Original Pair Research Articles

HCCL: Hierarchical Counterfactual Contrastive Learning for Robust Visual Question Answering

Despite most state-of-the-art models having achieved amazing performance in Visual Question Answering (VQA) , they usually utilize biases to answer the question. Recently, some studies synthesize counterfactual training samples to help the model to mitigate the biases. However, these synthetic samples need extra annotations and often contain noises. Moreover, these methods simply add synthetic samples to the training data to train the model with the cross-entropy loss, which cannot make the best use of synthetic samples to mitigate the biases. In this article, to mitigate the biases in VQA more effectively, we propose a Hierarchical Counterfactual Contrastive Learning (HCCL) method. Firstly, to avoid introducing noises and extra annotations, our method automatically masks the unimportant features in original pairs to obtain positive samples and create mismatched question-image pairs as negative samples. Then our method uses feature-level and answer-level contrastive learning to make the original sample close to positive samples in the feature space, while away from negative samples in both feature and answer spaces. In this way, the VQA model can learn the robust multimodal features and focus on both visual and language information to produce the answer. Our HCCL method can be adopted in different baselines, and the experimental results on VQA v2, VQA-CP, and GQA-OOD datasets show that our method is effective in mitigating the biases in VQA, which improves the robustness of the VQA model.

Nontrivial fusion of Majorana zero modes in interacting quantum-dot arrays

Motivated by recent experimental reports of Majorana zero modes (MZMs) in quantum-dot systems at the “sweet spot,” where the electronic hopping th is equal to the superconducting coupling Δ, we study the time-dependent spectroscopy corresponding to the fusion of MZMs. The term “nontrivial” refers to the fusion of Majoranas from different original pairs of MZMs, each with well-defined parities. We employ an experimentally accessible time-dependent real-space local density-of-states (LDOS) method to investigate the nontrivial MZM fusion outcomes in canonical chains and in a Y-shaped array of interacting electrons. In the case of quantum-dot chains where two pairs of MZMs are initially disconnected, after fusion we find equal-height peaks in the electron and hole components of the LDOS, signaling nontrivial fusion into both the vacuum I and fermion Ψ channels with equal weight. For π-junction quantum-dot chains, where the superconducting phase has opposite signs on the left and right portions of the chain, after the nontrivial fusion we observed the formation of an exotic two-site MZM near the center of the chain, coexisting with another single-site MZM. Furthermore, we also studied the fusion of three MZMs in the Y-shaped geometry. In this case, after the fusion we observed the novel formation of another exotic multisite MZM, with properties depending on the connection and geometry of the central region of the Y-shaped quantum-dot array. Published by the American Physical Society 2024

Multi-scale transformer network for super-resolution of visible and thermal air images

Reference image-based Super-Resolution (RefSR) is introduced to improve the quality of a Low-resolution (LR) input image by leveraging the additional information provided by a High-Resolution (HR) reference image (Ref). While existing RefSR methods focus on thermal or visible flows separately, they often struggle to enhance the resolution of small objects such as Mini/Micro UAVs (Unmanned Aerial Vehicle) due to the resolution disparities between the input and reference images. To cope with these challenges when dealing with UAV early detection in context of video surveillance, we propose ThermoVisSR, a multiscale texture transformer for enhancing the Super-Resolution (SR) of visible and thermal images of Mini/Micro UAVs. Our approach tries to reconstruct the fine details of these objects while preserving their approximation (the body form and color of the different scene objects) already contained in the LR image. Hence, our model is divided up into two streams dealing separately with approximation and detail reconstruction. In the first one, we introduce a Convolution Neural Network (CNN) fusion backbone to extract the Low-Frequency (LF) approximation from the original LR image pairs. In the second one and to extract the details from the Ref image, our approach involves blending features from both visible and thermal sources to make the most of what each offer. Subsequently, we introduce the High-Frequency Texture Transformer (HFTT) across various resolutions of the merged features to ensure an accurate correspondence matching and significant transfer of High-Frequency (HF) patches from Ref to LR images. Moreover, to adapt the injection to the different bands well, we incorporate the separable software decoder (SSD) into the HFTT allowing to capture channel-specific details during the reconstruction phase. We validated our approach using a newly created dataset of Air images of Mini/Micro UAVs. Experimental results demonstrate that the proposed model consistently outperforms the state-of-the-art approaches on both qualitative and quantitative assessments.

