Hybrid Optimization Strategy Research Articles

Design Optimization of the Lines of the Bulbous Bow of a Hull Based on Parametric Modeling and Computational Fluid Dynamics Calculation

To reduce the ship wave-making resistance, the lines of the bulbous bow of a hull are optimized by an automatic optimization platform at the ship design stage. Parametric modeling was applied to the hull by using non-uniform rational basis spline (NURBS). The Rankine-source panel method was used to calculate the wave-making resistance. A hybrid optimization strategy was applied to achieve the optimization goal. A Ro-Ro ship was taken as an example to illustrate the optimization method adopted, with the objective to minimize the wave-making resistance. The optimization results show that wave-making resistance obviously reduces and the wave-shape of the near bow becomes gentle after the lines of the bulbous bow of the hull are optimized, which demonstrates the validity of the proposed optimization design strategy.

Fast multi-objective optimization of multi-parameter antenna structures based on improved MOEA/D with surrogate-assisted model

For multi-objective design of multi-parameter antenna structures, optimization efficiency and computational cost are two major concerns. In this paper, an improved multi-objective evolutionary algorithm based on decomposition (MOEA/D) is proposed to improve global optimization capability by diversity detection operation and mixed population update operation. Further, in order to reduce the computational cost, a hybrid optimization strategy integrating a dynamically updatable surrogate-assisted model into the improved MOEA/D is proposed. The numerical results of test functions show that our algorithm outperforms original MOEA/D, modified MOEA/D (M-MOEA/D), and nondominated sorting genetic algorithm II (NGSA-II) in terms of diversity. Experimental validation of Pareto-optimal planar miniaturized multiband antenna designs is also provided, showing excellent convergence and considerable computational savings compared to those previously published approaches.

An empirical study on constraint optimization techniques for test generation

Constraint solving is a frequent, but expensive operation with symbolic execution to generate tests for a program. To improve the efficiency of test generation using constraint solving, four optimization techniques are usually applied to existing constraint solvers, which are constraint independence, constraint set simplification, constraint caching, and expression rewriting. In this paper, we conducted an empirical study, using these four constraint optimization techniques in a well known test generation tool KLEE with 77 GNU Coreutils applications, to systematically investigate how these optimization techniques affect the efficiency of test generation. The experimental results show that these constraint optimization techniques as well as their combinations cannot improve the efficiency of test generation significantly for ALL-SIZED programs. Moreover, we studied the constraint optimization techniques with respect to two static metrics, lines of code (LOC) and cyclomatic complexity (CC), of programs. The experimental results show that the ``constraint set simplification technique can improve the efficiency of test generation significantly for the programs with high LOC and CC values. The ``constraint caching optimization technique can improve the efficiency of test generation significantly for the programs with low LOC and CC values. Finally, we propose four hybrid optimization strategies and practical guidelines based on different static metrics.

Hybrid Optimization Algorithm to Combine Neural Network for Rainfall-Runoff Modeling

Rainfall-runoff modeling is very important for Water Resources Management because accurate and timely prediction can avoid accidents, such as the life risk, economic losses, etc. This paper proposed the novel hybrid optimization algorithm to combine Neural Network (NN) for rainfall-runoff modeling, namely HGASA-NN. Firstly, a novel and specialized hybrid optimization strategy by incorporating Simulated Annealing algorithm (SA) into Genetic Algorithm (GA) was used to train the initial connection weights and thresholds of NN. Secondly, the Back Propagation (BP) algorithm was adjusted the final weights and biases. Finally, a numerical example of daily rainfall-runoff data was used to elucidate the forecasting performance of the proposed HGASA-NN model. The HGASA-NN can make use of not only strong global searching ability of the GASA, but also strong local searching ability of the BP algorithm. The forecasting results indicate that the proposed model yields more accurate forecasting results than the BP-NN and pure GA training NN model. Therefore, the HGASA-NN model is a promising alternative for rainfall-runoff forecasting.

