Search

Storing only one unique copy of the same cloud data and guaranteeing its integrity are two main goals for cloud storage auditing and deduplication schemes. In such schemes, data owners can firmly believe the data integrity by periodically auditing and the cloud server can save lots of storage space by exploiting the duplication techniques. However, when a data owner deletes or modifies his outsourced data, he should lose the ownership for the original data and should not be able to successfully retrieve this data any more. For all we know, existing cloud storage auditing and deduplication literatures fail to support the modifications of ownership, which actually occur quite often in actual cloud storage scenarios. In this paper, we propose the first deduplicated data integrity auditing scheme supporting the ownership modification. It guarantees the integrity of the outsourced data and supports the dynamic access control over the outsourced data. We employ a re-encryption algorithm and the secure identity-based broadcast encryption technology, which prevent data from being disclosed to the revoked owners, even if they previously had prior ownership of these data. The security and efficiency of our proposed scheme have been validated by detailed analysis and experiments. [ABSTRACT FROM AUTHOR]

This paper proposes an enhanced crow search algorithm (ECSA) for solving numerical and real-life engineering problems. Novelties of the proposed method are fourfold: (1) addition of an archive component in the standard crow search algorithm (CSA) to incorporate past experience of finding solution (2) formulation of non-hideout position so that crow will remain near its hideout position, (3) Rechenberg's 1/5th rule is exploited to change the flight length (instead of fixed) to speed up optimization process and (4) awareness probability is regulated to set a trade-off between local and global exploration. The performance of proposed technique is investigated on 23 benchmark functions such as unimodal, multimodal and fixed-dimension multimodal benchmark functions. The results of ECSA are compared to other state-of-the-art metaheuristic algorithms, in which ECSA outperformed other algorithms in majority of the benchmark functions. Further, to validate the effectiveness of the proposed method, ECSA has been used for optimization of proportional–integral–derivative (PID) controller. Results of ECSA–PID have been compared with conventional CSA as well as with other state-of-the-art techniques like Ziegler–Nichols (Z–N), Kitamori, ACO, multi-objective ACO, multi-objective GA and fuzzy and space gravitational optimization algorithm. The proposed algorithm is implemented on the AVR system and tested under various conditions for robustness. Consistency in the results on benchmark systems as well as on their variants and AVR system and its variants prove the robustness of the proposed method. Also, the performance of the proposed algorithm is found to be better than the existing techniques. [ABSTRACT FROM AUTHOR]

This paper proposes an improved meta-heuristic algorithm called multivector particle swarm optimization (MVPSO) for solving single-objective optimization problems. MVPSO improves particle swarm optimization (PSO) algorithm by creating more possible solutions for each particle during the optimization process. It proposes a mathematical model and new position vectors for each particle that enhance the particle movement toward the global best value. This improvement emphasizes the exploration and exploitation of the particles in the search space during the optimization process. To test the performance of MVPSO, the algorithm is then benchmarked on 23 well-known test functions including unimodal, multimodal and fixed multimodal functions at different dimensions. These benchmark functions test the exploration, exploitation, local optima avoidance and convergence features of MVPSO. MVPSO has been compared to the state-of-the-art swarm optimization algorithms as well as PSO algorithm. Experimental results indicate that in terms of robustness, stability and quality of the solution obtained, MVPSO is better than original PSO algorithm, especially as the dimension increases. Further, it shows that a MVPSO based on the multivector mathematical model is competitive with the state-of-the-art swarm optimization algorithms. Moreover, the results of the tested benchmark functions, statistical analysis and performance metrics prove that the proposed algorithm is able to explore more solutions and regions in the search space, avoiding local optima points. [ABSTRACT FROM AUTHOR]

Particle swarm optimization (PSO) is an optimization method that is most widely used to solve a number of problems in various fields such as engineering, economics and computer systems. However, due to its scalability and unsatisfying performance particularly for large-scale optimization problems; numerous PSO variants have been suggested so far, in the literature. This paper also proposes a new variant of the canonical PSO algorithm ('N-state switching PSO—NS-SPSO') that uses the evolutionary factor information to update particles velocities and, therefore, further enhance its performance. The evolutionary factor is derived by using the population distribution and the mean distance of each particle from the global best. The population distribution and the mean distance are determined through Euclidean distance. Moreover, algorithmic parameters such as inertia weight, and acceleration coefficients are assigned appropriate values at N stages (derived from exploration, exploitation, convergence and jumping out states) that improves the search efficiency and convergence speed. The proposed algorithm is applied to 12 widely used mathematical benchmark functions that demonstrate its best performance in terms of minimum evaluation error, fast convergence and low computational time. Besides these, seven high-dimensional functions and few other algorithms for large-scale optimization were considered to test the scalability of NS-SPSO algorithm. Our comparative results show that NS-SPSO performs well on low-dimensional problems and is promising for solving large-scale optimization problems. Furthermore, the proposed NS-PSO algorithm almost outperforms its closest rivals for various benchmarks. [ABSTRACT FROM AUTHOR]

The initial population distribution is an essential issue in evolutionary computation performance. Population initialization methods for grammar-guided genetic programming have some difficulties generating a representative sample of the search space, which negatively affects the overall evolutionary process. This paper presents a grammatically uniform population initialization method to address this issue by improving the initial population uniformity: the equiprobability of obtaining any individual of the search space defined by the context-free grammar. The proposed initialization method assigns and updates probabilities dynamically to the production rules of the grammar to pursue uniformity and includes a code bloat control mechanism. We have conducted empirical experiments to compare the proposed algorithm with a standard initialization approach very often used in grammar-guided genetic programming. The results report that the proposed initialization method approximates very well a uniform distribution of the individuals in the search space. Moreover, the overall evolutionary process that takes place after the population initialization performs better in terms of convergence speed and quality of the final solutions achieved when the proposed method generates the initial population than when the usual approach does. The results also show that these performance differences are more significant when the experiments involve large search spaces. [ABSTRACT FROM AUTHOR]

