Multidisciplinary Optimization of Multibody Systems with Application to the Design of Rail Vehicles

A methodology for the design optimization of multibody systems is presented. The methodology has the following features: (1) multibody dynamics is employed to model and simulate complex systems; (2) multidisciplinary optimization (MDO) methods are used to combine multibody systems and additional systems in a synergistic manner; (3) using genetic algorithms (GAs) and other effective search algorithms, the mechanical and other design variables are optimized simultaneously. The methodology is shown to handle the conflicting requirements of rail vehicle design, i.e., lateral stability, curving performance, and ride quality, in an effective manner. By coordinating these conflicting requirements at the system level, three multibody models corresponding to each of these requirements for a rail vehicle are optimized simultaneously.     

A model for the integrated optimization of oil production systems

Typically, the optimization of oil production systems is conducted as a non-systematic effort in the form of trial and error processes for determining the combination of variables that leads to an optimal behavior of the system under consideration. An optimal or near optimal selection of oil production system parameters could significantly decrease costs and add value. This paper presents a solution methodology for the optimization of integrated oil production systems at the design and operational levels, involving the coupled execution of simulation models and optimization algorithms (SQP and DIRECT). The optimization refers to the maximization of performance measures such as revenue present value or cumulative oil production as objective functions, and tubing diameter, choke diameter, pipeline diameter, and oil flow rate as optimization variables. The reference configuration of the oil production system includes models for the reservoir, tubing, choke, separator, and business economics. The optimization algorithms Sequential Quadratic Programming (SQP) and DIRECT are considered as state-of-the-art in non-linear programming and global optimization methods, respectively. The proposed solution methodology effectively and efficiently optimizes integrated oil production systems within the context of synthetic case studies, and holds promise to be useful in more general scenarios in the oil industry.     

Software prototype for validation of machining optimization solutions obtained with meta-heuristic algorithms

Optimization of machining processes is of primary importance for increasing machining efficiency and economics. Determining optimal values of machining parameters is performed by applying optimization algorithms to mathematical models of relationships between machining parameters and machining performance measures. In recent years, there has been an increasing trend of using empirical models and meta-heuristic optimization algorithms. The use of meta-heuristic optimization algorithms is justified because of their ability to handle highly non-linear, multi-dimensional and multi-modal optimization problems. Meta-heuristic algorithms are powerful optimization tools which provide high quality solutions in a short amount of computational time. However, their stochastic nature creates the need to validate the obtained solutions. This paper presents a software prototype for single and multi-objective machining process optimization. Since it is based on an exhaustive iterative search, it guarantees the optimality of determined solution in given discrete search space. The motivation for the development of the presented software prototype was the validation of machining optimization solutions obtained by meta-heuristic algorithms. To analyze the software prototype applicability and performance, six case studies of machining optimization problems, both single and multi-objective, were considered. In each case study the optimization solutions that had been determined by past researchers using meta-heuristic algorithms were either validated or improved by using the developed software prototype.     

A Design of Genetically Oriented Fuzzy Relation Neural Networks (FrNNs) Based on the Fuzzy Polynomial Inference Scheme

In this paper, we introduce new architectures of genetically oriented fuzzy relation neural networks (FrNNs) and offer a comprehensive design methodology that supports their development. The proposed FrNNs are based on “if–then”-rule-based networks, with the extended structure of the premise and the consequence parts of the individual rules. We consider two types of the FrNN topologies, which are called FrNN-I and FrNN-II here, depending upon the usage of inputs in the premise and the consequence of fuzzy rules. Three different forms of regression polynomials (namely, constant, linear, and quadratic) are used to construct the consequence of the rules. In order to develop optimal FrNNs, the structure and the parameters are optimized using genetic algorithms (GAs). The proposed methodology is compared when the two development strategies, with separate and simultaneous optimization schemes that involve structure and parameters, are carried out. Given the large search space associated with these FrNN models, we enhance the search capabilities of the GAs by introducing the dynamic variants of genetic optimization. It fully exploits the processing capabilities of the FrNNs by supporting their structural and parametric optimization. To evaluate the performance of the proposed FrNNs, we exploit a suite of several representative numerical examples. A comparative analysis shows that the FrNNs exhibit higher accuracy and predictive capabilities as well as better modeling stability, when compared with some other models that exist in the literature.      

