Analytical Network Model Research Articles

A Sampling Strategy for High-Dimensional, Simulation-Based Transportation Optimization Problems

When tackling high-dimensional, continuous simulation-based optimization (SO) problems, it is important to balance exploration and exploitation. Most past SO research focuses on the enhancement of exploitation techniques. The exploration technique of an SO algorithm is often defined as a general-purpose sampling distribution, such as the uniform distribution, which is inefficient at searching high-dimensional spaces. This work is motivated by the formulation of exploration techniques that are suitable for large-scale transportation network problems and high-dimensional optimization problems. We formulate a sampling mechanism that combines inverse cumulative distribution function sampling with problem-specific structural information of the underlying transportation problem. The proposed sampling distribution assigns greater sampling probability to points with better expected performance as defined by an analytical network model. Validation experiments on a toy network illustrate that the proposed sampling distribution has important commonalities with the underlying and typically unknown true sampling distribution of the simulator. We study a high-dimensional traffic signal control case study of Midtown Manhattan in New York City. The results show that the use of the proposed sampling mechanism as part of an SO framework can help to efficiently identify solutions with good performance. Using the analytical information for exploration, regardless of whether it is used for exploitation, outperforms benchmarks that do not use it, including standard Bayesian optimization. Using the analytical information for exploration only yields solutions with similar performance than when the information is used for exploitation only, reducing the total compute times by 65%. This paper sheds light on the importance of developing suitable exploration techniques to enhance both the scalability and the compute efficiency of general-purpose SO algorithms. Funding: T. Tay thanks the Agency for Science, Technology and Research (A*STAR) Singapore for funding his work. Supplemental Material: The online appendix is available at https://doi.org/10.1287/trsc.2023.0110 .

Control of Intrinsic ThermoAcoustic instabilities for methane and hydrogen air flames: DNS and network analysis

Replacing methane by hydrogen in premixed burners is a typical decarbonation scenario. This is usually done by ensuring a constant flame speed, i.e. by using ultra-lean H2 mixtures. This change can modify the thermoacoustic properties of burners. This study focuses on one class of thermoacoustic instabilities, Intrinsic ThermoAcoustic (ITA) modes, for flames stabilized on a diaphragm. An acoustic network approach is built to construct an ITA stability criterion (called ITAS criteria) for a three-duct configuration and a fully analytical expression is obtained in the limit of thin diaphragms. It shows that instability will grow when the Flame Transfer Function (FTF) gain n−π, located at the frequency at which u′-q′ phase ϕ=−π, is greater than a critical gain nc expressed by the ITAS criteria. Two-dimensional DNS, with anechoic terminations, of methane–air and hydrogen–air flames stabilized on a diaphragm are simulated with similar laminar flame speed sL0 and theoretical flame length, with an equivalence ratio of 0.73 and 0.4 respectively. The theory developed allows to analyze the differences with these two strong unstable ITA modes. For ultra-lean hydrogen flame, preferential diffusion near the diaphragm lips and tip opening reduces the flame time delay and shifts the ITA frequency towards higher values compared to methane. The assessment of two new control strategies is possible with the help of the analytical network model by increasing nc. A first strategy involves the use of preheated leaner mixtures: increasing Tu and reducing ϕ while keeping the flame speed sL0 constant. A second strategy consists of modifying the combustor geometry by changing the section ratio S2/S0, between plenum and combustion chamber. Both strategies successfully mitigate the hydrogen–air ITA modes and seems promising strategies for methane cases despite the fact that instabilities are not fully mitigated for this fuel at the conditions of interest in this study.

A data-driven discrete simulation-based optimization algorithm for car-sharing service design