As an emerging technology, interval type-2 fuzzy logic systems (IT2 FLSs) have drawn great attentions in the past decade years. However, the computational intensive and time consuming type-reduction (TR) block may hinder the real applications IT2 FLSs. Unlike the most popular Karnik–Mendel (KM) iterative algorithms, the noniterative algorithms decrease the computational cost greatly. The comparison between the discrete and continuous algorithms is still an open problem. This paper compares the sum operations in discrete noniterative algorithms and the integral operations in continuous noniterative algorithms, and discovers the inner relations between the discrete and continuous noniterative algorithms. Then, three types of noniterative algorithms are adopted for performing the centroid TR of IT2 FLSs. Three computer simulations show the calculation results of sampling based discrete noniterative algorithms can accurately approximate the corresponding continuous noniterative algorithms as varying the number of sampling of primary variable appropriately, in addition, the computational efficiencies of former are much higher than the latter, which provide the potential value for designing IT2 FLSs. [ABSTRACT FROM AUTHOR]

Feature selection is one of the hottest machine learning topics in recent years. The main purposes of it are to simplify the original model, improve the readability of the model, and prevent over-fitting by searching for a suitable subset of features. There are many methods for this problem, including evolutionary algorithms and particle swarm optimization. Among them, the competitive swarm optimizer is a new optimization algorithm proposed in recent years, which is based on particle swarm optimization algorithm, and has achieved good results in high-dimensional feature selection problems, but it also has the problems of high computation time cost and easily being premature. Aiming at these problems, this paper proposes to add the crossover operator and mutation operator in the genetic algorithm to the competitive swarm optimization, so as to improve the generation speed of new individuals in the algorithm and prevent premature population. After testing on UC Irvine Machine Learning Repository, the new algorithm not only improves the computational efficiency, but also avoids the problem that the competitive swarm optimization algorithm is easy to fall into the local optimum, which greatly improves the calculation effect. [ABSTRACT FROM AUTHOR]

This paper proposes a new approach to integrating the cell formation, group layout of rectangle-shaped machines and routing selection problems. The problem is formulated as a mixed-integer program with the objective of minimizing the handling costs. Due to the computational complexity of the problem, a hybrid simulated annealing (SA) is employed to solve the problem. The sequence-pair representation, originally proposed for block placement, is utilized for solution encoding. Two placement algorithms are developed to evaluate the objective function value of an encoded solution in the SA. The first placement algorithm is based on solving a linear program embedded with a constraint reduction algorithm. The second one is a fast heuristic that can evaluate an encoded solution in much less computational time. Benchmarks selected from the literature are solved to verify the performance of the SA and to accomplish comparisons. The computational results demonstrated the high performance of both designed hybrid SA algorithms. The comparison against the literature also indicated that the flexibility incorporated into the model results in better layouts, even if the model is solved only by a solver such as CPLEX. [ABSTRACT FROM AUTHOR]

The efficiency and effectiveness of differential evolution (DE) greatly depend on the mutation operator due to the principle that different mutation operators are beneficial to different evolutionary states. However, it is not easy to automatically and effectively identify the evolutionary state. In this paper, we propose an evolutionary state estimation method (ESE) based on the correlation coefficient between the population's distributions in objective space (Δf) and solution space (Δx). To be specific, Δf consists of the distances between each individual and the current best individual based on their objective function values, while Δx includes the Euclidean distances between each individual and the current best individual based on their positions in the search space. Based on the correlation coefficient between Δx and Δf, the entire evolutionary process is classified into three kinds of state. At each generation, the evolutionary state is firstly determined according to the correlation coefficient, subsequently adaptively choosing a mutation operator from the corresponding candidate operator pool for each individual to generate its mutation vector. Moreover, a two-level selection mechanism (TLSM) is presented to get away from stagnation. The algorithm combines DE with ESE and TLSM (DEET for short) is proposed. Experimental results on twenty frequently used benchmark functions and the CEC2017 test problems show that DEET exhibits very competitive performance compared with other state-of-the-art DE variants. [ABSTRACT FROM AUTHOR]

The goal of this paper is to propose a new hybrid algorithm based on advanced quantum behaved particle swarm optimization (QPSO) technique and binary tournamenting for solving constrained optimization problems. In binary tournamenting, six different situations/options are considered and accordingly six variants of hybrid algorithms are proposed. Then to test the efficiency and performance of these algorithms and also to select the best algorithm among these, six benchmark optimization problems are selected and solved. Then the computational results are compared graphically as well as numerically. Finally, the best found algorithm is applied to solve three engineering design problems and the computational results are compared with the existing algorithms available in the literature. [ABSTRACT FROM AUTHOR]

A new meta-heuristic method, comprehensive teaching learning harmony search optimization algorithm (CTLHSO), is developed in this paper for the simultaneous reconfiguration and optimal allocation of distributed generation resources in radial distribution systems. The proposed method is a hybridization of the teaching–learning-based optimization (TLBO) and the harmony search (HS) algorithms. Primarily, eleven mathematical benchmark functions are used to test the performance of the CTLHSO algorithm. The results are then compared with that of global best artificial bee colony (G-ABC), particle swarm artificial bee colony (PS-ABC), TLBO and improved TLBO (I-TLBO) with identical parameters and initial conditions. The results show that the CTLHSO performance is better than the G-ABC, PS-ABC, TLBO and I-TLBO. Subsequently, CTLHSO is implemented on the IEEE 33-bus and 69-bus radial distribution systems for network reconfiguration and optimal placement of distributed generation resources to minimize the power losses and improve the voltage profiles. Five case studies at three different load levels are carried out. The results obtained are found to be better than those obtained with HS algorithm, genetic algorithm, refined genetic algorithm and fireworks algorithm. [ABSTRACT FROM AUTHOR]

In this paper, the consumption of energy is efficiently balanced using the high searching capability of harmony search algorithm (HSA). However, centrality in finding the route is considered as an utmost challenge in finding the important node in WSN. Hence, the path operator calculus centrality (SPOCC) is used to optimize the centrality problems in routing. The SPOCC finds the main routing path using HSA, and high centrality node is estimated by particle swarm optimization (PSO) algorithm, thereby ensuring optimal routing with reduced energy consumption. The utilization of PSO improves the lifetime of nodes using its dynamic capability. The performance of the proposed hybrid algorithm is evaluated using the various result metrics in smart grid outdoor transmission environment. The results show that the proposed hybrid algorithm obtains better improvement in terms of reduced delay and high residual energy than the existing algorithms. [ABSTRACT FROM AUTHOR]

