The Experts below are selected from a list of 360 Experts worldwide ranked by ideXlab platform
V D Tsoukalas - One of the best experts on this subject based on the ideXlab platform.
-
optimization of porosity formation in alsi9cu3 pressure die castings using genetic algorithm analysis
Materials & Design, 2008Co-Authors: V D TsoukalasAbstract:Abstract In this investigation, an effective approach based on multivariable linear regression (MVLR) and genetic algorithm (GA) methods has been developed to determine the optimum conditions leading to minimum porosity in AlSi9Cu3 aluminium alloy die castings. Experiments were conducted by varying holding furnace temperature, die temperature, plunger velocities in the first and second stage, and multiplied pressure in the third stage using L27 Orthogonal Array of Taguchi method. The experimental results from the Orthogonal Array were used as the training data for the MVLR model to map the relationship between process parameters and porosity formation of the die cast parts. With the fitness function based on this model, genetic algorithms were used for the process conditions optimization. By comparing the predicted values with the experimental data, it was demonstrated that the proposed model is a useful and efficient method to find the optimal process conditions in pressure die casting associated with the minimum porosity percent.
-
optimization of porosity formation in alsi9cu3 pressure die castings using genetic algorithm analysis
Materials & Design, 2008Co-Authors: V D TsoukalasAbstract:In this investigation, an effective approach based on multivariable linear regression (MVLR) and genetic algorithm (GA) methods has been developed to determine the optimum conditions leading to minimum porosity in AlSi9Cu3 aluminium alloy die castings. Experiments were conducted by varying holding furnace temperature, die temperature, plunger velocities in the first and second stage, and multiplied pressure in the third stage using L27 Orthogonal Array of Taguchi method. The experimental results from the Orthogonal Array were used as the training data for the MVLR model to map the relationship between process parameters and porosity formation of the die cast parts. With the fitness function based on this model, genetic algorithms were used for the process conditions optimization. By comparing the predicted values with the experimental data, it was demonstrated that the proposed model is a useful and efficient method to find the optimal process conditions in pressure die casting associated with the minimum porosity percent.
Boxin Tang - One of the best experts on this subject based on the ideXlab platform.
-
Orthogonal Array-Based Latin Hypercubes
Journal of the American Statistical Association, 1993Co-Authors: Boxin TangAbstract:Abstract In this article, we use Orthogonal Arrays (OA's) to construct Latin hypercubes. Besides preserving the univariate stratification properties of Latin hypercubes, these strength r OA-based Latin hypercubes also stratify each r-dimensional margin. Therefore, such OA-based Latin hypercubes provide more suitable designs for computer experiments and numerical integration than do general Latin hypercubes. We prove that when used for integration, the sampling scheme with OA-based Latin hypercubes offers a substantial improvement over Latin hypercube sampling.
Manqing Dong - One of the best experts on this subject based on the ideXlab platform.
-
deep neural network hyperparameter optimization with Orthogonal Array tuning
International Conference on Neural Information Processing, 2019Co-Authors: Xiang Zhang, Xiaocong Chen, Lina Yao, Manqing DongAbstract:Deep learning algorithms have achieved excellent performance lately in a wide range of fields (e.g., computer version). However, a severe challenge faced by deep learning is the high dependency on hyper-parameters. The algorithm results may fluctuate dramatically under the different configuration of hyper-parameters. Addressing the above issue, this paper presents an efficient Orthogonal Array Tuning Method (OATM) for deep learning hyper-parameter tuning. We describe the OATM approach in five detailed steps and elaborate on it using two widely used deep neural network structures (Recurrent Neural Networks and Convolutional Neural Networks). The proposed method is compared to the state-of-the-art hyper-parameter tuning methods including manually (e.g., grid search and random search) and automatically (e.g., Bayesian Optimization) ones. The experiment results state that OATM can significantly save the tuning time compared to the state-of-the-art methods while preserving the satisfying performance.
-
deep neural network hyperparameter optimization with Orthogonal Array tuning
arXiv: Learning, 2019Co-Authors: Xiang Zhang, Xiaocong Chen, Lina Yao, Manqing DongAbstract:Deep learning algorithms have achieved excellent performance lately in a wide range of fields (e.g., computer version). However, a severe challenge faced by deep learning is the high dependency on hyper-parameters. The algorithm results may fluctuate dramatically under the different configuration of hyper-parameters. Addressing the above issue, this paper presents an efficient Orthogonal Array Tuning Method (OATM) for deep learning hyper-parameter tuning. We describe the OATM approach in five detailed steps and elaborate on it using two widely used deep neural network structures (Recurrent Neural Networks and Convolutional Neural Networks). The proposed method is compared to the state-of-the-art hyper-parameter tuning methods including manually (e.g., grid search and random search) and automatically (e.g., Bayesian Optimization) ones. The experiment results state that OATM can significantly save the tuning time compared to the state-of-the-art methods while preserving the satisfying performance. The codes are open in GitHub (this https URL)
Emine şule Yazici - One of the best experts on this subject based on the ideXlab platform.
