Random Initialization Research Articles

A parallel chaotic sailfish optimization algorithm for estimation of DOA in wireless sensor array

Direction of Arrival (DOA) estimation by Maximum Likelihood (ML) method has been popular among researchers due to its effective performance. However, its practical applications are limited by its computational complexity in non-linear multidimensional solution search which requires optimization. In order to make accurate non-linear ML-DOA estimation, simultaneously reducing the computational complexity involved in the process evolutionary algorithms are employed in this manuscript. In this manuscript, a recently reported nature-inspired optimization algorithm termed as ‘sailfish optimizer’ has been applied for non-linear multidimensional DOA estimation. The sailfish optimization is based on the swarming behavior of a group of sailfish to catch the prey which are sardines. The algorithm has been designed with both random and chaotic sequence-based population initialization. Simulation studies reveal that the chaotic sequence-based algorithm has lower computational time and accurate DOA estimation compared to its random counterpart. A parallel implementation of the chaotic version of the algorithm has been carried out to reduce the computational aspect of the algorithm. Simulation studies reveal the parallel version has almost 50% lower run time than that of the sequential one for 2048 snapshots implementations. By varying the number of sources, sensors, and snapshots experimentation has been carried out which reveals accurate DOA estimation (more than 95% angle accuracy in most of the cases) with reduced computational time (due to parallel implementation) achieved with the proposed algorithm. The comparative analysis of the proposed algorithm has been carried out with that achieved by benchmark MUSIC (Multiple Signal Classification) algorithm and nature inspired algorithms like : Advanced Particle Swarm Optimization, Real Coded Genetic Algorithm, Differential Evolution, CLONAL Selection, and Artificial Bee Colony algorithm. Kruskal–Wallis non-parametric test is performed under various noise conditions to validate the superior performance of the proposed approach over the comparative algorithms.

Unsupervised feature learning for electrocardiogram data using the convolutional variational autoencoder.

Most existing electrocardiogram (ECG) feature extraction methods rely on rule-based approaches. It is difficult to manually define all ECG features. We propose an unsupervised feature learning method using a convolutional variational autoencoder (CVAE) that can extract ECG features with unlabeled data. We used 596,000 ECG samples from 1,278 patients archived in biosignal databases from intensive care units to train the CVAE. Three external datasets were used for feature validation using two approaches. First, we explored the features without an additional training process. Clustering, latent space exploration, and anomaly detection were conducted. We confirmed that CVAE features reflected the various types of ECG rhythms. Second, we applied CVAE features to new tasks as input data and CVAE weights to weight initialization for different models for transfer learning for the classification of 12 types of arrhythmias. The f1-score for arrhythmia classification with extreme gradient boosting was 0.86 using CVAE features only. The f1-score of the model in which weights were initialized with the CVAE encoder was 5% better than that obtained with random initialization. Unsupervised feature learning with CVAE can extract the characteristics of various types of ECGs and can be an alternative to the feature extraction method for ECGs.

Trainless model performance estimation based on random weights initialisations for neural architecture search

Neural architecture search has become an indispensable part of the deep learning field. Modern methods allow to find one of the best performing architectures, or to build one from scratch, but they typically make decisions based on the trained accuracy information. In the present article we explore instead how the architectural component of a neural network affects its prediction power. We focus on relationships between the trained accuracy of an architecture and its accuracy prior to training, by considering statistics over multiple initialisations. We observe that minimising the coefficient of variation of the untrained accuracy, $CV_{U}$, consistently leads to better performing architectures. We test the $CV_{U}$ as a neural architecture search scoring metric using the NAS-Bench-201 database of trained neural architectures. The architectures with the lowest $CV_{U}$ value have on average an accuracy of $91.90 \pm 2.27$, $64.08 \pm 5.63$ and $38.76 \pm 6.62$ for CIFAR-10, CIFAR-100 and a downscaled version of ImageNet, respectively. Since these values are statistically above the random baseline, we make a conclusion that a good architecture should be stable against weights initialisations. It takes about $190$ s for CIFAR-10 and CIFAR-100 and $133.9$ s for ImageNet16-120 to process $100$ architectures, on a batch of $256$ images, with $100$ initialisations.

