Proposed Algorithm Research Articles

Energy-efficient flexible job shop scheduling problem considering discrete operation sequence flexibility

The classical flexible job shop scheduling problem (FJSP) normally assumes that operations of each job have strict sequence constraints, i.e., each operation can be processed only after its previous operation is completed. However, in the actual production, the phenomenon that some operations of a job don't have any sequence constraints is very common. With regard to this, we firstly propose a FJSP with discrete operation sequence flexibility (FJSPDS) aiming at minimizing the makespan and total energy consumption, simultaneously. An effective mathematical model is established for the FJSPDS and its validity is proved by the CPLEX; and then an improved memetic algorithm (IMA) is designed to solve the FJSPDS. In the IMA, a new flexible sequencing method for determining process plan of each job and a right-leaning decoding method are proposed. And some effective crossover and mutation operators and an effective local search operator are designed to accelerate the convergence speed and expand the solution space of the algorithm. A total of 110 FJSPDS benchmark instances are constructed to conduct numerical simulation experiments. Experimental results show that our proposed IMA has superior performance in almost all of the instances compared with three well-known evolutionary algorithms. Our proposed model and algorithm can help the production managers who work with flexible manufacturing systems to obtain optimal scheduling schemes considering operations which have or don't have sequence constraints.

Two-step inertial forward–reflected–anchored–backward splitting algorithm for solving monotone inclusion problems

The main purpose of this paper is to propose and study a two-step inertial anchored version of the forward–reflected–backward splitting algorithm of Malitsky and Tam in a real Hilbert space. Our proposed algorithm converges strongly to a zero of the sum of a set-valued maximal monotone operator and a single-valued monotone Lipschitz continuous operator. It involves only one forward evaluation of the single-valued operator and one backward evaluation of the set-valued operator at each iteration; a feature that is absent in many other available strongly convergent splitting methods in the literature. Finally, we perform numerical experiments involving image restoration problem and compare our algorithm with known related strongly convergent splitting algorithms in the literature.

Discovering governing partial differential equations from noisy data

Partial differential equations (PDEs) derived from first principles knowledge have been indispensable tools for modelling many physical and chemical systems. However, the presence of complex terms in the PDEs may render traditional first principles modelling techniques inadequate. In such cases, data-driven methods such as PDE-FIND can be used to extract PDEs from the spatiotemporal measurements of the system. However, the PDE-FIND algorithm is sensitive to noise. Furthermore, while PDE-discovery models specifically developed for noisy data have been proposed, these models work best for low noise levels. Moreover, most of these models fail to discover the heat equation. We propose a smoothing-based approach for discovering the PDEs from noisy measurements. The framework is broadly data-driven, and its performance can be further improved by incorporating first principles knowledge (such as the order of the system). Our proposed algorithm effectively extracts partial differential equations (PDEs) from measurements with a low signal-to-noise ratio (SNR), outperforming existing techniques. Additionally, we have demonstrated the effectiveness of our algorithm in a real system (where collinear terms occur in the library) by using a new benchmark metric.

Towards accelerating model parallelism in distributed deep learning systems.

Modern deep neural networks cannot be often trained on a single GPU due to large model size and large data size. Model parallelism splits a model for multiple GPUs, but making it scalable and seamless is challenging due to different information sharing among GPUs with communication overhead. Specifically, we identify two key issues to make the parallelism being inefficient and inaccurate; an efficient pipelining technique is crucial to maximize GPU utilization and normalizations in deep neural networks may affect the performance due to different statistics sharing of mini-batch. In this work, we address these issues by investigating efficient pipelining for model parallelism and effective normalizations in model / data parallelisms when training a model with large mini-batch in multiple GPUs so that the model performance in accuracy can not be compromised. Firstly, we propose a novel method to search for an optimal micro-batch size considering the number of GPUs and memory size for model parallelism. For efficient pipelining, mini-batch is usually divided into smaller batches (called micro-batch). To maximize the utilization of GPU computing resources, training should be performed with the optimal micro-batch size. Our proposed micro-batch size search algorithm achieved increased image throughput by up to 12% and improved trainable mini-batch size by 25% as compared to the conventional model parallelism method. Secondly, we investigate normalizations in distributed deep learning training for different parallelisms. Our experiments using different normalization methods suggested that the performance with batch normalization can be improved by sharing the batch information among GPUs when performing data parallelism. It was also confirmed that group normalization helped minimizing accuracy degradation when performing model parallelism with pipelining and yielded consistent accuracies for diverse mini-batch sizes.