Parents and Children: Distinguishing Multimodal DeepFakes from Natural Images

Recent advancements in diffusion models have enabled the generation of realistic deepfakes from textual prompts in natural language. While these models have numerous benefits across various sectors, they have also raised concerns about the potential misuse of fake images and cast new pressures on fake image detection. In this work, we pioneer a systematic study on deepfake detection generated by state-of-the-art diffusion models. Firstly, we conduct a comprehensive analysis of the performance of contrastive and classification-based visual features, respectively extracted from CLIP-based models and ResNet or ViT-based architectures trained on image classification datasets. Our results demonstrate that fake images share common low-level cues, which render them easily recognizable. Further, we devise a multimodal setting wherein fake images are synthesized by different textual captions, which are used as seeds for a generator. Under this setting, we quantify the performance of fake detection strategies and introduce a contrastive-based disentangling method that lets us analyze the role of the semantics of textual descriptions and low-level perceptual cues. Finally, we release a new dataset, called COCOFake, containing about 1.2M images generated from the original COCO image-caption pairs using two recent text-to-image diffusion models, namely Stable Diffusion v1.4 and v2.0.

Reproductive behavior of Erythrodiplax abjecta (Rambur, 1842) from Andean Mountains

Research about the behavior of Neotropical species is crucial to understand how the rapid environmental changes in the Neotropics affect the reproduction of various organisms. The reproductive behavior of insects in tropical ecosystems, such as those belonging to the order Odonata, is as yet scarcely known. In this article, the reproductive behavior of Erythrodiplax abjecta is described from several localities in the Colombian Andean Mountains. Playing out between 10:00 and 14:30 h, male individuals of this species may exhibit either territorial or satellite behavior. Sexually mature males and females engage in tandem and copulation behaviors, which are brief and may occur either perched or in flight. Females of E. abjecta oviposit directly in the water at various times, alternating with periods of rest. During oviposition, males stand guard over females, perched or hovering, in a quest to protect them against other males. Conspecific males have been observed to engage in sieges of perching or ovipositing females, as well as mating pairs. Following a siege, either the female or the original pair flees, or the besieging male will succeed in outcompeting the original male, take the female in tandem, copulate with her, and then guard her while she oviposits. If the female flees, she may perch elsewhere, continue ovipositing, or may be taken by another male. The intense male guarding behavior during oviposition could represent an adaptation to temporarily high male densities at oviposition sites and a high level of competition during the short period in which active copulation takes place.

Types and Fecundity of Neotenic Reproductives Produced in 5-Year-Old Orphaned Colonies of the Drywood Termite, Cryptotermes domesticus (Blattodea: Kalotermitidae)

Orphaned colonies of Cryptotermes domesticus readily produce replacement reproductives and continue propagation. In this study, we aimed to investigate the production and fecundity of neotenic reproductives in 5-year-old colonies of C. domesticus after orphaning. All 15 experimental colonies were successfully re-established by the neotenic reproductive pair. Three types of neotenic reproductives with various wing-bud lengths were observed: type I with micro wing buds, type II with short wing buds, and type III with long wing buds. Four patterns of pairs made up of these neotenics, namely, type I + type II, type I + type III, type II + type II, and type II + type III, exhibited reproductive capacities similar to those of the primary reproductive pair. We speculated that these neotenic reproductives were derived from various nymphal instars. The 5-year-old colonies had three instars of nymphs, with the majority being in the second instar, followed by the first. Thus, the combination of neotenic reproductives with short wing buds and micro wing buds was the dominant differentiation pathway of the orphaned colonies. After the removal of the original primary reproductive pair, the nymphs matured into neotenic reproductives and took over reproduction in the colony in 107.40 ± 15.18 days. This study highlights the importance of quarantine and routine inspection of wood, as well as the significance of early prevention and control of C. domesticus infestation in wood. Moreover, this study confirms the high differentiation and reproductive capacities of C. domesticus.

