Improving the performance of simulated annealing in structural optimization

This study aims at improving the performance of simulated annealing (SA) search technique in real-size structural optimization applications with practical design considerations. It is noted that a standard SA algorithm usually fails to produce acceptable solutions to such problems associated with its poor convergence characteristics and incongruity with theoretical considerations. In the paper novel approaches are developed and incorporated into the standard SA algorithm to eliminate the observed drawbacks of the technique. The performance of the resulting (improved) algorithm is investigated in conjunction with two numerical examples (a 304-member braced planar steel frame, and 132-member unbraced space steel frame) designed according to provisions of the Allowable Stress Design (ASD) specification. In both examples, curves showing the variation of average acceptance probability parameter in standard and improved algorithms are plotted to verify usefulness and robustness of the integrated approaches.     

Serial and parallel implementations of model-based clustering via parsimonious Gaussian mixture models

Model-based clustering using a family of Gaussian mixture models, with parsimonious factor analysis like covariance structure, is described and an efficient algorithm for its implementation is presented. This algorithm uses the alternating expectation-conditional maximization (AECM) variant of the expectation-maximization (EM) algorithm. Two central issues around the implementation of this family of models, namely model selection and convergence criteria, are discussed. These central issues also have implications for other model-based clustering techniques and for the implementation of techniques like the EM algorithm, in general. The Bayesian information criterion (BIC) is used for model selection and Aitken’s acceleration, which is shown to outperform the lack of progress criterion, is used to determine convergence. A brief introduction to parallel computing is then given before the implementation of this algorithm in parallel is facilitated within the master-slave paradigm. A simulation study is then carried out to confirm the effectiveness of this parallelization. The resulting software is applied to two datasets to demonstrate its effectiveness when compared to existing software.     

Weight/shape structural optimization exploiting rigid movement

The paper explores the possibility of performing limited shape/size optimization to obtain a minimum weight design using rigid body movement. A two step approach is adopted where changes in the shape parameters are made separately from changes in the size parameters. The sizing step uses a conventional optimization, STARS, which is available commercially. For simplicity the method is demonstrated using planar arches subject to a variety of load and boundary conditions.     

Properties and performance modelling of finite buffer M/G/1/K networks

Finite buffer, single-server queueing systems and networks are difficult to analyze since the length of time a customer spends in the system does not follow the Markovian property. A two-moment approximation schema is developed for the probability distribution of M/G/1/K systems and extended to the analysis of M/G/1/K   queueing networks. The general purpose of this paper is to develop a flexible and practical transform-free approach for computing the probability distribution and performance measures of the system as well as identify the underlying properties of these systems. It is shown that for most performance measures, a sigmoid or S-shaped curve with an inflection point at ρ=1$\rho =1$ appears as K→∞$K\to \infty$. This has direct implications for the analysis and optimization of such systems. The performance modelling of the M/G/1/K queueing networks of general topologies along with extensive numerical results accompany the paper along with the linear concave performance measures for these systems.     

Improving classification performance of Support Vector Machine by genetically optimising kernel shape and hyper-parameters

Support Vector Machines (SVMs) deliver state-of-the-art performance in real-world applications and are now established as one of the standard tools for machine learning and data mining. A key problem of these methods is how to choose an optimal kernel and how to optimise its parameters. The real-world applications have also emphasised the need to consider a combination of kernels—a multiple kernel—in order to boost the classification accuracy by adapting the kernel to the characteristics of heterogeneous data. This combination could be linear or non-linear, weighted or un-weighted. Several approaches have been already proposed to find a linear weighted kernel combination and to optimise its parameters together with the SVM parameters, but no approach has tried to optimise a non-linear weighted combination. Therefore, our goal is to automatically generate and adapt a kernel combination (linear or non-linear, weighted or un-weighted, according to the data) and to optimise both the kernel parameters and SVM parameters by evolutionary means in a unified framework. We will denote our combination as a kernel of kernels (KoK). Numerical experiments show that the SVM algorithm, involving the evolutionary kernel of kernels (eKoK) we propose, performs better than well-known classic kernels whose parameters were optimised and a state of the art convex linear and an evolutionary linear, respectively, kernel combinations. These results emphasise the fact that the SVM algorithm could require a non-linear weighted combination of kernels.     

Improving the performance of k-means for color quantization

Color quantization is an important operation with many applications in graphics and image processing. Most quantization methods are essentially based on data clustering algorithms. However, despite its popularity as a general purpose clustering algorithm, k-means has not received much respect in the color quantization literature because of its high computational requirements and sensitivity to initialization. In this paper, we investigate the performance of k-means as a color quantizer. We implement fast and exact variants of k-means with several initialization schemes and then compare the resulting quantizers to some of the most popular quantizers in the literature. Experiments on a diverse set of images demonstrate that an efficient implementation of k-means with an appropriate initialization strategy can in fact serve as a very effective color quantizer.     

