A stochastic local search algorithm for constrained continuous global optimization

This paper presents a new stochastic local search algorithm known as feasible–infeasible search procedure (FISP) for constrained continuous global optimization. The proposed procedure uses metaheuristic strategies for combinatorial optimization as well as combined strategies for exploring continuous spaces, which are applied to an efficient process in increasingly refined neighborhoods of current points. We show effectiveness and efficiency of the proposed procedure on a standard set of 13 well‐known test problems. Furthermore, we compare the performance of FISP with SNOPT (sparse nonlinear optimizer) and with few successful existing stochastic algorithms on the same set of test problems.     

A function-based computational method for design concept evaluation

Concept generation is an indispensable step of innovation design. However, the limited knowledge and design thinking fixation of designers often impede the generation of novel design concepts. Computational tools can be a necessary supplement for designers. They can generate a big number of design concepts based on an existing knowledge base. For filtering these design concepts, this work presents a computational measurement of novelty, feasibility and diversity based on 500,000 granted patents. First, about 1700 functional terms (terminologies) are mapped to high dimensional vectors (100 dimensional space) by word embedding technique. The resulted database is knowledge base-I (KB-I). Then, we adopt circular convolution to convert patents into high dimensional vectors. The resulted database is KB-II. Based on the two knowledge bases, the computational definitions of novelty, feasibility and diversity are developed. We conduct six experiments based on KB-II, a random dataset and a real product dataset, and the results show that these metrics can be used to roughly filter a big number of design concepts, and then expert-based method can be further used. This work provides a computational framework for measuring the novelty, feasibility and diversity of design concept.     

More efficient two-mode stochastic local search for random 3-satisfiability

Stochastic local search (SLS) is a popular paradigm in incomplete solving for the Boolean satisfiability problem (SAT). Most SLS solvers for SAT switch between two modes, i.e., the greedy (intensification) mode and the diversification mode. However, the performance of these two-mode SLS algorithms lags far behind on solving random 3-satisfiability (3-SAT) problem, which is a significant special case of the SAT problem. In this paper, we propose a new hybrid scoring function called M C based on a linear combination of a greedy property m a k e and a diversification property C o n f T i m e s, and then utilize M C to develop a new two-mode SLS solver called CCMC. To evaluate the performance of CCMC, we conduct extensive experiments to compare CCMC with five state-of-the-art two-mode SLS solvers (i.e., Sparrow2011, Sattime2011, EagleUP, gNovelty+PCL and CCASat) on a broad range of random 3-SAT instances, including all large 3-SAT ones from SAT Competition 2009 and SAT Competition 2011 as well as 200 generated satisfiable huge random 3-SAT ones. The experiments illustrate that CCMC obviously outperforms its competitors, indicating the effectiveness of CCMC. We also analyze the effectiveness of the underlying ideas in CCMC and further improve the performance of CCMC on solving random 5-SAT instances.     

A fast random number generator for stochastic simulations

A discrete random number (DRN) generator for stochastic differential equations is proposed. The generator has exactly 8 states and thus 10 DRN's can be obtained from a single 32-bit random variable. This is advantageous when large numbers of DRN's are needed, as for example in fluctuating lattice-Boltzmann models. The moments of the discrete distribution match those of a Gaussian distribution (zero mean and unit variance) up to 5th order. Numerical tests show that satisfactory statistical properties can be obtained with several 32-bit pseudo random number (PRN) generators.     

一种GNSS信号兼容互用性评估方法

针对全球卫星导航系统(GNSS)兼容互用性评估问题,分析了频谱分离系数和等效载噪比衰减两种方法的区别,提出了一种基于等效载噪比衰减值的GNSS信号兼容性和互用性评估方法,通过仿真证明了此方法具有实用意义。     

Neumann dynamic stochastic finite element method of vibration for structures with stochastic parameters to random excitation

This paper presents a method for the dynamic analysis of structures with stochastic parameters to random excitation. A procedure to derive the statistical characteristics of the dynamic response for structure is proposed by using dynamic Neumann stochastic finite element method presented herein. Random equation of motion for structure is transformed into a quasi-static equilibrium equation for the solution of displacement in time domain. Neumann expansion method is developed and applied to the equation for deriving the statistical solution of the dynamic response of such a random structure system, within the framework of Monte Carlo simulation. Then, the results from Neumann dynamic stochastic finite element method are compared with those from the first- and second-order perturbation stochastic finite element methods and the direct Monte Carlo simulation with respect to accuracy, convergence and computational efficiency. Numerical examples are examined to show that the approach proposed in this paper has a very high accuracy and efficiency in the analysis of compound random vibration.     

