Search

Convolutional neural network (CNN) is one of the well-known deep learning algorithms that uses convolutional filters to extract the features in the images. The most important issue when training CNN is the overfitting which prevents the model from generalization to unseen data. This paper addressed this issue by proposing a novel hybrid approach that uses bees algorithm (BA) to optimize the regularization parameter and weight regularization factor to adjust the regularization value in each convolutional layer and fully connected layer resulting in a hybrid algorithm called bees algorithm regularized convolutional neural network (BA-RCNN). It was applied to three different datasets for classification or predictions purposes and showed an improvement in the validation and testing accuracy leading to a lower difference with the training accuracy which means the overfitting is reduced comparing to the original CNN. Applying the BA-RCNN algorithm to 'Cifar10DataDir' improved the validation accuracy from 80.34% for the original CNN to 82.80% for the hybrid BA-RCNN algorithm, in the electrocardiogram the improvement was from 87.80 to 90.47% and both datasets were used for classification. Furthermore, the hybrid BA-RCNN algorithm was applied to predict the porosity percentage based on artificial porosity images and the results showed that the validation accuracy was improved from 81.67% for the original CNN to 87.33% for the hybrid BA-RCNN algorithm. [ABSTRACT FROM AUTHOR]

The lemur optimizer (LO) is a new metaheuristic algorithm that is inspired by the two primary locomotive patterns of the lemurs. However, LO suffers from improper balance between exploration and exploitation phases and sub-optimal solution stagnation due to poor exploration capability. To address these limitations, an improved lemur optimizer (ILO) with dynamic scaling factors and Gaussian mutation is introduced in this paper. The dynamic scaling factors balance the exploration and exploitation transition by dynamically adjusting their values throughout the optimization process. The proposed algorithm incorporates Gaussian mutation to induce diversity in the population and hence improve the exploration capability. The efficiency of the proposed approach is tested on 33 functions including classical benchmark and CEC-2019 test functions, and a comprehensive comparison is made with six other well-known metaheuristic algorithms. The simulation and statistical results indicate superior performance of ILO in achieving optimal solutions compared to the competing algorithms. The algorithm is further employed to address two constrained classical engineering design problems, demonstrating its robustness. [ABSTRACT FROM AUTHOR]

In a complex environment with multiple feasible paths, planning a globally optimal path and performing dynamic path planning for robots is challenging. Therefore, this paper proposed a new adaptive Differential Evolution algorithm combined with K-modes clustering, BP neural network, and other strategies (KBPDE) to overcome this issue. The proposed KBPDE algorithm applies a new population initialization strategy based on K-modes clustering and a two-population strategy to enhance the population diversity and prevent the algorithm from falling into local optima. A BP neural network model is developed to obtain the appropriate mutation scale factor F at each generation G. Based on the positional relationships among individuals in the current generation, a novel mutation strategy is presented. Finally, an adaptive population size strategy is introduced to raise the algorithm's efficiency. In a complicated environment with several feasible paths, the experimental results demonstrate that KBPDE can obtain the globally optimal path in contrast to GA, DE, and the excellent versions of DE. In a partially known environment, the EKBPDE can plan the path successfully. [ABSTRACT FROM AUTHOR]

Tunicate Swarm Algorithm (TSA) is a new bio-based optimization technique that has proven not only to be able to compete with other methods but has also shown successful performance in classic design engineering problems/benchmark test problems. However, like some population-based methods, TSA tends to be trapped in local optima, converging to global optima in a long time, unbalanced exploitation/exploration, and the inability to effectively solve high-capacity engineering problems. In this paper, the M-TSA, which is a Modified version of the TSA, is proposed to overcome such problems. M-TSA was developed in three steps. The first is the new movement strategy that improves the movement of tunicates with a spiral movement, the second is the new herd strategy that improves the herd movement of tunics with the Levy movement, and the third is the consideration of the FAD effect. In this study, the efficiency and robustness of the M-TSA algorithm is tested on the CEC'17 test suite, six real-life design engineering problems, and two complex power system engineering problems. The test results were compared with other techniques reported in the literature and with the original TSA. Comparing the results from the M-TSA technique with other techniques proves the effectiveness of M-TSA with better exploration/exploitation balance and optimal solution finding. In this paper, MATLAB 2020b software is used for optimization problems simulation. [ABSTRACT FROM AUTHOR]

To alleviate the 'curse of dimensionality' and speed up the calculation of the total supply capability curve (TSC curve), this paper proposes a novel critical flow-based fast algorithm for distribution networks. Firstly, the critical flow and the TSC curve are found to have good correspondence through the comparison of their concepts and models. Secondly, a critical flow-based algorithm with three phases is proposed for calculating the TSC curve: (i). Determine the equivalent flow network of the distribution network; (ii). Solve critical flows of the equivalent flow network through a dimension reduction-sampling algorithm and obtain the corresponding TSC curve operating points; (iii). Calculate the flow value of the critical flow and plot the TSC curve. Finally, the proposed algorithm is verified by the expanded IEEE RBTS BUS4 case with soft open point (SOP). The application of the TSC curve in distribution network planning is demonstrated. In addition, the correctness and rapidity of the proposed algorithm are verified by comparing it with the existing sampling algorithm, and the reasons for the performance improvement are investigated. This paper applies the network flow theory to calculating the power supply capability of distribution networks for the first time, which simplifies the calculation of the TSC curve and increases the computational efficiency significantly. [ABSTRACT FROM AUTHOR]

This paper presents a range of novel reinforcement learning algorithms derived from the actor–critic approach. These modified algorithms effectively utilize the available information to enhance performance. Our proposed framework introduces several key components to the traditional actor–critic model, including an underlying model learner, selector, tuner, and estimator. The estimator employs an approximate value function and the learned underlying model to estimate the values of all actions at the next state. The selector approximates the optimal action at the next state, which is then utilized by the actor to optimize its policy. In contrast to the conventional actor–critic algorithm where the actor focuses solely on policy optimization and the critic performs value-function approximation and policy evaluation, our selector–actor–critic algorithm employs a selector to approximate the optimal action at the current state, thereby influencing the actor's policy updates. Furthermore, our tuner–actor–critic algorithm incorporates a critic and a model-learner to approximate the action-value function and the dynamics of the underlying environment, respectively. The tuner then utilizes this information to adjust the value of the current state–action pair. In the estimator–selector–actor–critic algorithm, we develop intelligent agents based on the concepts of lookahead and intuition. Lookahead is utilized in estimating the values of available actions at the next state, while intuition guides the maximization of the probability of selecting the approximate optimal action. Through simulation experiments, we evaluate the performance of these algorithms, and the results demonstrate the superiority of the estimator–selector–actor–critic approach over other existing algorithms. [ABSTRACT FROM AUTHOR]