Wireless sensor networks involved every part of application in today's life. This can detect the various type of information from environment and communicated to users in real time. Information is stored by using cloud technology and that can be accessed by the users. The information is prone to attacks as they are part of cooperative communication. Building algorithms at node level is not secured as they transfer data in open traffic. This has provided scope for this study to concentrate on algorithms that are available for denial-of-service attacks as they put down the network performance to a less minimum. The paper henceforth proposes optimized energy-based constraint DoS (denial-of-service) detection algorithm, i.e., OBES algorithm for handling denial-of-service attacks that learns the network traffic and manages the intruders. This has been compared with denial-of-service attack detection with energy constraint in WSN. The implementation has been done in NS2. The performance has been presented in terms of nodes, interval and lifetime. From results, it can be observed that the proposed OBES algorithm is efficient as it achieves more available energy, less delay, less packet loss and more lifetime for the network. From results, it can be observed that OBES algorithm performs well compared to denial-of-service attack detection with energy constraint algorithm. [ABSTRACT FROM AUTHOR]

This paper deals with the problem of finding the generators of the solution space for a system of inequalities A ⊗ x ≥ x in max-plus algebra. It provides an improved algorithm which can be used to find a smaller set of generators for the solution space by skipping a large number of invalid generators. [ABSTRACT FROM AUTHOR]

Neural network (NN) finds role in variety of applications due to combined effect of feature extraction and classification availability in deep learning algorithms. In this paper, we have chosen SVM, logistic regression machine learning algorithms and NN for EEG signal classification. Two-layer LSTM and four-layer improved NN deep learning algorithms are proposed to improve the performance in EEG classification. Novelty lies in one-dimensional gradient descent activation functions with radial basis operations used in the initial layers of improved NN which help in achieving better performance. Statistical features namely mean, standard deviation, kurtosis and skewness are extracted for input EEG collected from Bonn database and then applied for various classification techniques. Accuracy, precision, recall and F1 score are the performance metrics used for analyzing the algorithms. Improved NN and LSTM give better performance compared to all other architectures. The simulations are carried out with variety of activation functions, optimizers and loss models to analyze the performance using Python in keras. [ABSTRACT FROM AUTHOR]

The internet has provided numerous modes for secure data transmission from one end station to another, and email is one of those. The reason behind its popular usage is its cost-effectiveness and facility for fast communication. In the meantime, many undesirable emails are generated in a bulk format for a monetary benefit called spam. Despite the fact that people have the ability to promptly recognize an email as spam, performing such task may waste time. To simplify the classification task of a computer in an automated way, a machine learning method is used. Due to limited availability of datasets for email spam, constrained data and the text written in an informal way are the most feasible issues that forced the current algorithms to fail to meet the expectations during classification. This paper proposed a novel, spam mail detection method based on the document labeling concept which classifies the new ones into ham or spam. Moreover, algorithms like Naive Bayes, Decision Tree and Random Forest (RF) are used in the classification process. Three datasets are used to evaluate how the proposed algorithm works. Experimental results illustrate that RF has higher accuracy when compared with other methods. [ABSTRACT FROM AUTHOR]

During recent years, license plate recognition have been widely used as a core technology for security or traffic applications such as in traffic surveillance, parking lot access control, and information management. In this paper, Shadow Aware License Plate Recognition (SALPR) system is proposed to recognize Egyptian LP. This system achieves high recognition rate through applying shadow detection and removal, rotation adjustment and using Multilayer perceptron as a powerful tool to perform the recognition process. To show the efficiency of the proposed system, experiments have been done on numerous captured images including various types of vehicles with different lighting and noise effects. The experimental results yield 95.5 % recognition accuracy, the recognition process takes 1.6 s to recognize plate information. Most of the elapsed time used is for the license plate extraction and rotation adjustment. The results show the feasibility of the methodology followed in this paper. Performance comparison between SALPR and other LP recognition techniques shows that for most of the cases, SALPR performs better than other techniques under different lighting conditions and it shows the high robustness of the proposed algorithm. [ABSTRACT FROM AUTHOR]

This paper proposes a novel distributed differential evolution framework called distributed mixed variants (dynamic) differential evolution ( $$dmvD^{2}E)$$ . This novel framework is a heterogeneous mix of effective differential evolution ( DE) and dynamic differential evolution ( DDE) variants with diverse characteristics in a distributed framework to result in $$dmvD^{2}E$$ . The $$dmvD^{2}E$$ , discussed in this paper, constitute various proportions and combinations of DE/best/2/bin and DDE/best/2/bin as subpopulations with each variant evolving independently but also exchanging information amongst others to co-operatively enhance the efficacy of $$dmvD^{2}E$$ as whole. The $$dmvD^{2}E$$ variants have been run on 14 test problems of 30 dimensions to display their competitive performance over the distributed classical and dynamic versions of the constituent variants. The $$dmvD^{2}E$$ , when benchmarked on a different 13 test problems of 500 as well as 1,000 dimensions, scaled well and outperformed, on an average, five existing distributed differential evolution algorithms. [ABSTRACT FROM AUTHOR]

This paper proposes an improved memetic algorithm relying on ring neighborhood topology for constrained optimization problems based on our previous work in Cai et al. (Soft Comput (in press), ). The main motivation of using ring neighborhood topology is to provide a good balance between effective exploration and efficient exploitation, which is a very important design issue for memetic algorithms. More specifically, a novel variant of invasive weed optimization (IWO) as the local refinement procedure is proposed in this paper. The proposed IWO variant adopts a neighborhood-based dispersal operator to achieve more fine-grained local search through the estimation of neighborhood fitness information relying on the ring neighborhood topology. Furthermore, a modified version of differential evolution (DE), known as 'DE/current-to-best/1', is integrated into the improved memetic algorithm with the aim of providing a more effective exploration. Performance of the improved memetic algorithm has been comprehensively tested on 13 well-known benchmark test functions and four engineering constrained optimization problems. The experimental results show that the improved memetic algorithm obtains greater competitiveness when compared with the original memetic approach Cai et al. in (Soft Comput (in press), ) and other state-of-the-art algorithms. The effectiveness of the modification of each component in the proposed approach is also discussed in the paper. [ABSTRACT FROM AUTHOR]