Intelligent computing for numerical treatment of nonlinear prey–predator models

In this study, a new computing paradigm is presented for evaluation of dynamics of nonlinear prey–predator mathematical model by exploiting the strengths of integrated intelligent mechanism through artificial neural networks, genetic algorithms and interior-point algorithm. In the scheme, artificial neural network based differential equation models of the system are constructed and optimization of the networks is performed with effective global search ability of genetic algorithm and its hybridization with interior-point algorithm for rapid local search. The proposed technique is applied to variants of nonlinear prey–predator models by taking different rating factors and comparison with Adams numerical solver certify the correctness for each scenario. The statistical studies have been conducted to authenticate the accuracy and convergence of the design methodology in terms of mean absolute error, root mean squared error and Nash-Sutcliffe efficiency performance indices.     

改进的元启发式优化算法及其在图像分割中的应用

元启发式算法自20世纪60年代提出以后,由于其具有可以有效地减少计算量、提高优化效率等优点而得到了广泛应用.该类算法以模仿自然界中各类运行机制为特点,具有自我调节的特征,解决了诸如梯度法、牛顿法和共轭下降法等这些传统优化算法计算效率低、收敛性差等缺点,在组合优化、生产调度、图像处理等方面均有很好的效果.提出了一种改进的元启发式优化算法——NBAS算法.该算法通过将传统天牛须算法（BAS）离散化得到二进制离散天牛须算法（BBAS）,并与原始天牛须算法进行混合得出.算法平衡了局部与全局搜索,有效地弥补了算法容易陷入局部最优的不足.为了验证NBAS算法的有效性,将NBAS算法与二维K熵算法结合,提出了一种快速、准确的NBAS-K熵图像分割算法.该方法解决了优化图像阈值分割函数的优化算法易陷入局部最优、算法寻优个体数多、设计复杂度高所导致的计算量大、耗时长等问题.NBAS-K熵算法与BAS-K熵算法、BBAS-K熵算法、遗传K熵算法（GA-K熵）、粒子群K熵算法（PSO-K熵）和蚱蜢K熵算法（GOA-K熵）在Berkeley数据集、人工加噪图像以及遥感图像上的实验结果表明,该分割方法不仅具有较好的抗噪性能,而且具有较高的精度和鲁棒性,能够较为有效地实现复杂图像分割.     

Adaptive augmented Lagrangian methods: algorithms and practical numerical experience

In this paper, we consider augmented Lagrangian (AL) algorithms for solving large-scale nonlinear optimization problems that execute adaptive strategies for updating the penalty parameter. Our work is motivated by the recently proposed adaptive AL trust region method by Curtis et al. [An adaptive augmented Lagrangian method for large-scale constrained optimization, Math. Program. 152 (2015), pp. 201–245.]. The first focal point of this paper is a new variant of the approach that employs a line search rather than a trust region strategy, where a critical algorithmic feature for the line search strategy is the use of convexified piecewise quadratic models of the AL function for computing the search directions. We prove global convergence guarantees for our line search algorithm that are on par with those for the previously proposed trust region method. A second focal point of this paper is the practical performance of the line search and trust region algorithm variants in Matlab software, as well as that of an adaptive penalty parameter updating strategy incorporated into the Lancelot software. We test these methods on problems from the CUTEst and COPS collections, as well as on challenging test problems related to optimal power flow. Our numerical experience suggests that the adaptive algorithms outperform traditional AL methods in terms of efficiency and reliability. As with traditional AL algorithms, the adaptive methods are matrix-free and thus represent a viable option for solving large-scale problems.     

