presented on the basis of the model. It is powerful for the practical problem with small sampling, nonlinear and high dimension, which is very suitable for online fault diagnosis. Utilizing the character that principal components analysis algorithm can keep the discernability of original dataset after reduction, reduce of the original dataset is calculated and used to train individual SVR for ensemble, and consequently, increase the detection accuracy. The test results show that the proposed method is a promised method owning to its high diversity, high detection accuracy and faster speed in fault diagnosis. The experimental result shows that this fault detection method is feasible and effective.
We summarized recent research aimed at expanding the context of facility location decisions to incorporate additional features of a supply chain including variable construction cost, inventory management, transportati...
详细信息
ISBN:
(纸本)9781424473281
We summarized recent research aimed at expanding the context of facility location decisions to incorporate additional features of a supply chain including variable construction cost, inventory management, transportation cost, etc.. Authors expended location model of risk pooling with variable construstion cost (LMRPVCC) to construct location model of risking pool with variable construction cost with compensation policy. In a buyer market, logistics corporations may apply compensation policy in order to hold more customers by giving a discount transportation charge. Authors build a square nonlinear 0-1 integer-programming model and use particle swarm optimization algorithm to find suboptimum solutions. The numerical examples are given separately along with the models to evaluate the effectiveness of the model. According to the computational results, we can draw these conclusions: transport costs factor beta is positively related with the objective function value. Compensation cost factor W is positively related with the objective function value. Distribution radius D-r is negatively related with the objective function value.
The particleswarmoptimization (PSO) algorithm has been successfully applied to dynamic optimization problems with very competitive results. One of its best performing variants, the mQSO is based on an atomic model, ...
详细信息
ISBN:
(纸本)9781424481262
The particleswarmoptimization (PSO) algorithm has been successfully applied to dynamic optimization problems with very competitive results. One of its best performing variants, the mQSO is based on an atomic model, with quantum and trajectory particles. This work introduces a new version of this algorithm which uses heuristic rules for improving its performance. Two new rules are presented: one specifically designed for the mQSO, which locally bursts diversity after a change in the environment, and a second, more general one, which globally increases diversity in a precise way, without disturbing the intensification of the search. The new version with rules is tested against the original one using several variations of the Moving Peaks Benchmark and the Ackley function. The results show a drastic improvement in the performance of the algorithm.
In fermentation process, fuzzy neural networks (FNN) is a novel machine learning method of soft sensor modeling, while the typical algorithm of FNN is inefficient because they can not optimize fuzzy rules and has long...
详细信息
ISBN:
(纸本)9781424451944
In fermentation process, fuzzy neural networks (FNN) is a novel machine learning method of soft sensor modeling, while the typical algorithm of FNN is inefficient because they can not optimize fuzzy rules and has long training time. Biological parameters can be measured online in real time which is helpful for the control of process optimization. So this paper introduces the use of the particleswarmoptimization (PSO) for training FNN. Unlike the conventional back-propagation technique, the adaptation of the weights of the FNN approximator is done on-line using PSO. The PSO is based on the least squares error minimization with random initial condition and without any off-line pre-training. Experiment results show that, in contrast to the traditional fuzzy neural networks, the method has good prediction and is suitable to practical applications.
In this paper, we introduce optimization methods of Polynomial Radial Basis Function Neural Network (pRBFNN). The connection weight of proposed pRBFNN is represented as four kinds of polynomials, unlike in most conven...
详细信息
ISBN:
(纸本)9783642132773
In this paper, we introduce optimization methods of Polynomial Radial Basis Function Neural Network (pRBFNN). The connection weight of proposed pRBFNN is represented as four kinds of polynomials, unlike in most conventional RBFNN constructed with constant as connection weight. Input space in partitioned with the aid of kernel functions and each kernel function is used Gaussian type. Least Square Estimation (LSE) is used to estimate the coefficients of polynomial. Also, in order to design the optimized pRBFNN model, center value of each kernel function is determined based on C-Means clustering algorithm, the width of the RBF, the polynomial type in the each node, input variables are identified through particleswarmoptimization (PSO) algorithm. The performances of the NOx emission process of gas turbine power plant data and Automobile Miles per Gallon (MPG) data was applied to evaluate proposed model. We analyzed approximation and generalization of model.
暂无评论