With the rapid development of pure electric vehicles, how to improve the range has become the focus of research. Due to the large shell mass of the battery pack shell (BPE), it is necessary to optimize its structure in turn. In this paper, the BPE model was reconstructed by improving the upper shell material and adding reinforcements following the direction of the leaf vein fibers of the Magnolia Grandiflora. A quasi-Monte Carlo method based on Sobol sequences and a Latin hypercube design with variance sensitivity analysis was used to determine the seven design variables. The 122 data sets were trained and predicted using the basic gradient descent algorithm combined with the conjugate direction method, and the predictions were compared with static mechanical simulations for sharp cornering conditions on bumpy roads. The results showed that the BPE weight reduction ratio was 19.5%, and the maximum stress reduction ratio was 27.49%, and the displacement reduction ratio was 29.29% respectively, which satisfied the material requirements. It had a 20.88% increase in first-order mode frequency, which effectively prevented resonance. [ABSTRACT FROM AUTHOR]

It is of great theoretical value and practical significance to study the location of cascading failures in chemical material networks with large and complex characteristics. Previous mainly used the method of obtaining all information of nodes, maximum likelihood estimation function, and distance centrality to locate faults. These methods need to determine the status of all nodes in the network or perform operations on the nodes in the entire network related to the observation node, the localization process is more complicated and costly, and the results are not accurate enough. This paper proposes a positioning method based on the breadth-first search strategy (BFS)-time-reversal backpropagation algorithm to study the cascading fault location of chemical material networks. Firstly, the cascade fault propagation model of the chemical material network is constructed according to the relevant theory of the complex network, and the observation node selection strategy selects the observation node. Secondly, use the BFS to search out the fault area, then make assumptions about the time delay of the propagation process, and use the time-reversal backpropagation algorithm to filter the nodes in the fault area to realize the cascade fault location of the chemical material network. Finally, change the observation node selection strategy, analyze the positioning accuracy under each strategy, and select the optimal selection strategy for positioning using the method in this paper. Case analysis shows that this method can effectively realize the cascading fault location in the chemical material network. [ABSTRACT FROM AUTHOR]

Since the past few years, several metaheuristic algorithms, inspired by the natural processes, have been introduced to solve different complex optimization problems. Studying and comparing the convergence, computational burden and statistical significance of those metaheuristics are helpful for future algorithmic development and their applications. This paper focuses on comparing the optimization performance of classical dragonfly algorithm (DA) and its seven different variants, i.e., hybrid memory-based dragonfly algorithm with differential evolution (DADE), quantum-behaved and Gaussian mutational dragonfly algorithm (QGDA), memory-based hybrid dragonfly algorithm (MHDA), chaotic dragonfly algorithm (CDA), biogeography-based Mexican hat wavelet dragonfly algorithm (BMDA), hybrid Nelder–Mead algorithm and dragonfly algorithm (INMDA) and hybridization of dragonfly algorithm and artificial bee colony (HDA) while solving 80 CEC-2021 benchmark problems. It is observed that the convergence rates of different variants of DA algorithm vary, and the corresponding computational times for such variations are also evaluated. This paper finally ranks DA and its variants according to their convergence efficiency and Friedman test. The DADE, QGDA, BMDA and DA evolve out as the most efficient algorithms for solving the considered CEC-2021 benchmark problems. [ABSTRACT FROM AUTHOR]

To realize monitoring of trenching depth, this paper built an online monitoring system and measured depth at three levels of 25 cm, 30 cm, and 35 cm. The test results show that the measured data have large fluctuations and low precision and are smaller than the actual value. For the above problems, this paper proposes a comprehensive noise reduction filtering algorithm based on wavelet noise reduction and Kalman filtering, in which the output of the preprocessing part of the wavelet noise reduction is used as the input of the postprocessing part of the Kalman filter. This article first selects the optimal combination of wavelet threshold denoising parameters such as wavelet function, decomposition level, and threshold denoising rule. Then the optimal Q/R value and the combination of Q and R parameters for the Kalman filter reprocessing part are explored. Finally, the comprehensive filtering correction algorithm and Kalman filtering algorithm are used to process the above-measured data. By comparing the results of the two algorithms, we found that for the local data, the variance decreased by 0.0207, 0.0398, and 0.0609, and the mean square deviation decreased by 0.0609, 0.0854, and 0.0457. The mean values of the two algorithms are very close. Similarly, for the global data, the variance decreased by 0.0182, 0.0549, 0.0428, and the mean square deviation decreased by 0.0359, 0.041, 0.0323. There is no significant difference in means between the two algorithms. The results show that the comprehensive algorithm improves the quality of trenching depth data and has certain significance for precision agriculture. [ABSTRACT FROM AUTHOR]

The accurate model of the solar PV system is the principal organ that describes the performance of this resource. Several approaches based on optimizing algorithms were considered valuable tools to illustrate the I–V curve for improving the photovoltaic models. Their electrical parameters are estimated using optimization algorithms referring to the experimental database or manufacturer's datasheet. This paper proposes a novel developed a photovoltaic model based on improved arithmetic optimization algorithm (IAOA) to extract the solar cell parameters. Also, an experimental test bench is presented for obtaining the measured illustration of the I–V characteristics. Thus, the root mean square error value that describes the difference between measured and estimated results is considered the objective function for two different models, the simple-diode model and the one-diode model. The proposed IAOA results are compared with other research papers and optimization algorithms. Furthermore, the evaluation of the proposed IAOA has been discussed considering several statistical analysis tests. The presented results show that the effectiveness and accuracy of IAOA results are excellent, and their I–V characteristics coincide with experimental data. Moreover, the results obtained by the proposed algorithm show its high superiority in optimizing the solar cell parameters under a variety of operating conditions. [ABSTRACT FROM AUTHOR]

Estimating the parameters of photovoltaic (PV) models accurately is vital to increase the effectiveness of PV systems. During the past few years, many approaches have been developed to solve this problem. However, due to the presence of nonlinearity and multi-modality in the problem, the estimated parameters are usually not very accurate and reliable. Therefore, this paper proposes a novel hybrid algorithm called chaos game optimization-least squares (CGO-LS) algorithm. The novelty of CGO-LS is that it adopts a cascade estimation strategy based on parameter decomposition. By the aid of this decomposition, CGO-LS combines a sophisticated nonlinear optimization capability of chaos game optimization (CGO) and the power of the optimal linear least squares (LS) estimator. LS focuses directly on estimating linear parameters, thus reducing the workload of CGO and helping to increase its convergence speed. To validate the performance of CGO-LS, it is employed to estimate the parameters of four PV models, including single-diode, double-diode, three-diode models, and PV module model. The results obtained by CGO-LS are compared with those of CGO, six state-of-the-art metaheuristics, their hybridized versions with LS, as well as some reported results in the literature. The overall results show that CGO-LS possesses superior estimation performance and excellent robustness in emulating experimental datasets. [ABSTRACT FROM AUTHOR]