Perturbation based stochastic finite element method for homogenization of two-phase elastic composites

The main idea of the paper is to apply the second order perturbation and stochastic second central moment technique to solve the homogenization problem. In order to determine the effective elasticity tensor, the prevailing computational methodology discussed in the literature so far was the Monte-Carlo simulation providing appropriate expected values and higher order probabilistic moments of the effective tensor components. The technique applied in this paper aims at significantly reducing the computational cost of the simulation without sacrificing the solution accuracy. The numerical example substantiates this claim in the case of a periodic fiber-reinforced plane strain composite with random fiber and matrix Young’s moduli.     

A decentralized computational algorithm for the global Kalman filter

In this paper a new decentralized computational structure is developed for Optimal state estimation in large scale linear interconnected dynamical systems. The new filter uses a hierarchical structure to perform successive orthogoilalizations on the measurement subspaces of each sub-system in order to provide the optimal estimate. This ensures substantial savings in computation time. In addition, since only low-order subsystem equations are manipulated at each stage, numerical inaccuracies are reduced, and the filter remains stable for even high-order systems. This is illustrated on a multimachine example of a system comprising eleven interconnected machines.     

A computational procedure for response statistics-based optimization of stochastic non-linear FE-models

An approach for an efficient solution of response statistics-based optimization problems of non-linear FE systems under stochastic loading is presented. A sequential approximate optimization approach, where approximate stochastic analyses are used during portions of the optimization process, is implemented in the proposed formulation. In this approach, analytical approximations of the performance functions in terms of the design variables are considered during the optimization process. The analytical approximations are constructed by combining a mixed linearization approach with a stochastic response sensitivity analysis. The state of the system is defined in terms of the statistical second-moment characteristics of the structural response. The stochastic loading and the response of the system are represented by an orthogonal series expansion of the corresponding covariance matrices. In particular, a truncated Karhunen-Loève (K-L) expansion is applied. The system of non-linear equations is replaced by a statistical equivalent linear system. The evaluation of the K-L vectors is carried out by an efficient procedure that combines local linearization, modal analysis and static response of higher structural modes. An illustrative example is presented that shows the efficiency of the proposed methodology: it considers a building finite element model enforced with non-linear hysteretic devices and subject to a stochastic ground acceleration. Two types of problems are considered: a minimum structural weight design problem and an optimal non-linear device design problem.     

A computational algorithm for the optimal control of continuous-time stochastic systems

An algorithm for computing optimal controls for a class of continuous-time stochastic systems is presented. For a linear example, it is shown that the generated control is precisely the optimal control. Numerical results for a nonlinear example suggest that the generated control is at least locally optimal.     

融合局部与全局特征提取的虹膜识别方法

传统的虹膜识别系统需要将虹膜图像转换至极坐标系统并进行归一化,通过平移特征向量来达到旋转不变性。为了降低传统虹膜识别方法的复杂性,提出了一种融合局部与全局特征提取的虹膜识别方法,无须对预处理后的虹膜图像进行归一化。该方法首先对分割出的虹膜图像直接采用非张量积小波提取全局特征,接着采用SIFT方法提取选定区域的局部特征,最后对虹膜局部及全局特征采用不同的权值,进行相似性距离测试。结果表明该方法在等错误率为0.935%的情况下,正确识别率达到了99.065%。在不对虹膜图像归一化的情况下,可获得很好的识别性能。     

Time-variant global reliability sensitivity analysis of structures with both input random variables and stochastic processes

Structural and Multidisciplinary Optimization - The ubiquitous uncertainties presented in the input factors (e.g., material properties and loads) commonly lead to occasional failure of mechanical...