Topology Changes Research Articles

GrainGNN: A dynamic graph neural network for predicting 3D grain microstructure

We propose GrainGNN, a surrogate model for the evolution of polycrystalline grain structure under rapid solidification conditions in metal additive manufacturing. High fidelity simulations of solidification microstructures are typically performed using multicomponent partial differential equations (PDEs) with moving interfaces. The inherent randomness of the PDE initial conditions (grain seeds) necessitates ensemble simulations to predict microstructure statistics, e.g., grain size, aspect ratio, and crystallographic orientation. Currently such ensemble simulations are prohibitively expensive and surrogates are necessary.In GrainGNN, we use a dynamic graph to represent interface motion and topological changes due to grain coarsening. We use a reduced representation of the microstructure using hand-crafted features; we combine pattern finding and altering graph algorithms with two neural networks, a classifier (for topological changes) and a regressor (for interface motion). Both networks have an encoder-decoder architecture; the encoder has a multi-layer transformer long-short-term-memory architecture; the decoder is a single layer perceptron.We evaluate GrainGNN by comparing it to high-fidelity phase field simulations for in-distribution and out-of-distribution grain configurations for solidification under laser power bed fusion conditions. GrainGNN results in 80%–90% pointwise accuracy; and nearly identical distributions of scalar quantities of interest (QoI) between phase field and GrainGNN simulations compared using Kolmogorov-Smirnov test. GrainGNN's inference speedup (PyTorch on single x86 CPU) over a high-fidelity phase field simulation (CUDA on a single NVIDIA A100 GPU) is 150×–2000× for 100-initial grain problem. Further, using GrainGNN, we model the formation of 11,600 grains in 220 seconds on a single CPU core.

The influence of change in river networks on waterway connectivity and flooding processes

Urbanization disrupts river networks, alters waterway connectivity, and elevates flood control pressure. Quantitatively assessing river network changes and its impact on waterway connectivity and flooding processes is imperative. This study constructed a waterway connectivity evaluation model and a hydrologic-hydrodynamic model to assess the river network, the waterway connectivity and the flooding change in the Qinhuai River Basin, and used scenario analysis to get the effectiveness of different topology changes. The results showed a 0.09 decline in connectivity from 2000 to 2010, which led to a 1-hour earlier peak occurrence time, a 2.59 m3/s increase in peak discharge and a 51 mm rise in peak water level. Scenario analysis demonstrated that adding tributaries can enhance both connectivity and flood resilience. Increasing river width effectively reduces peak water levels while minimally affecting connectivity. However, extending river length has a negligible effect on peak water levels but may decrease waterway connectivity.

Study of Microstructure, Texture, and Cooking Qualities of Reformulated Whole Wheat Flour Pasta by Substituting Water with Stearic Acid–Candelilla Wax–Groundnut Oil Oleogel

Oleogels, which are traditionally utilized to reduce saturated and trans fats in bakery foods, have recently shown promising applications in non-bakery foods, particularly in the enhancement of their food texture and cooking qualities. This study investigates the impact of incorporating stearic acid-containing candelilla wax–groundnut oil oleogel in various proportions on the production of whole wheat pasta. Five different pasta samples were prepared by replacing water with oleogels in varying concentrations (2.5%, 5%, 10%, and 15%), and their physicochemical attributes were evaluated using a range of analytical methods for both cooked and uncooked pasta (like microscopy, colorimetry, dimensional analysis, texture, cooking qualities, moisture content, and FTIR). Significant differences in width, thickness, and color properties were observed between the control sample (0% oleogel) and those containing oleogel, with notable variations in surface texture and color intensities, particularly with the higher oleogel content (p < 0.05). Cooked pasta exhibited lower L* values and higher a* values than uncooked pasta. Stereo zoom microscope and field emission scanning electron microscope (FESEM) micrographs demonstrated a change in the pasta surface topology and microstructures. Dark spots on the pasta with greater oleogel concentrations (samples with 10% and 15% oleogel replacement) suggest the formation of starch–lipid complexes. Cooking induced pore formation, which was more pronounced when the oleogel content was increased, impacted the water absorption capacity, swelling index, and moisture content. The cooked samples exhibited higher moisture content and improved polymer network stability compared to the uncooked ones, indicating the potential of oleogel incorporation to modulate pasta properties in a concentration-dependent manner. These findings underscore the versatility of oleogels when their applications are diversified in non-bakery foods to enhance food texture and quality.