H performance of preview control systems

Performance of the preview control is investigated in terms of H^∞-criterion. In the infinite-horizon time setting, the limitation of H^∞ performance in preview control action is characterized with the Hamiltonian matrix. By employing the property of Hamiltonian matrix, the solvability and analytic solution for the preview full-information control problem are clarified based on an auxiliary introduced matrix Riccati equation. The property of resulting control law is illustrated with a numerical example.     

A performance index for topology and shape optimization of plate bending problems with displacement constraints

This paper presents a performance index for topology and shape optimization of plate bending problems with displacement constraints. The performance index is developed based on the scaling design approach. This performance index is used in the Performance-Based Optimization (PBO) method for plates in bending to keep track of the performance history when inefficient material is gradually removed from the design and to identify optimal topologies and shapes from the optimization process. Several examples are provided to illustrate the validity and effectiveness of the proposed performance index for topology and shape optimization of bending plates with single and multiple displacement constraints under various loading conditions. The topology optimization and shape optimization are undertaken for the same plate in bending, and the results are evaluated by using the performance index. The proposed performance index is also employed to compare the efficiency of topologies and shapes produced by different optimization methods. It is demonstrated that the performance index developed is an effective indicator of material efficiency for bending plates. From the manufacturing and efficient point of view, the shape optimization technique is recommended for the optimization of plates in bending. Received November 27, 1998?Revised version received June 6, 1999     

A hybrid deterministic/stochastic optimization approach for the shape configuration design of structures

This paper presents a computer-based shape configuration design methodology to generate optimum design of specified structures satisfying the structural performance requirements and the geometric connectivity of the model. Mathematically, this problem can be categorized as a large-scale, nonconvex and nonlinear problem. The solution methods, grouped into two main categories, deterministic and stochastic approaches, require enormous computational efforts to find global optimum designs, a matter of major importance since many local suboptimal solutions can exist. In this study, two popular methods belonging to each class are examined and compared. The methods understudied are selected as the enumeration technique for deterministic approach and the simulated annealing for the stochastic method. The advantages and disadvantages of each technique are investigated. Using the best properties of each method and an algorithm for phase change between the two, a hybrid global shape optimization approach is formulated. The hybrid method is structured to combine the enumeration method for local minimization process and the simulated annealing for global minimum search phase. The hybrid method can find global optimum designs in a robust and efficient way, contrary to the signle phase solutions examined in this study. To demonstrate the hybrid method and its effectiveness, several design examples are presented.     

Improving the I/O performance of intermediate multimedia storage nodes

The data retrieval operations in servers and proxies for Media-on-Demand applications represent a severe bottleneck, because a potentially (very) high number of users concurrently retrieve data with high data rates. In the Intermediate Storage Node Concept (INSTANCE) project, we have developed a new architecture for Media-on-Demand storage nodes that maximizes the number of concurrent clients a single node can support. We avoid the traditional bottlenecks, like copy operations, multiple copies of the same data element in main memory, and checksum calculation in communication protocols, by designing, implementing, and tightly integrating three orthogonal techniques: a zero-copy-one-copy memory architecture, network level framing, and integrated error management. In this paper, we describe the INSTANCE storage node, and present an evaluation of our mechanisms. Our experimental performance results show that the integration of these three techniques in NetBSD at least doubles the number of concurrent clients that a single storage node can serve in our testbed.     

Isogeometric shape optimization of vibrating membranes

We consider a model problem of isogeometric shape optimization of vibrating membranes whose shapes are allowed to vary freely. The main obstacle we face is the need for robust and inexpensive extension of a B-spline parametrization from the boundary of a domain onto its interior, a task which has to be performed in every optimization iteration. We experiment with two numerical methods (one is based on the idea of constructing a quasi-conformal mapping, whereas the other is based on a spring-based mesh model) for carrying out this task, which turn out to work sufficiently well in the present situation. We perform a number of numerical experiments with our isogeometric shape optimization algorithm and present smooth, optimized membrane shapes. Our conclusion is that isogeometric analysis fits well with shape optimization.     

Multiobjective shape optimization of speed humps

This paper presents a new approach to the shape optimization of road speed humps. The proposed approach is based on multiobjective genetic optimization of the hump profile while taking into account the separation phenomenon, which occurs when the front tires of the vehicle momentarily lose contact with the road surface. The optimization is carried out for speeds up to twice the authorized speed (throughout this article, the term authorized speed refers to speed limits enforced in speed reduction [bump] areas of the roads) rather than for illegally high speeds as adopted by many of the previous works. A 6-degree of freedom non-linear dynamic model is used to identify the speeds at which separation occurs, and hump profiles associated with these speeds are discarded as infeasible solutions. Three independent objective functions are selected for optimization. They include the maximum vertical acceleration experienced by the driver when crossing the hump below the authorized speed limit (to be minimized), the same vertical acceleration at speeds above the authorized speed (to be maximized), and the ascending ratio of the “speed—vertical acceleration” curve (to be maximized). These objective functions are evaluated for more than 10,000 humps of two popular profile types (sinusoidal and flat top with straight ramps) and optimum profiles for three speed limits of 20, 25, and 30km/h are determined using the multiobjective nondominated sorting genetic algorithm II. As a result, a Pareto front of at least ten optimal points is achieved for each of the two hump profile types. Furthermore, to incorporate the economical aspects of the real-world problem, Pareto optimal points for the two profile types were compared based on their lateral section areas (an indication of the manufacturing cost). The comparison shows that sinusoidal humps more often than not outdo their flat top rivals economically.     

