Practical Computation Time Research Articles

Multiple Parameter Replica Exchange Gaussian Accelerated Molecular Dynamics for Enhanced Sampling and Free Energy Calculation of Biomolecular Systems.

This study introduces a novel method named multiple parameter replica exchange Gaussian accelerated molecular dynamics (MP-Rex-GaMD), building on the Gaussian accelerated molecular dynamics (GaMD) algorithm. GaMD enhances sampling and retrieves free energy information for biomolecular systems by adding a harmonic boost potential to smooth the potential energy surface without the need for predefined reaction coordinates. Our innovative approach advances the acceleration power and energetic reweighting accuracy of GaMD by incorporating a replica exchange algorithm that enables the exchange of multiple parameters, including the GaMD boost parameters of force constant and energy threshold, as well as temperature. Applying MP-Rex-GaMD to the three model systems of dialanine, chignolin, and HIV protease, we demonstrate its superior capability over conventional molecular dynamics and GaMD simulations in exploring protein conformations and effectively navigating various biomolecular states across energy barriers. MP-Rex-GaMD allows users to accurately map free energy landscapes through energetic reweighting, capturing the ensemble of biomolecular states from low-energy conformations to rare high-energy transitions within practical computational time scales.

A software-agnostic benchmark for DEM simulation of cohesive and non-cohesive materials

Although DEM (Discrete Element Method) was introduced >40 years ago, there are still various challenges in applying it with a certain level of confidence. To develop a realistic material model, calibration, validation, particle shape representation and scaling are well-known challenges that have been studied by various researchers over the past two decades. In addition, once a realistic DEM material model is developed and published in line with open science principles, it is unknown to what extent the calibrated values can be used across software packages. Although a benchmark between 8 open source codes was recently published, it did not cover cohesive materials, rolling friction models, realistic calibrated material behaviour, and commercial software that is widely used in industries.The aim of this study is to investigate the portability of input parameters between different codes and to identify if software independent calibration is possible. We propose a framework to assist DEM users in industry and academia to reach a software independent DEM simulation when required.To compare the results between software packages, the framework considers the following aspects: 1) identical implementation of contact models, 2) single contact modelling, 3) bulk level simulation, and 4) identical post-processing. This framework is demonstrated for two contact models (Hertz-Mindlin and Edinburgh Elasto Plastic Adhesive) with rolling friction model and two commercial software packages (EDEM and PFC) but is applicable for any other (open-source) software. Moreover, two realistic calibrated material models are included to cover a range of material characteristics, from free-flowing incompressible materials to cohesive compressible materials. This study shows that individual particle level simulations give identical results, bulk level simulations show differences for both contact models. In general, users should use parameter values from other software packages with caution, especially where critical or sensitive applications are modelled. This paper also highlights the use of novel computation techniques, such as the GPU engine, to achieve practical computation times when modelling industrial applications.

The flexible airport bus and last-mile ride-sharing problem: Math-heuristic and metaheuristic approaches

Airport buses play a crucial role in addressing the last-mile problem of air travel, especially in cities and countries lacking inner-city rail transit systems. Nevertheless, airport buses are currently witnessing a decline in ridership due to drawbacks such as long departure intervals, inflexible stops, and considerable distances between stops. Consequently, delivering high-quality airport bus services has become a pressing concern for public transport operators. Motivated by new flexible buses and ride-sharing services, this paper explores a flexible airport bus service that integrates ride-sharing services for passengers traveling from bus stops to their destinations. This problem entails integrated decisions involving bus stop selection, passenger assignment to drop-off bus stops, as well as bus and ride-sharing routing. Accordingly, this problem presents more challenges in decision-making than traditional flexible bus or ride-sharing routing problems. We first develop an arc-based mixed-integer linear programming model. Subsequently, we design a double decomposition math-heuristic algorithm that builds upon logic-based Benders decomposition and column generation algorithms to obtain a near-optimal solution within practical computation time limits for practical-scale instances. Additionally, we implement an adaptive large neighborhood search algorithm to evaluate the solution quality of this math-heuristic algorithm and to solve large-scale instances. To validate the effectiveness of both the model and the algorithms, we conduct numerical experiments using instances derived from Shenzhen airport bus lines. The experimental results demonstrate that the flexible service mode offers significant advantages in reducing both passenger ride time and vehicle mileage over traditional airport bus or taxi modes.