Combined heat and power economic emission dispatch (CHPEED) can obtain good economic and environmental benefits, but the dispatch problem presents non-convex, nonlinear, multi-constrained and multi-objective characteristics. Thus, a two-stage cooperative multi-objective differential evolutionary algorithm (TCADEA) is proposed in this paper. The algorithm uses a two-stage framework: the first stage uses a two-population strategy to divide the population into an elite population and an ordinary population, where the elite population is used to obtain better target values and the ordinary population is used to search the target space to ensure the diversity of the population and update the two populations by different adaptive differential operators. In addition, the ε constraint processing technique is used to handle the constraints. The second stage combines two populations into one and generates offspring through the constrained dominance principle (CDP) and adaptive differential evolution to maintain well-distributed population. The actual case results show that the TCADEA algorithm reduces $0.034 × 106, $0.008 × 106, $0.07 × 106 and 113 × 105 lb, 102 × 105 lb, 155 × 105 lb in fuel cost and emissions compared to MODE, NSGA-II, and TOP, respectively. [ABSTRACT FROM AUTHOR]

To solve the problem that uneven illumination is inevitably to reduce the image quality, an adaptive illumination correction algorithm based on a combined model of bottom-hat and improved gamma transform is proposed to process and improve the image quality. In this paper, the multi-scale Gaussian function is adopted to estimate the image illumination components. Then, the two-dimensional gamma function is improved, and the coefficients of gamma transform are dynamically adjusted by using the obtained illumination components to realize the adaptive correction of the image. Finally, the corrected image is fused with the image transformed by the bottom-hat. Through a large number of experiments, it is verified that the proposed method can retain more detailed information, and the corrected picture does not produce small noise points with a better fidelity effect. [ABSTRACT FROM AUTHOR]

Several different setups may be introduced as candidate options in a process plan for manufacturing a workpiece. It is required to employ an efficient optimization method to decide on the most competent setups. In the existing methods for optimizing the process plan, the main problem is the large search space size, which reduces the speed of finding the optimal solution or causes getting caught in the local minima. To remedy this problem, an optimization method has been developed in the present study based on a combined approach by employing the simulated annealing and the permutation-based setup planning algorithms. A new cost function has been introduced in this approach. The proposed method also has other advantages and capabilities in line with minimizing the manufacturing error and obtaining a uniquely optimized solution. A comparative study has been conducted to illustrate the capability of the proposed method. The results indicate that the rate of convergence to the optimal solution increases and the performance in finding the solution improves by using the method described in this paper. [ABSTRACT FROM AUTHOR]

This paper presents a reliable algorithm for solving the inverse kinematic (IK) problem of continuum robots. Admittedly, there are an infinite number of configurations to get the robot's end point (EP) pose. Therefore, the proposed algorithm introduces a new solution that helps narrow the infinite number of the IK solutions down to a finite one by taking into account the first (n - 1) sections' sets of reachable EP poses, which are selected and created one-by-one from the first to (n - 1) th section, in which the Particle Swarm Optimization method is adopted to randomly pick up the first IK solution. However, a specific criterion has been accorded to select the appropriate IK solution (i.e., redundancy resolution) among the existing varieties that perfectly fits within the targeted trajectory. Furthermore, numerical experiments are performed on a Cable-Driven Continuum Robot for tracking a trajectory in a free and confined environment, pointing up on IK, multiple IK solutions and redundancy resolution. The obtained results demonstrated that the proposed algorithm is computationally accurate and efficient in obtaining all-inclusive IK solutions and redundancy resolution of continuum robots in comparison to previously proposed approaches. It is noteworthy to say that the proposed algorithm can be applied to any kind of continuum robots and most importantly regardless of its number of sections, which makes the developed algorithm a sovereign remedy for the IK problem of these robots. [ABSTRACT FROM AUTHOR]

Functional operation of integrated network during circumstances like disturbance and improper operation has to be addressed properly to improve the integrated network performance and safety. One of the situations is a island state and requires to secure human beings and device components that are working in integrated network. This paper presents a k-means clustering machine learning (kmc-ML) to obtain upper limit (UL) and lower limit (LL) parameters for continuous change in fuzzy variable. Machine learning (ML) will support to change the position of fuzzy set towards universe of discourse (UoD) either in LL or UL, and hence, mismatch observation can make for better estimation of integrated network state. Identification of state at every abnormal mode (AbM) operation will depend on UoD set to achieve effective outcome. ML-based soft fuzzy set control vector (S-FSCV) decides the integrated network state, i.e. island state or not. [ABSTRACT FROM AUTHOR]

Spectrum scarcity is one issue that needs to be tackled for the uninterrupted growth of the Internet of Things (IoT). Limited licensed spectrum available and hugely populated unlicensed bands are becoming a bottleneck for the expansion of IoT. Cognitive radio networks (CRNs) have been widely researched and proven to solve the issue of spectrum scarcity and spectrum underutilization. In CRNs, users make use of the vacant portion of the spectrum intelligently. Users in CRN may find different sets of available channels based on their location. To initiate the communication process, a common channel is required between the users to exchange handshake information and data. This process is termed rendezvous and can be achieved either by using a centralized controller such as a common control channel (CCC) or by the channel hopping (CH) technique. The CH technique is preferred due to many drawbacks of the CCC approach. In the CH technique, users visit their available channels according to a pre-defined method with the aim of achieving guaranteed rendezvous in a minimum time. In this paper, an asymmetric synchronous CH algorithm has been proposed for guaranteed rendezvous between the users that provide fair rendezvous opportunity and full rendezvous diversity. Compared with the existing CH rendezvous algorithms, evaluation of the proposed algorithm reveals that the proposed algorithm achieves better time to rendezvous (both in terms of maximum and average time) and provides rendezvous fairness among the users and full rendezvous diversity. [ABSTRACT FROM AUTHOR]

The wireless sensor network is a collection of diverse sensor nodes in which data is communicated via physical sensors and relayed using self-configured protocols. The network becomes congested because myriad nodes are involved in the sensing and data transmission. Congestion decreases the packet delivery ratio (PDR) virtue of significant packet drop and increases transmission delay. The lost packets are retransmitted at the cost of additional energy and time, which provokes the reduction in network performance and the extravagance of energy resources. Traditional techniques for congestion control are inflexible and lack adaptive capability. To overcome these constraints, this paper proposes a PFP-PID controller (Progressive Fuzzy Particle Swarm Optimization (PSO) Proportional Integral Derivative (PID)). The initial position of the reference particle is determined using fuzzy logic, which accelerates PSO convergence by escaping numerous initial iterations. Additionally, this hybrid mechanism based on PSO and fuzzy logic controller is employed to generate the ideal PID controller design with rapid convergence, resulting in an optimized transmission rate for sensor nodes. The PFP-PID is implemented to alleviate congestion by simulating it in network simulator (NS3) and comparing it to cuckoo fuzzy PID (CFPID), PSO-neural PID, fuzzy-PID, and traditional PID. The simulation results demonstrate that the PFP-PID is scalable and has outclassed existing mechanisms and significantly improved performance, increasing PDR by 4.89%, minimizing packet drop, delay, and active queue length deviation by 36.83%, 25.32%, and 10.84%, respectively, against CFPID. [ABSTRACT FROM AUTHOR]