Noise-induced barren plateaus in variational quantum algorithms

Variational Quantum Algorithms (VQAs) may be a path to quantum advantage on Noisy Intermediate-Scale Quantum (NISQ) computers. A natural question is whether noise on NISQ devices places fundamental limitations on VQA performance. We rigorously prove a serious limitation for noisy VQAs, in that the noise causes the training landscape to have a barren plateau (i.e., vanishing gradient). Specifically, for the local Pauli noise considered, we prove that the gradient vanishes exponentially in the number of qubits n if the depth of the ansatz grows linearly with n. These noise-induced barren plateaus (NIBPs) are conceptually different from noise-free barren plateaus, which are linked to random parameter initialization. Our result is formulated for a generic ansatz that includes as special cases the Quantum Alternating Operator Ansatz and the Unitary Coupled Cluster Ansatz, among others. For the former, our numerical heuristics demonstrate the NIBP phenomenon for a realistic hardware noise model.

Computational modeling of mRNA degradation dynamics using deep neural networks.

messenger RNA (mRNA) degradation plays critical roles in post-transcriptional gene regulation. A major component of mRNA degradation is determined by 3'-UTR elements. Hence, researchers are interested in studying mRNA dynamics as a function of 3'-UTR elements. A recent study measured the mRNA degradation dynamics of tens of thousands of 3'-UTR sequences using a massively parallel reporter assay. However, the computational approach used to model mRNA degradation was based on a simplifying assumption of a linear degradation rate. Consequently, the underlying mechanism of 3'-UTR elements is still not fully understood. Here, we developed deep neural networks to predict mRNA degradation dynamics and interpreted the networks to identify regulatory elements in the 3'-UTR and their positional effect. Given an input of a 110 nt-long 3'-UTR sequence and an initial mRNA level, the model predicts mRNA levels of eight consecutive time points. Our deep neural networks significantly improved prediction performance of mRNA degradation dynamics compared with extant methods for the task. Moreover, we demonstrated that models predicting the dynamics of two identical 3'-UTR sequences, differing by their poly(A) tail, performed better than single-task models. On the interpretability front, by using Integrated Gradients, our convolutional neural networks (CNNs) models identified known and novel cis-regulatory sequence elements of mRNA degradation. By applying a novel systematic evaluation of model interpretability, we demonstrated that the recurrent neural network models are inferior to the CNN models in terms of interpretability and that random initialization ensemble improves both prediction and interoperability performance. Moreover, using a mutagenesis analysis, we newly discovered the positional effect of various 3'-UTR elements. All the code developed through this study is available at github.com/OrensteinLab/DeepUTR/. Supplementary data are available at Bioinformatics online.

K-Means-Based Nature-Inspired Metaheuristic Algorithms for Automatic Data Clustering Problems: Recent Advances and Future Directions

K-means clustering algorithm is a partitional clustering algorithm that has been used widely in many applications for traditional clustering due to its simplicity and low computational complexity. This clustering technique depends on the user specification of the number of clusters generated from the dataset, which affects the clustering results. Moreover, random initialization of cluster centers results in its local minimal convergence. Automatic clustering is a recent approach to clustering where the specification of cluster number is not required. In automatic clustering, natural clusters existing in datasets are identified without any background information of the data objects. Nature-inspired metaheuristic optimization algorithms have been deployed in recent times to overcome the challenges of the traditional clustering algorithm in handling automatic data clustering. Some nature-inspired metaheuristics algorithms have been hybridized with the traditional K-means algorithm to boost its performance and capability to handle automatic data clustering problems. This study aims to identify, retrieve, summarize, and analyze recently proposed studies related to the improvements of the K-means clustering algorithm with nature-inspired optimization techniques. A quest approach for article selection was adopted, which led to the identification and selection of 147 related studies from different reputable academic avenues and databases. More so, the analysis revealed that although the K-means algorithm has been well researched in the literature, its superiority over several well-established state-of-the-art clustering algorithms in terms of speed, accessibility, simplicity of use, and applicability to solve clustering problems with unlabeled and nonlinearly separable datasets has been clearly observed in the study. The current study also evaluated and discussed some of the well-known weaknesses of the K-means clustering algorithm, for which the existing improvement methods were conceptualized. It is noteworthy to mention that the current systematic review and analysis of existing literature on K-means enhancement approaches presents possible perspectives in the clustering analysis research domain and serves as a comprehensive source of information regarding the K-means algorithm and its variants for the research community.