Behavior of microdroplets in diffuser/nozzle structures

This paper investigates the behavior of microdroplets flowing in microchannels with a series of diffuser/nozzle structures. Depending on the imposed flow direction, the serial structures can act either as a series of diffusers or nozzles. Different serial diffuser/nozzle microchannels with opening angles ranging from 15° to 45° were considered. A 2D numerical model was employed to study the dynamics of the microdroplet during its passage through the diffuser/nozzle structures. The deformation of the microdroplet was captured using a level set method. On the experimental front, test devices were fabricated in polydimethylsiloxane using soft lithography. T-junctions for droplet formation, diffuser/nozzle structures and pressure ports were integrated in a single device. Mineral oil with 2% w/w surfactant span 80 and de-ionized water with fluorescent worked as the carrier phase and the dispersed phase, respectively. The deformation of the water droplet and the corresponding pressure drop across the diffuser/nozzle structures were measured in both diffuser and nozzle configurations at a fixed flow rate ratio between oil and water of 30. The results show a linear relationship between the pressure drop and the flow rate. Furthermore, the rectification effect was observed in all tested devices. The pressure drop in the diffuser configuration is higher than that of the nozzle configuration. Finally, the pressure measured results with droplet and without droplet were analyzed and compared.     

引入测评机制的综合学习粒子群优化算法

综合学习粒子群算法（CLPSO）能够改善粒子群算法多样性差且易局部收敛的问题,相比传统PSO算法能够一定程度避免算法早熟,但却存在收敛速度慢的问题。对此,提出一种CLPSO的改进算法（CLPSO‐II）,为每个粒子随机构造两个学习粒子,引入测评机制,择优学习。实验结果表明,CLPSO‐II能有效提高CLPSO的搜索效率,在处理多峰函数时,其性能优于传统粒子群算法（PSO）、全面学习粒子群算法（FIPS）和综合学习粒子群算法（CLPSO）。     

On the modelling of ribbed plates for shape optimization

The development of a homogenized plate model suitable for shape optimization is presented. The development is based on a homogenization method for layered materials with a periodic microstructure. A particular advantage of the approach is that it accommodates without significant changes the modelling of ribbed, honeycomb and perforated plates. The model is compared with others that have appeared in the literature and that are also useful in the context of the optimization of the shape and layout of plates and plate-like structures. The results indicate that the model presented here is useful in the optimization of both thick and thin plates.     

Improving the performance of LSSVM model in predicting the safety factor for circular failure slope through optimization algorithms

Circular failure can be seen in weak rocks, the slope of soil, mine dump, and highly jointed rock mass. The challenging issue is to accurately predict the safety factor (SF) and the behavior of slopes. The aim of this study is to offer advanced and accurate models to predict the SF of slopes through machine learning methods improved by optimization algorithms. To this view, three different methods, i.e., trial and error (TE) method, gravitational search algorithm (GSA), and whale optimization algorithm (WOA) were used to investigate the proper control parameters of least squares support vector machine (LSSVM) method. In the constructed LSSVM-TE, LSSVM-GSA and LSSVM-WOA methods, six effective parameters on the SF, such as pore pressure ratio and angle of internal friction, were used as the input parameters. The results of the error criteria indicated that both GSA and WOA can improve the performance prediction of the LSSVM method in predicting the SF. However, the LSSVM-WOA method, with root mean square error of 0.141, performed better than the LSSVM-GSA with root mean square error of 0.170.

     

Improving the performance of kurtosis estimator

In this communication, sample measures of kurtosis adapted by various software packages are compared for data from normal and non-normal populations. Further, two improved estimators of population kurtosis are proposed and their performance is compared with the currently used measures. The suggested estimators have considerably lower mean squared error (MSE) for various sampling designs in our simulation study. Two empirical examples are given to illustrate the usefulness of suggested estimators in practice.     

Improving the performance of reverse debugging

Reverse debugging is the software development technique that effectively helps fix bugs occurring at the nondeterministic program behavior. It allows one to examine the past states of the program without rerunning it. An implementation of reverse debugging based on deterministic replay in the QEMU 2.0 emulator is described. A number of techniques improving the debugging performance due to reducing the amount of saved data, optimized storage of system snapshots, indexing, and compressing of the event log are proposed. The emulator can work together with the interactive GDB debugger, which makes it possible to use the reverse-continue, reverse-nexti, reverse-stepi and reverse-finish commands in the course of debugging. The execution time of these commands depends on the frequency of recording the system’s state snapshots. An estimate of the optimal frequency for the reverse-continue command is obtained.