This paper formulates a discrete simulation-based optimization (SO) algorithm for a family of large-scale car-sharing service design problems. We focus on the profit-optimal assignment of vehicle fleet across a network of two-way (i.e., round-trip) car-sharing stations. The proposed approach is a metamodel SO approach. A novel metamodel based on a mixed-integer program (MIP) is formulated. The metamodel is embedded within a general-purpose discrete SO algorithm. The proposed algorithm is validated with synthetic toy network experiments. The algorithm is then applied to a high-dimensional Boston case study using reservation data from a major US car-sharing operator. The method is benchmarked versus several algorithms, including stochastic programming. The experiments indicate that the analytical network model information, provided by the MIP to the SO algorithm, is useful both at the first iteration of the algorithm and across subsequent iterations. The solutions derived by the proposed method are benchmarked versus the solution deployed in the field by the car-sharing operator. Via simulation, the proposed solutions improve those deployed with an average improvement of profit of 6% and of vehicle utilization of 3%.The combination of the problem-specific analytical MIP with a general-purpose SO algorithm enables the discrete SO algorithm to: (i) address high-dimensional problems, (ii) become computationally efficient (i.e., it can identify good quality solutions within few simulation observations), (iii) become robust to the quality of the initial points and of the stochasticity of the simulator. More generally, the information provided by the MIP to the SO algorithm enables it to exploit problem-specific structural information. This leads to an algorithm with both asymptotic convergence guarantees as well as good short term performance (i.e., performance given few simulation observations). We view this general idea of combining analytical MIP formulations with general-purpose SO algorithms, or more broadly with general-purpose sampling strategies of high-resolution data, as an innovative and promising area of future research.

Modeling and structural optimization design of switched reluctance motor based on fusing attention mechanism and CNN-BiLSTM

Six-degrees of freedom (6-DOF) parallel mechanisms driven by switched reluctance motors (SRMs) can realize flexible control with high precision. Efficiency is an important indicator to measure the speed control system of SRMs. There are many characteristic factors affecting efficiency and strong nonlinear relationships between different characteristic parameters, which makes it difficult for analytical models and traditional neural network models to express their spatial correlation. For this reason, a convolutional neural network (CNN)-bidirectional long short-term memory network (BiLSTM) efficiency regression prediction model (CNN-BiLSTM-SENet) that integrates the attention mechanism (SENet) is proposed. Firstly, customize a formula for sensitivity analysis to screen characteristic parameters of efficiency. Secondly, build the CNN-BiLSTM-SENet model and use sparrow search algorithm (SSA) for hyperparameter optimization, input data to CNN to extract high-dimensional feature vectors that reflect complex changing relationships between features and efficiency while establishing feature channels. Embed the SENet to adaptively perceive and assign different weights to the feature channels, enhancing the influence of key features. Input the feature vectors outputted by the front-end network to BiLSTM to bidirectionally learn coupling relationships between sequences and complete regression prediction. Finally, propose improved northern goshawk optimization (MNGO) to solve the regression model to obtain the maximum efficiency and corresponding characteristic parameters. The results proved that the SSA-optimized CNN-BiLSTM-SENet model has higher fitting exactness and better prediction effect for efficiency regression prediction, and the MNGO also has stronger search ability and faster convergence.

A novel coupled framework for detecting hotspots of methane emission from the vulnerable Indian Sundarban mangrove ecosystem using data-driven models

Coastal mangroves have been lost to deforestation for anthropogenic activities such as agriculture over the past two decades. The genesis of methane (CH4), a significant greenhouse gas (GHG) with a high potential for global warming, occurs through these mangrove beds. The mangrove forests in the Indian Sundarban deltaic region were studied for pre-monsoonal and post-monsoonal variations of CH4 emission. Considering the importance of CH4 emission, a process-based spatiotemporal (PBS) and an analytical neural network (ANN) model were proposed and used to estimate the amount of CH4 emission from different land use land cover classes (LULC) of mangroves. The field work was performed in 2020, and gas samples of various LULC were directly collected from the mangrove bed using the enclosed box chamber method. Historical climatic data (1960–1989) were used to predict future climate scenarios and associated CH4 emissions. The analysis and estimation activities were carried out utilizing satellite images from the pre-monsoonal and post-monsoonal seasons of the same year. The study revealed that pre-monsoonal CH4 emission was higher in the south-west and northern parts of the deforested mangrove of the Indian Sundarban. A sensitivity study of the anticipated models was conducted using a variety of environmental input parameters and related main field observations. The measured precision area under curve of receiver operating characteristics was 0.753 for PBS and 0.718 for ANN models, respectively. The temperature factor (Tf) was the most crucial variable for CH4 emissions. Based on the PBS model with coupled model intercomparison project-6 temperature data, a global circulation model was run to predict increasing CH4 emissions up to 2100. The model revealed that the agricultural lands were the prime emitters of CH4 in the Sundarban mangrove ecosystem.