An intelligent system for crop identification and classification from UAV images using conjugated dense convolutional neural network

Crop identification and classification is an important aspect for modern agricultural sector. With development of unmanned aerial vehicle (UAV) systems, crop identification from RGB images is experiencing a paradigm shift from conventional image processing techniques to deep learning strategies because of successful breakthrough in convolutional neural networks (CNNs). UAV images are quite trustworthy to identify different crops due to its higher spatial resolution. For precision agriculture crop identification is the primal criteria. Identifying a specific type of crop in a land is essential for performing proper farming and that also helps to estimate the net yield production of a particular crop. Previous works are limited to identify a single crop from the RGB images captured by UAVs and have not explored the chance of multi-crop classification by implementing deep learning techniques. Multi crop identification tool is highly needed as designing separate tool for each type of crop is a cumbersome job, but if a tool can successfully differentiate multiple crops then that will be helpful for the agro experts. In contrast with the previous existing techniques, this article elucidates a new conjugated dense CNN (CD-CNN) architecture with a new activation function named SL-ReLU for intelligent classification of multiple crops from RGB images captured by UAV. CD-CNN integrates data fusion and feature map extraction in conjunction with classification process. Initially a dense block architecture is proposed with a new activation function, called SL-ReLU, associated with the convolution operation to mitigate the chance of unbounded convolved output and gradient explosion. Dense block architecture concatenates all the previous layer features for determining the new features. This reduces the chance of losing important features due to deepening of the CNN module. Later, two dense blocks are conjugated with the help of a conversion block for obtaining better performance. Unlike traditional CNN, CD-CNN omits the use of fully connected layer and that reduces the chance of feature loss due to random weight initialization. The proposed CD-CNN achieves a strong distinguishing capability from several classes of crops. Raw UAV images of five different crops are captured from different parts of India and then small candidate crop regions are extracted from the raw images with the help of Arc GIS 10.3.1 software and then the candidate regions are fed to CD-CNN for proper training purpose. Experimental results show that the proposed module can achieve an accuracy of 96.2% for the concerned data. Further, superiority of the proposed network is established after comparing with other machine learning techniques viz. RF-200 and SVM, and standard CNN architectures viz. AlexNet, VGG-16, VGG-19 and ResNet-50.

Predicting Days to Maturity, Plant Height, and Grain Yield in Soybean: A Machine and Deep Learning Approach Using Multispectral Data