The non-smooth dynamic model can effectively handle unilateral constraints in multi-body systems, but its complex mathematical definition is difficult to be applied in engineering applications. To reduce the difficulty of this model's application and to improve its calculation accuracy, an implicit integration algorithm is proposed in this paper to solve non-smooth dynamic models more effectively. Using velocity and impulse as independent variables, the linear complementary form of non-smooth dynamic equations, which are expressed as a set of differential–algebraic equations (DAEs), is derived based on the complementary contact law. By constructing an approximate velocity of a system at the next point in time, the proposed algorithm obtains the displacement in the next time step by weighting the approximate velocity constructed before and the real velocity at current point in time, and then updates the other state variables of the next time step. The accuracy of the proposed algorithm and the stability of the improved time-stepping method are verified by three numerical experiments. The results show that the time-stepping method based on the proposed implicit integration algorithm has higher accuracy and less time cost than Moreau's midpoint method for solving linear complementarity based non-smooth dynamic models. The third study case examines the computational ability of the algorithm proposed in this paper for contact phenomena with friction. The algorithm proposed in this paper can be applied in engineering applications that require an accurate solution of contact force. [ABSTRACT FROM AUTHOR]

This paper proposes an effective wrapper approach by integrating the Levy flight and evolutionary selection operators into the Moth Flame Optimization (MFO) algorithm. The main purpose is to solve the Feature Selection (FS) problem in medical applications. FS is used as a preprocessing step in a data mining process to improve the performance of the classification system by eliminating irrelevant and redundant features from a dataset. This simplifies the generalization process and reduces the complexity of the generating models. Furthermore, it speeds up the learning process and reduces the cost for additional hardware resources. However, FS is a challenging NP-hard problem because the search space grows exponentially with an increase in the number of features. In this paper, the swarming behavior of the moths is utilized by using the MFO optimizer as a search strategy within a wrapper approach. The Levy flight operator is proposed to enhance the exploratory behavior of the MFO and mitigate the stagnation in local minima. Different selection mechanisms: random selection (RS), tournament selection (TS), and roulette wheel selection (RWS) methods, are investigated to decrease the bias of the MFO algorithm toward exploitation. These selection operators are proposed in the combination of the Levy flight in the form of four different FS methods LBMFO-R1, LBMFO-R2, LBMFO-TS, and LBMFO-RWS. The proposed methods are validated using 23 medical data sets from well-regarded data repositories. The comprehensive results and various comparisons reveal that the Levy flight and selection operators have a great positive impact on the performance of the MFO. They enhance the exploration, convergence trends, and diversity of solutions. [ABSTRACT FROM AUTHOR]

To reveal the complete limits of a flow network, this paper proposes the concept of critical flow. Firstly, it is found that a flow network is in a critical state when it reaches its maximum flow, and the network cannot afford any flow increase. In fact, not only maximum flow but many other network flows, i.e., blocking flows, indicate the critical states of a flow network. From a new viewpoint, the concept of critical flow is proposed, which unifies the maximum flow and the blocking flow. Secondly, the mathematical model, solution algorithms, and visualization method for critical flow are proposed. An analytical algorithm and a numerical algorithm are proposed to calculate the complete critical flows. In addition, a critical flow curve is designed to visualize the complete limits of a network, with the top point being the maximum flow. Finally, case studies are conducted to illustrate the critical flow and its typical applications in flow networks and real networks. It is concluded that the complete limits of a flow network are the range of critical flows, not a single maximum flow value. The proposed critical flows describe the complete limits of a flow network for the first time. Critical flow can be used for network monitoring and network planning. In monitoring, it reflects the margin of the flow network more accurately than maximum flow. In planning, it evaluates the flow transmission capability more accurately than maximum flow. [ABSTRACT FROM AUTHOR]

In recent years, distributed energy has been gradually applied in residential electricity consumption, and smart devices have been rapidly developed among residential households. This paper establishes a model of optimal scheduling system for building load, taking into account the needs of grid side and customer side, and takes the total cost of electricity consumption and total deviation of electricity consumption of building residents as the optimization objectives. Then, a multi-objective equilibrium optimizer algorithm integrated competition mechanism (CMOEO) is proposed to solve the multi-objective optimization scheduling problem of building loads with multi-decision variables. The algorithm is improved from the equilibrium optimizer algorithm by introducing the competition mechanism, replacing the original exploration and exploitation constant coefficients with dynamic factors and integrating differential evolution. Through the comparison experiments of different algorithms and the different capacity of photovoltaic energy storage system (PESS), the results show that the CMOEO has fast convergence speed, strong merit-seeking ability and better population diversity. In addition, the configuration of PESS is conducive to reducing the cost of electricity for building residents and playing the role of peak-shaving and valley-filling. [ABSTRACT FROM AUTHOR]

This paper presents the handling of nonlinear system identification problem based on Volterra model. Gradient-based algorithms are generally applied to solve system identification problems. However, these algorithms have the limitation of getting trapped in the local minimum. In the presented work, a novel population-based optimization algorithm popularly known as sine cosine algorithm (SCA) is being utilized for the identification of nonlinear discrete-time system. The SCA uses mathematical sine and cosine functions for the purpose of optimization. SCA is responsible for the creation of multiple random solutions and moving them towards best solution while maintaining proper balance between the exploitation and exploration phases of optimization. The performance evaluation of the applied SCA is carried out in terms of coefficient evaluation, mean square error and convergence profile. Two different examples for nonlinear system are presented in this work so as to demonstrate the validity of the employed algorithm. Performance analysis of the proposed approach with the existing state-of-the-art algorithms proves that the SCA outperforms the other algorithms. [ABSTRACT FROM AUTHOR]

In general, conventional differential scheme (CDS) is used for the three-terminal transmission line protection; however, because of CT saturation under external faults, the relay loses its security. Further, adaptive direct under reaching transfer trip (DUTT)-based algorithms are reliable until the overlappings between zone-1 settings are ensured. However, the overlapping of zone-1 may vanish due to infeed current and under high resistance fault, resulting in an unprotected region in the line. Hence, in this paper, a new differential current circle diagram-based protection algorithm is proposed by utilizing the magnitude and phase angle of three-terminal currents instead of the only magnitude as in CDS. The effectiveness of the proposed algorithm is tested on 400 kV, 50 Hz using PSCAD/EMTDC and MATLAB, and the corresponding results show the authenticity of this algorithm to discriminate and classify the faults correctly. Finally, the comparative section reveals that the proposed algorithm provides reliable results compared to the CDS and adaptive DUTT protection schemes. [ABSTRACT FROM AUTHOR]