Measurement and mitigation of the “wild goose chase” phenomenon in taxi services

This paper examines the so-called wild goose chase phenomenon in taxi systems and proposes a new mitigation strategy based on dynamic vehicle swaps. An analytic queuing network model, including a system of differential equations, is derived to yield approximate formulas for the expected system performance in the steady-state equilibria under the vehicle swap strategy. This model is solved numerically to yield equilibrium points and the expected system performance metrics. To corroborate the theoretical predictions, a new data-processing method based on dynamic programming is proposed to identify and measure multiple distinct equilibrium states from observed data. Agent-based simulations are conducted to generate examples of such data, and to verify the formulas. Numerical results show that (i) the proposed measurement method works well in identifying the distinct equilibria, and (ii) the vehicle swap strategy works quite effectively in mitigating the wild goose chase phenomenon.

Low-cost quasi-global optimization of expensive electromagnetic simulation models by inverse surrogates and response features

Conceptual design of contemporary high-frequency structures is typically followed by a careful tuning of their parameters, predominantly the geometry ones. The process aims at improving the relevant performance figures, and may be quite expensive. The reason is that conventional design methods, e.g., based on analytical or equivalent network models, often only yield rough initial designs. This is especially the case for miniaturized components featuring considerable electromagnetic (EM) cross couplings, or antenna systems with non-negligible radiator coupling (e.g., MIMO, closely-spaced arrays). For reliability reasons, parametric optimization is carried out using EM simulation tools, which is a time-consuming task. In many cases, designer needs to resort to a global search, especially when handling several objectives and constraints is necessary, or the high-frequency structure under design is overly complex. Combination of both aforementioned factors makes it no longer possible to rely on engineering insight, even to detect a promising region of the design space. Unfortunately, nature-inspired algorithms, commonly employed for solving these tasks typically exhibit significant computational expenditures. This paper proposes a simple yet efficient method for globalized search using a response feature approach and inverse regression surrogates. Owing to less nonlinear dependence of the feature point coordinates on the system variables (as compared to the original responses, e.g., S-parameter frequency characteristics), our methodology permits a rapid identification of the most appropriate regions of the parametric space, and further design tuning by means of local routines. At the same time, the overall optimization cost is comparable to the cost of local procedures. The proposed approach is validated using several high-frequency structures (a dual-band antenna, a microstrip coupler, an impedance matching transformer) optimized under different design scenarios. Global search capability and computational efficiency are demonstrated through comprehensive comparisons with multiple-start local search, as well as particle swarm optimizer, a representative nature-inspired algorithm.

Interplay of Clusters of Acoustic and Intrinsic Thermoacoustic Modes in Can-Annular Combustors

Abstract Thermoacoustic systems can exhibit self-excited instabilities of two nature, namely cavity modes or intrinsic thermoacoustic (ITA) modes. In heavy-duty land-based gas turbines with can-annular combustors, the cross-talk between cans causes the cavity modes of various azimuthal order to create clusters, i.e., ensembles of modes with close frequencies. Similarly, in systems exhibiting rotational symmetry, ITA modes also have the peculiar behavior of forming clusters. In the present study, we investigate how such clusters interplay when they are located in the same frequency range. We first consider a simple Rijke tube configuration and derive a general analytical low-order network model using only dimensionless numbers. We investigate the trajectories of the eigenmodes when changing the downstream length and the flame position. In particular, we show that ITA and acoustic modes can switch nature and their trajectories are strongly influenced by the presence of exceptional points. We then study a generic can-annular combustor. We show that such configuration can be approximated by an equivalent Rijke tube. We demonstrate that, in the absence of mean flow, the eigenvalues of the system necessarily lie on specific trajectories imposed by the upstream conditions.