In this paper, multiplier-less nearly perfect reconstruction tree structured non-uniform filter banks (NUFB) are proposed. When sharp transition width filter banks are to be implemented, the order of the filters and hence the complexity will become very high. The filter banks employ an iterative algorithm which adjusts the cut off frequencies of the prototype filter, to reduce the amplitude distortion. It is found that the proposed design method, in which the prototype filter is designed by the frequency response masking method, gives better results when compared to the earlier reported results, in terms of the number of multipliers when sharp transition width filter banks are needed. To reduce the complexity and power consumption for hardware realization, a design method which makes the NUFB totally multiplier-less is also proposed in this paper. The NUFB is made multiplier-less by converting the continuous filter bank coefficients to finite precision coefficients in the signed power of two space. The filter bank with finite precision coefficients may lead to performance degradation. This calls for the use of suitable optimization techniques. The classical gradient based optimization techniques cannot be deployed here, because the search space consists of only integers. In this context, meta-heuristic algorithm is a good choice as it can be tailor made to suit the problem under consideration. Thus, this design method results in near perfect NUFBs which are simple and multiplier-less and have linear phase and sharp transition width with very low aliasing. Also, different non-uniform bands can be obtained from the tree structured filter bank by rearranging the branches. [ABSTRACT FROM AUTHOR]

Penalized regression-based clustering model (PRClust) is an extension of "sum-of-norms" clustering model. Three previously proposed heuristic algorithms for solving PRClust are: (1) DC-CD, which combines the difference of convex programming (DC) and a coordinate-wise descent algorithm (CD), (2) DC-ADMM, which combines DC with the alternating direction method of multipliers (ADMM), and (3) ALT, which uses alternate optimization. DC-CD uses p × n × n - 1 / 2 scalar slack variables to solve PRClust, where n is the number of data and p is the number of their features. In each iteration of DC-CD, these slack variables and cluster centers are updated using a second-order cone programming (SOCP). DC-ADMM uses p × n × n - 1 scalar slack variables. In each iteration of DC-ADMM, these slack variables and cluster centers are updated with a standard ADMM. In this paper, first, PRClust is reformulated into an equivalent model. Then, a novel heuristic algorithm is proposed to solve the reformulated model. Our proposed algorithm needs only n × n - 1 / 2 scalar slack variables which are much less than those of DC-CD and DC-ADMM, and updates them using a simple equation in each iteration of the algorithm. Therefore, updating slack variables in our proposed algorithm is less time-consuming than that of DC-CD and DC-ADMM. Our proposed algorithm updates only cluster centers using an unconstrained convex quadratic problem. Therefore, our proposed unconstrained convex quadratic problem is much smaller than the SOCP of DC-CD which is used to update both cluster centers and slack variables. Meanwhile, ALT updates cluster centers using a SOCP, while our proposed algorithm updates cluster centers using an unconstrained convex quadratic problem with the same number of variables. Solving an unconstrained convex quadratic problem is less time-consuming than a SOCP with the same number of variables. Our experimental results on 12 datasets confirm that the runtime of our proposed algorithm is better than that of DC-ADMM and DC-CD. [ABSTRACT FROM AUTHOR]

It is well known that electric generators consume huge amounts of energy every year. Nowadays, research for the unit commitment problem (UCP) has become a very important task in a power plant. However, the existing optimal methods for solving UCP are very easy to fall into local optimum, resulting in poor performance. Moreover, as no separate layering of economic load distribution, the existing algorithms are very inefficient. Toward this end, a new algorithm named improved simulated annealing particle swarm optimization (ISAPSO) is proposed in this paper. The proposed algorithm consists of a two-layer structure which is designed to simplify the complex problem of UCP. Specifically, in the upper layer, the algorithm based on elitist strategy PSO and SA is much easier to jump out of the local optimum when solving UCP and thus gets a better solution. In the lower layer, convex optimization approach is used to improve the search efficiency of ISAPSO. Furthermore, several methods are also designed to solve the problem-related constraints, which can save a lot of computing resources. Finally, the experimental results show that the cost performance of ISAPSO is better than that of the existing algorithms. [ABSTRACT FROM AUTHOR]

In this paper, quadripartitioned neutrosophic numbers (QNNs) are introduced, operations over them have been defined and some of their properties have been studied. QNNs have been implemented in defining the quadripartitioned neutrosophic weighted arithmetic averaging operator and the quadripartitioned neutrosophic weighted geometric averaging operator for ranking the final scores of the alternatives and choosing the most suitable alternative among them. The concept of interval-valued possibility quadripartitioned single-valued neutrosophic soft sets has been utilized to propose an algorithm for a multi-criteria group decision making problem. In this approach, entropy-based weights are allocated to the elements of the universe of discourse under consideration. Finally, the obtained results are compared with existing ones by means of comparative studies. [ABSTRACT FROM AUTHOR]

With the advent of neural networks, fuzzy logic, genetic algorithms, and other soft computing methodologies, many researchers have demonstrated successful designs for intelligent control strategies that appear to outperform traditional linear feedback controls; however, industry has mostly ignored the new technology due to the lack of stability guarantees (required in most formal engineering risk-management processes). In this paper, we offer a Lyapunov-stability framework where one can place any arbitrary computational algorithm and still get guarantees of uniformly ultimately bounded (UUB) signals. We would expect an intelligent algorithm to be well-designed such that the proposed framework would not come into play unless unanticipated disturbances affect the system. But even if the intelligent algorithm was poorly designed, the resulting performance (inside our framework) would just look similar to that of a typical nonlinear neural-adaptive control. In our strategy, the intelligent algorithm trains a cerebellar model articulation controller (CMAC) neural network with arbitrarily bounded weights in order to achieve good performance, while a second CMAC trains in parallel using direct-adaptive-control laws in order to provide stability (even in the case of the first CMAC weights reaching their imposed bound). We test the strategy using a previously proposed ad hoc CMAC weight smoothing strategy, serving as the intelligent algorithm, with simulations and experiment controlling a two-link flexible-joint robot. [ABSTRACT FROM AUTHOR]

In this paper, an extraction algorithm of color disease spot image based on Otsu and watershed is proposed to overcome the problem of excessive segmentation in the traditional watershed algorithm. The proposed algorithm converts the color space of the color spot image to calculate the component gradient that is not interfered by the reflected light in the new color space. Then, final gradient image is obtained from the gradient image reconstructed by open and close operation under the different-sized structural elements. The label from the final gradient image is extracted by Otsu algorithm, and then, the H-minima transform is used to modify the labeled image. The modified gradient image is transformed with a label by the watershed algorithm. Finally, the extraction of the disease spot is implemented. The experimental results show that the proposed approach obtains accurate and continuous target contour and reaches the requirement of human visual characteristics. Compared with other similar algorithms, the proposed algorithm can effectively suppress the impact of reflected light, optimize the extraction results of disease spot, better maintain the information of disease spot image, and improve the robustness and the applicability. [ABSTRACT FROM AUTHOR]