Unravelling route choices of large trucks using trajectory clustering and conditional Logit models

The mobility of sizable trucks is often limited by their large size. They thus may have additional requirements on road types, road widths, and the turning radius at the intersection when travelling. Therefore, this study explores the unique needs and preferences of large truck drivers’ route choice with a focus on trip and road network characteristics. GPS trajectory data from the central Kansai area of Japan with numerous ports and freight terminals are used. Trajectories are considered having the same origin (destination) if their starting (ending) coordinates are in the same 500 m × 500 m mesh. For the trajectories of the same pair of origin–destination (OD) meshes, several route clusters are obtained based on geographical configuration using a QuickBundles algorithm. Sampling techniques are employed to equalize the number of input points for each vehicle trajectory and the optimal number of clusters is determined automatically by our algorithm based on the Silhouette Coefficient. By taking the clusters as route choice options for an OD pair, a Conditional Logit Model is used to identify the factors that influence the route choice considering both vehicle- and trip-specific attributes. The results quantify the preference of trucks for wider roads and toll routes, as well as aversion to long distances and turns. Heterogeneity in route choice based on vehicle type, trip time (date), and trip purpose is also evident. The findings of this study can provide insights for freight road network design and optimization.

Long-Term Effects of Pre-Weaning Individual or Pair Housing of Dairy Heifer Calves on Subsequent Growth and Feed Efficiency.

Our objective in this exploratory study was to evaluate the long-term impacts of pre-weaning social isolation vs. contact on subsequent growth and feed efficiency of Holstein heifers. As pre-weaned calves, 41 heifers were housed individually (n = 15 heifers) or in pairs (n = 13 pairs; 26 heifers). At 18 months of age, heifers were blocked by body weight and randomly assigned to one of three pens within a block (six to eight heifers per pen; six pens total), with original pairs maintained. Body weight (BW), hip height and width, and chest girth were measured at the start and end of the study. Each pen was given 3 days of access to a GreenFeed greenhouse gas emissions monitor to assess potential physiological differences between treatments in enteric methane emissions or behavioral differences in propensity to approach a novel object. During the 9-week study, heifers were fed a common diet containing 62.3% male-sterile corn silage, 36.0% haylage, 0.7% urea, and 1.0% mineral (DM basis). To calculate daily feed intake, as-fed weights and refusals were recorded for individual heifers using Calan gates. Feed samples were collected daily, composited by week, and dried to calculate dry matter intake (DMI). Feed refusal and fecal samples were collected on 3 consecutive days at 3 timepoints, composited by heifer, dried, and analyzed to calculate neutral detergent fiber (NDF), organic matter (OM), and DM digestibility. Feed efficiency was calculated as feed conversion efficiency (FCE; DMI/average daily gain [ADG]) and residual feed intake (RFI; observed DMI-predicted DMI). Paired and individually housed heifers did not differ in DMI, ADG, FCE, or RFI. Although no differences were found in initial or final hip height, hip width, or chest girth, heifers which had been pair-housed maintained a greater BW than individually housed heifers during the trial. Methane production, intensity, and yield were similar between treatments. Pre-weaning paired or individual housing did not impact the number of visits or latency to approach the GreenFeed; approximately 50% of heifers in each treatment visited the GreenFeed within 8 h of exposure. Digestibility of OM, DM, and NDF were also similar between housing treatments. In conclusion, pre-weaning pair housing had no adverse effects on growth, feed efficiency, or methane emissions at 18 to 20 months of age.