Direct constraint control for EM-based miniaturization of microwave passives

Handling constraints imposed on physical dimensions of microwave circuits has become an important design consideration over the recent years. It is primarily fostered by the needs of emerging application areas such as 5G mobile communications, internet of things, or wearable/implantable devices. The size of conventional passive components is determined by the guided wavelength, and its reduction requires topological modifications, e.g., transmission line folding, or utilization of compact cells capitalizing on the slow-wave phenomenon. The resulting miniaturized structures are geometrically complex and typically exhibit strong cross coupling effects, which cannot be adequately accounted for by analytical or equivalent network models. Consequently, electromagnetic (EM)-driven parameter tuning is necessary, which is computationally expensive. When the primary objective is size reduction, the optimization task becomes far more challenging due to the presence of constraints related to electrical performance figures (bandwidth, power split ratio, etc.), which are all costly to evaluate. A popular solution approach is to utilize penalty functions. Therein, possible violations of constraints degrade the primary objective, thereby enforcing their satisfaction. Yet, the appropriate setup of penalty coefficients is a non-trivial problem by itself, and is often associated to extra computational expenses. In this work, we propose an explicit approach to constraint handling, which is combined with the trust-region gradient-search procedure. In our technique, the decision about the adjustment of the search radius is determined based on the reliability of rendering the feasible region boundary by linear approximation models of the constraints. Comprehensive numerical experiments conducted using three miniaturized coupler structures demonstrate superiority of the presented method over the penalty function paradigm. Apart from the efficacy, its appealing features include algorithmic simplicity, and no need for tailoring the procedure for a particular circuit to be optimized.

Synergic use of neural networks model and remote sensing algorithms to estimate water clarity indicators in Khanpur reservoir, Pakistan

Freshwater reservoirs are limited and facing issues of over-exploitation, climate change effects and poor maintenance which have serious consequences for water quality. Developing countries face the challenge of collecting in situ information on ecological status and water quality of these reservoirs due to constraints of cost, time and infrastructure. In this study, a practical method of retrieval of two water clarity indicators, total suspended matter and secchi disk depth, using Sentinel-2 satellite data is adopted for preliminary assessment of water quality and trophic conditions in Khanpur reservoir, Pakistan. The study explores the synergy of utilizing two independent models, i.e., case 2 regional coast color analytical neural network model and semiempirical remote sensing algorithms to understand the spatiotemporal dynamics of water clarity patterns in the dammed reservoir, in the absence of ground measurements. The drinking water quality and trophic state of the reservoir water is determined based purely on satellite measurements. Out of the five months studied, the reservoir water has high turbidity and poor eutrophic status in three months. The results from both computational models are compared, which exhibit a high degree of statistical agreement. The study demonstrates the effective utilization of relevant analytical and semiempirical methods on satellite data to map water clarity indicators and understand their dynamics in both space and time. This solution is particularly useful for regions where routine ground sampling and observation of environmental variables are absent.

Modeling and optimization of hybrid ground source heat pump with district heating and cooling

Hybrid heating systems with ground source heat pumps (GSHP) and district heating and cooling offer flexibility in operation to both building owners and energy providers. The flexibility can be used to make the heating system more economical and environmentally friendly. However, due to the lack of suitable models that can accurately predict the long-term performance of the GSHP, there is uncertainty in their performance and concerns about the long-term stability of the ground temperature, which has limited the utilization of such hybrid heating systems. This work presents a hybrid model of a GSHP system that uses analytical and artificial neural network models to accurately represent a GSHP system's long-term behavior. A method to improve the operation of a hybrid GSHP is also presented. The method was applied to hospital buildings in northern Sweden. It was shown that in the improved case, the cost of providing heating to the building can be reduced by 64 t€, and the CO2 emissions can be reduced by 92 tons while maintaining a stable ground temperature.

Differentiated Toll Optimization on Goods Vehicles Based on Prediction of Demand for Large-Scale Network

The area rate of the toll road network is designed to influence the regional road network traffic flow configuration and an important factor of goods vehicle path selection, transportation, and other related departments in formulating the related standard. The toll rate is usually combined with the local passenger and cargo loading conditions. Adjusting and optimizing the charge policy, optimization of processes and methods is still facing many difficulties. In this paper, the model parameters and toll status of various goods vehicles are considered, and an optimization algorithm of toll rates for different vehicles in a large-scale road network is proposed. By constructing a traffic demand prediction model based on multiple truck classification, this paper analyzes the distribution status of network trucks and the final charging under the influence of different rates. At the bottom of the algorithm is an analytical model composed of nonlinear equations whose dimensions scale linearly with the scale of the road network. The scale of the road network is independent of the path set and link attribute. The model is verified and the parameters are calibrated by a small network. Finally, the Road network is selected for empirical analysis in Xinjiang Autonomous Region. A case study shows that the analytical structure information provided by the analytic network model enables the algorithm to quickly identify high quality solutions, and the algorithm has better simulation accuracy and premium rate design advantages.