Hybrid-optimization strategy for the communication of large-scale Kinetic Monte Carlo simulation

The parallel Kinetic Monte Carlo (KMC) algorithm based on domain decomposition has been widely used in large-scale physical simulations. However, the communication overhead of the parallel KMC algorithm is critical, and severely degrades the overall performance and scalability. In this paper, we present a hybrid optimization strategy to reduce the communication overhead for the parallel KMC simulations. We first propose a communication aggregation algorithm to reduce the total number of messages and eliminate the communication redundancy. Then, we utilize the shared memory to reduce the memory copy overhead of the intra-node communication. Finally, we optimize the communication scheduling using the neighborhood collective operations. We demonstrate the scalability and high performance of our hybrid optimization strategy by both theoretical and experimental analysis. Results show that the optimized KMC algorithm exhibits better performance and scalability than the well-known open-source library—SPPARKS. On 32-node Xeon E5-2680 cluster (total 640 cores), the optimized algorithm reduces the communication time by 24.8% compared with SPPARKS.

Mars atmospheric entry trajectory optimization via particle swarm optimization and Gauss pseudo-spectral method

In this paper, a hybrid optimization strategy using particle swarm optimization and Gauss pseudo-spectral method is proposed to generate the optimal entry trajectory of Mars pin-point landing mission. This hybrid optimization strategy merges global optimization with local optimization. Coarse optimization is first conducted to provide a global relative suitable initial guess for subsequent local accurate optimization, which can greatly improve optimization efficiency and robustness and produce the optimal Mars atmospheric entry trajectory from the present position to the designed terminal state satisfying the path constraints. The validity of the hybrid optimization strategy developed in this paper is confirmed by computer simulation in the presence of aerodynamics uncertainties.

Global/local optimization strategies combined for waterflooding problems

In oil Reservoir Engineering application, a problem of great interest is the dynamic optimization of waterflooding management. In this work, the net present value (NPV) is taken as the function to be maximized, in which the allocated rates of wells are considered as design variables. Alternatively, the switching times of the control cycles can also be considered as design variables. This assumption increases flexibility to the management. Despite this, the formulation of this problem leads to a highly nonlinear, multimodal objective function. Therefore, to conduct the management, a hybrid optimization strategy is proposed here considering surrogate models. The hybrid strategy combines different methods at two different stages, a global and local. In this sense, the global search is driven by the genetic algorithm (GA) and the local search is driven by the sequential approximation optimization (SAO) method. The proposed methodology was successful in identifying wells that should be late started or shut-in before the end of the concession period and in handling different kinds of production strategies. It was also verified that an increasing on operation flexibility results in NPV improvement. Cycle duration variables proved to be useful in decreasing the number of design variables while maintaining recovery efficiency.

Optimum design of frame structures using the Eagle Strategy with Differential Evolution

Modern metaheuristic algorithms are in general suited for global optimization. This paper combines the recently developed eagle strategy algorithm with differential evolution. The new algorithm, denoted as the ES–DE, is implemented by interfacing SAP2000 structural analysis code and MATLAB mathematical software. The performance of the ES–DE is evaluated by solving four benchmark problems where the objective is to minimize the weight of steel frames. The optimized designs obtained by the proposed algorithm are better than those found by the standard differential evolution algorithm and also very competitive with literature. The overall convergence behavior is significantly enhanced by the hybrid optimization strategy.

Automatic estimation of kinetic and isotherm parameters from ZLC experiments

In order to automate and improve the analysis of the experimental Zero Length Column (ZLC) data a general adsorption simulator is applied to the simulation of the ZLC system. The output of this ZLC simulation is linked to a hybrid optimisation strategy: combining a global Genetic Algorithm (GA) with a local Nelder–Mead algorithm ensures an efficient and accurate optimisation procedure. By minimising the deviation between the experimental data and the simulation the kinetic and equilibrium parameters of the adsorbent can be estimated. In the first step the parameters of the blank experiment are fitted; thus taking the dynamics of the detector and the system piping into account. In the second step several experiments at different flow rates and temperatures are fitted simultaneously. This increases the accuracy and reliability of the results. The automated parameter estimation was applied to ZLC runs with carbon dioxide on zeolite 13X pellets. The results agree with the traditional analysis methods and with independent measurements. In addition it is demonstrated that it is possible to approximate the correct adsorption isotherm from ZLC runs under kinetic control. The automatic analysis tool allows the inclusion of nonlinear effects in the fitting and lowers the access barrier to the application of the ZLC method.