In soybean, there is a lack of research aiming to compare the performance of machine learning (ML) and deep learning (DL) methods to predict more than one agronomic variable, such as days to maturity (DM), plant height (PH), and grain yield (GY). As these variables are important to developing an overall precision farming model, we propose a machine learning approach to predict DM, PH, and GY for soybean cultivars based on multispectral bands. The field experiment considered 524 genotypes of soybeans in the 2017/2018 and 2018/2019 growing seasons and a multitemporal–multispectral dataset collected by embedded sensor in an unmanned aerial vehicle (UAV). We proposed a multilayer deep learning regression network, trained during 2000 epochs using an adaptive subgradient method, a random Gaussian initialization, and a 50% dropout in the first hidden layer for regularization. Three different scenarios, including only spectral bands, only vegetation indices, and spectral bands plus vegetation indices, were adopted to infer each variable (PH, DM, and GY). The DL model performance was compared against shallow learning methods such as random forest (RF), support vector machine (SVM), and linear regression (LR). The results indicate that our approach has the potential to predict soybean-related variables using multispectral bands only. Both DL and RF models presented a strong (r surpassing 0.77) prediction capacity for the PH variable, regardless of the adopted input variables group. Our results demonstrated that the DL model (r = 0.66) was superior to predict DM when the input variable was the spectral bands. For GY, all machine learning models evaluated presented similar performance (r ranging from 0.42 to 0.44) for each tested scenario. In conclusion, this study demonstrated an efficient approach to a computational solution capable of predicting multiple important soybean crop variables based on remote sensing data. Future research could benefit from the information presented here and be implemented in subsequent processes related to soybean cultivars or other types of agronomic crops.

Discussion on "distributional independent component analysis for diverse neuroimaging modalities" by Ben Wu, Subhadip Pal, Jian Kang, and Ying Guo.

Wu etal. have made an important contribution to the methodology for data-driven analysis of MRI data. However, we wish to challenge the authors on new potential applications of their approach beyond diffusion tensor imaging data, and to think carefully about the impact of random initialization implicit in their method. We illustrate the variability found from re-analyzing the supplied demonstration data multiple times, finding that the discovered independent components have a wide range of reliability, from nearly perfect overlap to no overlap atall.

Modelling EMG driven wrist movements using a bio-inspired neural network

The hidden pattern within the sEMG signal has wide applications in human-robot interaction. Decoding the patterns from sEMG signal tends to use black box models, which limits the further analysis of the mechanism of human musculoskeletal system. Therefore, a bio-inspired neural network (BNN) is proposed to model the information propagation procedures from nerve-related information (i.e. EMG signal) to muscle activation to joint activation to extremity movements. Instead of random parameter initialisation, the priori knowledge, such as muscle-electrode relationship, and muscles’ functionality, are fully considered to initialise the parameters. Besides, an interpretability constraint error back propagation algorithm (ICBP) is proposed to fine-tune the model for movement prediction, without scarifying model’s interpretability. An open sEMG database ISRMyo-I is utilised to verify the proposed methods for the classification of six wrist movements. With the only input of mean absolute value (MAV) feature, the proposed approach achieves an accuracy of >82%, which outperforms the support vector machine (78%), linear discriminant analysis (80%), k-nearest neighbors (78%), multi-layer perceptron (69%), random forest (74%), and convolutional neural network (74%).

Deep/Transfer Learning with Feature Space Ensemble Networks (FeatSpaceEnsNets) and Average Ensemble Networks (AvgEnsNets) for Change Detection Using DInSAR Sentinel-1 and Optical Sentinel-2 Satellite Data Fusion