Analytical model for solute transport in discrete fracture networks: 2D spatiotemporal solution with matrix diffusion

A computationally efficient analytical network (AN) model simulating solute transport in discrete fracture networks (DFNs) is developed. The model simulates two-dimensional spatial and temporal distribution of a solute considering advection and hydrodynamic dispersion within the fractures, matrix diffusion, sorption onto the fracture walls and in the matrix, and first order decay for one constituent. The AN model is based on previously developed analytical solutions for a single fracture, and the convolution method is applied to extend these solutions to a DFN. Mass sharing at fracture intersections is calculated using the complete mixing and stream-tube methods. The AN model was verified against numerical models based on the time domain random walk and random walk particle tracking methods using two different fracture networks with a range of properties. In all cases, the AN model solutions showed excellent agreement with the numerical model solutions. The sensitivity of the mass sharing methods to the dominant transport mechanism (i.e., advection or diffusion) was investigated for Peclet numbers ranging from Pe = 3 × 10−6 - 380. Both mass sharing methods give the same results when the transport processes are advection-dominated and matrix diffusion is considered; however, attention must be paid to the mass sharing method employed under other conditions, particularly when the fracture density is small. The AN model was at least 97% more efficient than the numerical models used in this work, and this efficiency will increase with network complexity. The AN model provides a useful reference tool for the verification of numerical dual-porosity fracture network simulations.

Practical Model for Evaluating Project Manager Competencies in Construction Projects

Due to the importance role of the behavioural characteristics of the project manager in construction projects and the need for effective and useful communication with other stakeholders of the project, studies have not comprehensively examined the behavioural competencies of the project manager and the lack of a practical model in the Iranian construction industry to select a capable project manager feels tangible. This research is on a more systematic approach to address the behavioural capabilities of the project manager by considering the factors related to his behavioural variables to evaluate the behavioural capabilities of the project manager. Using a multi-stage research process, including extensive literature review, analysis and a combination of important factors in the behavioural competencies of project managers, the development of an analytical network model (AHP, TOPSIS) was considered. This model would be a decision support tool for the employer organization to compare project managers according to their behavioural competencies by creating a database. This protects construction companies from the effective challenges posed by the performance of project managers and helps them to evaluate the performance of their project managers. It also enables project managers to focus on their efforts to address behavioural practices to improve their project performance.

Efficient Simulation-Based Toll Optimization for Large-Scale Networks

This paper proposes a simulation-based optimization technique for high-dimensional toll optimization problems of large-scale road networks. We formulate a novel analytical network model. The latter is embedded within a metamodel simulation-based optimization (SO) algorithm. It provides analytical and differentiable structural information of the underlying problem to the SO algorithm. Hence, the algorithm no longer treats the simulator as a black box. The analytical model is formulated as a system of nonlinear equations that can be efficiently evaluated with standard solvers. The dimension of the system of equations scales linearly with network size. It scales independently of the dimension of the route choice set and of link attributes such as link length. Hence, it is a scalable formulation suitable for the optimization of large-scale networks. For instance, the model is used in the case study of the paper for toll optimization of a Singapore network with more than 4,050 OD (origin-destination) pairs and 18,200 feasible routes. The corresponding analytical model is implemented as a system of 860 nonlinear equations. The analytical network model is validated based on one-dimensional toy network problems. It captures the main trends of the simulation-based objective function and, more importantly, accurately locates the global optimum for all experiments. The proposed SO approach is then used to optimize a set of 16 tolls for the network of expressways and major arterials of Singapore. The proposed method is compared with a general-purpose algorithm. The proposed method identifies good quality solutions at the very first iteration. The benchmark method identifies solutions with similar performance after 2 days of computation or similarly after more than 30 points have been simulated. The case study indicates that the analytical structural information provided to the algorithm by the analytical network model enables it to (i) identify good quality solutions fast and (ii) become robust to both the quality of the initial points and to the stochasticity of the simulator. The final solutions identified by the proposed algorithm outperform those of the benchmark method by an average of 18%.