Range-Free Localization Approaches Based on Intelligent Swarm Optimization for Internet of Things

Recently, the precise location of sensor nodes has emerged as a significant challenge in the realm of Internet of Things (IoT) applications, including Wireless Sensor Networks (WSNs). The accurate determination of geographical coordinates for detected events holds pivotal importance in these applications. Despite DV-Hop gaining popularity due to its cost-effectiveness, feasibility, and lack of additional hardware requirements, it remains hindered by a relatively notable localization error. To overcome this limitation, our study introduces three new localization approaches that combine DV-Hop with Chicken Swarm Optimization (CSO). The primary objective is to improve the precision of DV-Hop-based approaches. In this paper, we compare the efficiency of the proposed localization algorithms with other existing approaches, including several algorithms based on Particle Swarm Optimization (PSO), while considering random network topologies. The simulation results validate the efficiency of our proposed algorithms. The proposed HW-DV-HopCSO algorithm achieves a considerable improvement in positioning accuracy compared to those of existing models.

Balanced Clique Computation in Signed Networks: Concepts and Algorithms

Clique is one of the most fundamental models for cohesive subgraph mining in network analysis. Existing clique model mainly focuses on unsigned networks. However, in real world, many applications are modeled as signed networks with positive and negative edges. As the signed networks hold their own properties different from the unsigned networks, the existing clique model is inapplicable for the signed networks. Motivated by this, we propose the balanced clique model that considers the most fundamental and dominant theory, structural balance theory, for signed networks. Following the balanced clique model, we study the <underline xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">m</u> aximal <underline xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">b</u> alanced <underline xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">c</u> lique <underline xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">e</u> numeration problem (MBCE) which computes all the maximal balanced cliques in a given signed network. Moreover, in some applications, users prefer a unique and representative balanced clique with maximum size rather than all balanced cliques. Thus, we also study the <underline xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">m</u> aximum <underline xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">b</u> alanced <underline xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">c</u> lique <underline xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">s</u> earch problem (MBCS) which computes the balanced clique with maximum size. We show that MBCE problem and MBCS problem are both NP-Hard. For the MBCE problem, a straightforward solution is to treat the signed network as two unsigned networks and leverage the off-the-shelf techniques for unsigned networks. However, such a solution is inefficient for large signed networks. To address this problem, in this paper, we first propose a new maximal balanced clique enumeration algorithm by exploiting the unique properties of signed networks. Based on the new proposed algorithm, we devise two optimization strategies to further improve the efficiency of the enumeration. For the MBCS problem, we first propose a baseline solution. To overcome the huge search space problem of the baseline solution, we propose a new search framework based on search space partition. To further improve the efficiency of the new framework, we propose multiple optimization strategies regarding to redundant search branches and invalid candidates. We conduct extensive experiments on large real datasets. The experimental results demonstrate the efficiency, effectiveness and scalability of our proposed algorithms for MBCE problem and MBCS problem.

The accuracy analysis of different algorithms for 4-Bit absolute value detector

This paper proposes and compares three different algorithms for implementing a 4-bit absolute value detector (AVD) using Complementary Metal Oxide Semiconductor (CMOS) technology. The demand for low power, low delay and high-performance drive the need for AVDs. The proposed algorithms are analysed based on logical effort theory and simulation results for delay and energy performance. The circuits employ a range of logical gates, including multiplexers, NAND gates, XNOR gates, XOR gates, and inverters. The chosen proposed algorithm achieves the lowest delay and energy cost while maintaining high accuracy. Sizing and Vdd optimization can be used to optimize energy, and a 53.7 FO4(1V), 39.91Eu(1V) 4-bit Absolute-Value Detector can be achieved. Overall, this paper provides valuable insights into the design and optimization of AVDs using CMOS technology, which can have important applications in developing electronic systems.

