Since it is difficult to find the optimal solution directly by the traditional CFD optimization method due to its strong dependence on the designer's experience, an automatic aerodynamic optimization design platfo...
详细信息
Since it is difficult to find the optimal solution directly by the traditional CFD optimization method due to its strong dependence on the designer's experience, an automatic aerodynamic optimization design platform for automotive shape was built based on mesh deformation technology, surrogate model and optimization algorithm in this paper. A parameterized model of an automotive was established. Latin hypercube method was adopted to select sample points. The drag coefficients corresponding to sample points were calculated by CFD simulation, whereby the influence of each parameter on drag coefficient was obtained. By comparing the calculation time, optimization effect and optimization accuracy of 9 combinations of surrogate models and optimization algorithms, the combination of RBF model and NLPQL algorithm was selected as the optimal one which is the most appropriate for the aerodynamic optimization design for automotive shape.
Aimed at the problem of the calculation value tends to distribute at both ends in the traditional chaos optimization algorithm, the paper proposes a method to improve the efficiency of the optimization problem by impr...
详细信息
ISBN:
(纸本)9781538683088
Aimed at the problem of the calculation value tends to distribute at both ends in the traditional chaos optimization algorithm, the paper proposes a method to improve the efficiency of the optimization problem by improving the Chebyshev mapping distribution characteristics through a homogenizing regulator. The homogenization regulator based on the probability density function in this method can improve the ergodic properties of Chebyshev mapping. This paper lists two classical examples to test the optimization algorithm based on homogenization regulator, and compares it with the optimization results based on traditional Tent mapping and Logistic mapping. The test results show that the algorithm significantly reduces the solution step length on the basis of maintaining accuracy. Therefore, this algorithm is feasible and effective in solving optimization problems.
In this paper, a novel Gravitational Artificial Bee Colony (GABC) optimization algorithm was proposed and utilized to the non-supervised pattern recognition problems. In this approach, the gravitational search strateg...
详细信息
ISBN:
(纸本)9781538682463
In this paper, a novel Gravitational Artificial Bee Colony (GABC) optimization algorithm was proposed and utilized to the non-supervised pattern recognition problems. In this approach, the gravitational search strategy was introduced into the artificial bee colony algorithm, and a gravitational bee colony was established. The gravitational bee could search the global optimal result under the influence of both gravitational force and colony cooperation, which makes the optimization process more effectively and efficiently. Based on GABC algorithm, an intelligent kernel clustering model was established, in which the clustering center and kernel parameters were combined to be the optimal variable, while the clustering index was used as the objective function. GABC was utilized to find the optimal result of the clustering model. The standard testing functions were used to test the proposed algorithm, and GABC showed high accuracy and convergence speed. Then the testing data and fault samples were utilized to test the performance of GABC based clustering model, and its superiority on effectiveness and efficiency was demonstrated.
Multiple numbers of Building Energy Simulation (BES) programs have been improved and implemented during the last decades. BES models play a crucial role in understanding building energy demands and accelerating the ma...
详细信息
Multiple numbers of Building Energy Simulation (BES) programs have been improved and implemented during the last decades. BES models play a crucial role in understanding building energy demands and accelerating the malfunction diagnosis. However, due to the very high number of interacting parameters, most of the developed energy simulation programs do not accurately predict building energy performance under a known condition. Even the energy models which are developed with the very precise assignment of parameters, there is always significant discrepancies between the simulation results and the real-time data measurements. Current study develops an optimization-based framework to calibrate the whole building energy model. The optimization algorithm attempts to set the identified parameters to minimize the error between the simulation results and the real-time measurements. Due to the high number of parameters, the developed optimization algorithm utilizes a Harmony Search algorithm as its search engine coupled with the energy simulation model to accelerate the calibration process. Moreover, to illustrate the efficiency of using the developed framework, a case study of the office building is modeled and calibrated and the statistical analysis was conducted to assess the accuracy of the results. The results of the calibration process show the reliability of the framework. (C) 2019 Elsevier B.V. All rights reserved.
Design optimization of moderately thick hexagonal honeycomb sandwich plate has been investigated via employing an improved multi-objective particle swarm optimization with genetic algorithm (MOPSOGA). Based on the fir...
详细信息
Design optimization of moderately thick hexagonal honeycomb sandwich plate has been investigated via employing an improved multi-objective particle swarm optimization with genetic algorithm (MOPSOGA). Based on the first-order shear deformation theory (FSDT), governing equations of the plate are obtained. The equations are solved analytically. Total weight and maximum deflection of the plate under static gravity loads are considered to be objective functions of the problem. Core height, faces thickness, cell walls thickness, vertical and inclined cell wall length and the angle between inclined cell wall and horizontal line are set to be design variables of the problem. The geometrical and failure constrains are chosen to have desirable performance and stability of the sandwich plate. In the used multi-objective optimization technique, the optimum velocity parameter, inertia weight and acceleration coefficients for next iteration of the MOPSO are obtained by employing the genetic algorithm via minimizing generational distance between the sets of dominated and non-dominated particles in the previous iteration. Efficiency and accuracy of the proposed solution procedure are demonstrated and effects of different parameters on design optimization of the plate are studied. Also, TOPSIS multi-criteria decision-making method has been selected to report appreciate results from the Pareto-front curve of the MOPSOGA.