Integrated Topology Optimization of Structure/Vibration Control for Piezoelectric Cylindrical Shell Based on the Genetic Algorithm

A hybrid optimization strategy for integrated topological optimization design of piezoelectric cylindrical flat shell structure is proposed. The method combines the genetic algorithm (GA) and linear-quadratic-regulator (LQR) theory to optimize the performance of coupling structure/control system. The GA is used to choose the optimal structure topology and number and placements of actuators and control parameters; meanwhile, the LQR is used to design control system to suppress vibration of optimal structure under sinusoidal excitation, which is based on the couple-mode space control. In addition, the mathematical morphology operators are used for repairs of disconnected structure topology. The results of numerical simulation and computations show that the proposed method is effective and feasible, with good performance for the optimal and coupling piezoelectric cylindrical shell structure/control system.

Experiment Driven Ann-GA Based Technique for Optimal Distribution of Discrete Heat Sources Under Mixed Convection

This article reports the results of mixed convection heat transfer studies from five heat sources (aluminum) mounted at different positions on a substrate board (Bakelite). The goal is to determine the optimal arrangement, such that, the maximum temperature excess is minimum among all the possible configurations. For accomplishing this, a completely experimental driven hybrid optimization strategy, that combines Artificial neural network (ANN) with Genetic algorithm (GA) is used. Initial optimization studies are carried out by employing a heuristic non-dimensional geometric parameter λ, which is identified to be the key parameter to decide the maximum temperature in the system.

Evolving RBF neural networks for rainfall prediction using hybrid particle swarm optimization and genetic algorithm

In this paper, an effective hybrid optimization strategy by incorporating the adaptive optimization of particle swarm optimization (PSO) into genetic algorithm (GA), namely HPSOGA, is used for determining the parameters of radial basis function neural networks (number of neurons, their respective centers and radii) automatically. While this task depends upon operator׳s experience with trial and error due to lack of prior knowledge, or based on gradient algorithms which are highly dependent on initial values. In this paper, hybrid evolutionary algorithms are used to automatically build a radial basis function neural networks (RBF-NN) that solves a specified problem, related to rainfall forecasting in this case. In HPSOGA, individuals in a new generation are created through three approaches to improve the global optimization performance, which are elitist strategy, PSO strategy and GA strategy. The upper-half of the best-performing individuals in a population are regarded as elites, whereas the half of the worst-performing individuals are regarded as a swarm. The group constituted by the elites are enhanced by selection, crossover and mutation operation on these enhanced elites. HPSOGA is applied to RBF-NN design for rainfall prediction. The performance of HPSOGA is compared to pure GA in these basis function neural networks design problems, showing that the hybrid strategy is of more effective global exploration ability and to avoid premature convergence. Our findings reveal that the hybrid optimization strategy proposed here may be used as a promising alternative forecasting tool for higher forecasting accuracy and better generalization ability.

Design and Optimization of PSD Housing Using a MIGA-NLPQL Hybrid Strategy Based on a Surrogate Model

A new power-split device (PSD) is designed as the core component of the multi-power coupling system in a hybrid electric vehicle. It is important to consider the influence of the factors of weight and stiffness on performance of PSD when designing PSD housing. In this paper, the overall arrangement of the power distribution system and PSD housing are redesigned. The finite element analysis is conducted to test PSD housing stiffness. According to the results of the analysis, this study adopts a hybrid optimization strategy based on surrogate models to obtain the least weight under the constraint of PSD housing stiffness. The surrogate models are established using a responsive surface method based on the data obtained by the optimal Latin hypercube design (OLHD). The hybrid optimization strategy combines a multi-island genetic algorithm (MIGA) with a nonlinear programming quadratic line search (NLPQL), which ensures obtaining optimal design parameters for PSD housing. In comparison with optimization using a single MIGA, this hybrid optimization strategy is more efficient and feasible for optimizing housing.