Weighted K-nearest neighbor algorithm (WKNN) is the most commonly used algorithm based on Wi-Fi. However, the accuracy of traditional algorithm is not very high due to the inflexible strategy and the ignorance of distance corresponding relationship. In this paper, an improved WKNN algorithm using fingerprint method for indoor positioning called DC-WKNN is proposed, which based on flexible K selection strategy and distance compensation. The fingerprint database is filled by fuzzy C-Means (FCM) in the offline training stage, and the online positioning stage consists of two positioning phases named coarse positioning and fine positioning. An improved K selection strategy based on relative density is firstly proposed to choose the value of K flexibly according to different test point in coarse positioning phase. Then, an WKNN algorithm based on distance compensation (DC) for adjusting the characteristic distance and realizing fine positioning is proposed. The experimental results show that the average positioning error of proposed algorithm is less than 1.1m, which reduce by 11.1%, 10% and 6.5%. [ABSTRACT FROM AUTHOR]

Due to the complexity of indoor environments, there are still some problems in indoor positioning, such as low position calculation efficiency and positioning accuracy. To solve the above two problems, an improved indoor positioning algorithm based on backpropagation neural network (BPNN) is proposed in this paper. Computation efficiency is increased by area partition and area screening; at the same time, positioning accuracy is improved by using BPNN algorithm considering the environment factors. For comparison, an improved K-nearest neighbor (KNN) algorithm based on area partition and area screening is also proposed. Experiment results show that the improved BPNN-based algorithm can obtain better positioning performance compared with both the traditional KNN method and the improved KNN method. [ABSTRACT FROM AUTHOR]

Feature selection is a problem that continues to be explored by researchers until an exact and fast method is found. A variety of criteria based on statistics, consistency, information, distance and similarity were used to propose new feature selection methods. In this paper, a new criterion inspired by the crowding distance used in multi-objective optimization problems is introduced. A crowding distance is computed for each feature based on its neighbors. All computed distances are sorted to assess the variability of each feature. Features with the highest values exhibit more diversity and capture some pattern in the data; these features are the most relevant ones. Based on this criterion, a simple novel supervised and unsupervised filter feature selection method called Crowding Clustering Algorithm (CCA) is proposed. Mainly, crowding distance is used to determine relevant features, and clustering is used to eliminate redundant features. Several variants of each method are also proposed in order to allow user to select the most suitable method based on specific needs and constraints of the application. We conduct an extensive comparative study on diverse datasets to evaluate the effectiveness and efficiency of the proposed methods. Our proposed unsupervised filter method achieved the best performance on 18 out of 26 datasets, with a mean accuracy above 81%, significantly outperforming five state-of-the-art methods, which range from 67 to 73%. This reflects an 8–14% accuracy improvement. Similarly, our supervised filter methods achieved the highest accuracy on 18 out of 26 datasets, exceeding 87%, representing a 2–5% improvement over six state-of-the-art methods (82–85% accuracy). In terms of processing time, both our unsupervised and supervised methods show a slight increase compared to state-of-the-art methods, with differences ranging from 0.02 to 0.2 s, which are typically negligible. Overall results demonstrate that the proposed methods outperform or achieve comparable results to commonly used filter feature selection techniques in terms of accuracy and computational time. The findings provide valuable insights and guidance for further research in this domain. [ABSTRACT FROM AUTHOR]

In this paper, we propose EOSMA, an equilibrium optimizer (EO)-guided slime mould algorithm (SMA), to improve efficiency by optimizing the search of SMAs. Firstly, the exploration and exploitation abilities of SMA are adapted to enhance the exploration in the early stage and the exploitation in the later stage. Secondly, the anisotropic search operator of SMA is replaced with the search operator of the equilibrium optimizer (EO) to guide search space of SMAs (EOSMA). Finally, the stochastic difference mutation operator is added to help the SMA algorithm escape from local optima after population location updates and to increase the diversity of the population at the late iteration. The performance of EOSMA is evaluated on CEC2019 and CEC2021 functions and nine engineering design problems. The results show that for the challenging CEC2019 functions, EOSMA significantly outperforms the 15 well-known comparison algorithms in terms of Mean and Min metrics. For the latest CEC2021 functions, EOSMA significantly outperforms IMODE, LSHADE, and LSHADE_cnEpSin in terms of Min metrics. In particular, for all nine engineering problems, EOSMA is able to find feasible solutions satisfying all constraints and significantly outperforms the advanced comparison algorithms such as SMA, EO, MPA, IGWO, AGPSO, MTDE in terms of solution accuracy, speed, and robustness. [ABSTRACT FROM AUTHOR]

Today, all smartphones, notebooks, or other communication devices could connect to the cloud, so the data are accessible everywhere. When these devices are interconnected through the internet, they make an Internet of Things (IoT) network that exchanges data among network nodes and other services. IoT has a broad application area from smart applications to various industrial usages. However, the high volume of data transferred in the IoT network makes it crucial to implement mechanisms to transfer the data safe and secure. Enciphering is one of the best techniques to offer end-to-end security. Considering an IoT network, nodes have restricted resources, and applying classical cryptography methods are costly and not efficient, so lightweight block ciphers are one of the sophisticated solutions to overcome security drawbacks in this scope. In this paper, ten lightweight algorithms involve AES, PRESENT, LBlock, Skipjack, SIMON, XTEA, PRINCE, Piccolo, HIGHT, RECTANGLE tested to evaluate their performance for key factors such as memory usage (RAM and ROM), energy consumption, throughput, and execution time for both encryption and decryption modes over cloud transmission. We have done simulations using Raspberry Pi 3 and Arduino Mega 2560 as the leading devices in the IoT scope. As a result, this paper will help IoT developers to choose the right platform and enciphering algorithm to set up a secure network due to multiple factors like energy and memory usage, especially for software platforms. [ABSTRACT FROM AUTHOR]

In order to accurately detect the false data injection attack for node voltage in the cyber-physical power system, an attack detection method based on recurrent neural network (RNN) is proposed in this paper. The features of voltage data of power topology nodes and the construction method of attack vector are studied. Based on the fast regression algorithm, the best strategy for the data integrity attack of the specific node voltage is solved. The RNN is used to reconstruct time series of the node voltage, and set the threshold of error between the input data and the output data. Then, by calculating whether the error between the reconstructed output data and the original input data exceeds the threshold, it is determined whether the system has suffered data integrity attack. Finally, the feasibility and effectiveness of the detection method proposed in this paper were verified by simulated attack experiments. [ABSTRACT FROM AUTHOR]