Differential interferometric synthetic aperture radar (DInSAR), coherence, phase, and displacement are derived from processing SAR images to monitor geological phenomena and urban change. Previously, Sentinel-1 SAR data combined with Sentinel-2 optical imagery has improved classification accuracy in various domains. However, the fusing of Sentinel-1 DInSAR processed imagery with Sentinel-2 optical imagery has not been thoroughly investigated. Thus, we explored this fusion in urban change detection by creating a verified balanced binary classification dataset comprising 1440 blobs. Machine learning models using feature descriptors and non-deep learning classifiers, including a two-layer convolutional neural network (ConvNet2), were used as baselines. Transfer learning by feature extraction (TLFE) using various pre-trained models, deep learning from random initialization, and transfer learning by fine-tuning (TLFT) were all evaluated. We introduce a feature space ensemble family (FeatSpaceEnsNet), an average ensemble family (AvgEnsNet), and a hybrid ensemble family (HybridEnsNet) of TLFE neural networks. The FeatSpaceEnsNets combine TLFE features directly in the feature space using logistic regression. AvgEnsNets combine TLFEs at the decision level by aggregation. HybridEnsNets are a combination of FeatSpaceEnsNets and AvgEnsNets. Several FeatSpaceEnsNets, AvgEnsNets, and HybridEnsNets, comprising a heterogeneous mixture of different depth and architecture models, are defined and evaluated. We show that, in general, TLFE outperforms both TLFT and classic deep learning for the small dataset used and that larger ensembles of TLFE models do not always improve accuracy. The best performing ensemble is an AvgEnsNet (84.862%) comprised of a ResNet50, ResNeXt50, and EfficientNet B4. This was matched by a similarly composed FeatSpaceEnsNet with an F1 score of 0.001 and variance of 0.266 less. The best performing HybridEnsNet had an accuracy of 84.775%. All of the ensembles evaluated outperform the best performing single model, ResNet50 with TLFE (83.751%), except for AvgEnsNet 3, AvgEnsNet 6, and FeatSpaceEnsNet 5. Five of the seven similarly composed FeatSpaceEnsNets outperform the corresponding AvgEnsNet.

COVID-19 X-ray image segmentation by modified whale optimization algorithm with population reduction

Coronavirus disease 2019 (COVID-19) has caused a massive disaster in every human life field, including health, education, economics, and tourism, over the last year and a half. Rapid interpretation of COVID-19 patients' X-ray images is critical for diagnosis and, consequently, treatment of the disease. The major goal of this research is to develop a computational tool that can quickly and accurately determine the severity of an illness using COVID-19 chest X-ray pictures and improve the degree of diagnosis using a modified whale optimization method (WOA). To improve the WOA, a random initialization of the population is integrated during the global search phase. The parameters, coefficient vector (A) and constant value (b), are changed so that the algorithm can explore in the early stages while also exploiting the search space extensively in the latter stages. The efficiency of the proposed modified whale optimization algorithm with population reduction (mWOAPR) method is assessed by using it to segment six benchmark images using multilevel thresholding approach and Kapur's entropy-based fitness function calculated from the 2D histogram of greyscale images. By gathering three distinct COVID-19 chest X-ray images, the projected algorithm (mWOAPR) is utilized to segment the COVID-19 chest X-ray images. In both benchmark pictures and COVID-19 chest X-ray images, comparisons of the evaluated findings with basic and modified forms of metaheuristic algorithms supported the suggested mWOAPR's improved performance.

Absence of Barren Plateaus in Quantum Convolutional Neural Networks

Quantum neural networks (QNNs) have generated excitement around the possibility of efficiently analyzing quantum data. But this excitement has been tempered by the existence of exponentially vanishing gradients, known as barren plateau landscapes, for many QNN architectures. Recently, quantum convolutional neural networks (QCNNs) have been proposed, involving a sequence of convolutional and pooling layers that reduce the number of qubits while preserving information about relevant data features. In this work, we rigorously analyze the gradient scaling for the parameters in the QCNN architecture. We find that the variance of the gradient vanishes no faster than polynomially, implying that QCNNs do not exhibit barren plateaus. This result provides an analytical guarantee for the trainability of randomly initialized QCNNs, which highlights QCNNs as being trainable under random initialization unlike many other QNN architectures. To derive our results, we introduce a novel graph-based method to analyze expectation values over Haar-distributed unitaries, which will likely be useful in other contexts. Finally, we perform numerical simulations to verify our analytical results.9 MoreReceived 12 March 2021Revised 13 July 2021Accepted 2 August 2021DOI:https://doi.org/10.1103/PhysRevX.11.041011Published by the American Physical Society under the terms of the Creative Commons Attribution 4.0 International license. Further distribution of this work must maintain attribution to the author(s) and the published article’s title, journal citation, and DOI.Published by the American Physical SocietyPhysics Subject Headings (PhySH)Research AreasMachine learningQuantum algorithmsQuantum computationQuantum Information