Joint service placement and user assignment model in multi‐access edge computing networks against base‐station failure

AbstractMulti‐access edge computing (MEC) enables users to exploit the resources of cloud computing at a base station (BS) in proximity to the users where an MEC server is hosted. While we have advantage of being able to communicate with low latency and small network load in MEC networks, the resources in BSs are limited. One challenge is where to provide users with services from to make efficient use of resources. Furthermore, to enhance the reliability of MEC system, the case that a BS fails needs to be considered. This paper proposes a service placement and user assignment model with preventive start‐time optimization against a single BS failure in MEC networks. The proposed model preventively determines the service placement and user assignment in each BS failure pattern to minimize the worst‐case penalty which is the largest penalty among all failure patterns. We formulate the proposed model as an integer linear programming problem. We prove that the considered problem is NP‐hard. When the problem size becomes large, it may not be solved in a practical computation time. To solve larger size problems, we introduce two algorithms: one is the greedy algorithm with allocation upgrade and the other is with allocation upgrade and preemption. The results show that the introduced algorithms obtain solutions with smaller worst‐case penalty than the benchmark in a practical time.

Binaural Auralization of Room Acoustics with a Highly Scalable Wave-Based Acoustics Simulation

This paper presents a proposal of an efficient binaural room-acoustics auralization method, an essential goal of room-acoustics modeling. The method uses a massively parallel wave-based room-acoustics solver based on a dispersion-optimized explicit time-domain finite element method (TD-FEM). The binaural room-acoustics auralization uses a hybrid technique of first-order Ambisonics (FOA) and head-related transfer functions. Ambisonics encoding uses room impulse responses computed by a parallel wave-based room-acoustics solver that can model sound absorbers with complex-valued surface impedance. Details are given of the novel procedure for computing expansion coefficients of spherical harmonics composing the FOA signal. This report is the first presenting a parallel wave-based solver able to simulate room impulse responses with practical computational times using an HPC cloud environment. A meeting room problem and a classroom problem are used, respectively, having 35 million degrees of freedom (DOF) and 100 million DOF, to test the parallel performance of up to 6144 CPU cores. Then, the potential of the proposed binaural room-acoustics auralization method is demonstrated via an auditorium acoustics simulation of up to 5 kHz having 750,000,000 DOFs. Room-acoustics auralization is performed with two acoustics treatment scenarios and room-acoustics evaluations that use an FOA signal, binaural room impulse response, and four room acoustical parameters. The auditorium acoustics simulation showed that the proposed method enables binaural room-acoustics auralization within 13,000 s using 6144 cores.

Spatial evolution of human cultures inferred through Bayesian phylogenetic analysis.

Spatial distribution of human culture reflects both descent from the common ancestor and horizontal transmission among neighbouring populations. To analyse empirically documented geographical variations in cultural repertoire, we will describe a framework for Bayesian statistics in a spatially explicit model. To consider both horizontal transmission and mutation of the cultural trait in question, our method employs a network model in which populations are represented by nodes. Using algorithms borrowed from Bayesian phylogenetic analysis, we will perform a Markov chain Monte Carlo (MCMC) method to compute the posterior distributions of parameters, such as the rate of horizontal transmission and the mutation rates among trait variants, as well as the identity of trait variants in unobserved populations. Besides the inference of model parameters, our method enables the reconstruction of the genealogical tree of the focal trait, provided that the mutation rate is sufficiently small. We will also describe a heuristic algorithm to reduce the dimension of the parameter space explored in the MCMC method, where we simulate the coalescent process in the network of populations. Numerical examples show that our algorithms compute the posterior distribution of model parameters within a practical computation time, although the posterior distribution tends to be broad if we use uninformative priors.

