用遗传算法提高协同优化方法的可靠性

分析了协同优化方法可靠性不好的根源。提出了用遗传算法来替代系统级优化问题中的基于梯度的优化算法的策略。实例表明所提出的策略是可行有效的。     

Optimization of slider air bearing contours using the combined genetic algorithm-subregion approach

A hybrid optimization approach, the combined genetic algorithm-subregion method, which combines the advantage of the genetic algorithm and the subregion approach, is presented. Using a binary string to represent a selected design space, the combined genetic algorithm-subregion method adopts the genetic algorithm to perform the optimization process. Starting from a pico slider design originally flying at 14 nm, optimized designs were obtained for sliders with target flying heights of 7, 5 and 3.5 nm, respectively. The results show that the combined genetic algorithm-subregion method has good convergence with a substantial reduction of computation time.     

微谐振式压力传感器的系统级建模与仿真

具有微型化、多域耦合等特点的MEMS传感器对设计具有高精度、高效率、迭代设计、结构与电路联合设计等要求。基于组件库或集总参数的MEMS系统级设计方法虽然满足了这些要求,但在面向特定器件的定制化设计方面却略显不足。以一MEMS谐振式压力传感器为例,提出了一种全参数化系统级建模与仿真方法。在建立的详细数学模型基础上,通过硬件描述语言对传感器谐振子按照详细拓扑结构进行了建模与仿真。仿真结果和实验结果的对比显示这是一种满足MEMS设计要求的建模与仿真方法。     

基于遗传算法的微机械陀螺的多学科设计优化

基于micro-electro-mechanical system(MEMS)技术的微机械陀螺是集传感器、致动器、检测与控制等于一体的复杂多学科交叉系统，其整体特性是各个子系统综合作用的结果。在充分考虑工艺、结构、电路、工作环境等多学科或因素的约束条件下，提出微机械陀螺的多学科概念设计模型。以陀螺的灵敏度最大为优化目标，利用遗传算法对设计模型进行全局优化，获得初步的最优设计方案，并采用有限元软件ANSYS验证优化结果的正确性。     

Optimal design of four-bar mechanisms using a hybrid multi-objective GA with adaptive local search

Responding to an increasing demand for mechanism synthesis tools that are both efficient and accurate, this paper presents a novel approach to the multi-objective optimal design of four-bar linkages for path-generation purposes. Three, often conflicting criteria including the mechanism's tracking error, deviation of its transmission angle from 90° and its maximum angular velocity ratio are considered as objectives of the optimization problem. To accelerate the search in the highly multimodal solution space, a hybrid Pareto genetic algorithm with a built-in adaptive local search is employed which extends its exploration to an adaptively adjusted neighborhood of promising points. The efficiency of the proposed algorithm is demonstrated by applying it to a classical design problem for one, two and three objective functions and comparing the results with those reported in the literature. The comparison shows that the proposed algorithm distinctly outperforms other algorithms both quantitatively and qualitatively (from a practical point of view).     

Applications of soft computing techniques in response surface based approximate optimization

The paper describes the construction of global function approximation models for use in design optimization via global search techniques such as genetic algorithms. Two different approximation methods referred to as evolutionary fuzzy modeling (EFM) and neuro-fuzzy modeling (NFM) are implemented in the context of global approximate optimization. EFM and NFM are based on soft computing paradigms utilizing fuzzy systems, neural networks and evolutionary computing techniques. Such approximation methods may have their promising characteristics in a case where the training data is not sufficiently provided or uncertain information may be included in design process. Fuzzy inference system is the central system for of identifying the input/output relationship in both methods. The paper introduces the general procedures including fuzzy rule generation, membership function selection and inference process for EFM and NFM, and presents their generalization capabilities in terms of a number of fuzzy rules and training data with application to a three-bar truss optimization.     

混合遗传-模拟退火算法应用于液压集成块优化设计的研究

以液压集成块布局布孔优化设计问题为背景,给出了在立体空间中带性能约束的多目标优化数学模型。针对遗传算法在求解该组合优化问题时局部搜索能力不强的情况,将模拟退火算法与之结合,设计了多参数级联的编码方法,且加入记忆装置,构造了一种具有记忆功能的混合遗传-模拟退火算法。实例计算表明,改进后的新算法能很好地改善对解空间的局部搜索能力,从而得到较高质量的解。     