Special Issue on Programming Models and Applications for Multicores and Manycores 2020

This special issue focuses on all aspects of parallel programming on multicore and manycore architectures such as programming models and systems, applications algorithms, performance analysis, and debugging and performance tools. It includes selected articles from the 2020 International Workshop on Programming Models and Applications for Multicores and Manycores (PMAM 2020), which was colocated with the PPoPP 2020 conference, in San Diego, California, February 22–26, 2020. From high-end servers to mobile phones, multicore and manycore systems are steadily entering every aspect of the information technology landscape. While this advanced technology enables new opportunities for scientific discovery, it also opens new challenges. In particular, although program developers have been trained in parallel programming, most existing parallel programs are still prone to errors such as data race and deadlock and suffer from suboptimal performance. In order to fully explore the computing power of multicore and manycore hardware architectures, innovations in parallel programming models as well as the associated development and execution eco-systems are urgently needed to enable delivery of error-free and high-performance parallel programs. The PMAM 2020 workshop provided a discussion forum for people interested in programming environments, models, tools, and applications specifically designed for parallel multicore and manycore hardware environments. The four outstanding contributions for this special issue were selected from eight workshop presentations. The accepted publications cover several emerging research directions, specifically focusing on programming models and runtime techniques on manycore and heterogeneous computing systems for graph and streaming applications, as summarized below. In “A Self-Adjusting Task Granularity Mechanism for the Java Lifeline-Based Global Load Balancer Library,” Finnerty et al. present a tuning mechanism that dynamically adjusts the sensitive task granularity in the load balancing scheme of X10 for Java parallel programs and demonstrate the improved performance of four applications on two manycore supercomputers. In “A Performance Predictor for Implementation Selection of Parallelized Static and Temporal Graph Algorithms,” Rehman et al. analyze the performance of graph workload with different algorithm and hardware conditions. They propose an inter-implementation predictor to choose the best performing parallel implementation for static and temporal graph benchmarks and inputs when executing on a CPU or a GPU architecture by leveraging analytical and neural network models. In “Sharing Non-Cache-Coherent Memory with Bounded Incoherence,” Ren et al. introduce a memory consistency approach, called bounded incoherence, that enables cached access to shared data-structures in non-cache-coherent memory. The proposed model ensures that updates to memory on one node are visible within at most a bounded amount of time on all other nodes. Test results show 30% performance improvement of the PowerGraph graph processing framework over state-of-the-art distributed approaches. In “Code Generation for Energy-Efficient Execution of Dynamic Streaming Task Graphs on Parallel and Heterogeneous Platforms,” Litzinger et al. enhance the programming model of streaming applications by introducing dynamic elements into the task graphs so that the runtime system can remap tasks and adapt the change of streaming-in task structures at runtime. The authors provide a prototype implementation of the toolchain and demonstrate low overhead and low energy consumption of the proposed remapping technique. We hope that the readers of the PMAM 2020 special issue will find the articles insightful in the field of parallel programming for multicore and manycore systems. The guest editors would like to thank all the reviewers and the Concurrency and Computation Practice and Experience Wiley office staff who worked hard to make this high-quality journal issue happen, while facing challenges from the COVID-19 pandemic. Data sharing is not applicable to this article as no new data were created or analyzed in this study.

Bridging the analytical and artificial neural network models for keyhole formation with experimental verification in laser melting deposition: A novel approach

Recent scientific and technological developments demonstrated that laser melting deposition (LMD) could yield near-net-shape parts by additive manufacturing. Under context, it was noticed that the primary operating parameters (i.e., laser power, scanning speed, and powder flow rate) significantly influence the generated keyhole's dimensions, ultimately affecting the deposited clad characteristics. In this study, a novel approach is proposed to generate a pool of datasets to implement artificial neural networking (ANN) in manufacturing process automation. A mathematical model was developed to approximate the keyhole's top and bottom widths and penetration depth, depending on the primary operating parameters. Single-layers of AISI 304 stainless steel were deposited via LMD to verify the mathematical results. It was shown that the mathematical model had predictions close to the experimental results. The validated model was used in correlation with the ANN model based upon 3-10-3 (no. of inputs-no. of neurons in hidden layer-no. of outputs) architecture. The outputs based on the given inputs, via the verified mathematical model, were used for the ANN model training. The results of experiments were compared with mathematical and ANN models. It was found that an increase in the laser scanning speed decreases the deposited layer's width, and a direct correlation has been inferred between powder flow rate and laser power with layer width. An increment in the scanning speed reduces the layer height and keyhole's dimensions. A direct correlation has been observed between powder flow rate and layer height, irrespective of the keyhole's dimensions. Laser power was in a direct relationship with layer height and keyhole dimensions. The developed mathematical model can estimate the keyhole's dimensions with an accuracy of (5–9) %. However, the output predicted for keyhole's dimensions by ANN, in a range of (2–3.5) %, is much closer to the experimental results, which identifies the ANN as a potential tool for the 3D printing process automation.