SFTM: Fast matching of web pages using Similarity-based Flexible Tree Matching

Tree matching techniques have been investigated in many fields, including web data mining and extraction, as a key component to analyze the content of web pages. However, when applied to existing web pages, traditional tree matching approaches, covered by algorithms like Tree-Edit Distance (TED) or XyDiff, either fail to scale beyond a few hundred nodes or exhibit a relatively low accuracy.In this article, we therefore propose a novel algorithm, named Similarity-based Flexible Tree Matching (SFTM), which enables high accuracy tree matching on real-life web pages, with practical computation times. We approach tree matching as an optimization problem and leverage node labels and local topology similarity in order to avoid any combinatorial explosion. Our practical evaluation demonstrates that SFTM significantly improves the state of the art in terms of accuracy, while allowing computation times significantly lower than the most accurate solutions. By gaining on these two dimensions, SFTM therefore offers an affordable solution to match complex trees in practice.

A Modified Canny Detector-based Active Contour for Segmentation

In the present work, an integrated modified canny detector and an active contour were proposed for automated medical image segmentation. Since the traditional canny detector (TCD) detects only the edge’s pixels, which are insufficient for labelling the image, a shape feature was extracted to select the initial region of interest ‘IROI’ as an initial mask for the active contour without edge (ACWE), using a proposed modified canny detector (MCD). This procedure overcomes the drawback of the manual initialization of the mask location and shape in the traditional ACWE, which is sensitive to the shape of region of region of interest (ROI). The proposed method solves this problem by selecting the initial location and shape of the IROI using the MCD. Also, a post-processing stage was applied for more cleaning and smoothing the ROI. A practical computational time is achieved as the proposed system requires less than 5 minutes, which is significantly less than the required time using the traditional ACWE. The results proved the ability of the proposed method for medical image segmentation with average dice 87.54%.

A novel optimization method combining metaheuristics and machine learning for daily optimal operations in building energy and storage systems

In recent years, research on operational optimization of buildings and regional energy systems has been actively conducted. There are several groups that utilized linear approximations, considered nonlinearity, conducted scenario-based research, and used an optimization algorithm to find an optimum solution. In terms of real-world implementation in buildings, the nonlinearity of machine characteristics should be considered within practical computation time because linearization incurs modeling costs, and computational resources are limited. Hence, the authors propose a hybrid algorithm that consists of metaheuristics and machine learning for optimizing daily operating schedules in building energy systems. The deep neural network machine learning technique was used to predict optimal operations of integrated cooling tower systems, and metaheuristics were used to optimize the operation of the other components. The proposed method may reduce daily operating costs by more than 13.4%. In addition, the integrated cooling tower system evaluated in this study reduced cost and energy requirements compared to an individual cooling tower system.

Validation of a deep learning reconstruction framework for 3D delayed myocardial enhancement imaging