An integrated method of particle swarm optimization and differential evolution

Particle swarm optimization (PSO) and differential evolution (DE) have their similarities and compatibility in the design update process, such that a new design vector is determined by using neighborhood designs under algorithm control parameters. The paper deals with an integrated method of a hybrid PSO (HPSO) algorithm combined with DE in order to refine the optimization performance. PSO and DE also possess common characteristics compared with genetic algorithm (GA). The crossover- and mutation-like operators are suggested in the HPSO. A bounce back method is also implemented to prevent the design from locating out of design spaces during the optimization process. For the purpose of further enhancing the search capabilities, such HPSO is combined with the Q-learning that is one of efficient reinforcement learning methods. Using a number of nonlinear multimodal functions and engineering optimization problems, the proposed algorithms of HPSO and HPSO with Q-learning are compared with PSO DE and GA. This paper was recommended for publication in revised form by Associate Editor Tae Hee Lee Jongsoo Lee received a B.S. degree in Mechanical Engineering from Yonsei University in 1988. He then went on to receive his M.S. degree from University of Minnesota in 1992 and Ph.D. degree from Rensselaer Polytechnic Institute in 1996. Dr. Lee is currently a Professor at the School of Mechanical Engineering at Yonsei University in Seoul, Korea. He is currently serving as a committee member of the division of CAE and Applied Mechanics in the Korean Society of Mechanical Engineers. Dr. Lee’s research interests are in the area of engineering design optimization, fluidstructure interactions, and reliability based robust product design.     

Conceptual Design of Fixtures using Genetic Algorithms

Fixture design is a complex and intuitive process, which demands rich experience from the designer. Multiple acceptable designs are possible for a given workpiece and hence the solution space is large. Recent advances in CAD/CAM, especially in artificial intelligence, have alleviated this problem by intelligently restricting the search space considered, thus opening the possibility of obtaining better designs. Researchers have used various techniques under the general rubric of artificial intelligence to solve the fixture design problem. The most common of these have been case-based reasoning and expert systems. However, these two common methods do not ensure that the resulting solution is efficient or optimal. In this paper we propose to combine complementarily the strengths of genetic algorithms and neural networks to develop a fixture design system. Results obtained using this combined multi-agent approach for the design of fixtures are promising.     

Hydraulic Optimization of a Double-channel Pump’s Impeller Based on Multi-objective Genetic Algorithm

Computational fluid dynamics (CFD) can give a lot of potentially very useful information for hydraulic optimization design of pumps, however, it cannot directly state what kind of modification should b...     

Performance enhancement of axial fan blade through multi-objective optimization techniques

This paper presents an axial fan blade design optimization method incorporating a hybrid multi-objective evolutionary algorithm (hybrid MOEA). In flow analyses, Reynolds-averaged Navier-Stokes (RANS) equations were solved using the shear stress transport turbulence model. The numerical results for the axial and tangential velocities were validated by comparing them with experimental data. Six design variables relating to the blade lean angle and the blade profile were selected through Latin hypercube sampling of design of experiments (DOE) to generate design points within the selected design space. Two objective functions, namely, total efficiency and torque, were employed, and multi-objective optimization was carried out, to enhance the performance. A surrogate model, Response Surface Approximation (RSA), was constructed for each objective function based on the numerical solutions obtained at the specified design points. The Non-dominated Sorting of Genetic Algorithm (NSGA-II) with local search was used for multi-objective optimization. The Pareto-optimal solutions were obtained, and a trade-off analysis was performed between the two conflicting objectives in view of the design and flow constraints. It was observed that, by the process of multi-objective optimization, the total efficiency was enhanced and the torque reduced. The mechanisms of these performance improvements were elucidated by analysis of the Pareto-optimal solutions.     

Optimal design and development of a 2-DOF PKM-based machine tool

This paper presents multiobjective optimization of a typical 2-degree-of-freedom (DOF) parallel kinematic machine (PKM) tool that has only single DOF joints. Nondimensional indices, namely global stiffness index (GSI), global conditioning index (GCI), and workspace index, are considered as the objectives for optimization. The indices GSI and GCI depict the variation of stiffness and dexterity of PKM within the workspace. The leg length and distance between two rails on which actuators slide are treated as design variables as these greatly influence the characteristics of PKM. A multiobjective genetic algorithm (MOGA) approach is implemented in MATLAB to find an efficient solution to this complex optimization problem. Fitness function includes inverse kinematics equations, Jacobian and stiffness matrices to compute and optimize the nondimensional indices. First, the optimal value of each index is obtained by single-objective GA. To further improve the results, a hybrid function PATTERNSEARCH is used. This helps to select appropriate boundary conditions for MOGA. To obtain the optimal values of all the three indices, a multiobjective GA is carried out. The results are compared with a conventional exhaustive search method of optimization. The obtained results show that the use of MOGA enhances the quality of the optimization outcome. Secondly, a prototype has been designed and developed with the optimal dimensions. The actual workspace of the PKM and influence of leg collision on the workspace are studied. Finally, a preliminary experimentation was carried out. A comparison between PKM and the three-axis serial kinematic machine tool is presented.     