Globally Optimal Superconducting Homogeneous Magnet Design for an Asymmetric 3.0 T Head MRI Scanner

This paper describes a hybrid numerical method with a combination of linear programming (LP) and nonlinear programming to design an asymmetric magnetic resonance imaging (MRI) system for head imaging. The method was successfully employed to a 0.7 T open-style MRI. Recently, we have updated the LP mathematical model in the hybrid optimization strategy by adding maximum axial and radial magnetic field strength limitation to make the magnet safe. The whole magnet consists of eight coaxial coils asymmetry about the z-axis. The six innermost coils are the main coils with positive current direction and the two outermost coils are shielding coils with negative current direction. All coils contribute their efforts to produce a central magnetic field strength of 3.0 T over a 20 cm diameter of spherical volume, and the peak to peak homogeneity is 12 ppm. The temperature bore for the magnet is formed with two different diameters cylindrical bores which makes head and shoulder access to the imaging volume easier. A 5 Gauss footprint area is restricted larger than an elliptical region with the sizes of 3.5 m in radial and 4.0 m in axial direction. The results show that the methodology is very flexible and efficient for asymmetric MRI magnet design.

Modelling and multi-response optimization of hole sinking electrical discharge micromachining of titanium alloy thin sheet

Thin sheets of titanium alloys are widely used in aerospace and automotive industries for specific applications. The creation of micro holes with requisite hole quality in thin sheets of these alloys using energy of electric discharge is a challenging task for manufacturing engineers. Hole sinking electrical discharge micromachining (HS-EDMM) is one of the most promising micromachining processes to create symmetrical and non-symmetrical micro holes. The present paper is related to selection of optimum parameter settings for obtaining maximum material removal, minimum tool wear and minimum hole taper in HS-EDMM. In this paper an attempt has been made to develop an integrated model (ANN-GRA-PCA) of single hidden layer back propagation neural network (BPNN) for prediction and grey relational analysis (GRA) coupled with principal component analysis (PCA) hybrid optimization strategy with multiple responses of HSEDMM of Ti-6Al-4V. Experiments have been conducted to generate dataset for training and testing of the network where input parameters consist of gap voltage, capacitance of capacitor and the resulting performance parameters are represented by material removal rate (MRR), tool wear rate (TWR), and hole taper (Ta). The results indicate that the integrated model is capable to predict and optimize process performance with reasonable accuracy under varied operating conditions of HS-EDMM. The proposed approach would be extendable to other configurations of EDMM processes for different materials.

Derivative-based hybrid heuristics for continuous-time simulation optimization

The topic of simulation–optimization has not been fundamentally tackled by many continuous-time modeling and simulation tools, yet. Common simulation-based optimization problems are usually coupled with standard optimization algorithms like any other simulation-free nonlinear optimization problems. While such couplings are usually based on many state-of-the-art software engineering concepts with a high-level user interface for flexible incorporation of simulation and optimization, the design of specialized optimization strategies targeting simulation-based objective functions is lacked within many simulation–optimization tools. In this work, new redefinition of Non Linear Programming (NLP) problems in the context of continuous-time simulation optimization is presented. Then, the modified optimization problems are efficiently tackled using derivative-based hybrid heuristics. In order to specify, illustrate and implement such heuristics, a new terminology is proposed. According to the proposed terminology, derivative-based hybrid strategies are implemented by hybridizing naive multistart derivative-based optimization methods with population-based metaheuristics. It is shown that the adoption of derivative-based optimization methods within hybrid optimization strategies significantly improves the solution quality of continuous-time simulation optimization problems.