A Two-Stage Registration Strategy for Thermal–Visible Images in Substations

The analysis of infrared video images is becoming one of the methods used to detect thermal hazards in many large-scale engineering sites. The fusion of infrared thermal imaging and visible image data in the target area can help people to identify and locate the fault points of thermal hazards. Among them, a very important step is the registration of thermally visible images. However, the direct registration of images with large-scale differences may lead to large registration errors or even failure. This paper presents a novel two-stage thermal–visible-image registration strategy specifically designed for exceptional scenes, such as a substation. Firstly, the original image pairs that occur after binarization are quickly and roughly registered. Secondly, the adaptive downsampling unit partial-intensity invariant feature descriptor (ADU-PIIFD) algorithm is proposed to correct the small-scale differences in details and achieve finer registration. Experiments are conducted on 30 data sets containing complex power station scenes and compared with several other methods. The results show that the proposed method exhibits an excellent and stable performance in thermal–visible-image registration, and the registration error on the entire data set is within five pixels. Especially for multimodal images with poor image quality and many detailed features, the robustness of the proposed method is far better than that of other methods, which provides a more reliable image registration scheme for the field of fire safety.

Effects of Different Media and Their Strengths in In Vitro Culture of Three Different Cistus creticus L. Populations and Their Genetic Assessment Using Simple Sequence Repeat Molecular Markers

Cistus creticus L. (rockrose), a species of ecological and medicinal significance, constitutes a valuable component of the Mediterranean ecosystem. The present study investigated the effect of the inorganic salt concentration of Murashige and Skoog medium (MS), woody plant medium (WPM), and Driver and Kuniyaki Walnut medium (DKW) at several strengths (1/8×, 1/4×, 1/2×, 1×, and 2×) on the in vitro growth and organogenesis of rockrose. Significant interactions were observed throughout the experiments between pairs of plant origins, medium types, and strengths, and we also examined the extent to which they affected the studied traits was examined. The types of nutrient medium affected all studied traits except shoot and root percentages. The maximum growth percentage (143.49%) was gained using full-strength WPM. The best performance in shoot percentage was obtained using MS (100%) at several strengths along with 1× WPM (100%). The topmost rooting percentage values (98.61%) were obtained using 1× WPM and 1/2× DKW. The highest number of shoots and roots were observed using full-strength MS (9.39) and half-strength WPM (6.49), respectively. The maximum values for shoot and root length were achieved using 1/2× MS (0.78 cm) and 1/8× WPM (1.55 cm), respectively. The origin of the plant material did not influence any studied trait. Moreover, the genetic relations among the populations used in the in vitro culture were assessed using simple sequence repeats (SSR) markers. Twenty-eight alleles were identified across all five STR loci. The different and effective alleles per locus were 5.60 and 4.72, respectively. The average observed and expected heterozygosity was estimated at 0.52 and 0.72, respectively. Shannon’s information index and the inbreeding coefficient (F) were assessed at 1.48 and 0.30, respectively, revealing a narrow genetic base and high genetic similarity among origins, suggesting that they belong to the same population.

Quantum optical induced-coherence tomography by a hybrid interferometer

Abstract Quantum interferometry based on induced-coherence phenomena has demonstrated the possibility of undetected-photon measurements. Perturbation in the optical path of probe photons can be detected by interference signals generated by quantum mechanically correlated twin photons propagating through a different path, possibly at a different wavelength. To the best of our knowledge, this work demonstrates for the first time a hybrid-type induced-coherence interferometer that incorporates a Mach–Zehnder-type interferometer for near-visible photons and a Michelson-type interferometer for infrared (IR) photons, based on double-pass-pumped spontaneous parametric down-conversion. This configuration enables IR optical measurements via the detection of near-visible photons and provides methods for optimizing the quality of measurements by identifying photon pairs of different origins. We theoretically identify that the induced-coherence interference visibility is approximately the same as the heralding efficiency between twin photons along the relevant spatial modes, and experimentally maximize the visibility by setting up a common reference spatial mode for IR photons. Applications to both time-domain and frequency-domain quantum optical induced-coherence tomography for three-dimensional test structures are demonstrated. The results prove the feasibility of practical undetected-photon sensing and imaging techniques based on the presented structure.