Learned Load Balancing

Effectively balancing traffic in datacenter networks is a crucial operational goal. Most existing load balancing approaches are handcrafted to the structure of the network and/or network workloads. Thus, new load balancing strategies are required if the underlying network conditions change, e.g., due to hard or grey failures, network topology evolution, or workload shifts. While we can theoretically derive the optimal load balancing strategy by solving an optimization problem given certain traffic and topology conditions, these problems take too much time to solve and makes the derived solution stale to deploy. In this paper, we describe a load balancing scheme Learned Load Balancing (LLB), which is a general approach to finding an optimal load balancing strategy for a given network topology and workload, and is fast enough in practice to deploy the inferred strategies. LLB uses deep supervised learning techniques to learn how to handle different traffic patterns and topology changes, and adapts to any failures in the underlying network. LLB leverages emerging trends in network telemetry, programmable switching, and “smart” NICs. Our experiments show that LLB performs well under failures and can be expanded to more complex, multi-layered network topologies. We also prototype neural network inference on smartNICs to demonstrate the workability of LLB.

A two-field formulation for surfactant transport within the algebraic volume of fluid method

Surfactant transport plays an important role in many technical processes and industrial applications such as chemical reactors, microfluidics, printing and coating technology. High fidelity numerical simulations of two-phase flow phenomena reveal rich insights into the flow dynamics, heat, mass and species transport. In the present study, a two-field formulation for surfactant transport within the algebraic volume of fluid method is presented. The slight diffuse nature of representing the interface in the algebraic volume of fluid method is utilized to track the concentration of surfactant at the interface as a volumetric concentration. Transport of insoluble and soluble surfactants is investigated by tracking two different concentrations of the surfactant, one within the bulk of the liquid and the other one at the interface. These two transport equations are in turn coupled by source terms considering the ad-/desorption processes at a liquid-gas interface. Appropriate boundary conditions at a solid-fluid interface are formulated to ensure surfactant conservation, while also enabling to study the ad-/desorption processes at a solid-fluid interface. The developed numerical method is verified by comparing the numerical simulations with well-known analytical and numerical reference solutions and applied to a surfactant-laden drop in shear flow. The presented numerical methodology offers a seamless integration of surfactant transport into the algebraic volume of fluid method, where the latter has many advantages such as volume conservation and an inherent ability of handling large interface deformations and topological changes.

The effect of NaF and ZnO on regulating the network topological structure of CsPbBr3 quantum dots-embedded sodium borosilicate glass for energy-efficient dust indicator

All-inorganic CsPbX3 (X = Cl, Br, I) perovskite quantum dots (QDs) exhibit substantial potential in optoelectronics, rendering them valuable materials for next-generation lighting. However, the high cost and substantial usage of Cs present challenges. This study investigates the impact of NaF and/or ZnO modifiers on the network topological structure of CsPbBr3 QDs-embedded sodium borosilicate (SBS) glass. The incorporation of NaF and ZnO not only reduces Cs consumption and energy requirements during heat treatment but also enhances the stability and optical properties of CsPbBr3 QDs. Through a combination of experimental analysis and molecular dynamics (MD) simulations, we explore the changes in the SBS glass network topology induced by these modifiers. The results demonstrate that the addition of these modifiers effectively alters the topological structure of SBS glass, providing outstanding water and heat resistance for the embedded CsPbBr3 QDs. This suggests that the NaF and ZnO co-modified CsPbBr3 QDs-embedded SBS glass, with improved stability and optical characteristics, exhibits great promise for applications in optoelectronics.