基于结构组件库的MEMS概念设计方法

提出了一种基于结构组件库的MEMS概念设计方法。给出了结构组件的表征形式,探讨了基于功能分解的结构组件模型获取方法,借以形成支持概念设计的结构组件模型库。阐明了MEMS概念设计时如何通过行为匹配自动检索出相关组件及其行为链组合、通过对行为链组合进行功能结构细化获得备选的器件方案。结合航空及空气动力学领域急需的MEMS微型剪应力传感器实例对该概念设计方法进行了阐述。研究表明,该方法对于提高MEMS概念设计效率及方案创新能力、促进MEMS概念设计工具化有积极作用。     

复杂微机电系统多学科设计方法研究

针对复杂微机电系统设计涉及多个学科领域的特点,提出了应用多学科设计优化方法来设计复杂微机电工程产品,归纳分析了复杂微机电系统在设计过程中可能涉及的学科,探讨了多学科设计方法在复杂微机电系统设计中的应用。     

Solving a multi-objective job shop scheduling problem with sequence-dependent setup times by a Pareto archive PSO combined with genetic operators and VNS

In this paper, we present a combination of particle swarm optimization (PSO) and genetic operators for a multi-objective job shop scheduling problem that minimizes the mean weighted completion time and the sum of the weighted tardiness/earliness costs, simultaneously. At first, we propose a new integer linear programming for the given problem. Then, we redefine and modify PSO by introducing genetic operators, such as crossover and mutation operators, to update particles and improve particles by variable neighborhood search. Furthermore, we consider sequence-dependent setup times. We then design a Pareto archive PSO, where the global best position selection is combined with the crowding measure-based archive updating method. To prove the efficiency of our proposed PSO, a number of test problems are solved. Its reliability based on some comparison metrics is compared with a prominent multi-objective genetic algorithm (MOGA), namely non-dominated sorting genetic algorithm II (NSGA-II). The computational results show that the proposed PSO outperforms the above MOGA, especially for large-sized problems.     

基于进化算法和模拟退火算法的混合调度算法

将进化算法与模拟退火算法相结合,提出四种有效的混合调度算法,即遗传退火算法、改进遗传算法、改进进化规划和并行模拟退火算法。两种算法搜索机制的互补增强了全局探索能力,基于关键路径的邻域函数运用提高了算法的效率。仿真结果表明:混合算法在求解质量和求解效率方面均有优势,优于国外同类研究成果;基于模拟退火的变异算子的搜索能力优于交叉算子;改进进化规划优于其他混合算法。     

遗传算法在多层次结构机械产品智能方案设计中应用

通过把机械方案设计过程看作是一个状态空间的求解问题,用遗传算法控制其搜索过程,构建完善了新的遗传编码体系,为了适应新的编码体系重新构建了交叉和变异等遗传操作,并利用复制、交换和变异等操作进行一次次迭代,最终自动生成一组最优的设计方案。实例验证,该系统有效、准确和实用。     

Hybrid evolutionary algorithm with marriage of genetic algorithm and extremal optimization for production scheduling

This paper presents a hybrid evolutionary algorithm with marriage of genetic algorithm (GA) and extremal optimization (EO) for solving a class of production scheduling problems in manufacturing. The scheduling problem, which is derived from hot rolling production in steel industry, is characterized by two major requirements: (i) selecting a subset of orders from manufacturing orders to be processed; (ii) determining the optimal production sequence under multiple constraints, such as sequence-dependant transition costs, non-execution penalties, earliness/tardiness (E/T) penalties, etc. A combinatorial optimization model is proposed to formulate it mathematically. For its NP-hard complexity, an effective hybrid evolutionary algorithm is developed to solve the scheduling problem through combining the population-based search capacity of GA and the fine-grained local search efficacy of EO. The experimental results with production scale data demonstrate that the proposed hybrid evolutionary algorithm can provide superior performances in scheduling quality and computation efficiency.     

混合离散变量优化设计的复合遗传算法

目前，对混合离散变量的遗传算法研究较少，而且现有算法对设计变量的处理不能很好地满足工程设计要求。为此，提出了一种面向设计、制造的设计变量工程化处理方法，能合理地处理优化设计中混合离散变量的取值问题。引入了混沌移民算子对基本遗传算法进行了改进，并开发了混合离散变量优化的复合遗传算法程序LSGA。工程设计实例表明，该算法对优化设计问题的特性无特殊要求，具有较好的普适性，而且程序运行可靠，全局收敛能力强。     

A novel particle swarm optimization approach for product design and manufacturing

This paper presents a novel optimization approach that is a new hybrid optimization approach based on the particle swarm optimization algorithm and receptor editing property of immune system. The aim of the present research is to develop a new optimization approach and then to apply it in the solution of optimization problems in both the design and manufacturing areas. A single-objective test problem, tension spring problem, pressure vessel design optimization problem taken from the literature and two case studies for multi-pass turning operations are solved by the proposed new hybrid approach to evaluate performance of the approach. The results obtained by the proposed approach for the case studies are compared with a hybrid genetic algorithm, scatter search algorithm, genetic algorithm, and integration of simulated annealing and Hooke-Jeeves pattern search.