Jeans modelling of weakly flattened ellipsoidal systems

ABSTRACT In the homoeoidal expansion, a given ellipsoidally stratified density distribution and its associated potential are expanded in the (small) density flattening parameter η, and usually truncated at the linear order. The truncated density–potential pair obeys exactly the Poisson equation, and it can be interpreted as the first-order expansion of the original ellipsoidal density–potential pair, or as a new autonomous system. In the first interpretation, in the solutions of the Jeans equations the quadratic terms in η must be discarded (‘η-linear’ solutions), while in the second (‘η-quadratic’) all terms are retained. In this work, we study the importance of the quadratic terms by using the ellipsoidal Plummer model and the Perfect Ellipsoid, which allow for fully analytical η-quadratic solutions. These solutions are then compared with those obtained numerically for the original ellipsoidal models, finding that the η-linear models already provide an excellent approximation of the numerical solutions. As an application, the η-linear Plummer model (with a central black hole) is used for the phenomenological interpretation of the dynamics of the weakly flattened and rotating globular cluster NGC 4372, confirming that this system cannot be interpreted as an isotropic rotator, a conclusion reached previously with more sophisticated studies.

Римите в поезията на Константи Илдефонс Галчински (особености при превода на български език)

The article is devoted to the rhymes in the poetry of Polish author Konstanty Ildefons Gałczyński and the ways they have been translated into the Bulgarian language. First, distinguishing features of the poetic text and the concept of rhyme are discussed, after which basic characteristics of rhymes in Polish poetry are noted. An overview of Gałczyński’s translations into Bulgarian leads to the main focus of the paper which concerns translators’ difficulties in converting the original rhyming pairs into the target language. The analysis is based on three translations of one poem: “Rozmowa liryczna” (“Lyrical Dialogue”), made by two prominent translators, Parvan Stefanov and Dimitar Pantеleev between 1960 and 1984. The observations indicate that the Bulgarian versions of the poem modify the original rhyming patterns due to the specificity of each language, as well as the translators’ individual aesthetic preferences. Almost all rhyming lexical units in the Polish text are adjacent and feminine, while there is a significantly more frequent use of cross and masculine rhymes in the three Bulgarian versions. Also, the translators refuse to preserve several irregular rhymes. The article concludes that the approaches of Stefanov and Panteleev are related to general trends in translating rhymed poetry from Polish into Bulgarian. Among them, the increased presence of masculine rhymes, largely dictated by the analytic nature of the Bulgarian language, seems prominent.

Discriminative multi-label feature selection with adaptive graph diffusion

Feature selection can alleviate the problem of the curse of dimensionality by selecting more discriminative features, which plays an important role in multi-label learning. Recently, embedded feature selection methods have received increasing attentions. However, most existing methods learn the low-dimensional embeddings under the guidance of the local structure between the original instance pairs, thereby ignoring the high-order structure between instances and being sensitive to noise in the original features. To address these issues, we propose a feature selection method named discriminative multi-label feature selection with adaptive graph diffusion (MFS-AGD). Specifically, we first construct a graph embedding learning framework equipped with adaptive graph diffusion to uncover a latent subspace that preserves the higher-order structure information between four tuples. Then, the Hilbert–Schmidt independence criterion (HSIC) is incorporated into the embedding learning framework to ensure the maximum dependency between the latent representation and labels. Benefiting from the interactive optimization of the feature selection matrix, latent representation and similarity graph, the selected features can accurately explore the higher-order structural and supervised information of data. By further considering the correlation between labels, MFS-AG is extended to a more discriminative version,i.e., LMFS-AG. Extensive experimental results on various benchmark data sets validate the advantages of the proposed MFS-AGD and LMFS-AGD methods.