COMPUTATIONAL INTELLIGENCE BASED ROUTING SOLUTIONS FOR MOBILE AD HOC NETWORKS

Mobile Ad Hoc Networks (MANETs) refers to a wireless network where nodes dynamically establish and maintain connections with each other, directly without the need for a centralized infrastructure. MANETs are characterized by their decentralized nature, presenting challenges for routing protocols concerning rapid data transmission, minimized delay and Quality of Service (QoS) provisioning. Traditional routing protocols may struggle to perform optimally in MANETs due to frequent changes in network topology and resource constraints. The development of intelligent routing protocols tailored to the specific characteristics of MANETs is vital for dynamic and resource-constrained environments. Computational intelligence (CI) techniques have gained attention for designing intelligent routing protocols for MANETs. This paper provides a comprehensive overview of majorly used CI methodologies in MANETs. The paper reviews existing research on intelligent routing protocols based on CI techniques in MANETs and identifies future research directions in this field to address the unique challenges of MANETs and improve network performance and reliability.

Bilinear stiffness and bimodular Poisson's ratio in cylindrical sinusoidal lattices through topology morphing

Bilinear elastic behaviour allows structural designs to respond in either a stiff or compliant manner depending on the load. Here a cylindrical sinusoidal lattice structure is described that stiffens beyond a certain load. When subjected to axial compression, the lattice can undergo a topological transformation by forming contact connections. This topology change involves a transition from rectangular-like unit cells to kagome-like unit cells, associated with an approximately fourfold increase in stiffness. The lattice exhibits negative Poisson's ratio with a step-change from ≈−0.66 to ≈−0.23 prior to and during contact formation, respectively. After contact formation, it displays a nonlinear Poisson's ratio behaviour. The mechanics underpinning these behaviours are analysed using a combination of experiments and numerical modelling. A comparison with similar planar lattices reveals the effect of the global topology of the lattice (e.g. planar, cylindrical) on the unit cell-level topology morphing. The proposed topology-morphing cylindrical sinusoidal lattice introduces new design possibilities in the application-rich context of tubular structures with nonlinear mechanical properties.

The impact of relationship stickiness and memory on the evolution of individual behavior

This study investigates the dynamics of individual behavior and decision-making processes shaped by changes in personal relationships and preferences, using the prisoner’s dilemma game as a framework. The research focuses on the effects of relationship stickiness and memory on behavior evolution and network structural changes, considering two types of initial relationship networks. The findings indicate that as the number of games increases, there is a general trend of gradual increases in the average relationship stickiness, topological changes, and network heterogeneity. Notably, networks with greater initial heterogeneity, such as scale-free networks, better facilitate individual cooperation. Contrary to previous studies, the stable density of cooperators does not consistently increase with higher temptation gains. This suggests that individuals’ adaptive strategies, which include changing neighbors, can foster cooperation even under greater temptations. Additionally, an increase in the sensitivity coefficient correlates with increases in the final cooperator density, network heterogeneity, and topological changes. As individual memory and temptation gain rise, network heterogeneity measures tend to increase, while measures of topological change generally decrease. A sharp decline in the stable density of cooperators coincides almost synchronously with a rapid increase in network heterogeneity and a decrease in topological changes. This research reveals how relationship stickiness and memory impact individual cooperation, enhancing our understanding of the underlying mechanisms driving cooperative behaviors in complex networks.

On the transient and equilibrium features of growing fractal complex networks

Complex networks have certain properties that distinguish them from their respective uniform or regular counterparts. One of these properties is the variation of topological properties along different hierarchical levels. In this work, we study how networks that are constructed by repeatedly incorporating a given motif exhibit this property. A motif is henceforth understood as a small subgraph with a reference node where the incorporation respectively occurs. We generate fractal networks using different motifs and observe how their structures change along the growth stages. Two regimes are respectively identified: transient and equilibrium. The former is characterized by significant topological changes that depend on the motif topology, while the equilibrium regime shows more stable, parallel trajectories. A more systematic analysis revealed that the betweenness centrality and the average shortest path lengths were the main topological properties accounting for the changes along the network growth.