Bi-level programming problems are a special class of optimization problems with two levels of optimization tasks. These problems have been widely studied in the literature and often appear in many practical problem solving tasks. Although many applications fit the bi-level framework, however, real-life implementations are scarce, due mainly to the lack of efficient algorithms able to handle effectively this NP-hard problem. Several solution approaches have been proposed to solve these problems; however, most of them are restricted to the continuous case. Motivated by this observation, we have recently proposed a Co-evolutionary Decomposition-based Algorithm (CODBA) to solve bi-level combinatorial problems. CODBA scheme has been able to bring down the computational expense significantly as compared to other competitive approaches within this research area. In this paper, we further improve CODBA approach by incorporating a local search procedure to make the search process more efficient. The proposed extension called CODBA-LS includes a variable neighborhood search to the lower-level task to help in faster convergence of the algorithm. Further experimental tests based on the bi-level production–distribution problems in supply chain management model on a set of artificial and real-life data turned out to be effective on both computation time and solution quality. [ABSTRACT FROM AUTHOR]

In this paper, a clustering algorithm based on emotional preference and migratory behavior (EPMC) is proposed for data clustering. The algorithm consists of four models: the migration model, the emotional preference model, the social group model and the inertial learning model. First, the migration model calculates the probability of individuals being learned, so that individuals can learn from the superior. Second, the emotional preference model is introduced to help individuals find the most suitable neighbor for learning. Third, the social group model divides the whole population into different groups and enhances the mutual cooperation between individuals under different conditions. Finally, the inertial learning model balances the exploration and exploitation during the optimization, so that the algorithm can avoid falling into the local optimal solution. In addition, the convergence of EPMC algorithm is verified by theoretical analysis, and the algorithm is compared with four clustering algorithms. Experimental results validate the effectiveness of EPMC algorithm. [ABSTRACT FROM AUTHOR]

In this paper, we study the improvement of a storage location strategy through the use of big data technology, including data collection, cluster analysis and association analysis, to improve order picking efficiency. A clustering algorithm is used to categorize the types of goods in orders. Classification is performed based on the turnover of goods, value, sales volume, favorable commodity ratings, whether free shipping is provided and whether cash on delivery is supported. An association algorithm is used to determine the relationships among goods by studying the habits of consumers who buy them. A method for improving the class-based storage strategy is proposed. The picking distance of the improved storage strategy is compared with that of the traditional strategy via simulation experiments. The picking efficiency is shown to be enhanced by the improved strategy. [ABSTRACT FROM AUTHOR]

This paper proposes a subsystem-based structure evolution algorithm for stable IIR digital filter design. The method designs the IIR digital filter through optimizing the filter structure. A filter structure is defined as the connection of the subsystems. A subsystem is a 2-order IIR digital structure with uncertain parameters. Subsystems are randomly connected with the constraints of no feedback branches between subsystems. The subsystem's parameters and the connections between subsystems are optimized by evolutionary algorithms. An adaptive multiple-elites-guided composite differential evolution algorithm (AMECoDEs) is used to optimize this problem. Four classic types of filters, lowpass, highpass, bandpass and bandstop filters are designed. Five state-of-the-art evolutionary algorithms are compared. The simulation results show that AMECoDEs holds the first place on comprehensive performance and convergence rate. At the same time, the poles of filters all reside within the unit circle, which validates the stability of the IIR digital filters. [ABSTRACT FROM AUTHOR]

This paper presents a new variant of particle swarm optimization (PSO) algorithm named guided adaptive search-based particle swarm optimizer (GuASPSO). In this algorithm, the personal best particles are all divided into a linearly decreasing number of clusters. Then, the unique global best guide of a given particle located at a cluster is obtained as the weighted average calculated over other clusters' best particles. Since the clustered particles are being well-distributed over the whole search space in the clustering process, there would be a moderate distance between each particle and its unique global best guide, contributing the particles neither to be trapped in local optima nor engaged in a drift leading to lose diversity in the search space. In this approach, the number of clusters is high at the early iterations and is gradually decreased by lapse of iterations to less stress the diversity factor and further stress the fitness role to cause the particles to better converge to the optimal point. Holding this balance between global and personal bests' role to attract the particles, on the one hand and between convergence and diversity, on the other hand, can hold a better exploration–exploitation balance in the proposed algorithm. To test the performance of GuASPSO, four popular meta-heuristic algorithms, including genetic algorithm, gravitational search algorithm, gray wolf optimizer, and PSO algorithm as well as 23 standard benchmark functions as the test beds, are employed. The experimental results validated GuASPSO as a robust well-designed algorithm to handle various optimization problems. [ABSTRACT FROM AUTHOR]

In high-dimensional data, many of the features are either irrelevant to the machine learning task or are redundant. These situations lead to two problems, firstly overfitting and secondly high computational overhead. The paper proposes a feature selection method to identify the relevant subset of features for the machine-learning task using wrapper approach. The wrapper approach uses the Binary Bat algorithm to select the set of features and One-pass Generalized Classifier Neural Network (OGCNN) to evaluate the selected set of features using a novel fitness function. The proposed fitness function accounts for the entropy of sensitivity and specificity along with accuracy of classifier and fraction of selected features. The fitness function is compared using four classifiers (Radial Basis Function Neural Network, Probabilistic Neural Network, Extreme Learning Machine and OGCNN) on six publicly available datasets. One-pass classifiers are chosen as these are computationally faster. The results suggest that OGCNN along with the novel fitness function performs well in the majority of cases. [ABSTRACT FROM AUTHOR]

Reconstructing gene regulatory networks (GRNs) from gene expression data is an important and challenging problem in system biology. In general, the problem of reconstructing GRNs can be modeled as an optimization problem. Recurrent neural network (RNN) has been widely used for GRNs. However, in a real GRN, the number of genes is very large and the relationships between genes are usually very sparse. In this paper, we design a memetic algorithm to learn partial parameters of RNN, and develop a framework based on the least absolute shrinkage and selection operator (LASSO) to reconstruct GRNs based on RNN, which is termed as MALASSORNN-GRN. In the LASSO, the task of reconstructing GRNs is decomposed into a sparse signal reconstructing problem. In the experiments, MALASSORNN-GRN is applied on synthetic data and well-known benchmark datasets DREAM3 and DREAM4. The effect of parameters on MALASSORNN-GRN is discussed, and MALASSORNN-GRN is compared with three other algorithms which are all state-of-the-art RNN learning algorithms. The results show that MALASSORNN-GRN performs best and is capable of reconstructing large-scale GRNs. [ABSTRACT FROM AUTHOR]