Performance Optimization and Fault-Tolerance of Highly Dynamic Systems Via Q-Learning With an Incrementally Attached Controller Gain System.

High-performance and reliable control of systems that are highly dynamic and open-loop unstable is challenging but of considerable practical interest. Thus, this article investigates the performance optimization and fault tolerance of highly dynamic systems. First, an incremental control structure is proposed, where a controller gain system is attached to the predesigned controller, and by reconfiguring the controller gain system, the performance can be equivalently optimized as configuring the predesigned one. The incremental attachment of the controller gain system does not modify the existing control system, and it can be easily attached via various communication channels. Second, a structure integrating fault-tolerance strategy and hardware redundancy is proposed. Under this structure, command fusion and fault-tolerance strategies are developed where the control commands from different control units are optimally fused, and each control unit can be reconfigured w.r.t. the performance of the other ones. Furthermore, Q -learning algorithms are developed to realize the proposed structures and strategies in real-time model-freely. As such, varying operational conditions of the highly dynamic system can be tackled. Finally, the proposed structures and algorithms are validated case by case to show their effectiveness.

Adaptive Set-Membership State Estimation for Nonlinear Systems Under Bit Rate Allocation Mechanism: A Neural-Network-Based Approach.

In this article, the adaptive neural-network-based (NN-based) set-membership state estimation problem is studied for a class of nonlinear systems subject to bit rate constraints and unknown-but-bounded noises. The measurement output signals are transmitted from sensors to a remote estimator via a bit rate constrained communication channel. To relieve the communication burden and ameliorate the state estimation accuracy, a bit rate allocation mechanism is put forward for the sensor nodes by solving a constrained optimization problem. Subsequently, through the NN learning method, an NN-based set-membership estimator is designed to determine an ellipsoidal set that contains the system state, where the proposed estimator relies upon a prediction-correction structure. With the help of the mathematical induction technique and the set theory, sufficient conditions are obtained to ensure the existence of both the adaptive tuning parameters and the set-membership estimators, and then, the corresponding parameters and estimator gains are calculated by solving a set of optimization problems. In addition, the monotonicity of the upper bound on the squared estimation error with respect to the bit rate and the convergence of the NN weight are analyzed, respectively. Finally, an illustrative example is given to demonstrate the effectiveness of the proposed state estimation algorithm.

Real-time Lightweight Target Detection Network under Autonomous Driving

Target detection in traffic scenes has been a focal point concerning computer vision, especially in the context of autonomous driving, including the recognition of vehicles, pedestrians, and traffic signs. However, the computational power of devices operating in traffic scenarios is limited, placing stringent demands on computational resources and latency. To address these challenges, this study proposed a lightweight detection algorithm, GV2-YOLO, which prioritizes both speed and accuracy. First, we apply the GhostNetv2 architecture and incorporate the Ghost module to decrease the parameters of the backbone feature extraction network. Second, the algorithm also integrates SPPF and Slim-neck by GSConv techniques, which effectively reduces the computational complexity while ensuring the resulting significant reduction in detection accuracy. Our proposed algorithm achieves a mAP of 84.7% on the CCTSDB dataset. The algorithm has a total parameter count of 9.1 M, making it ideal for deployment in embedded autonomous driving platforms.