In estimating the performance of parallel microchannels, non-uniformity of flow distribution (ϕ) and pressure drop (Δp) are critical parameters. These parameters are affected by flow rate, the inlet flow configuration and microchip geometric factors. This paper, in a constant rate, and configuration of fluid entrance into the chip, examines the effect of geometric—channels and manifolds—parameters on ϕ and Δp. For this purpose, by employing a set of CFD models modeled in COMSOL, a finite element software, the effect of each geometric factor is examined separately. The results show that increasing the number of channels, the distance between channels, the concavity, and reducing the width of the channel has a positive effect on the distribution of flow between the channels. Moreover, reducing the convexity, increasing the number of channels, increasing the distance between channels, and declining the width of the channels leads to an increase in Δp. Eventually, a hybrid artificial neural network (MLP-ICA) is trained by means of 245 CFD models to derive the model of changing of ϕ and Δp with the geometric parameters. The MLP-ICA predictive model of ϕ the R2 of 0.9912 and the MSE of 4.1 × 10–4 demonstrate the higher accuracy of this model than MLP Model with R2 of 0.9466 MSE of 7.9 × 10–3. [ABSTRACT FROM AUTHOR]

In this paper, a look-up table-based direct torque control (DTC) of interior permanent magnet synchronous motor (IPMSM) drive is reported. An approach is made to develop a sensorless four-quadrant electric drive for a wide operating speed range. In order to ensure efficient operation of IPMSM, maximum torque per ampere and field weakening algorithms are employed. The constraints associated with pure integrator in DTC are discussed, and an adaptive controller-based approach to estimating stator flux is proposed. A bidirectional DC–DC converter is implemented to allow four-quadrant operation of drive. In order to realize sensorless operation, stator current-based model reference adaptive control is employed to estimate the motor speed. Speed reversal, torque reversal, above-base-speed operation and low-speed operation of IPMSM drive are considered in this paper. The proposed algorithm is implemented in the MATLAB and Simulink, and the soundness of the suggested estimator is presented by simulation results. [ABSTRACT FROM AUTHOR]

The device-to-device (D2D) communication is a candidate technology to implement 5G standards commercially. To initiate D2D, device discovery is a primary issue and very few algorithms have been proposed for device discovery. A discovery algorithm has many parameters to discover the accurate position of the devices in walking and velocity scenarios. Due to rapid changes in the environment, LOS and NLOS algorithms become complex and accurate discovery ventures. Therefore, it is needed to evaluate the performance of the discovery algorithms. In this paper, a methodological approach is introduced for the performance evaluation of discovery algorithms. The performance evaluation for discovery estimation errors and complexity is evaluated using metrics and parameters, and analysis is made for range-based RSS technique using performance metrics. Discussion of performance evaluation metrics and criteria is analyzed followed by numerical/experimental, simulation models, and the parameters which affect performance and assessment. The metrics and criteria are defined in terms of a discovery signal success ratio, average residual energy, accuracy, and root-mean-square error (RMSE). Two differentiating discovery studies, Hamming and Cosine, are given and contrasted with reference RMSE for evaluation. This paper concludes with a discovery algorithm improvement cycle overview from simulation to implementation. It decreases discovery error and enhances RMSE accuracy by an average of 21%. It also reduces the complexity of 12 pairs by Euclidean distance by 29%. [ABSTRACT FROM AUTHOR]

Today, an undeniable interest is given to the development of socially intelligent robotic systems and efficient crowd- and socially aware navigation strategies. In this paper, we introduce a novel crowd-aware navigation algorithm that combines the A* path planner with the double deep Q-network (DDQN) method. The algorithm, named Social A*, aims to allow safe navigation for mobile robots in social dynamics and crowded environments. In order to facilitate future real-world implementation, a new learning environment compatible with the robot operating system (ROS) is developed. This allows expert teleoperation to help train the DDQN agent and refine the reward function. We conducted extensive simulations to compare the performance of Social A* with Socially attentive Reinforcement Learning (SARL*) and intention-aware robot crowd navigation with attention-based interaction graph (IARL). The obtained simulation results demonstrate that Social A* not only surpasses SARL* and IARL but also shows enhanced performance in handling static obstacles. These results showcase the excellent crowd-aware navigation performance, the efficiency, and the significant potential of the algorithm. [ABSTRACT FROM AUTHOR]

Sampling-based path planning algorithms are popularly used in autonomous mobile robot navigation applications. Optimal Rapidly exploring Random Trees (RRT*) is one of the well-known sampling-based single-query path planning algorithms and it is asymptotically optimal, but its convergence is slow. To address the slow convergence problem of the RRT* algorithm, this paper proposes a directional RRT* algorithm called D-RRT*. The key idea of D-RRT* is to reduce the sampling space. This is achieved in this proposed work by focusing on the direction of the goal from the starting configuration through a simple elliptical heuristic formed between them. The proposed methodology is validated in two different cluttered 2D environments and compared with existing algorithms. The proposed D-RRT* path planning algorithm outperforms the RRT* in three performance measures: the initial cost solution, convergence time, and the number of nodes visited. The convergence rate of the proposed D-RRT* is improved over RRT* by 8.5% and 14.7% in the two cluttered environments considered. Also, the proposed D-RRT* algorithm is validated in a real-time environment using the TurtleBot3 robot. [ABSTRACT FROM AUTHOR]

To address the voltage-based power quality issues in medium voltage distribution networks, this paper presents Quick-prop control algorithm driven Dynamic Voltage Restorer (DVR)-based on PV integrated reduced switch count multilevel inverter. For medium voltage applications, multilevel inverters (MLIs) are widely used due to their low voltage stress and reduced total harmonic distortion (THD). However, the major concern in this regard is the reliability of MLI as it uses large number of components as compared to two-level inverters. Therefore, a newly developed topology of multilevel inverters (MLIs) named as Transistor Clamped H-Bridge (TCHB) inverter is used in this work as DVR. The proposed topology has lesser component count and relatively lesser power losses as compared to other conventional topologies with same output quality. A 5-level PV integrated TCHB-MLI based DVR is delineated and analysed for 11 kV line in this work. To improve the robustness of the system against varying parameters, the Quick-prop based control algorithm is proposed that offers enhanced accuracy in estimation process, faster weight convergence speed, simplicity and optimal operation of the system. The proposed system improves the output power quality and enhances the transition of maximum power from solar PV array to grid and the load. The performance of proposed DVR is evaluated under different voltage-based power quality issues through MATLAB/SIMULINK software and the behaviour of proposed system is compared with conventional approaches. The results obtained show that the proposed system performs better than previous approaches. Experimental validation of results is obtained by using Digital Real-Time Simulator. [ABSTRACT FROM AUTHOR]

