ABSTRACT Intermodal transport is an important part of international trade, and poor connectivity between transport modes can significantly increase costs. To minimize these costs, we propose a novel two-stage optimization model that combines transport mode selection and route planning. However, previous studies often ignore the uncertainty of node detention service time. To address this issue, we extend the robust optimization theory and introduce a two-stage robust planning model. The model is illustrated based on a simulation case. The study shows that even if the transport mode remains unchanged, uncertain parameter changes can lead to a significant drop in service levels. Based on a deep reinforcement learning algorithm, two improvement strategies are combined to improve the algorithm efficiency. In addition, optimizing transport mode selection can reduce carbon emissions without sacrificing transportation efficiency. This study provides valuable theoretical support for the low-carbon transformation of intermodal transport enterprises.

Making personalized recommendation for cold-start users, who only have a few interaction histories, is a challenging problem in recommendation systems. Recent works leverage hypernetworks to directly map interaction histories to user-specific parameters, which are then used to modulate predictor by certain modulation structure. These works obtain the state-of-the-art performance. However, there lacks a general approach to design the modulation structure. Instead of using a fixed modulation function and deciding modulation position by expertise, we propose to determine proper modulation structure, including function and position, via neural architecture search. We propose two approaches. We first design a symbolic search space which covers broad models and theoretically prove that this search space can be transformed to a much smaller space, enabling an efficient and robust one-shot search algorithm, called ColdNAS. Since recommendation systems are a special case of bipartite matching problems, the proposed methods can be generalized to a wide range of cold-start tasks, such as disease-gene association prediction for emerging diseases. However, diverse scenarios introduce new challenges in both the flexibility of the search algorithm and the search space. To address these limitations, we further propose ColdNAS$_+$, where we employ neural networks to model modulation functions to extend search space and design a two-stage decoupled stochastic search algorithm to enable non-differentiable targets in continuous spaces. Extensive experimental results on benchmark datasets show that modulation structures obtained by ColdNAS and ColdNAS$_+$ consistently outperform hand-designed cold-start techniques for recommending items for new users and predicting associated genes for new disease. We observe that different modulation functions lead to the best performance on different datasets or under different metrics, which validates the necessity of designing the modulation structure in a data-driven way.

Modern Systems-on-Chip (SoCs) employ undocumented linear address-scrambling functions to obfuscate DRAM addressing, which complicates DRAM-aware performance optimizations and hinders proactive security analysis of DRAM-based attacks; most notably, Rowhammer. Although previous work tackled the issue of reversing physical-to-DRAM mapping, existing heuristic-based reverse-engineering approaches are partial, costly, and impractical for comprehensive recovery. This paper establishes a rigorous theoretical foundation and provides efficient practical algorithms for black-box, complete physical-to-DRAM address-mapping recovery. We first formulate the reverse-engineering problem within a linear algebraic model over the finite field GF(2). We characterize the timing fingerprints of row-buffer conflicts, proving a relationship between a bank addressing matrix and an empirically constructed matrix of physical addresses. Based on this characterization, we develop an efficient, noise-robust, and fully platform-agnostic algorithm to recover the full bank-mask basis in polynomial time, a significant improvement over the exponential search from previous works. We further generalize our model to complex row mappings, introducing new hardware-based hypotheses that enable the automatic recovery of a row basis instead of previous human-guided contributions. Evaluations across embedded and server-class architectures confirm our method's effectiveness, successfully reconstructing known mappings and uncovering previously unknown scrambling functions. Our method provides a 99% recall and accuracy on all tested platforms. Most notably, Knoc-Knock runs in under a few minutes, even on systems with more than 500GB of DRAM, showcasing the scalability of our method. Our approach provides an automated, principled pathway to accurate DRAM reverse engineering.

The determination of lamination parameters feasible region is extremely important for design and optimization of composite structures. While the feasible region for unconstrained laminates is well-established, applying mandatory design rules alters both its size and shape. This is especially relevant in aeronautical composite structures, where design rules are quite often mandatory. However, existing studies do not specify the exact boundaries of the feasible region, which is significant since the optimum for critical loading conditions, such as buckling, often lies on or near the boundary. This work proposes an innovative approach for finding the exact feasible region with several design rules. The initial step is to generate laminate databases containing all feasible stacking sequences with the prescribed design rules, up to 36 layers. These data provide a rigorous insight into the distribution patterns. Based on this information, a simple and computationally efficient procedure, based on the exact determination of a few key boundary points, allows for the exact determination of the feasible region for any number of layers.

Anchor point segmentation based multi-view clustering.