Multiparty distance minimization: Problems and an evolutionary approach

Multiparty multiobjective optimization problems (MPMOPs) have been proposed to represent situations in which involves multiple decision makers, each decision maker concerns on a multiobjective optimization problem (MOP) and their MOPs are different. To study multiparty multiobjective evolutionary algorithms in depth, this paper constructs a series of MPMOPs based on distance minimization problems (DMPs). These MPMOPs, called MPDMPs, can easily represent the solutions in the decision space. Thus, the behaviors of evolutionary algorithms performing on MPDMPs can be conveniently studied including the movement of the solutions and the distribution of the final solutions. To address MPDMPs, the new proposed algorithm OptMPNDS3 uses a multiparty initialization method to initialize the population and the JADE2 operator to generate the offspring. OptMPNDS3 is compared with OptAll, OptMPNDS and OptMPNDS2 on the problem suite. The results show that the performance of OptMPNDS3 is strong and comparable to that of other algorithms.

Communication-efficient distributed statistical inference on zero-inflated Poisson models

Zero-inflated count outcomes are common in many studies, such as counting claim frequency in the insurance industry in which identifying and understanding excessive zeros are of interest. Moreover, with the progress of data collecting and storage techniques, the amount of data is too massive to be stored or processed by a single node or branch. Hence, to develop distributed data analysis is blossoming. In this paper, several communication-efficient distributed zero-inflated Poisson regression algorithms are developed to analyse such kind of large-scale zero-inflated data. Both asymptotic properties of the proposed estimators and algorithm complexities are well studied and conducted. Various simulation studies demonstrate that our proposed method and algorithm work well and efficiently. Finally, in the case study, we apply our proposed algorithms to car insurance data from Kaggle.

Diagnosis of Alzheimer’s Disease Based on Accelerated Mirror Descent Optimization and a Three-Dimensional Aggregated Residual Network

Alzheimer’s disease (AD), a neuropsychiatric disorder, continually arises in the elderly. To date, no targeted medications have been developed for AD. Early and fast diagnosis of AD plays a pivotal role in identifying potential AD patients, enabling timely medical interventions, and mitigating disease progression. Computer-aided diagnosis (CAD) becomes possible with the burgeoning of deep learning. However, the existing CAD models for processing 3D Alzheimer’s disease images usually have the problems of slow convergence, disappearance of gradient, and falling into local optimum. This makes the training of 3D diagnosis models need a lot of time, and the accuracy is often poor. In this paper, a novel 3D aggregated residual network with accelerated mirror descent optimization is proposed for diagnosing AD. First, a novel unbiased subgradient accelerated mirror descent (SAMD) optimization algorithm is proposed to speed up diagnosis network training. By optimizing the nonlinear projection process, our proposed algorithm can avoid the occurrence of the local optimum in the non-Euclidean distance metric. The most notable aspect is that, to the best of our knowledge, this is the pioneering attempt to optimize the AD diagnosis training process by improving the optimization algorithm. Then, we provide a rigorous proof of the SAMD’s convergence, and the convergence of SAMD is better than any existing gradient descent algorithms. Finally, we use our proposed SAMD algorithm to train our proposed 3D aggregated residual network architecture (ARCNN). We employed the ADNI dataset to train ARCNN diagnostic models separately for the AD vs. NC task and the sMCI vs. pMCI task, followed by testing to evaluate the disease diagnostic outcomes. The results reveal that the accuracy can be improved in diagnosing AD, and the training speed can be accelerated. Our proposed method achieves 95.4% accuracy in AD diagnosis and 79.9% accuracy in MCI diagnosis; the best results contrasted with several state-of-the-art diagnosis methods. In addition, our proposed SAMD algorithm can save about 19% of the convergence time on average in the AD diagnosis model compared with the gradient descent algorithms, which is very momentous in clinic.