The cluster-sparse proportionate affine projection algorithm (CS-PAPA) exhibits a good solution for estimating the unknown echo path in network echo cancellation. However, the algorithm does not take into account the past proportionate factors for updating the filter coefficients. In this paper, a modification to the CS-PAPA is proposed in the context of echo cancellation. The proposed algorithm named memory cluster-sparse proportionate affine projection algorithm (MCS-PAPA) incorporates the history of proportionate factors into the CS-PAPA. Based on the energy conservation arguments, a rigorous performance analysis of the proposed algorithm is presented, which expresses the steady-state mean square error in terms of the projection order and other parameters. Moreover, the condition for the mean stability is derived. Experimental results for the steady-state mean square error corroborate with the theoretical expressions. Simulation experiments show that the proposed algorithm outperforms some existing algorithms in terms of normalised misalignment, convergence rate, and tracking. As the history of proportionate gain elements is included recursively, the proposed algorithm also exhibits a reduction in computational complexity in terms of the number of multiplications required. [ABSTRACT FROM AUTHOR]

In this paper, we analyze the performance of the Differential Evolution (DE) compared with the Genetic Algorithm (GA) and the Particle Swarm Optimization (PSO) applied to the optimization of the detection thresholds of distributed Ordered Statistics Constant False Alarm Rate (OS-CFAR) and Censored Mean Level Detector (CMLD-CFAR) systems using "OR" and "AND" fusion rules at the fusion center. The global systems' detection performance is analyzed in a Gaussian clutter considering identical and non-identical CFAR detectors. The obtained results showed that the DE optimization technique gives better performance than PSO and GA in both cases, either in identical or non-identical CFAR detectors. [ABSTRACT FROM AUTHOR]

Modern communication services are developing rapidly, and indoor positioning technologies are diverse. However, the accuracy of single algorithm cannot meet the actual requirements. To address this issue, an hybrid indoor positioning algorithm for cellular and Wi-Fi networks is proposed in this paper. The proposed algorithm consists of two phases, namely the offline phase and the online phase. In the offline phase, a fingerprint database is reconstructed by principal component analysis (PCA) and interpolation methods to reduce the costs of time. Then, in the online phase, the back propagation (BP) neural network positioning algorithm optimized by adaptive genetic algorithm (AGA-BP) is used for positioning. Moreover, the algorithm uses cellular network positioning to divide sub-regions and then uses Wi-Fi networks to further improve accuracy. The experimental results show that the average positioning error of the proposed hybrid algorithm is 1.70 m, which is 56.0% lower than using Wi-Fi network only. [ABSTRACT FROM AUTHOR]

Placement of heterogeneous FPGAs has been a complex problem for large designs. In addition, the extensive use of FPGAs in many applications, they have become very popular since last two decades. Hence, robust but simple steps algorithms are must needed for solving different layout problems for heterogeneous FPGAs. Like other design problems, the placement problem of heterogeneous FPGAs has become very hard with the increasing number of the various components and their inter connections. In addition to that routability-aware placement is a must needed solution, since placement without routability awareness may lead to failure of routing. In this paper, a Boolean satisfiability (SAT)-based packing and legalization free placement algorithm called SatPhF has been introduced for Xilinx UltraScale FPGA architectures. The proposed technique reduces the effort given for packing and legalization in the state-of-the-art placement algorithms. Additionally, the proposed approach takes care of congestion to produce an optimized routability-aware placement solutions in terms of wirelength and congestion. This technique performs a novel placement row selection method for a set of cells connected by a net using 2-SAT formulation which is able to optimize wirelength without use of any expensive analytical optimization framework. A new legalized location finding SAT approach has been introduced to find suitable locations for different cells without performing separate legalization step like other state-of-the-art placement algorithms. Finally, a SAT-based placement refinement procedure is introduced to reduce the congestion and making the placement routable. The proposed technique claims that most of the SAT instances created to formulate the placement problem is 2-SAT. The experimental results obtained by the presented approach show its ability to perform placement reasonably well in terms of wirelength and congestion. The results have been compared with four state-of-the-art techniques and show that SatPhF achieves a reasonably well reduction in total wirelength. [ABSTRACT FROM AUTHOR]

For a fatigue testing machine, when the aerospace flexible thermal connection components are subjected to tension and compression testing on the machine, the actual vibration frequency and amplitude are lower than the setting value. In order to solve this problem, this paper proposes an optimized PID controller based on improved non-dominated sorting genetic algorithm (reference point-based non-dominated sorting genetic algorithm, NSGA-III), to promote the test speed and efficiency of the test system. First, the stability in the frequency domain is taken as the constraint condition, the overshoot, adjustment time and ITAE of the system are taken as the optimization targets, and the parameters Kp and Ki are used as the design variables to establish a multi-objective optimization model. Secondly, in view of the fixed rate of crossover and mutation operators used by NSGA-III algorithm, which is prone to problems such as premature convergence and poor search ability, an adaptive crossover and mutation operator improved non-dominated sorting genetic algorithm (NSGA-III) is proposed. Finally, MATLAB/Simulink conducts system simulation and compares the PID controller, NSGA-III optimized PID controller and improved NSGA-III optimized PID controller. The results show that the NSGA is improved when the input step response imposes disturbance. Compared with PID controller and NSGA-III optimized PID controller, the adjustment time of NSGA-III optimized PID controller is reduced by 0.21 s and 0.03 s, respectively, which shows that the improved NSGA-III optimized PID controller has stable position output and is more fast adjustment response and better anti-interference ability. Under the input sinusoidal response, the maximum position error of the improved NSGA-III is 0.05, which is 0.38 compared to the maximum position error of the PID controller, and the maximum position error of the NSGA-III optimized PID controller is 0.1, which reduces 86% and 50%, respectively. Improving NSGA-III to optimize the PID controller can significantly improve the dynamic tracking accuracy. [ABSTRACT FROM AUTHOR]

The wireless sensor network (WSN) infrastructure is deployed with sensor nodes that are used to collect the sensed information and transmit to the central station, namely base station. The power limitations of sensors are commonly known in WSN. Due to limitation in power supply, energy utilization becomes a challenge in the networks. Various categories of protocols have been developed and addressed the energy issues, but the longevity of network lifetime is a crucial factor. In this paper, fuzzy-based energy-efficient cluster head selection algorithm is introduced to prolong the network lifetime. While the formation of the cluster, k-means algorithm is used to form a cluster in an efficient way and cluster head is selected based on fuzzy logic system. Simulation results show the proposed algorithm outperforms the existing algorithms. [ABSTRACT FROM AUTHOR]