Engineered nanoparticle network models for autonomous computing.

Materials that exhibit synaptic properties are a key target for our effort to develop computing devices that mimic the brain intrinsically. If successful, they could lead to high performance, low energy consumption, and huge data storage. A 2D square array of engineered nanoparticles (ENPs) interconnected by an emergent polymer network is a possible candidate. Its behavior has been observed and characterized using coarse-grained molecular dynamics (CGMD) simulations and analytical lattice network models. Both models are consistent in predicting network links at varying temperatures, free volumes, and E-field (E⃗) strengths. Hysteretic behavior, synaptic short-term plasticity and long-term plasticity-necessary for brain-like data storage and computing-have been observed in CGMD simulations of the ENP networks in response to E-fields. Non-volatility properties of the ENP networks were also confirmed to be robust to perturbations in the dielectric constant, temperature, and affine geometry.

A performance predictor for implementation selection of parallelized static and temporal graph algorithms

AbstractTask‐based execution of graph workloads allows various ordered and unordered implementations, with tasks representing dependencies between graph vertices and edges. This work explores graph algorithms in the context of ordered and unordered task‐based implementations, that trade‐off work‐efficiency with parallelism. The monotonicity of convergent graph solutions is the reason behind the trade‐off between work‐efficiency and parallelism. This trade‐off results in variable performance‐based choices within and across different machines (CPUs and GPUs), graph algorithms, implementations (ordered, relaxed, and unordered). Input graphs also augment this choice space, with this work analyzing temporally changing graphs in addition to the static graphs explored by prior works. These algorithmic and architectural choices are first explored in this work, and it is seen that different graph workload‐input combinations perform optimally on diverse architectural configurations. The resulting choice space is analyzed and this work represents it in the form of characteristic variables that correlate with each choice space. Using these characteristic variables, this work proposes analytical and neural network models to correlate these choice spaces to find the best performing implementation. The variables and the prediction models proposed in this work are also integrated with a state‐of‐the‐art performance predictor on a multiaccelerator setup, and shows geometric performance gains of 54% on a CPU, 14% on a GPU, and 31.5% in a multiaccelerator setup over baseline implementations without performance prediction.

Modeling Multi-Year Customers’ Considerations and Choices in China’s Auto Market Using Two-Stage Bipartite Network Analysis

Choice modeling is important in transportation planning, marketing and engineering design, as it can quantify the influence of product attributes and customer demographics on customers’ choice behaviors. Consumer studies suggest that customers’ choice-making process often consists of two different stages: customers first consider subsets of available products on the market, and then make the final choice from the subsets. As existing preference modeling is mostly focused on the choice stage, there is a need to develop methods for understanding customer preferences at both stages, and investigate how customer preferences change from “consideration” to “choice”, and whether such changes will be consistent over time. In this paper, we study customers’ consideration and purchase behaviors in China’s auto market using multi-year survey datasets. We demonstrate how descriptive network analysis and analytic network models (bipartite Exponential Random Graph Model (ERGM)) capture the change of customers’ preferences from the consideration stage to the choice stage in multiple consecutive years. Our results show that factors such as fuel consumption per unit power, car make origin, and place of production influence customers’ considerations and final purchase decisions in different ways, and this difference between consideration and purchase is consistent over time. The main contribution of this study is that we validate the two-stage network-based modeling approach and its utility in preference elicitation using multiple-year dataset, which sheds lights on understanding the trend of customers’ consideration and choice behaviors across years. Our study also contributes to a refined interpretation of the ERGM results with categorization of continuous variables into ranges, which shows that customer choice decisions may be more qualitatively influenced by product attributes rather than quantitatively. Our approach is generic and thus can be applied to solving broader choice modeling problems, such as the transportation mode selection and the adoption of clean technology (e.g., electric vehicles).