The Steiner minimal tree (SMT) problem is an NP-hard problem, which is the best connection model for a multi-terminal net in global routing problem. This paper presents a unified algorithm for octagonal and rectilinear SMT construction based on hybrid transformation strategy (HTS) and self-adapting particle swarm optimization. Firstly, an effective HTS is proposed to enlarge the search space and improve the convergence speed. Secondly, the proposed HTS in the evolutionary process may produce an ineffective solution, and consequently the crossover and mutation operators of genetic algorithm (GA) based on union-find sets is proposed. Thirdly, a self-adapting strategy that can adjust the acceleration coefficients is proposed to further improve the convergence and the quality of the proposed algorithm. Finally, the hybrid transformation can be applied to GA and the proposed algorithm can be applied to rectilinear architecture. To our best knowledge, the proposed algorithm is the first unified algorithm to solve the SMT construction under both octagonal and rectilinear architecture. The experimental results show that the proposed algorithm can efficiently provide a better solution for SMT problem both in octagonal and rectilinear architectures than others. Moreover, the algorithm can obtain several topologies of SMT, which is beneficial for optimizing congestion in VLSI global routing stage. [ABSTRACT FROM AUTHOR]

The nonparametric evidential regression (EVREG) method provides flexible forms of prediction regarding the value of output, allowing the output of training instances to be partially unknown. However, the superfluous training instances still have negative effects on the parameter learning in EVREG. To relax this limitation, this paper introduces an interactive nonparametric evidential regression (IEVREG) algorithm with instance selection. More specifically, the significance of an instance is firstly measured by defining the evaluation functions, taking into account both the prediction accuracy of regression model and the spatial information between that instance with other ones. According to a search strategy, the instances with high degree of significance are then selected to maximize an objective function. Different from existing instance selection methods, the selection of training instances is synchronously accomplished with the parameter learning in IEVREG, rather than just a separated data preprocessing operation as traditional methods do. Furthermore, the noise and redundant instances can be simultaneously removed and the performance of IEVREG is robust to the order of presentation of instances in raw data set. Experimental results show that the proposed IEVREG algorithm has appropriate prediction accuracy, while performing well selection of the representative training instances from the raw data set. Simulations on synthetic and UCI real-world data sets validate our conclusions. [ABSTRACT FROM AUTHOR]

The identification of ideal values for algorithm-specific parameters required for the functioning of metaheuristic approaches at their optimum performance is a difficult task. This paper presents the application of a recently proposed teaching–learning-based optimisation (TLBO) algorithm to determine the lowest factor of safety (FS) along a critical slip surface for soil slope. TLBO is a nature-inspired search algorithm based on the teaching–learning phenomenon of a classroom. Four benchmark slopes are reanalysed to test the performance of the TLBO approach. The results indicate that the present technique can detect the critical failure surface and can be easily implemented by practitioners without fine-tuning the parameters that affect the convergence of results. Statistical analyses indicate a drastic decrease in uncertainty and the number of function evaluations in the estimation of the FS over previous approaches. [ABSTRACT FROM AUTHOR]

Enterprise-wide supply chain planning problems naturally exhibit a multi-level decision network structure, where the upper level of a hierarchy may have his objective function and decision space partly determined by other levels. In addition, each planner's control instruments may allow him to influence the policies at other levels and thereby to improve his own objective function. As a tool, bi-level programming is applied for modeling decentralized decisions in which two decision makers make decisions successively. In this paper, we specifically address bi-level decision-making problems with budget constraint as an attractive feature in the context of enterprise-wide supply chain. We first describe the typical bi-level linear programming problem (BLLPP) and its optimal solution to the penalty function problem, and then, a cooperative decision-making problem in supply chain is modeled as BLLPP. A particle swarm optimization-based computational algorithm is designed to solve the problem, and the numerical example is presented to illustrate the proposed framework. [ABSTRACT FROM AUTHOR]

Recently, many evolutionary algorithms (EAs) have been used to solve clustering problem. However, compared to K-means which is a simple and fast clustering algorithm, these EA-based clustering algorithms take too much computation time. In addition, the parameters of most EAs are fixed or dynamical adjustment by a simple method on different datasets, and it will cause that the performance of these algorithms is good on some datasets but bad on others. In order to overcome these disadvantages, a novel parallel object-tracking behavior algorithm (POTBA) based on dynamics is proposed in this paper. The proposed algorithm consists of three different models which are parallel object-tracking model, parameters self-learning model and energy model, respectively. First, the parallel object-tracking model is designed to accelerate the computation speed and avoid local minima. Second, the parameters of POTBA are self-adjusted by the parameters self-learning model. Third, the energy model is introduced to depict energy changes of POTBA during the evolutionary process. The correctness and convergence properties of POTBA are analyzed theoretically. Moreover, the effectiveness and parallelism of POTBA are evaluated through several standard datasets, and the experimental results demonstrate that POTBA exhibits superior overall performance than five other state-of-the-art algorithms. In the aspect of search performance, the results of POTBA are better than other comparison algorithms on most used datasets. In the aspect of time performance, the time overhead of POTBA is significantly reduced through parallel computing. When the number of processors increases to 32, the computation time of POTBA is less or close to K-means which is the fastest comparison algorithm. [ABSTRACT FROM AUTHOR]