-
estimates of the coverage of parameter space by latin hypercube and Orthogonal Array based sampling
Applied Mathematical Modelling, 2017Co-Authors: Diane Donovan, Kevin Burrage, Pamela Burrage, T A Mccourt, Bevan Thompson, Emine şule YaziciAbstract:Abstract In this paper we use counting arguments to prove that the expected percentage coverage of a d dimensional parameter space of size n when performing k trials with either Latin Hypercube sampling or Orthogonal Array-based Latin Hypercube sampling is the same. We then extend these results to an experimental design setting by projecting onto a t
Ying-pin Chang - One of the best experts on this subject based on the ideXlab platform.
-
optimal design of discrete value tilt angle of pv using sequential neural network approximation and Orthogonal Array
Expert Systems With Applications, 2009Co-Authors: Ying-pin ChangAbstract:This report presents a method for combining sequential neural-network approximation and Orthogonal Arrays (SNAOA) to determine the tilt angle for photovoltaic (PV) modules. An Orthogonal Array is first conducted to obtain the initial solution set. The set is then treated as the initial training sample. Next, a back-propagation sequential neural network is trained to simulate the feasible domain for seeking the optimal tilt angle of PV modules. The size of the training sample is greatly reduced due to the use of the Orthogonal Array. In addition, a restart strategy is also incorporated into the SNAOA so that the searching process may have a better opportunity to reach a near global optimum. The objective is to maximize the output power energy of the modules. In this study, seven Taiwanese areas were selected for analysis. The sun's position at any time and location was predicted by the mathematical procedure of Julian dating; then, the solar irradiation was obtained at each site under a clear sky. To confirm the computer simulation results, an experimental system is conducted for determining the optimal tilt angle of the PV modules. The results show that the annual optimal angle for the Taipei area is 23.25^o; for Taichung, 22.25^o; for Tainan, 21.25^o; for Kaosiung, 20.75^o; for Hengchung, 20.25^o; for Hualian, 22.25^o; and for Taitung, 21^o in Taiwan, and the actual best annual tilt angles are close to the computer simulation results. Additional results related to SNAOA are also reported and discussed as well.
-
an Orthogonal Array based particle swarm optimizer with nonlinear time varying evolution
Applied Mathematics and Computation, 2007Co-Authors: Ying-pin ChangAbstract:Particle swarm optimization (PSO) is a population-based heuristic optimization technique. It has been developed to be a prominent evolution algorithm due to its simplicity of implementation and ability to quickly converge to a reasonable solution. However, it has also been reported that the algorithm has a tendency to get stuck in a near-optimal solution in multi-dimensional spaces. To overcome the stagnation in searching a globally optimal solution, a PSO method with nonlinear time-varying evolution (PSO-NTVE) is proposed to approach the optimal solution closely. When determining the parameters in the proposed method, matrix experiments with an Orthogonal Array are utilized, in which a minimal number of experiments would have an effect that approximates the full factorial experiments. To demonstrate the performance of the proposed PSO-NTVE method, five well-known benchmarks are used for illustration. The results will show the feasibility and validity of the proposed method and its superiority over several previous PSO algorithms.
-
Optimal Design of Discrete-Value Passive Harmonic Filters Using Sequential Neural-Network Approximation and Orthogonal Array
IEEE Transactions on Power Delivery, 2007Co-Authors: Ying-pin Chang, Chinyao LowAbstract:This paper presents a method for combining sequential neural-network approximation and Orthogonal Arrays (SNAOA) in the planning of large-scale passive harmonic filters. An Orthogonal Array is first conducted to obtain the initial solution set. The set is then treated as the initial training sample. Next, a back-propagation sequential neural network is trained to simulate the feasible domain for seeking the optimal filter design. The size of the training sample is greatly reduced due to the use of the Orthogonal Array. In addition, a restart strategy is also incorporated into the SNAOA so that the searching process may have a better opportunity to reach a near global optimum. To illustrate the performance of the SNAOA, a practical harmonic mitigation problem in a chemical plant is studied. The results show that the SNAOA performs better than the original scheme and satisfies the harmonic limitations with respect to the objective of minimizing the total demand distortion of harmonic currents and total harmonic distortion of voltages. Filter loss, reactive power compensation, and the constraints of individual harmonics are also considered. Additional results related to SNAOA are also reported and discussed as well.