Abstract Funding Acknowledgements Type of funding sources: None. Introduction Myocardial delayed enhancement (MDE) MRI plays an important role in the identification of several cardiac conditions, both ischemic and non-ischemic (e.g. myocarditis, IDC, amyloidosis). 3D imaging offers increased resolution, full heart coverage and better depiction of complex pathologies, but its image quality is limited by long acquisition times. Deep learning (DL) models enable advanced reconstruction algorithms that yield regularized images in practical computation times. In this study we evaluate a novel 3D-DL reconstruction to overcome the trade-off between reconstructed quality and acquisition time on MDE data. Methods A group of 14 subjects referred for CMR (5 F / 9 M, 59 ± 11 y.o., 78 ± 13 kg) were scanned with a 3D MDE sequence prototype: SPGR with IR preparation, fat & spatial saturation, respiratory navigator, ARC 2x, FOV 40x40cm, ST 1.4-2.4mm, matrix 280²-320², FA 20deg, BW 62.5 kHz, TE 2.1 ± 0.1ms, TI based on a CINE IR scout. All were retrospectively reconstructed using a 3D DL algorithm, trained on a database of over 700 datasets to reconstruct high-quality images with adjustable noise reduction. The images were compared with standard 3D Cartesian reconstruction by two experienced cardiologists, to identify alterations in morphology or contrast distribution. Noise was estimated using the intensity standard deviation on a blood pool ROI. Feature preservation was estimated using the structural similarity index (SSI). Results The new method improved perceived image quality without loss of structural information or resolution (fig 1). Quantitative analysis (fig 2) confirmed these results: The average coefficient of variation in the blood was 0.08 ± 0.02 in the reference and 0.05 ± 0.02 with the new method; Given a target image noise level, DL reconstruction yielded up to 10% better SSI, compared to anisotropic filtering. The clinical review didn’t reveal diagnostically significant alterations of structure or uptake pattern. A perceived reduction of sharpness was initially reported but individual examination of landmarks (e.g. pulmonary and coronary arteries) confirmed that no relevant features were being lost with the new reconstruction. Discussion The 3D MDE images obtained with DL reconstruction improved the trade-off between image noise -estimated by the blood pool intensity deviation- and feature preservation -estimated by SSI-. Consistent improvement of image quality without morphological alterations of diagnostic relevance indicates that the new method can be considered for clinical practice. The next step in the validation process will require testing the robustness over a large set of cases with heterogeneous acquisition settings. Conclusion We presented the preliminary evaluation of a deep learning reconstruction method with 3D myocardial delayed enhancement data. The results show systematic improvement of overall image quality without loss of relevant diagnostic information. Abstract Figure.

Lossless Image Compression by Joint Prediction of Pixel and Context Using Duplex Neural Networks

This paper presents a new lossless image compression method based on the learning of pixel values and contexts through multilayer perceptrons (MLPs). The prediction errors and contexts obtained by MLPs are forwarded to adaptive arithmetic encoders, like the conventional lossless compression schemes. The MLP-based prediction has long been attempted for lossless compression, and recently convolutional neural networks (CNNs) are also adopted for the lossy/lossless coding. While the existing MLP-based lossless compression schemes focused only on accurate pixel prediction, we jointly predict the pixel values and contexts. We also adopt and design channel-wise progressive learning, residual learning, and duplex network in this MLP-based framework, which leads to improved coding gain compared to the conventional methods. Experiments show that the proposed method performs better than the conventional non-learning algorithms and also recent learning-based compression methods with practical computation time.

Distribution line downsizing based on a set of non-dominated solutions for a network reconfiguration problem of an electric distribution system with many objectives

To provide electric power supply to consumers even in abnormal situations, the equipment used in electric distribution systems, including distribution lines, have sufficient capacity margins. Maintenance and/or replacement of these lines involving network reconfigurations are routinely performed. Since the monetary costs of equipment investments are significant, downsizing the equipment is important from the standpoint of cost savings. In this paper, we propose a method to downsize distribution lines while maintaining the feasibility of the network reconfigurations. The proposed method consists of three steps. First, (1) sets of non-dominated solutions approximating the Pareto fronts of the network reconfiguration problems are found by a many-objective evolutionary algorithm. Second, (2) the distribution system operator specifies the preferred solutions among the sets found in (1). Finally, (3) a set of downsized lines is chosen from among the set of lines under the condition that the specified reconfiguration procedures remain feasible even after downsizing the lines. A computational experiment using a large-scale distribution system model demonstrates that the proposed method successfully identifies within a practical computational time those distribution lines that provide possibilities for downsizing and their downsized capacities.

Fully automated multiorgan segmentation in abdominal magnetic resonance imaging with deep neural networks.