In this paper, a swarm-based optimization algorithm, normative fish swarm algorithm (NFSA) is proposed as an effective global and local search technique to obtain effective global optima at superior convergence speed. Artificial fish swarm algorithm is a recent swarm-based algorithm that imitates the behavior of fish swarm in the real environment. Many improvements and modifications have been proposed regularly on fish swarm algorithm to improve the performance of optimization, but to date, existing fish swarm algorithms have not yet obtained a global optimum at extremely superior convergence rates. Hence, there still remains a huge potential for the development of fish swarm algorithm. NFSA hybridizes the characteristics of PSOEM-FSA with the normative knowledge as the complementary guidelines for more accurate and precise global optimum approaching. NFSA further improves the adaptive parameters in term of visual and step to balance the contradiction between the exploration and exploitation processes. Random initialization of the initial population is introduced to spread out the solution candidates of artificial fishes over the solution space. For the purpose of experiments, ten benchmark functions have been used in the evaluation process. The proposed algorithm is then compared with other related algorithms published in the literature. The results proved that the proposed NFSA achieved superior results in terms of convergence rate and best optimal solution on a majority of the tested benchmark functions in comparison with other comparative algorithms. [ABSTRACT FROM AUTHOR]

The set-union knapsack problem (SUKP) is a variation of the 0–1 knapsack problem (KP) in which each item is a set of elements, each item has a nonnegative value, and each element has a nonnegative weight. The weight of one item is given by the total weight of the elements in the union of the items' sets. The SUKP accommodates a number of real-life applications and is more complicated and computationally difficult than the 0–1 KP. In this paper, we propose a novel hybrid Jaya algorithm with double coding (DHJaya) to solve the SUKP. In the DHJaya, double coding is used to represent the individual, which includes the solutions for solving the SUKP by adopting a mapping function. The Jaya algorithm and differential evolution algorithm are combined to improve the exploration ability. To enhance the exploitation ability, the Cauchy mutation is performed on some individuals. Meanwhile, an improved repairing and optimization algorithm (MS-GROA) is proposed to repair the infeasible solutions and optimize the feasible solutions. We test the DHJaya using three sets of SUKP instances to demonstrate its efficiency, and the obtained results are compared with those in the previous study. Extensive experiments show a remarkable performance of the proposed approach. [ABSTRACT FROM AUTHOR]

Spectral clustering algorithm suffers from high computational complexity due to the eigen decomposition of Laplacian matrix and large similarity matrix for large-scale datasets. Some researches explore the possibility of deep learning in spectral clustering and propose to replace the eigen decomposition with autoencoder. K-means clustering is generally used to obtain clustering results on the embedding representation, which can improve efficiency but further increase memory consumption. An efficient spectral algorithm based on stacked autoencoder is proposed to solve this issue. In this paper, we select the representative data points as landmarks and use the similarity of landmarks with all data points as the input of autoencoder instead of similarity matrix of the whole datasets. To further refine clustering result, we combine learning the embedding representation and performing clustering. Clustering loss is used to update the parameters of autoencoder and cluster centers simultaneously. The reconstruction loss is also included to prevent the distortion of embedding space and preserve the local structure of data. Experiments on several large-scale datasets validate the effectiveness of the proposed method. [ABSTRACT FROM AUTHOR]

Many iterative algorithms like Picard, Mann, Ishikawa are very useful to solve fixed point problems of nonlinear operators in real Hilbert spaces. The recent trend is to enhance their convergence rate abruptly by using inertial terms. The purpose of this paper is to investigate a new inertial iterative algorithm for finding the fixed points of nonexpansive operators in the framework of Hilbert spaces. We study the weak convergence of the proposed algorithm under mild assumptions. We apply our algorithm to design a new accelerated proximal gradient method. This new proximal gradient technique is applied to regression problems. Numerical experiments have been conducted for regression problems with several publicly available high-dimensional datasets and compare the proposed algorithm with already existing algorithms on the basis of their performance for accuracy and objective function values. Results show that the performance of our proposed algorithm overreaches the other algorithms, while keeping the iteration parameters unchanged. [ABSTRACT FROM AUTHOR]

Meta-heuristic algorithms are widely viewed as feasible techniques to solve continuous large-scale numerical optimization problems. Grey wolf optimizer (GWO) is a relatively new stochastic algorithm with only a few parameters to adjust that can be easily used for global optimization. This paper presents an efficient and robust GWO (ERGWO) variant to solve large-scale numerical optimization problems. Inspired by particle swarm optimization, a nonlinearly adjustment strategy for parameter control is designed to balance exploration and exploitation. Additionally, a modified position-updating equation is presented to improve convergence speed. The performance of ERGWO is verified on 18 benchmark large-scale numerical optimization problems with dimensions ranging from 30 to 10,000, 30 benchmarks from CEC 2014, 30 functions in CEC 2017, respectively. Numerical experiments are performed to compare ERGWO to the basic GWO algorithm, other GWO variants, and other well-known meta-heuristic search techniques. Simulations demonstrate that the proposed ERGWO algorithm can find high quality solutions with low computational cost and very fast convergence. [ABSTRACT FROM AUTHOR]

To provide a new numerical algorithm for solving elliptic partial differential equations (PDEs), the Legendre neural network (LNN) and improved extreme learning machine (IELM) algorithm are introduced to propose a Legendre improved extreme learning machine (L-IELM) method, which is applied to solving elliptic PDEs in this paper. The product of two Legendre polynomials is chosen as basis functions of hidden neurons. Single hidden layer LNN is used to construct approximate solutions and its derivatives of differential equations. IELM algorithm is used for network weights training, and the algorithm steps of the proposed L-IELM method are summarized. Finally, in order to evaluate the present algorithm, various test examples are selected and solved by the proposed approach to validate the calculation accuracy. Comparative study with the earlier methods in literature is described to verify the superiority of the presented L-IELM method. Experiment results show that the proposed L-IELM algorithm can perform well in terms of accuracy and execution time, which in addition provides a new algorithm for solving elliptic PDEs. [ABSTRACT FROM AUTHOR]

The widely used smartphone with powerful positioning capability makes it easy for a user to find his precise physical location. However, this may reveal a user's geo-location information, making the real-time tracking of the user possible. For example, on the basis of a sequence of numbers (i.e., Cell-IDs) received in the Cell-ID positioning, an entity can gain access to a person's movement routes without his consent. We argue that if the trajectory of a person is traced, then all his visits may be exposed. Therefore, trajectory cloaking against the mobile positioning is urgently necessary. In this paper, we propose a dummy base station replacement (DBSR) algorithm. It mainly uses the idea of dummy trajectory anonymity and is achieved by replacing the true Cell-ID provided by the network with a fake but nearby Cell-ID. We also implement our DBSR algorithm on an Android-based smartphone to evaluate its performance. Experimental results show that the DBSR algorithm can efficiently tackle the privacy breach caused by the single-base-station positioning in cellular network. [ABSTRACT FROM AUTHOR]