Parameter estimation applied to gray-box modeling approaches is of great importance for advanced building control and demand response. Stochastic optimization algorithms based on the swarm intelligence, e.g. generic a...
详细信息
Parameter estimation applied to gray-box modeling approaches is of great importance for advanced building control and demand response. Stochastic optimization algorithms based on the swarm intelligence, e.g. generic algorithm, have been applied widely. However, these algorithms are time-consuming to accomplish optimization due to the large populations employed in random searching. To overcome this problem, this paper presents, for the first time, a novel efficient optimization algorithm so-called Beetle Swarm Antennae Search (BSAS) to estimate parameters of a resistance-capacitance(RC) model using simulated data collected from EnergyPlus. Furthermore, this paper also investigates the application of BSAS and four other widely used optimization algorithms including genetic algorithm(GA), particle swarm optimization(PSO), differential evolution(DE) and beetle antennae search(BAS) on parameter estimation. A case study is conducted on a single-room building simulated by EnergyPlus. BSAS and four other algorithms are employed to estimate the parameters of a 4R3C model for this tested building. The results demonstrate that the BSAS can achieve good performance on parameter estimation with sufficient accuracy and less computational cost by comparison with other algorithms. More specifically, the computational cost of the proposed algorithm is only about one third of the GA and a quarter of the DE while the mean fitness index is very close.
This paper introduces a hybrid genetic algorithm and particle swarm optimization algorithm (GA-PSO) dedicated to pattern synthesis of conformal antenna arrays. Taking advantages of both methods, the proposed hybrid GA...
详细信息
This paper introduces a hybrid genetic algorithm and particle swarm optimization algorithm (GA-PSO) dedicated to pattern synthesis of conformal antenna arrays. Taking advantages of both methods, the proposed hybrid GA-PSO optimization algorithm has fast convergence speed and high convergence accuracy when applied to antenna array pattern synthesis. To show the performance of the hybrid optimization algorithm, several typical test functions and optimization examples of a linear array pattern synthesis are illustrated. Finally, a 4x2 cylindrical conformal microstrip antenna array as a practical synthesis example is studied to demonstrate the proposed algorithm. The simulated and measured results have shown the proposed method is effective and reliable for conformal antenna array pattern synthesis.
Plasmonic nano particles can be greatly enhancing the optical absorption coefficient spectrum. Since optical properties of these particles strongly depends on the size and shape of the nano particles, in this paper pa...
详细信息
Plasmonic nano particles can be greatly enhancing the optical absorption coefficient spectrum. Since optical properties of these particles strongly depends on the size and shape of the nano particles, in this paper particle swarm optimization algorithm (PSO) is used to optimize the nano particles shape and size in order to amplification of the absorption coefficient. In PSO a swarm consists of a matrix with decimal numbers, controls the particles shape and size in order to increase the absorption coefficient in the visible part of light spectrum. It is found that significant plasmonic enhancement of above 100000 can be obtained by optimize selection of particle shapes and sizes. (C) 2016 Elsevier GmbH. All rights reserved.
The coverage rate of the underwater sensor networks directly influences on the monitoring efficiency in underwater environment, and it can be effectively improved by adjusting the positions of the mobile nodes reasona...
详细信息
The coverage rate of the underwater sensor networks directly influences on the monitoring efficiency in underwater environment, and it can be effectively improved by adjusting the positions of the mobile nodes reasonably for a 3D underwater sensor network which consists of mobile nodes as underwater robots like Autonomous Underwater Vehicles. An optimal deployment method can quickly set up a reasonable topology of the sensor networks and achieve a higher efficiency for detecting or investigating. An optimal algorithm of coverage enhancing for 3D Underwater sensor networks based on improved Fruit Fly optimization algorithm (UFOA) is proposed in this paper. This method realizes the global optimal coverage based on foraging behavior of fruit flies, and it has the features of higher speed of convergence, few parameters to set up and stronger global searching ability. Simulation result indicates that the proposed UFOA method can significantly improve the effective coverage rate of the sensor networks compared with some widely studied PSO and IPSO algorithms.
Process modeling with activated sludge models (ASMs) is useful for the design and operational improvement of biological nutrient removal (BNR) processes. Effective utilization of ASMs requires the influent fraction an...
详细信息
Process modeling with activated sludge models (ASMs) is useful for the design and operational improvement of biological nutrient removal (BNR) processes. Effective utilization of ASMs requires the influent fraction analysis (IFA) of the wastewater treatment plant (WWTP). However, this is difficult due to the time and cost involved in the design and operation steps, thereby declining the simulation reliability. Harmony Search (HS) algorithm was utilized herein to determine the relationships between composite variables and state variables of the model IWA ASM1. Influent fraction analysis was used in estimating fractions of the state variables of the WWTP influent and its application to 9 wastewater treatment processes in South Korea. The results of influent Ss and Xs+X-BH, which are the most sensitive variables for design of activated sludge process, are estimated within the error ranges of 8.9-14.2% and 3.8-6.4%, respectively. Utilizing the chemical oxygen demand (COD) fraction analysis for influent wastewater, it was possible to predict the concentrations of treated organic matter and nitrogen in 9 full scale BNR processes with high accuracy. In addition, the results of daily influent fraction analysis (D-IFA) method were superior to those of the constant influent fraction analysis (C-IFA) method.
暂无评论