Low Earth Orbit Satellite Network Routing Algorithm Based on Graph Neural Networks and Deep Q-Network

Low Earth orbit (LEO) satellite networks are characterized by rapid topological changes, numerous network nodes and varying states of node resource constraints, which have resulted in traditional routing algorithms no longer being suitable for LEO satellite network routing. Therefore, this paper proposes an inductive learning architecture based on Graph Sample and Aggregate (GraphSAGE), which can significantly reduce the number of topology nodes to be trained, thereby reducing the computational complexity of the nodes. Then deep reinforcement learning (DRL) is employed for the continuous learning optimization of routing algorithms, and its generalization is improved by selecting GraphSAGE to construct the DRL agent. In the proposed graph neural-network-based routing optimization algorithm for LEO satellite networks, each Deep Q-Network (DQN) agent independently generates the hidden states of the nodes through the GraphSAGE model and uses them as inputs to the DRL model to make routing decisions. After a simulation and comparison, the proposed algorithm not only improves the overall network throughput, but also reduces the average end-to-end delay. The average throughput of the proposed algorithm increases by 29.47% and 18.42% compared to that of Dijkstra and the DQN, respectively. The average end-to-end delay is reduced by 39.76% and 15.29%, respectively, and can also adapt to changing topologies.

Topographical and morphological changes of the single‐crystal (La0.3Sr0.7)(Al0.65Ta0.35)O3 (001) surface treated with high‐purity deionized water

The topological and morphological changes of the (001) surface of single‐crystal (La0.3Sr0.7)(Al0.65Ta0.35)O3 during treatment with high‐purity deionized water (DI) have been investigated by atomic force microscopy and scanning transmission electron microscopy (STEM). Etching by DI treatment at room temperature proceeded mainly in anisotropic etching mode, resulting in rectangular etch pits with {001} inner surfaces forming in the surface. The dominant etching mode changed from anisotropic to isotropic with increasing DI‐treatment temperature, and the lowest surface roughness occurred for DI treatment at 50°C for 60 min. Direct observation of the atomic structure of the surface by STEM showed that the etched surface was the B‐site‐terminated (001) surface after anisotropic etching at 50°C for 60 min, in which the B‐site‐terminated atomic layer was found to maintain the B‐site ordered structure existing in the crystal. This is because DI treatment, which does not require high‐temperature heat treatment of the substrate, suppresses element diffusion, sublimation, and formation of point defects on and in the vicinity of the surface.

DMXP: A De Novo Small-Molecule 3D Structure Predictor with Graph Attention Networks.

Generating the three-dimensional (3D) structure of small molecules is crucial in both structure- and ligand-based drug design. Structure-based drug design needs bioactive conformations of compounds for lead identification and optimization. Ligand-based drug design techniques, such as 3D shape similarity search, 3D pharmacophore model, 3D-QSAR, etc., all require high-quality small-molecule ligand conformations to obtain reliable results. Although predicting a small molecular bioactive conformer requires information from the receptor, a crystal structure of the molecule is a proper approximation to its bioactive conformer in a specific receptor because the binding pose of a small molecule in its receptor's binding pockets should be energetically close to the crystal structures. This study presents a de novo small molecular structure predictor (dMXP) with graph attention networks based on crystal data derived from the Cambridge Structural Database (CSD) combined with molecular electrostatic information calculated by density-functional theory (DFT). Two featuring strategies (topological and atomic partial change features) were employed to explore the relation between these features and the 3D crystal structure of a small molecule. These features were then assembled to construct the holistic 3D crystal structure of a molecule. Molecular graphs were encoded using a graph attention mechanism to deal with the issues of the inconsistencies of local substructures contributing to the entire molecular structure. The root-mean-square deviation (RMSDs) of approximately 80% dMXP predicted structures and the native binding poses within receptors are less than 2.0 Å.