For a given set of spatial locations (e.g., criminal activities, disease outbreak, etc.), circular hotspot detection (CHD) identifies circular zones (or hotspots) of significantly high concentration of activity points in the given space. Furthermore, the prevalence of activities inside these zones is considerably higher than outside them. CHD is essential for numerous societal applications, including criminology, epidemiology, etc. The existing methods for CHD consider one of the given activity points as the center of the circular zone (or hotspot). Hence, these methods are incapable of ignoring small gaps (like mountains, rivers, etc.) in the spatial contiguity of hotspots and are not geographically robust. Geographically robust CHD (GR-CHD) requires enumeration of those candidate circles where the hotspot center is not necessarily an activity point. Unfortunately, this enumeration unreasonably increases the enumerated circles and makes GR-CHD a computationally expensive problem. Therefore, in this paper, GR-CHD is modeled as a single-objective optimization problem. Further, three swarm intelligence (SI) algorithms, namely particle swarm optimization, grey wolf optimizer, and salp swarm algorithm, are applied to the proposed model to detect geographically robust hotspots. Finally, the performances of the presented SI-based schemes and state-of-the-art cubic grid circle (CGC) algorithm have been compared. They are evaluated over the controlled synthetic datasets and the crime dataset of Chicago city for the year 2019. The obtained results indicate that the computational times required by the SI-based schemes are significantly less than the CGC algorithm. Also, the quality of hotspots detected by SI-based algorithms is either improved or at par with the CGC algorithm. [ABSTRACT FROM AUTHOR]

In this paper, the pattern synthesis of concentric elliptical antenna arrays (CEAAs) for effective antenna design is presented. The current amplitudes and angular positions values of the CEAA elements are optimally found by using the Marine Predators Algorithm (MPA) in order to obtain the optimum radiation pattern. The MPA method is a novel nature-based optimization algorithm that is inspired by ocean predators foraging strategy. The antenna arrays with reduced maximum side lobe levels (MSL) and narrowed half power beam widths (HPBW) are an efficient antenna with optimum radiation pattern to suppress interference and increase directionality. In addition, CEAAs with different numbers of elements are examined to show the flexibility and performance of the proposed algorithm in this study. For 18-element CEAA, MPA obtained the HPBW between 29.18° and 31.34°, and MSL between − 20.70 and − 25.89 dB. For 30-element CEAA, the optimum values of HPBW are between 20.13° and 21.38°, and the optimum values of MSL are between − 18.90 and − 20.50 dB. The results of pattern synthesis obtained by MPA are compared with other metaheuristic optimization methods, and it is seen that the results of the proposed method are generally better. [ABSTRACT FROM AUTHOR]

This paper aims to present the concept of interval-valued picture uncertain linguistic set (IVPULS), which composes the grade of truth, abstinence, and falsity in the form of a subset of the unit interval. IVPULS is an extensive capable theory to manage awkward and unreliable information. To explore the theory, we firstly stated some basic operational laws of them and investigated their properties. Based on these stated laws, we defined several weighted and ordered weighted generalized Hamacher aggregation (GHA) operators. Several special cases are deduced from the proposed operators. Furthermore, a multi-attribute decision-making approach algorithm is stated by using the concept of proposed GHA operators under the IVPULSs environment. The presented algorithm has been explored with a numerical example and compares their results with the existing studies to examine and improve the quality and feasibility of the discovered theory. [ABSTRACT FROM AUTHOR]

In long haul communication environments, speech data transmission is severely affected by echoes. This phenomenon results in high bit errors as well as in degraded and annoying performance. Traditionally these problems, including hybrid and acoustic echoes, have been controlled through the use of echo suppressors. These suppressors were subsequently replaced by line echo cancellers using adaptive Finite Impulse Response filters. Fractional calculus has been applied successfully for fixed filtering with constant coefficients and in discrete time adaptive filtering that adjusts the weights according to the environment. This paper presents the Fractional Least Mean Square (FLMS) and Fractional Normalized LMS (FNLMS) algorithms for application in echo cancellation. Moreover, the performances of the FLMS and FNLMS are compared with those provided by the standard LMS, NLMS and Block Discrete Fourier Transform solutions. The mean square error criterion is used as the performance comparison criterion for two types of voice signals namely real and synthetic. The simulation results show a performance improvement of about 50% over the traditional counterparts. [ABSTRACT FROM AUTHOR]

This paper presents an adaptive-slope squashing-function (ASF)-based artificial neural network (ANN) for efficient estimation of smoothly time-varying multipath fading channels, in a 4 × 1 space-frequency-block-coded orthogonal-frequency-division-multiplexing (SFBC-OFDM) system using 64 subcarriers. The channel-state-information (CSI) estimated at first stage is further used for OFDM information symbol detection (through minimum mean square error criterion-based detection) at second stage. To combat the impact of smoothly time-varying environment, we emphasize on the utilization of ASF-ANN using backpropagation (BP) algorithm for the estimation of channel tap coefficients in frequency domain. The underlying ANN is modeled as feedforward multi-layered perceptron that updates the network weights. The major focus is on the gradient-descent algorithm-based adaptation of the slope of squashing-function (SF) along with other ANN parameters, which enhances the training efficiency of ASF-ANN in terms of the lower mean-squared channel estimation error in comparison with the traditional fixed-slope squashing-function (FSF) ANN technique. Simulation results corresponding to the underlying 4 × 1 SFBC-OFDM system are presented to depict that ASF-ANN-based approach outperforms the FSF-ANN technique by providing lower bit-error-rate (BER) due to the usage of well-estimated CSI. At 15 dB SNR and fade rate = 0.001, the average BER reduces to 2.85 × 10 - 4 for the ASF-ANN based approach, due to improved CSI estimation, which accounts for approximately 5% improvement in the detection success rate as compared to the FSF-ANN-based approach. [ABSTRACT FROM AUTHOR]

To address the problem of performance degradation caused by the blockchain backup mechanism as well as the high throughput and low latency requirements characteristic of blockchain applications, this paper optimizes the Raft consensus algorithm for the Hyperledger Fabric platform in terms of both log replication and leader election. In the log replication phase, the improvement of log replication based on apportionment idea aims to reduce the communication complexity of leader node by involving peer node in distributing log information. In the leader election phase, the improvement of leader election based on vote change mechanism changes the vote affiliation of peer node based on a comparison of votes from candidate nodes, aims to reach consensus in a round of election and reduces the election time. Through the performance test of the blockchain, it is demonstrated that the improved AdRaft algorithm improves 5.8% in throughput and reduces 1.3% in latency over the original Raft algorithm. [ABSTRACT FROM AUTHOR]

Scheduling workflows in IaaS cloud infrastructure has become a research issue. Considering multiple objectives for optimizing in scheduling workflows makes it difficult to adapt the cloud computing features like resource heterogeneity and elasticity on pay-per-use business model. This paper highlights such problems and suggests a workflow scheduling model to minimize the makespan and execution cost as a multi-objective optimization problem in cloud computing. We presented non-dominated sort-based hybrid wind-driven optimization (WDO) algorithm to address the desired goal on infrastructure as a service platform. It is combination of WDO and genetic algorithm that produces a Pareto optimal solution set with optimized makespan, cost and reliability. The user is offered to select the best solution out of produced optimal set. The evaluation of the proposed method is ascertained using three metrics: inverted generational distance, hypervolume and spacing. The results obtained by conducting experiments using four different scientific workflows validate that the solutions generated with the proposed hybrid algorithm are improved in terms of diversity and accuracy over different multi-objective solutions presented in the literature. [ABSTRACT FROM AUTHOR]