Memory intensive AND/OR search for combinatorial optimization in graphical models

In this paper we explore the impact of caching during search in the context of the recent framework of AND/OR search in graphical models. Specifically, we extend the depth-first AND/OR Branch-and-Bound tree search algorithm to explore an AND/OR search graph by equipping it with an adaptive caching scheme similar to good and no-good recording. Furthermore, we present best-first search algorithms for traversing the same underlying AND/OR search graph and compare both algorithms empirically. We focus on two common optimization problems in graphical models: finding the Most Probable Explanation (MPE) in belief networks and solving Weighted CSPs (WCSP). In an extensive empirical evaluation we demonstrate conclusively the superiority of the memory intensive AND/OR search algorithms on a variety of benchmarks.     

Learning cooperative linguistic fuzzy rules using the best–worst ant system algorithm

Within the field of linguistic fuzzy modeling with fuzzy rule‐based systems, the automatic derivation of the linguistic fuzzy rules from numerical data is an important task. In the last few years, a large number of contributions based on techniques such as neural networks and genetic algorithms have been proposed to face this problem. In this article, we introduce a novel approach to the fuzzy rule learning problem with ant colony optimization (ACO) algorithms. To do so, this learning task is formulated as a combinatorial optimization problem. Our learning process is based on the COR methodology proposed in previous works, which provides a search space that allows us to obtain fuzzy models with a good interpretability–accuracy trade‐off. A specific ACO‐based algorithm, the Best–Worst Ant System, is used for this purpose due to the good performance shown when solving other optimization problems. We analyze the behavior of the proposed method and compare it to other learning methods and search techniques when solving two real‐world applications. The obtained results lead us to remark the good performance of our proposal in terms of interpretability, accuracy, and efficiency. © 2005 Wiley Periodicals, Inc. Int J Int Syst 20: 433–452, 2005.     

Weighted heuristic anytime search: new schemes for optimization over graphical models

Weighted heuristic search (best-first or depth-first) refers to search with a heuristic function multiplied by a constant w [31]. The paper shows, for the first time, that for optimization queries in graphical models the weighted heuristic best-first and weighted heuristic depth-first branch and bound search schemes are competitive energy-minimization anytime optimization algorithms. Weighted heuristic best-first schemes were investigated for path-finding tasks. However, their potential for graphical models was ignored, possibly because of their memory costs and because the alternative depth-first branch and bound seemed very appropriate for bounded depth. The weighted heuristic depth-first search has not been studied for graphical models. We report on a significant empirical evaluation, demonstrating the potential of both weighted heuristic best-first search and weighted heuristic depth-first branch and bound algorithms as approximation anytime schemes (that have sub-optimality bounds) and compare against one of the best depth-first branch and bound solvers to date.     

Modelisation d'algorithmes d'optimisation a strategie aleatoire

Conceptual algorithms for random search in optimization. I am proposing two conceptual algorithms for extending results about almost certainly convergence of stochastic algorithms for optimization described as follows. Let f be a map from the vector space E to the set of real number R; f is to be minimized; x⁰ is an arbitrary point of E and (ξ ^k ) a family of random vectors, if f(x^k+ξ ^k)≥f(x^k) then x^k+1=x^k or else x^k+1=x^k+ξ ^k. The inspiration for the two conceptual algorithms came from Polak's conceptual algorithm [11] for deterministic search in optimization.      

Learning Bayesian Networks: The Combination of Knowledge and Statistical Data

We describe a Bayesian approach for learning Bayesian networks from a combination of prior knowledge and statistical data. First and foremost, we develop a methodology for assessing informative priors needed for learning. Our approach is derived from a set of assumptions made previously as well as the assumption of likelihood equivalence, which says that data should not help to discriminate network structures that represent the same assertions of conditional independence. We show that likelihood equivalence when combined with previously made assumptions implies that the user's priors for network parameters can be encoded in a single Bayesian network for the next case to be seen—a prior network—and a single measure of confidence for that network. Second, using these priors, we show how to compute the relative posterior probabilities of network structures given data. Third, we describe search methods for identifying network structures with high posterior probabilities. We describe polynomial algorithms for finding the highest-scoring network structures in the special case where every node has at most k = 1 parent. For the general case (k > 1), which is NP-hard, we review heuristic search algorithms including local search, iterative local search, and simulated annealing. Finally, we describe a methodology for evaluating Bayesian-network learning algorithms, and apply this approach to a comparison of various approaches.     