Inspired by the advantages of hybrid intelligent optimization methods, this paper at first proposes a hybrid differential evolution with particle swarm optimization (DEPS) to solve a two-stage hybrid flow shops scheduling problem. On the basis of analyzing the convergence and optimization scheme of DEPS, the Kalman filter algorithm and a multi-stage learning strategy are then creatively fused into DEPS, namely KLDEPS, to enhance the running performance of the algorithm. The introduction of the Kalman filter enriches the diversity of individuals and enhances the neighborhood search ability of the algorithm, and the combination with the multi-stage learning strategy has beneficial effect on jumping out of the local optimal scheme. To make the proposed KLDEPS more suitable for a real manufacturing environment, the constraints of queueing time between two stages, different job sizes and processing time are imposed on the scheduling problem. The performance of the proposed KLDEPS is evaluated by comparing with two other high-performing intelligent optimization algorithms. The computational results reveal that the proposed KLDEPS outperforms the other two algorithms both in solutions' quality and convergence rate. [ABSTRACT FROM AUTHOR]

The firefly algorithm (FA) is a powerful optimization tool. However, the existing FA and its variants seldom take advantage of intermediate data generated during algorithm iteration. In this paper, the centroid opposition-based learning with a two-point full crossover (CCOBL) is proposed to make full use of the favor information of the candidate solutions. It adopts a centroid opposition computing for considering the search information of population and a two-point full crossover for using the favor information in the candidate solution and its opposite. Then, the CCOBL is incorporated into the partially attracted firefly algorithm. The proposed algorithm is tested on the CEC' 2013 benchmark suite and a real-world optimization problem and is compared with some state-of-the-art FA algorithms and other up-to-date opposition-based evolutionary algorithms. The experimental results demonstrate the effectiveness of the CCOBL and the better performance of the proposed algorithm. [ABSTRACT FROM AUTHOR]

An effective way to analyze and apprehend structural properties of networks is to find their most critical nodes. This paper studies a bi-objective critical node detection problem, denoted as Bi-CNDP. In this variant, we do not make any assumptions on the psychology of decision makers and seek to find a set of solutions which minimize the pairwise connectivity of the induced graph and the cost of removing these critical nodes at the same time. After explicitly stating the formulation of Bi-CNDP, we first prove the NP-hardness of this problem for general graphs and the existence of a polynomial algorithm for constructing the ε -approximated Pareto front for Bi-CNDPs on trees. Then different approaches of determining the mating pool and the replacement pool are proposed for the decomposition-based multi-objective evolutionary algorithms. Based on this, two types of decomposition-based multi-objective evolutionary algorithms (MOEA/D and DMOEA- ε C ) are modified and applied to solve the proposed Bi-CNDP. Numerical experiments on sixteen famous benchmark problems with random and logarithmic weights are firstly conducted to assess different types of the mating pool and the replacement pool. Besides, computational results between two improved algorithms, i.e., I-MOEA/D and I-DMOEA- ε C , demonstrate that they behave differently on these instances and I-DMOEA- ε C shows better performance on the majority of test instances. Finally, a decision-making process from the perspective of minimizing the pairwise connectivity of the induced graph given a constraint on the cost of removing nodes is presented for helping decision makers to identify the most critical nodes for further protection or attack. [ABSTRACT FROM AUTHOR]

Teaching–learning-based optimization (TLBO) has been widely used to solve global optimization problems. However, the optimization problems in various fields are becoming more and more complex. The canonical TLBO is easy to be trapped in the local optimum when dealing with these problems. In this paper, a new TLBO algorithm with collective intelligence concept introduced is proposed, namely collective information-based TLBO (CIBTLBO). CIBTLBO uses the information from the top learners to form CITeachers and uses the neighborhood information of each learner to form NTeachers, and these teachers help other learners learn in the teacher phase. Furthermore, CITeacher also helps in the learner phase. To demonstrate superiority of the proposed algorithm, experiments on 28 benchmark functions from CEC2013 are carried out, and the benchmark functions are set to 10, 30, 50 and 100 dimensions, respectively. The results show that the proposed CIBTLBO algorithm outperforms the other previous related algorithms. [ABSTRACT FROM AUTHOR]

The median problem on a graph is an important field of facility location problem in which the aim is to find the best place among vertices and interior points of edges such that the summation of weighted distances are minimized. This point is called absolute median point. Beforehand, it was shown that the absolute median of a graph is always at a vertex of it. This theorem is called vertex optimality theorem. In this paper, a graph with fuzzy weights is considered. Then, the fuzzy vertex optimality theorem is proved for this problem. That is, it is proved that an absolute median of this fuzzy graph is always at a vertex of it. Thereafter, this theorem is proved for the general form problem where the weights of vertices and the lengths of the arcs are all fuzzy numbers. It is worthwhile to note that this is the first time that this theorem is proved for fuzzy location problems. Also, by proving this fuzzy theorem and applying it on existing algorithms for obtaining median point, the computational complexity is reduced. Finally, by solving some numerical examples, the concept of these proved theorems and their proofs are illustrated. [ABSTRACT FROM AUTHOR]

Unlike traditional analytical optimization techniques, bio-inspired multi-objective optimization has recently enjoyed an intense interest regarding wireless sensor network (WSN) issues. Network lifetime and target coverage are among the major concerns in many well-established scenarios of WSNs, such as routing and node deployment. For set covers scenario in WSNs, however, little attention has been paid to the role of multi-objective requirements. In this paper, we take a step toward exploring the feasibility of such bio-inspired algorithms for solving multi-objective set covers problem in WSNs. The main contributions of this study are twofold. First, we extend the set covers problem and address it with three issues: network lifetime, target coverage, and network connectivity as a multi-objective set covers (MOSC) formulation. To the best of our knowledge, this is the first effort to define such a general multi-objective set covers problem. Second, we design and elaborate four well-known multi-objective optimization algorithms from evolutionary and swarm intelligence community to tackle the formulated MOSC problem. All characteristic components of the adopted algorithms are tailored specifically to handle the formulated problem. Further, a self-adaptive heuristic mutation operator is proposed to attain and emphasize the strength of the algorithms in terms of network lifetime and coverage probability. Extensive simulations are performed to test and demonstrate the performance of the designed algorithms to tackle the problem appropriately. [ABSTRACT FROM AUTHOR]