Forex market directional trends forecasting with Bidirectional-LSTM and enhanced DeepSense network using all member-based optimizer

This study focuses on successful Forex trading by emphasizing the importance of identifying market trends and utilizing trend analysis for informed decision-making. The authors collected low-correlated currency pair datasets to mitigate multicollinearity risk. Authors developed a two-stage predictive model that combines regression and classification tasks, using the predicted closing price to determine entry and exit points. The model incorporates Bi-directional long short-term memory (Bi-LSTM) for improved price forecasting and higher highs and lower lows (HHs-HLs and LHs-LLs) to identify trend changes. They proposed an enhanced DeepSense network (DSN) with all member-based optimization (AMBO-DSN) to optimize decision variables of DSN. The performance of the models was compared to various machine learning, deep learning, and statistical approaches including support vector regressor (SVR), artificial neural network (ANN), auto-regressive integrated moving average (ARIMA), vanilla-LSTM (V-LSTM), and recurrent neural network (RNN). The optimized form of DSN using genetic algorithm (GA), particle swarm optimization (PSO), and differential evolution (DE) was compared with AMBO-DSN, yielding satisfactory results that demonstrated comparable quality to the observed trends on the original currency pairs. The effectiveness and reliability of the AMBO-DSN approach in forecasting trends for USD/EUR, AUD/JPY, and CHF/INR currency pairs were validated through statistical analysis while considering computational cost.

Disturbed Augmentation Invariance for Unsupervised Visual Representation Learning

Contrastive learning has gained great prominence recently, which achieves excellent performance by simple augmentation invariance. However, the simple contrastive pairs suffer from lacking of diversity due to the mechanical augmentation strategies. In this paper, we propose Disturbed Augmentation Invariance (DAI for abbreviation), which constructs disturbed contrastive pairs by generating appropriate disturbed views for each augmented view in the feature space to increase the diversity. In practice, we establish a multivariate normal distribution for each augmented view, whose mean is corresponding augmented view and covariance matrix is estimated from its nearest neighbors in the dataset. Then we sample random vectors from this distribution as the disturbed views to construct disturbed contrastive pairs. In order to avoid extra computational cost with the increase of disturbed contrastive pairs, we utilize an upper bound of the trivial disturbed augmentation invariance loss to construct the DAI loss. In addition, we propose Bottleneck version of Disturbed Augmentation Invariance (BDAI for abbreviation) inspired by the Information Bottleneck principle, which further refines the extracted information and learns a compact representation by additionally increasing the variance of the original contrastive pair. In order to make BDAI work effectively, we design a statistical strategy to control the balance between the amount of the information shared by all disturbed contrastive pairs and the compactness of the representation. Our approach gets a consistent improvement over the popular contrastive learning methods on a variety of downstream tasks, e.g. image classification, object detection and instance segmentation.

Summarization of Lengthy Legal Documents via Abstractive Dataset Building: An Extract-then-Assign Approach

Development of effective automatic summarization approaches for legal documents suffer from several challenges like extremely long document-summary pairs, lack of large scale training datasets with tractable document-summary token lengths, etc. In this work, we deal with the problem of legal document summarization by building a modified abstractive dataset from the original dataset. This ensures that the length of each document-summary pair is manageable and can be processed by the state of the art summarization approaches (such as BART). Secondly, we deal with the data scarcity problem by creating more number of training samples, from each of the original document-summary pair. This is done by creating multiple extractive summaries from each sample in the original dataset, following which ground-truth summary sentences are assigned to each of the extractive summary to generate new training samples. This results in a larger training dataset that can be utilized for fine-tuning summarization models. Our proposed approach has been evaluated on two different legal datasets- BillSum and Forum of Information Retrieval Evaluation (FIRE). With respect to the ROUGE metrics, the proposed approach is able to outperform pre-trained BART model fine-tuned on original dataset by (3−8)% for FIRE test sets, and by (1−3)% for the BillSum test sets. Considering the BERTScore metrics, the proposed approach obtains (1−2)% improvements on the FIRE test sets, while for the BillSum test sets (3−8)% improvements are observed. Such improvements suggest that the proposed dataset building approach can help achieve improved abstractive summarization of lengthy legal documents.