ABSTRACT The automated guided vehicle (AGV) failures cause operational interruptions in automated container terminals. This study investigates the multi-equipment cooperative scheduling problem involving quay cranes, AGVs, and yard cranes. A new mathematical programming model is formulated to minimize the operational makespan and energy consumption. Considering AGV failures, the deviation of rescheduling plan is concerned to propose a rescheduling optimization model. An improved moss growth optimization (IMGO) algorithm is designed with the Logistic-Tent mapping, adaptive disturbance strategy, and parallel computing method. Experimental results demonstrate the models could solve small-scale cases optimally. By comparison, the IMGO algorithm exhibits significantly faster solving time, and the average deviations from exact solutions are quite small. For medium/large-scale cases, the IMGO algorithm outperformed the standard MGO algorithm and co-learning imperial competition algorithm in solution quality and efficiency. Furthermore, the impacts of different operation modes and equipment configuration ratios on the operational efficiency and energy consumption are analyzed.

This study presents a comprehensive multidisciplinary investigation aimed at enhancing the performance of a Savonius vertical axis hydrokinetic turbine by integrating experimental testing, computational fluid dynamics (CFD) analysis employing dynamic mesh model and hybrid optimization through a genetic algorithm integrated machine learning (GA-ML) framework. A novel dual-deflector arrangement is proposed and tested specifically for an elliptical Savonius rotor. The hybrid optimization framework merges the global search efficiency of genetic algorithms with the predictive capabilities of machine learning to identify optimal deflector configurations effectively. The implementation of dynamic meshing enables a realistic representation of the turbine's unsteady rotational motion, thereby enhancing the accuracy and reliability of the simulation predictions. The results of the study show that using two deflector plates with the elliptical shaped Savonius turbine improves the power factor by 30% compared to the conventional Savonius turbine without deflector plates. The Savonius turbine integrated with a novel and optimized channel configuration has been found to be more efficient and suitable and is recommended for power generation from flowing water especially in canals and hydropower farms.

An efficient genetic algorithm method to extend and upscale direct solar irradiance spectra measured with spectroradiometers

Efficient phase field structural design algorithm for reliability-based topology optimization with material uncertainties

An efficient MD-GRao algorithm for quickest change detection in sensor networks with unknown post-change parameters

Efficient order-based algorithms for core maintenance in weighted graph

An efficient two-stage diagnostic algorithm for assessing system reliability

An efficient algorithm for vertex enumeration of arrangement

Working with the one-dimensional Schrödinger equation in impedance form, we derive an exact inverse scattering formula that expresses impedance in terms of the reflection coefficient, and we prove injectivity of the scattering map for impedance functions of lower regularity than previously analyzed. The inverse scattering formula translates directly into an efficient numerical algorithm that accurately transforms digital scattering data into impedance. The results apply to acoustic imaging of layered media, as well as to inverse quantum scattering.

ABSTRACT In recent years, autonomous underwater vehicle (AUV) has been widely used in various fields, which has attracted increasing attention from scholars for its motion control. In this study, we mainly address the fault‐tolerant tracking control problem of AUV with actuator faults. Because AUV system is a complex nonlinear system, it exhibits coupling and nonaffine properties. These features and actuator faults pose challenges to the fault‐tolerant tracking control of AUV. Considering the coupling and nonaffine properties and actuator faults, we use the mean value theorem to transform the fully coupled nonaffine AUV (FCNAUV) model into an affine AUV model. Different from offline reinforcement learning (RL), we proposed a novel online‐policy‐iteration reinforcement learning (OPIRL) to construct a fault‐tolerant tracking control scheme. Using a single‐layer critic network, the performance index function is approximated, which effectively reduces the computing burden in the online training process. In addition, a novel weight update law is designed to improve the algorithm efficiency. The simulation results show that the proposed method achieves better system performance and has a better convergence speed for FCNAUV with rudder faults and with propeller faults.

The emergence of artificial intelligence in retail advertising has allowed the optimization of earnings on an unprecedented scale, and raises a serious ethical concern: the fairness, transparency, and consumer trust at the same time. This article will discuss the conflict between the efficiency of algorithms and the responsible implementation of AI in retail advertising practices, and how machine learning systems may unintentionally reproduce biases and discriminatory results even with a high level of technical complexity. The discussion has discussed several aspects of ensuring fairness in advertising algorithms, such as demographic parity, equalized odds, and individual fairness, and has admitted that it is mathematically impossible to have all fairness dimensions met at the same time. Explainability and transparency appear to be the main features of compliance with the regulations and consumer trust, but explainability is not enough when there is no clear communication plan to address the interests of the various stakeholder groups. To establish sustainable trust, it is necessary to integrate technical protection mechanisms like the differentiation of privacy and federated learning with effective organizational governance infrastructure, including ethics committees, human-in-the-loop and consumer control. The article provides useful implementation models including data collection, model development, deployment architecture, and post-deployment governance, as examples of how ethical AI practices are not only compliance requirements but competitive advantages. Companies that are able to effectively incorporate the issue of fairness in optimization goals have the potential to attain excellent long-term business performance and fulfill the growing demands of society to hold technology use in business responsibly.