A Novel Hybrid Approach for Similarity-Based Link Prediction in Complex Networks

Link prediction in complex networks is the challenging task of predicting missing or future connections between nodes. Complex networks, such as social networks, biological networks, and online recommendation systems, are often incomplete, with unknown or unestablished relationships between nodes. Link prediction algorithms fill in these missing links by leveraging the existing network structure and properties. By analyzing network patterns, connectivity, and characteristics, these algorithms can predict the likelihood of future connections, enabling applications such as recommender systems, network analysis, and understanding the dynamics of complex systems. Link prediction is a crucial task for providing relevant recommendations in e-commerce recommender systems. However, it is challenging due to the sparsity of user-item interaction data and the dynamic nature of user preferences. In this paper, we propose a novel hybrid link prediction algorithm that combines the Jaccard Coefficient Similarity Index, Adamic Adar Index, and MapSim Similarity based Index methods. Our algorithm leverages the complementary strengths of each individual method to improve the overall prediction accuracy. The Jaccard Coefficient Similarity Index measures the similarity between two users or items based on the number of shared items or users. The Adamic Adar Index considers the common neighbors between two users or items to predict the link probability. The MapSim Similarity based Index method incorporates the geographic location of users and items to predict the link probability. We evaluate our proposed hybrid algorithm on two real-world e-commerce datasets, and the results show that it outperforms several state-of-the-art link prediction algorithms in terms of accuracy and precision.

Scalable angular spectrum propagation

Coherent field propagation is an essential computational tool in optics with applications ranging from computational optics and optical design to iterative field reconstructions. An improvement in the computational speed of current propagation methods is therefore highly desired. We describe a scalable angular spectrum (SAS) algorithm with zoom capability for numerical propagation of scalar wave fields in homogeneous media. It allows for propagation models where the destination pixel pitch is larger than the source pixel pitch, requires a computational complexity proportional to the cost of three successive fast Fourier transform operations of the input field, and it is valid for high numerical aperture (NA) propagation geometries. We find that SAS propagation approaches the precision of the computationally far more expensive angular spectrum method in conjunction with zero-padding. This was computationally confirmed by propagation examples. Finally, we discuss the validity of the proposed SAS method, derive practical bandlimit criteria, and state a limit for the propagation distance. The scalability, efficiency, and accuracy at high NA of our proposed wave propagation algorithm yield benefits for a large variety of forward and inverse modeling problems with the ability to apply automatic differentiation.

Optimal service provisioning in IoT fog-based environment for QoS-aware delay-sensitive application

This paper addresses the escalating challenges posed by the ever-increasing data volume, velocity, and the demand for low-latency applications, driven by the proliferation of smart devices and Internet of Things (IoT) applications. To mitigate service delay and enhance Quality of Service (QoS), we introduce a hybrid optimization of Particle Swarm (PSO) and Chemical Reaction (CRO) to improve service delay in FogPlan, an offline framework that prioritizes QoS and enables dynamic fog service deployment. The method optimizes fog service allocation based on incoming traffic to each fog node, formulating it as an Integer Non-Linear Programming (INLP) problem, considering various service attributes and costs. Our proposed algorithm aims to minimize service delay and QoS degradation. The evaluation using real MAWI Working Group traffic data demonstrates a substantial 29.34% reduction in service delay, a 66.02% decrease in service costs, and a noteworthy 50.15% reduction in delay violations compared to the FogPlan framework.

Secure and efficient prediction of electric vehicle charging demand using [formula omitted]-LSTM and AES-128 cryptography

In recent years, there has been a significant surge in demand for electric vehicles (EVs), necessitating accurate prediction of EV charging requirements. This prediction plays a crucial role in evaluating its impact on the power grid, encompassing power management and peak demand management. In this paper, a novel deep neural network based on α2 -LSTM is proposed to predict the demand for charging from electric vehicles at a 15-minute time resolution. Additionally, we employ AES-128 for station quantization and secure communication with users. Our proposed algorithm achieves a 9.2% reduction in both the Root Mean Square Error (RMSE) and the mean absolute error compared to LSTM, along with a 13.01% increase in demand accuracy. We present a 12-month prediction of EV charging demand at charging stations, accompanied by an effective comparative analysis of Mean Absolute Percentage Error (MAPE) and Mean Percentage Error (MPE) over the last five years using our proposed model. The prediction analysis has been conducted using Python programming.