A machine-order search space for job-shop scheduling problems

This paper proposes a methodology for real‐time job‐shop scheduling problems. It introduces a new classification of the scheduling methods for JSSPs with emphasis on the search methods and the significance of the search space. Subsequently, a machine‐order search space is proposed as a new framework in which different single‐machine scheduling algorithms and search methods can be incorporated to solve JSSPs. An optimization model relating makespan minimization and the proposed machine‐order search space is also described. The proposed methodology finds an optimal solution by searching a proper machine order in the machine‐order search space and scheduling the machines one by one in this order. Such an approach significantly reduces the size of the search space, and hence the computing efforts. As a result, scheduling of large JSSPs in real‐time becomes practicable.     

More efficient PEST compatible model independent model calibration

This article describes some of the capabilities encapsulated within the Model Independent Calibration and Uncertainty Analysis Toolbox (MICUT), which was written to support the popular PEST model independent interface. We have implemented a secant version of the Levenberg–Marquardt (LM) method that requires far fewer model calls for local search than the PEST LM methodology. Efficiency studies on three distinct environmental model structures (HSPF, FASST, and GSSHA) show that we can find comparable local minima with 36–84% fewer model calls than a conventional model independent LM application. Using the secant LM method for local search, MICUT also supports global optimization through the use of a slightly modified version of a stochastic global search technique called Multi-Level Single Linkage [Rinnooy Kan, A.H.G., Timmer, G., 1987a. Stochastic global optimization methods, part I: clustering methods. Math. Program. 39, 27–56; Rinnooy Kan, A.H.G., Timmer, G., 1987b. Stochastic global optimization methods, part ii: multi level methods. Math. Program. 39, 57–78.]. Comparison studies with three environmental models suggest that the stochastic global optimization algorithm in MICUT is at least as, and sometimes more efficient and reliable than the global optimization algorithms available in PEST.     

Deriving operating policies for multi-objective reservoir systems: Application of Self-Learning Genetic Algorithm

Optimal multi-reservoir operation is a multi-objective problem in nature and some of its objectives are nonlinear, non-convex and multi-modal functions. There are a few areas of application of mathematical optimization models with a richer or more diverse history than in reservoir systems optimization. However, actual implementations remain limited or have not been sustained.Genetic Algorithms (GAs) are probabilistic search algorithms that are capable of solving a variety of complex multi-objective optimization problems, which may include non-linear, non-convex and multi-modal functions. GA is a population based global search method that can escape from local optima traps and find the global optima. However GAs have some drawbacks such as inaccuracy of the intensification process near the optimal set.In this paper, a new model called Self-Learning Genetic Algorithm (SLGA) is presented, which is an improved version of the SOM-Based Multi-Objective GA (SBMOGA) presented by Hakimi-Asiabar et al. (2009) [45]. The proposed model is used to derive optimal operating policies for a three-objective multi-reservoir system. SLGA is a new hybrid algorithm which uses Self-Organizing Map (SOM) and Variable Neighborhood Search (VNS) algorithms to add a memory to the GA and improve its local search accuracy. SOM is a neural network which is capable of learning and can improve the efficiency of data processing algorithms. The VNS algorithm can enhance the local search efficiency in the Evolutionary Algorithms (EAs).To evaluate the applicability and efficiency of the proposed methodology, it is used for developing optimal operating policies for the Karoon-Dez multi-reservoir system, which includes one-fifth of Iran's surface water resources. The objective functions of the problem are supplying water demands, generating hydropower energy and controlling water quality in downstream river.     