Latent factor models are valuable in bioinformatics for accounting for unmeasured variation alongside observed covariates. Yet many methods struggle to separate known effects from latent structure and to handle losses beyond standard regression. We present a unified framework that augments row and column predictors with a low-rank latent component, jointly modeling measured effects and residual variation. To remove ambiguity in estimating observed and latent effects, we impose a carefully designed set of orthogonality constraints on the coefficient and latent factor matrices, relative to the spans of the predictor matrices. These constraints ensure identifiability, yield a decomposition in which the latent term captures only variation unexplained by the covariates, and improve interpretability. An efficient algorithm handles general non-quadratic losses via surrogates with monotone descent. Each iteration updates the latent term by truncated singular value decomposition of a doubly projected residual and refines coefficients by projections. The number of latent factors is selected by applying an elbow rule to a degrees-of-freedom-adjusted information criterion. A parametric bootstrap provides valid inference on feature-outcome associations under the regularized low-rank structure. Applied to real pharmacogenomic data, the method recovers biologically coherent gene-drug associations missed by standard factor models, such as the EGFR-inhibitor link, highlights novel candidates with plausible mechanisms, and reveals gene programs aligned with compound modes of action, including a latent unfolded-protein-response module affecting drug sensitivity. These results support the framework’s utility for precision oncology, yielding stronger biomarkers for patient stratification and deeper insight into drug resistance mechanisms.

Purpose: Dental caries can emerge anywhere in the mouth, particularly in the interior of the cheeks and the gums. Some of the indications are patches on the inner lining of the mouth, along with bleeding, toothache, numbness, and an unusual red and white staining. Hence, it is important to predict the presence of a cavity at an early stage. The currently available manual method is inefficient and hence we provide an advanced method by using the deep learning concepts. Materials and Methods: In this work, different types of algorithms such as Res Net, Deeper Google Net, and mini VGG Net are to be used to predict the class of cavity at an early stage. Results: A comparison between the accuracy of three different algorithms is given in this paper. Thus, by using efficient deep learning algorithms, it will be able to predict the presence of the cavity and the class of the cavity at an early stage and take the necessary steps to overcome it. Conclusion: In this work, a comparison between three different algorithms is given and proved that the efficient algorithm is the inception algorithm among the other algorithms that achieves an accuracy of about 98%, which is suitable for use in hospitals.

In this paper, we investigate the lexicographic and colexicographic orderings of m-ary vectors of length n, as well as the mirror (left-recursive) reflected Gray code, complementing the classical m-ary reflected Gray code. We present efficient algorithms for generating vectors in each of these orders, each achieving constant amortized time per vector. Additionally, we propose algorithms implementing the four fundamental functions in generating combinatorial objects—successor, predecessor, rank, and unrank—each with time complexity Θ(n). The properties and the relationships between these orderings and the set of integers {0,1,…,mn−1} are examined in detail. We define explicit transformations between the different orders and illustrate them as a digraph very close to the complete symmetric digraph. In this way, we provide a unified framework for understanding ranking, unranking, and order conversion. Our approach, based on emulating the execution of nested loops, proves powerful and flexible, leading to elegant and efficient algorithms that can be extended to the generation of submultisets, the generation of numbers in mixed-radix number systems, and related problems. The mirror m-ary Gray code introduced here has potential applications in coding theory and related areas. By providing an alternative perspective on m-ary Gray codes, we aim to inspire further research and applications in combinatorial generation and coding theory.

The joint optimization of transportation and inventory replenishment related decisions promises to yield significant cost savings coupled with higher customer satisfaction levels. This paper investigates a supply chain system consisting of a single supplier replenishing a single retailer, where the primary focus is on the retailer's decision-making process, aimed at determining the most efficient operational policy. This includes identifying the optimal replenishment quantity from the supplier and selecting the appropriate mix and size of the truck fleet under different situations. At first, the scenario whereby the retailer exclusively operates its limited fleet of trucks for inbound transportation is considered. An efficient solution procedure along with closed form expressions for the optimal ordering quantity and the number of trucks are devised. Subsequently, the problem is extended to incorporate environmental considerations under carbon tax and carbon cap policies. We propose a computationally efficient algorithm for generating the optimal operational policy following the carbon cap policy. Finally, to better resemble reality, the scope of the operational optimization model is extended via allowing the retailer the option to lease trucks from the external market. The conducted numerical experiments demonstrate that this flexibility can lead to significant cost reductions that are increasing with demand.