A Combined Optimal Sensor Placement Strategy for the Structural Health Monitoring of Bridge Structures

Optimal sensor placement is an important part in the structural health monitoring of bridge structures. However, some defects are present in the existing methods, such as the focus on a single optimal index, the selection of modal order and sensor number based on experience, and the long computation time. A hybrid optimization strategy named MSE-AGA is proposed in this study to address these problems. The approach firstly selects modal order using modal participation factor. Then, the modal strain energy method is adopted to conduct the initial sensor placement. Finally, the adaptive genetic algorithm (AGA) is utilized to determine the optimal number and locations of the sensors, which uses the root mean square of off-diagonal elements in the modal assurance criterion matrix as the fitness function. A case study of sensor placement on a numerically simulated bridge structure is provided to verify the effectiveness of the MSE-AGA strategy, and the AGA method without initial placement is used as a contrast experiment. A comparison of these strategies shows that the optimal results obtained by the MSE-AGA method have a high modal strain energy index, a short computation time, and small off-diagonal elements in the modal assurance criterion matrix.

Hybrid Optimization Framework with Proper-Orthogonal-Decomposition-Based Order Reduction and Design-Space Evolution Scheme

This paper describes the development of an innovative hybrid-optimization framework that employs a proper-orthogonal-decomposition-based design-space order-reduction scheme combined with the evolutionary-algorithm technique. The proper-orthogonal-decomposition method of extracting dominant modes from an ensemble of candidate configurations is used for the design-space order reduction wherein the proper-orthogonal-decomposition coefficients act as the new shape design variables. The snapshot of the candidate population is updated iteratively using an evolutionary-algorithm technique of fitness-driven retention. Established with the goal of developing a robust optimization scheme, the developed strategy capitalizes on the advantages of the evolutionary algorithm as well as the proper-orthogonal-decomposition-based reduced-order modeling, while overcoming the shortcomings inherent with these techniques. When linked with a streamlined multidisciplinary optimization framework, this hybrid-optimization strategy offers a computationally efficient methodology for problems having a high level of complexity and a challenging design space with a large number of design variables. The developed framework is demonstrated for its robustness on a nonconventional supersonic tailless air-vehicle wing-shape optimization problem.

Globally Optimal Algorithm for Design of 0.7 T Actively Shielded Whole-Body Open MRI Superconducting Magnet System

A 0.7-T open magnetic resonance imaging superconducting magnet system with actively shielded technique was designed and will be fabricated. The design methodology of the open structure magnetic resonance superconducting magnet is employed by a hybrid optimization strategy with combination with linear programming and nonlinear optimization algorithm. The linear programming was employed to search the coils' initial position and shape with advantage of high efficient and global optimization, and the nonlinear optimization algorithm was to discrete the initial data into solenoids to reach the system requirements, such as field strength and homogeneity over an interesting region, the scope of 5 Gauss fringe field, maximum magnetic field and maximum hoop stress within solenoids. The hybrid optimization is flexible and efficient to design conventional cylindrical system and split open system.

HYBRID OF GENETIC ALGORITHM AND SIMULATED ANNEALING FOR SUPPORT VECTOR REGRESSION OPTIMIZATION IN RAINFALL FORECASTING

Accurate forecasting of rainfall has been one of the most important issues in hydrological research such as river training works and design of flood warning systems. Support vector regression (SVR) is a popular regression method in rainfall forecasting. Type of kernel function and kernel parameter setting in the SVR traing procedure, along with the input feature subset selection, significantly influence regression accuracy. In this paper, an effective hybrid optimization strategy by combining the strengths of genetic algorithm (GA) and simulated annealing (SA), is employed to simultaneously optimize the input feature subset selection, the type of kernel function and the kernel parameter setting of SVR, namely GASA–SVR. The developed GASA–SVR model is being applied for monthly rainfall forecasting in Guilin of Guangxi. The GA is carried out as a main frame of this hybrid algorithm while SA is used as a local search strategy to help GA jump out of local optima and avoid sinking into the local optimal solution early. Compared with SVR, pure GA–SVR and HGA–SVR, results show that the hybrid GASA–SVR model can correctly select the discriminating input features subset, successfully identify the optimal type of kernel function and all the optimal values of the parameters of SVR with the lowest prediction error values in rainfall forecasting, can also significantly improve the rainfall forecasting accuracy. Experimental results reveal that the predictions using the proposed approach are consistently better than those obtained using the other methods presented in this study in terms of the same measurements. Those results show that the proposed GASA–SVR model provides a promising alternative to monthly rainfall prediction.