Geometric optimization of yield-line patterns using a direct search method

A class of problems in the geometric optimization of yield-line patterns, for which the currently advocatedconjugate gradient andsequential linear programming geometric optimization algorithms fail is investigated. TheHooke-Jeeves direct search method is implemented and is demonstrated to solve such problems robustly.     

A developed system based on nature-inspired algorithms for DNA motif finding process

In this paper, recent algorithms are suggested to repair the issue of motif finding. The proposed algorithms are cuckoo search, modified cuckoo search and finally a hybrid of gravitational search and particle swarm optimization algorithm. Motif finding is the technique of handling expressive motifs successfully in huge DNA sequences. DNA motif finding is important because it acts as a significant function in understanding the approach of gene regulation. Recent results of existing motifs finding programs display low accuracy and can not be used to find motifs in different types of datasets. Practical tests are implemented first on synthetic datasets and then on benchmark real datasets that are based on nature-inspired algorithms. The results revealed that the hybridization of gravitational search algorithm and particle swarm algorithms provides higher precision and recall values and provides average enhancement of F-score up to 0.24, compared to other existing algorithms and tools, and also that cuckoo search and modified cuckoo search have been able to successfully locate motifs in DNA sequences.

     

Self-adaptive differential evolution with multi-trajectory search for large-scale optimization

In this paper, self-adaptive differential evolution (DE) is enhanced by incorporating the JADE mutation strategy and hybridized with modified multi-trajectory search (MMTS) algorithm (SaDE-MMTS) to solve large-scale continuous optimization problems. The JADE mutation strategy, the “DE/current-to-pbest” which is a variation of the classic “DE/current-to-best”, is used for generating mutant vectors. After the mutation phase, the binomial (uniform) crossover, the exponential crossover as well as no crossover option are used to generate each pair of target and trial vectors. By utilizing the self-adaptation in SaDE, both trial vector generation strategies and their associated control parameter values are gradually self-adapted by learning from their previous experiences in generating promising solutions. Consequently, suitable offspring generation strategy along with associated parameter settings will be determined adaptively to match different phases of the search process. MMTS is applied frequently to refine several diversely distributed solutions at different search stages satisfying both the global and the local search requirement. The initialization of step sizes is also defined by a self-adaption during every MMTS step. The success rates of both SaDE and the MMTS are determined and compared; consequently, future function evaluations for both search algorithms are assigned proportionally to their recent past performance. The proposed SaDE-MMTS is employed to solve the 19 numerical optimization problems in special issue of soft computing on scalability of evolutionary algorithms for large-scale continuous optimization problems and competitive results are presented.     

A study on hybrid random signal-based learning and its applications

In this paper, we propose three effective hybrid random signal-based learning (RSL) algorithms which are a combination of RSL with simulated annealing (SA) and a genetic algorithm (GA) to obtain a global solution that can be used in combinatorial optimization problems. GAs are becoming more popular because of their relative simplicity and robustness. GAs are global search techniques for non-linear optimization, but they are not good at fine-tuning solutions. RSL is similar to the reinforcement learning of neural networks using random signals. It can find an accurate solution in local search space. However, it is poor at hill-climbing, whereas simulated annealing has the ability to perform probabilistic hill-climbing. Therefore, combining them yields effective hybrid algorithms, i.e. hybrid RSL algorithms, with the merits of both. To check the generalization ability of the proposed algorithms, the optimizations of several benchmark test functions are considered, while the optimization of a fuzzy logic controller for the inverted pendulum is detailed to show the applicability of the proposed algorithms to fuzzy control.     

Evolution of neural networks for classification and regression

Although Artificial Neural Networks (ANNs) are importantdata mining techniques, the search for the optimal ANN is a challenging task: the ANN should learn the input–output mapping without overfitting the data and training algorithms may get trapped in local minima. The use of Evolutionary Computation (EC) is a promising alternative for ANN optimization. This work presents two hybrid EC/ANN algorithms: the first evolves neural topologies while the latter performs simultaneous optimization of architectures and weights. Sixteen real-world tasks were used to test these strategies. Competitive results were achieved when compared with a heuristic model selection and other Data Mining algorithms.