Biphase routing scheme for optimal throughput in large-scale optical satellite networks

In the large-scale optical satellite network (LS-OSN), hundreds to thousands of low Earth orbit (LEO) satellites will be interconnected via laser links, offering global coverage characterized by high throughput and low latency. LS-OSNs present an attractive strategy to cultivate a comprehensively connected, intelligent world. However, the dynamic nature of the satellites, as they orbit the Earth, results in frequent changes in the LS-OSN topology. Thus, there is a pressing need for efficient routing algorithms that not only cater to massive traffic demands but also swiftly adapt to these constant topological changes. Traditional routing algorithms for services with specific bandwidth requirements often compromise on either computational speed or throughput efficiency. In response, this study introduces a routing scheme based on flow optimization and decomposition (FOND). This seeks to shorten the computation time while preserving optimal network throughput. Expanding upon the FOND scheme, we further devised two heuristic algorithms: the flow-based greedy path (FGP) and the flow-based greedy width (FGW). Simulation results from a 288-satellite constellation network indicate that both the FGP and FGW outpace contemporary methods in terms of the routing computation time while maintaining a consistent throughput equal to 100% of the network capacity. Notably, the FGP has exhibited an impressive capability, reducing the routing computation time to 0.23% compared to the baseline incremental-widest-path (IWP) algorithm, which operates on Dijkstra’s algorithm principles.

ER-OCN: Toward efficient network routing in ocean city based on deep reinforcement learning

Smart ocean cities are crucial to future city development, and offshore networks are an essential part of smart ocean cities. Nowadays, communication in offshore networks is becoming increasingly important. However, the offshore networks are highly heterogeneous, with low reliability, narrow bandwidth, and dramatic dynamics. Offshore networks are more challenging to manage than traditional terrestrial networks, and the standard quality of service routing faces significant challenges. Currently, artificial intelligence provides a way to meet the requirements. This paper proposes ER-OCN (Efficient Routing in Ocean City Networks), a two-stage routing optimization scheme driven by artificial intelligence to cope with the complex offshore networks environment. The first stage will calculate the flow paths based on Lagrangian relaxation. Besides, the ocean environment is under dramatic dynamics, leading to unpredictable network topology or link bandwidth changes. Thus, the second stage deployed a deep reinforcement learning algorithm to monitor the environment in real-time and update the routing strategy when needed. We conducted experiments based on actual network topology and traffic data sets of offshore platforms’ network systems. The results prove that our scheme improves communication cost performance by 36.61% compared to traditional methods.

New multiplicity results in prescribing Q-curvature on standard spheres

Abstract In this paper, we study the problem of prescribing Q-Curvature on higher dimensional standard spheres. The problem consists in finding the right assumptions on a function K so that it is the Q-Curvature of a metric conformal to the standard one on the sphere. Using some pinching condition, we track the change in topology that occurs when crossing a critical level (or a virtually critical level if it is a critical point at infinity) and then compute a certain Euler-Poincaré index which allows us to prove the existence of many solutions. The locations of the levels sets of these solutions are determined in a very precise manner. These type of multiplicity results are new and are proved without any assumption of symmetry or periodicity on the function K.

DCFH: A dynamic clustering approach based on fire hawk optimizer in flying ad hoc networks

In flying ad hoc networks (FANETs), unmanned aerial vehicles (UAVs) communicate with each other without any fixed infrastructure. Because of frequent topological changes, instability of wireless communication, three-dimensional movement of UAVs, and limited resources, especially energy, FANETs deal with many challenges, especially the instability of UAV swarms. One solution to address these problems is clustering because it maintains network performance and increases scalability. In this paper, a dynamic clustering scheme based on fire hawk optimizer (DCFH) is proposed for FANETs. In DCFH, each cluster head calculates the period of hello messages in its cluster based on its velocity. Then, a fire hawk optimizer (FHO)-based dynamic clustering operation is carried out to determine the role of each UAV (cluster head (CH) or cluster member (CM)) in the network. To calculate the fitness value of each fire hawk, a fitness function is suggested based on four elements, namely the balance of energy consumption, the number of isolated clusters, the distribution of CHs, and the neighbor degree. To improve cluster stability, each CH manages the movement of its CMs and adjusts it based on its movement in the network. In the last phase, DCFH defines a greedy routing process to determine the next-hop node based on a score, which consists of distance between CHs, energy, and buffer capacity. Finally, DCFH is simulated using the network simulator version 2 (NS2), and its performance is compared with three methods, including the mobility-based weighted cluster routing scheme (MWCRSF), the dynamic clustering mechanism (DCM), and the Grey wolf optimization (GWO)-based clustering protocol. The simulation results show that DCFH well manages the number of clusters in the network. It improves the cluster construction time (about 55.51%), cluster lifetime (approximately 11.13%), energy consumption (about 15.16%), network lifetime (about 2.6%), throughput (approximately 3.9%), packet delivery rate (about 0.61%), and delay (approximately 14.29%). However, its overhead is approximately 8.72% more than MWCRSF.