Strawberry Defect Identification Using Deep Learning Infrared–Visible Image Fusion

Feature detection of strawberry multi-type defects and the ripeness stage faces huge challenges because of color diversity and visual similarity. Images from hyperspectral near-infrared (NIR) information sources are also limited by their low spatial resolution. In this study, an accurate RGB image (with a spatial resolution of 2048×1536 pixels) and NIR image (ranging from 700–1100 nm in wavelength, covering 146 bands, and with a spatial resolution of 696×700 pixels) fusion method was proposed to improve the detection of defects and features in strawberries. This fusion method was based on a pretrained VGG-19 model. The high-frequency parts of original RGB and NIR image pairs were filtered and fed into the pretrained VGG-19 simultaneously. The high-frequency features were extracted and output into ReLU layers; the l1-norm was used to fuse multiple feature maps into one feature map, and area pixel averaging was introduced to avoid the effect of extreme pixels. The high- and low-frequency parts of RGB and NIR were summed into one image according to the information weights at the end. In the validation section, the detection dataset included expanded 4000 RGB images and 4000 NIR images (training and testing set ratio was 4:1) from 240 strawberry samples labeled as mud contaminated, bruised, both defects, defect-free, ripe, half-ripe, and unripe. The detection neural network YOLOv3-tiny operated on RGB-only, NIR-only, and fused image input modes, achieving the highest mean average precision of 87.18% for the proposed method. Finally, the effects of different RGB and NIR weights on the detection results were also studied. This research demonstrated that the proposed fusion method can greatly improve the defect and feature detection of strawberry samples.

A Gray-Box Surrogate Vehicle Energy Consumption Model Capable of Real-Time Updating

Vehicle energy consumption model, as a function of its operational environment, plays a significant role in real-time optimization of vehicle route and speed for a given pair of origin and destination with a desired arrival time for minimal energy consumption. In this article, a gray-box vehicle energy consumption model is developed based on a high-fidelity vehicle dynamic model with environmental influence based on the kriging modeling method, which includes rolling resistance, aerodynamics, gravity, and energy consumption of air conditioning and heater (HVAC), along with environmental conditions, such as temperature, wind speed, etc. The data-driven model, trained based on Gaussian process assumption, ensures the accuracy of the resulting model with a modeling error below 2.5%. The real-time model updating is based on recursive least-squares optimization using current driving data so that the model used for route and speed optimization represents the current vehicle status. The proposed gray-box model is validated in computer-in-the-loop simulations using SUMO and MATLAB with less than 2% error of energy consumption, which is a significant improvement over the vehicle dynamic model with up to 35% error in certain cases. A case study also indicates energy consumption reduction for vehicle route-speed optimization.

Convoy movement problem for combat supply distribution network on the battlefield

The convoy movement problem involves the routing and scheduling of military convoys. While existing research has focused only on the routing and scheduling of convoys, this paper proposes a novel convoy movement problem that incorporates the location-allocation problem, thus making it significantly more comprehensive and useful for real-world scenarios. Our framing of the problem also has a significant impact on military logistics, enabling the more efficient transportation of military supplies during wartime. Our problem is formulated as a mixed-integer linear programming model, and we develop a matheuristic algorithm consisting of two phases that are executed iteratively to solve it. In the first phase, an optimisation solver is used to generate an origin and destination pair for a convoy and the quantity of each supply type to be transported. In the second phase, we employ an ant colony optimisation algorithm to determine the final solution based on the results of the first phase. Computational experiments show that the proposed matheuristic algorithm outperforms a commercial solver in terms of performance and computational time. Furthermore, we analyse the impact of changes to the headway time, time windows, and network density on the problem, thus producing valuable insights for military decision-makers.