Spatiotemporal Fusion Model of Remote Sensing Images Combining Single-Band and Multi-Band Prediction

In recent years, convolutional neural network (CNN)-based spatiotemporal fusion (STF) models for remote sensing images have made significant progress. However, existing STF models may suffer from two main drawbacks. Firstly, multi-band prediction often generates a hybrid feature representation that includes information from all bands. This blending of features can lead to the loss or blurring of high-frequency details, making it challenging to reconstruct multi-spectral remote sensing images with significant spectral differences between bands. Another challenge in many STF models is the limited preservation of spectral information during 2D convolution operations. Combining all input channels’ convolution results into a single-channel output feature map can lead to the degradation of spectral dimension information. To address these issues and to strike a balance between avoiding hybrid features and fully utilizing spectral information, we propose a remote sensing image STF model that combines single-band and multi-band prediction (SMSTFM). The SMSTFM initially performs single-band prediction, generating separate predicted images for each band, which are then stacked together to form a preliminary fused image. Subsequently, the multi-band prediction module leverages the spectral dimension information of the input images to further enhance the preliminary predictions. We employ the modern ConvNeXt convolutional module as the primary feature extraction component. During the multi-band prediction phase, we enhance the spatial and channel information captures by replacing the 2D convolutions within ConvNeXt with 3D convolutions. In the experimental section, we evaluate our proposed algorithm on two public datasets with 16x resolution differences and one dataset with a 3x resolution difference. The results demonstrate that our SMSTFM achieves state-of-the-art performance on these datasets and is proven effective and reasonable through ablation studies.

Optimal Placement of μPMUs in Distribution Networks with Adaptive Topology Changes

With the increasing integration of energy sources and the growing complexity of distribution networks, it is crucial to monitor and early detection of topological changes to ensure grid stability and resilience. Current methods, for optimizing the placement of micro Phasor Measurement Units (μPMUs) focus on achieving observability and efficient monitoring. These algorithms aim to minimize the number of μPMUs needed while maintaining system observability or meeting criteria for observability. However, they may not consider all real-world constraints and uncertainties. In this study, we introduce a strategy for placing μPMUs with the objective of enhancing observability and monitoring capabilities. Our proposed algorithm employs a technique that makes optimal decisions at each step to approximate the global optimum. To determine the locations for μPMUs our algorithm takes into account parameters such as network structure, key nodes, and system stability. One distinguishing feature is its adaptability to distribution networks, including changes, in topology or potential device failures. Unlike classical approaches, our algorithm can continuously provide optimal placement solutions even in evolving network conditions. We have demonstrated that our suggested method achieves better results in terms of observability value and the required number of μPMUs compared to the state-of-the-art. By strategically placing μPMUs, operators can improve system observability, quickly detect and locate faults, and make informed decisions for effective network operations. This research helps improve optimal placement strategies for μPMUs by providing practical and effective solutions to improve distribution network reliability, resilience, and performance in the face of changing dynamics.

Hybrid Energy Regulated Constant Gain Kalman-Filter for Optimized Target Detection and Tracking in Wireless Sensor Networks

Wireless Sensor Networks (WSNs) are one of the most researched areas worldwide as the wide-scale networks possess low cost, are small in size, consume low power, and can be deployed in various environments. Among various applications of WSNs, target tracking is a highly demanding and broadly investigated application of wireless sensor networks. The parameter of accurate tracking is restricted because of the limited resources present in the wireless sensor networks, noise of the network, environmental factors, and faulty sensor nodes. Our work aims to enhance the accuracy of the tracking process as well as energy utilization by combing the mechanism of clustering with the prediction. Here, we present a hybrid energy-regulated constant gain Kalman filter-based target detection and tracking method, which is an algorithm to make the best use of energy and enhance precision in tracking. Our proposed algorithm is compared with the existing approaches where it is observed that the proposed technique possesses efficient energy utilization by decreasing the transference of unimportant data within the sensor network, achieving accurate results.