An Information Theoretic Approach to Analyze Irregularity of Graph Signals and Network Topological Changes Based on Bubble Entropy

An Information Theoretic Approach to Analyze Irregularity of Graph Signals and Network Topological Changes Based on Bubble Entropy

An edge-based interface tracking (EBIT) method for multiphase-flow simulation with surface tension

We present a novel Front-Tracking method, the Edge-Based Interface Tracking (EBIT) method for multiphase flow simulations. In the EBIT method, the markers are located on the grid edges and the interface can be reconstructed without storing the connectivity of the markers. This feature makes the process of marker addition or removal easier than in the traditional Front-Tracking method. The EBIT method also allows almost automatic parallelization due to the lack of explicit connectivity.In a previous journal article we have presented the kinematic part of the EBIT method, that includes the algorithms for piecewise linear reconstruction and advection of the interface. Here, we complete the presentation of the EBIT method and combine the kinematic algorithm with a Navier–Stokes solver. A circle fit is now implemented to improve the accuracy of mass conservation in the reconstruction phase. Furthermore, to identify the reference phase and to distinguish ambiguous topological configurations, we introduce a new feature: the Color Vertex. For the coupling with the Navier–Stokes equations, we first calculate volume fractions from the position of the markers and the Color Vertex, then viscosity and density fields from the computed volume fractions and finally surface tension stresses with the Height-Function method. In addition, an automatic topology change algorithm is implemented into the EBIT method, making it possible the simulation of more complex flows. The two-dimensional version of the EBIT method has been implemented in the free Basilisk platform, and validated with seven standard test cases: stagnation flow, translation with uniform velocity, single vortex, Zalesak's disk, capillary wave, Rayleigh-Taylor instability and rising bubble. The results are compared with those obtained with the Volume-of-Fluid (VOF) method already implemented in Basilisk.

The structure of turbulence in unsteady flow over urban canopies

The topology of turbulent coherent structures is known to regulate the transport of energy, mass and momentum in the atmospheric boundary layer (ABL). While previous research has primarily focused on characterizing the structure of turbulence in stationary ABL flows, real-world scenarios frequently deviate from stationarity, giving rise to nuanced and poorly understood changes in the turbulence geometry and associated transport mechanisms. This study sheds light on this problem by examining topological changes in ABL turbulence induced by non-stationarity and their effects on momentum transport. Results from a large-eddy simulation of pulsatile open channel flow over an array of surface-mounted cuboids are examined. The analysis reveals that the flow pulsation triggers a phase-dependent shear rate, and the ejection-sweep pattern varies with the shear rate during the pulsatile cycle. From a turbulence structure perspective, it is attributed to the changes in the geometry of hairpin vortices. An increase (decrease) in the shear rate intensifies (relaxes) these structures, leading to an increase (decrease) in the frequency of ejections and an amplification (reduction) of their percentage contribution to the total momentum flux. Furthermore, the size of the hairpin packets undergoes variations, which depend on the geometry of the constituting hairpin vortices, yet the packet inclination preserves its orientation throughout the pulsatile cycle. These observations reinforce the important role non-stationarity holds in shaping the structure of ABL turbulence and the momentum transport mechanisms it governs.