Segmentation of multiple organs-at-risk (OARs) is essential for magnetic resonance (MR)-only radiation therapy treatment planning and MR-guided adaptive radiotherapy of abdominal cancers. Current practice requires manual delineation that is labor-intensive, time-consuming, and prone to intra- and interobserver variations. We developed a deep learning (DL) technique for fully automated segmentation of multiple OARs on clinical abdominal MR images with high accuracy, reliability, and efficiency. We developed Automated deep Learning-based abdominal multiorgan segmentation (ALAMO) technique based on two-dimensional U-net and a densely connected network structure with tailored design in data augmentation and training procedures such as deep connection, auxiliary supervision, and multiview. The model takes in multislice MR images and generates the output of segmentation results. 3.0-Tesla T1 VIBE (Volumetric Interpolated Breath-hold Examination) images of 102 subjects were used in our study and split into 66 for training, 16 for validation, and 20 for testing. Ten OARs were studied, including the liver, spleen, pancreas, left/right kidneys, stomach, duodenum, small intestine, spinal cord, and vertebral bodies. An experienced radiologist manually labeled each OAR, followed by reediting, if necessary, by a senior radiologist, to create the ground-truth. The performance was measured using volume overlapping and surface distance. The ALAMO technique generated segmentation labels in good agreement with the manual results. Specifically, among the ten OARs, nine achieved high dice similarity coefficients (DSCs) in the range of 0.87-0.96, except for the duodenum with a DSC of 0.80. The inference completed within 1min for a three-dimensional volume of 320×288×180. Overall, the ALAMO model matched the state-of-the-art techniques in performance. The proposed ALAMO technique allows for fully automated abdominal MR segmentation with high accuracy and practical memory and computation time demands.

Efficient diabatic quantum algorithm in number factorization

Adiabatic quantum algorithm is developed to reduce the time complexity of a classical nondeterministic polynomial problem to a polynomial one. However, our numerical and analytical results reveal that the evolution time of the factorization with a standard adiabatic quantum algorithm increases exponentially with the system size. By introducing the expectation of practical computation time TE, we find the optimal TE on the single evolution time requires the quantum algorithm to work in the diabatic regime, contrast to the traditional adiabatic quantum algorithm.

A Supervised-Learning-Based Strategy for Optimal Demand Response of an HVAC System in a Multi-Zone Office Building

The thermal capacity of buildings enables heating, ventilating, and air-conditioning (HVAC) systems to be exploited as demand response (DR) resources. Optimal DR of HVAC units is challenging, particularly for multi-zone buildings, because this requires detailed physics-based models of zonal temperature variations for HVAC system operation and building thermal conditions. Using supervised learning (SL), this paper proposes a new strategy for optimal DR of an HVAC system in a multi-zone building. Artificial neural networks (ANNs) are trained with data obtained under normal building operating conditions. The ANNs are replicated using piecewise linear equations, which are explicitly integrated into an optimization problem for price-based DR. The problem is solved for various electricity prices and building thermal conditions. The solutions are used to train a deep neural network (DNN) to directly determine the optimal DR schedule, which is termed as SL-aided meta-prediction (SLAMP) here; the DNN can work as a price-and-optimal-demand curve. Case studies are performed using three different methods: explicit ANN replication, SLAMP, and physics-based modeling. The case study results verify that the proposed SL-based strategy is effective in terms of both practical applicability and computational time, while also ensuring the thermal comfort of occupants and cost-effective operation of the HVAC system.

A Comprehensive Study of the Efficiency of Type-Reduction Algorithms

Improving the efficiency of type-reduction algorithms continues to attract research interest. Recently, there has been some new type-reduction approaches claiming that they are more efficient than the well-known algorithms such as the enhanced Karnik–Mendel (EKM) and the enhanced iterative algorithm with stopping condition (EIASC). In a previous paper, we found that the computational efficiency of an algorithm is closely related to the platform, and how it is implemented. In computer science, the dependence on languages is usually avoided by focusing on the complexity of algorithms (using big O notation). In this article, the main contribution is the proposal of two novel type-reduction algorithms. Also, for the first time, a comprehensive study on both existing and new type-reduction approaches is made based on both algorithm complexity and practical computational time under a variety of programming languages. Based on the results, suggestions are given for the preferred algorithms in different scenarios depending on implementation platform and application context.