Modified ant colony optimization based PID controller design for coupled tank system

Most of the optimization algorithm suffers from slow convergence and poor solution accuracy due to nonlinearity characteristics of the problem. This paper proposes a meta-heuristic method, namely the modified Ant Colony Optimization (m-ACO) technique to address these issues. The conventional ant colony optimization algorithm is improved by using multiple random initializations to increase the probability of getting a better start population and hence increase the possibility to achieve near-global optimum. The efficacy of the proposed algorithm is tested on classical benchmark functions and on the PID controllers. The performance of controllers depends on the selection of tuning parameters that are optimized to obtain an optimum solution. The m-ACO technique is used to optimize parameters of PID controller for Coupled Tank System. It is one of the benchmark control problems due to its nonlinear characteristics. The obtained results of the proposed algorithm are compared with the Zeigler Nichols tuned PID controller, and PSO tuned PID controller in terms of transient performances including rise-time, peak overshoot and performance parameters such as Integral Square Error (ISE) and Integral Absolute Error (IAE). The proposed method is also tested for showing effectiveness in set-point tracking and disturbance rejection.

An approach combining a new weight initialization method and constructive algorithm to configure a single Feedforward Neural Network for multi-class classification

This paper presents a new method for initializing weights in a Feedforward Neural Network (FNN) with a single hidden layer combined with a constructive approach to define the number of hidden units associated with the best classification performance. The strategy consists of defining an initial number of hidden units according to the classification problem, the linearization of the whole network around an equilibrium point and the determination of the initial weights and bias through the maximum approximation of the linearized model to the Optimal Linear Classifier (OLC) whose solution can be obtained analytically. The constructive algorithm comprises a gradual increase in the number of hidden units in such a way that at each training only the weights and bias associated with the new hidden units are initiated randomly while the weights and bias obtained from previous training are used as initial guesses. Additionally, the constructive algorithm seeks to ensure that the loss function of the trained networks decreases with the successive additions of hidden units. The proposed approach (Weight Initialization based on the Linearization of the Whole Neural Network combined with a new Constructive Algorithm, WILWNN-CA) is applied to synthetic and real datasets widely used as benchmark for multi-class classification problems. The comparison with conventional random weight initialization and other approaches involving different network topologies (and initialization strategies) shows that the proposed method is efficient and capable of providing success rates (correct classification rates) higher or similar to those achieved with existing methods.

ANALYSIS OF THE APPLICABILITY CRITERION FOR K MEANS CLUSTERING ALGORITHM RUN TEN NUMBER OF TIMES ON THE FIRST 25 NUMBERS OF THE FIBONACCI SERIES

In this research investigation Analysis Of The Applicability Criterion For K Means Clustering Algorithm Run Ten Number Of Times On The First 25 Numbers Of The Fibonacci Series is performed. For this analysis RCB Model Of Applicability Criterion For K Means Clustering Algorithm is used. K-means is one of the simplest unsupervised learning algorithms that solve the well-known clustering problem. K- Means clustering algorithm is a scheme for clustering continuous and numeric data. As K-Means algorithm consists of scheme of random initialization of centroids, every time it is run, it gives different or slightly different results because it may reach some local optima. Quantification of such aforementioned variation is of some importance as this sheds light on the nature of the Discrete K-Means Objective function with regards its maxima and minima. The K-Means Clustering algorithm aims at minimizing the aforementioned Objective function. The RCB Model Of Applicability Criterion for K-Means Clustering aims at telling us if we can use the K-Means Clustering Algorithm on a given set of data within acceptable variation limits of the results of the K-Means Clustering Algorithm when it is run several times. KEY WORDS: K-means clustering algorithm, RCB model and Cluster evaluation.