Profile-Splitting Linearized Bregman Iterations for Trend Break Detection Applications

Trend break detection is a fundamental problem that materializes in many areas of applied science, where being able to identify correctly, and in a timely manner, trend breaks in a noisy signal plays a central role in the success of the application. The linearized Bregman iterations algorithm is one of the methodologies that can solve such a problem in practical computation times with a high level of accuracy and precision. In applications such as fault detection in optical fibers, the length N of the dataset to be processed by the algorithm, however, may render the total processing time impracticable, since there is a quadratic increase on the latter with respect to N. To overcome this problem, the herewith proposed profile-splitting methodology enables blocks of data to be processed simultaneously, with significant gains in processing time and comparable performance. A thorough analysis of the efficiency of the proposed methodology stipulates optimized parameters for individual hardware units implementing the profile-splitting. These results pave the way for high performance linearized Bregman iteration algorithm hardware implementations capable of efficiently dealing with large datasets.

Liquid-solid interaction sound synthesis

Liquid-solid interaction produces a characteristic sound which is different from the sound of bubbly water flows. This paper explores liquid-solid interaction sound (LSIS) synthesis. The resulting LSIS consists of two components: bubble sound and impact sound, which corresponding to liquid-liquid collision and liquid-solid collision, respectively. To improve the quality of LSIS, we propose a novel sound enrichment method called Feature Transfer Synthesis (FTS), which is designed to compensate for the differences between the real-world recording and the synthesized sound. We also greatly resolve the synchronization problem during blending the two components of LSIS through a key frame algorithm with normal force and grid surface. Moreover, a generalized dipole model for sound radiation is performed to estimate the LSIS pressure at a listener position. We illustrate our approach through a series of experiments and a perceptual user study, demonstrating the utility of our LSIS synthesis pipeline in producing realistic sounds at practical computational times.

Robust Classification

Motivated by the fact that there may be inaccuracies in features and labels of training data, we apply robust optimization techniques to study in a principled way the uncertainty in data features and labels in classification problems and obtain robust formulations for the three most widely used classification methods: support vector machines, logistic regression, and decision trees. We show that adding robustness does not materially change the complexity of the problem and that all robust counterparts can be solved in practical computational times. We demonstrate the advantage of these robust formulations over regularized and nominal methods in synthetic data experiments, and we show that our robust classification methods offer improved out-of-sample accuracy. Furthermore, we run large-scale computational experiments across a sample of 75 data sets from the University of California Irvine Machine Learning Repository and show that adding robustness to any of the three nonregularized classification methods improves the accuracy in the majority of the data sets. We observe the most significant gains for robust classification methods on high-dimensional and difficult classification problems, with an average improvement in out-of-sample accuracy of robust versus nominal problems of 5.3% for support vector machines, 4.0% for logistic regression, and 1.3% for decision trees.

Resource optimization of optical aggregation network for efficient software-defined datacenters

Software-defined datacenters (SDDCs) have recently received much attention for rapid delivery of cloud-based applications. In this paper, we consider the design of an optical aggregation network for cost-efficient software-defined datacenters (SDDCs). Here, an optical aggregation network connects a server accommodating virtual network functions (VNFs) with access gateway nodes (AGNs) via a time-division-multiplexing (TDM)-based point-to-multipoint (P2MP) wavelength path to aggregate traffic from geographically distributed end users to cloud datacenters. Each VNF must be placed on an adequate server in consideration of the efficiency of wavelength paths to reduce network cost for cloud access. However, existing VNF placement algorithms determine VNF placement without considering the efficiency of wavelength paths, which deteriorates network performance and increases network cost. To solve this problem, VNFs must be placed so that a P2MP wavelength path can be efficiently shared by multiple AGNs to reduce network cost. For this purpose, we propose a VNF placement algorithm, called the decomposition-based VNF placement algorithm (DVA), in a TDM-wavelength-division-multiplexing (WDM)-based optical aggregation network for cost-efficient SDDCs. DVA can find approximate solutions of sufficient quality within practical computation time.