Energy optimization of internet of things in wireless sensor network models using type‐2 fuzzy neural systems

SummaryIn this paper, type‐2 fuzzy logic design is employed to find the weight values of the radial basis function (RBF) neural network model, and thereby, the trained RBF neural network (RBFNN) model is intended to perform network energy optimization of the cloud‐assisted internet of things in wireless sensor networks (WSNs). RBF neural model comes under the class of feed forward neural network architecture and is a network with better generalization capability. RBFNN employs Gaussian activation function to determine the output of the network and the special feature of this activation function is that it follows a normal probability density function; hence it is a continuous activation function and provides better solutions than the other discrete activation functions. Due to which, in this work RBFNN is employed to determine network energy and to increase the life time of the network by selecting best cluster heads and also the network route. In RBFNN modelling, basically, the weights are initialized in a random manner, and this random initialization of weights at time results in the occurrences of global and local minima. The weights are tuned for their optimal values using type‐2 fuzzy model due to their capability of handling uncertainties, and since this problem of identifying optimal weight values of RBF neural model possesses highest level of uncertainties, type‐2 fuzzy system is applied here. Simulation process is done, and the results prove the effectiveness of the proposed approach in comparison with that of the existing approaches from previous literatures for the energy optimization problem.

Unsupervised feature learning-based encoder and adversarial networks

In this paper, we propose a novel deep learning-based feature learning architecture for object classification. Conventionally, deep learning methods are trained with supervised learning for object classification. But, this would require large amount of training data. Currently there are increasing trends to employ unsupervised learning for deep learning. By doing so, dependency on the availability of large training data could be reduced. One implementation of unsupervised deep learning is for feature learning where the network is designed to “learn” features automatically from data to obtain good representation that then could be used for classification. Autoencoder and generative adversarial networks (GAN) are examples of unsupervised deep learning methods. For GAN however, the trajectories of feature learning may go to unpredicted directions due to random initialization, making it unsuitable for feature learning. To overcome this, a hybrid of encoder and deep convolutional generative adversarial network (DCGAN) architectures, a variant of GAN, are proposed. Encoder is put on top of the Generator networks of GAN to avoid random initialisation. We called our method as EGAN. The output of EGAN is used as features for two deep convolutional neural networks (DCNNs): AlexNet and DenseNet. We evaluate the proposed methods on three types of dataset and the results indicate that better performances are achieved by our proposed method compared to using autoencoder and GAN.

Ant Colony Algorithm for Time Optimal Load Balancing Scheduling Problem Strategy

In order to optimize the task execution efficiency of FPGA high-performance network clusters, this paper proposes a time-optimal load balancing scheduling problem algorithm and initialization strategy. Taking the task completion time as the optimization direction, the load balancing scheduling problem is abstracted as solving the placement matrix problem. On this basis, a node selection strategy using critical matrix is proposed, and the algorithm design is completed based on the ant colony algorithm. Aiming at the problem of low efficiency of random initialization, this paper proposes two ant colony initialization methods, which are ranked greedy initialization strategy and polling decentralized initialization strategy. Experiments show that the algorithm and initialization strategy designed in this paper can adapt to different computing scales, and the two initialization strategies can achieve a 15%∼55% improvement in the iterative convergence times and load balancing compared to random initialization.

Optimization of district heating production with thermal storage using mixed-integer nonlinear programming with a new initialization approach

Non-convex scheduling of energy production allows for more complex models that better describe the physical nature of the energy production system. Solutions to non-convex optimization problems can only be guaranteed to be local optima. For this reason, there is a need for methodologies that consistently provide low-cost solutions to the non-convex optimal scheduling problem. In this study, a novel Monte Carlo Tree Search initialization method for branch and bound solvers is proposed for the production planning of a combined heat and power unit with thermal heat storage in a district heating system. The optimization problem is formulated as a non-convex mixed-integer program, which is incorporated in a sliding time window framework. Here, the proposed initialization method offers lower-cost production planning compared to random initialization for larger time windows. For the test case, the proposed method lowers the yearly operational cost by more than 2,000,000 DKK per year. The method is one step in the direction of more reliable non-convex optimization that allows for more complex models of energy systems.