<?xml version="1.0" encoding="utf-8"?>
    <rss version="2.0">
      <channel xmlns:content="http://purl.org/rss/1.0/modules/content/">
        <title>Frontiers in Applied Mathematics and Statistics | Optimization section | New and Recent Articles</title>
        <link>https://www.frontiersin.org/journals/applied-mathematics-and-statistics/sections/optimization</link>
        <description>RSS Feed for Optimization section in the Frontiers in Applied Mathematics and Statistics journal | New and Recent Articles</description>
        <language>en-us</language>
        <generator>Frontiers Feed Generator,version:1</generator>
        <pubDate>2026-08-22T20:55:09.429+00:00</pubDate>
        <ttl>60</ttl>
        <item>
        <guid isPermaLink="true">https://www.frontiersin.org/articles/10.3389/fams.2026.1853423</guid>
        <link>https://www.frontiersin.org/articles/10.3389/fams.2026.1853423</link>
        <title><![CDATA[EduDonateBlock: a blockchain-enabled crowdfunding for transparent educational smart supply chain]]></title>
        <pubdate>2026-08-10T00:00:00Z</pubdate>
        <category>Original Research</category>
        <author>Tapasi Bhattacharjee</author><author>Amalendu Singha Mahapatra</author><author>Dipika Pramanik</author>
        <description><![CDATA[Educational crowdfunding has emerged as a promising approach to provide educational resources to underprivileged communities. Conventional systems often suffer from a lack of transparency, weak accountability, inefficient allocation of funds, and inadequate traceability of resource use. To address these issues, the present study proposes an intelligent and efficient educational supply chain management system, “EduDonateBlock.” It uses a blockchain-based crowdfunding framework to ensure transparency, accountability, and efficiency. Decentralization, immutability, and verifiable transactions are supported in educational campaigns. The entire workflow is decomposed into modular smart contracts. These are the identity and access contract (IAC), campaign and donation contract (CDC), verification and allocation contract (VAC), and supply chain and tracking contract (SCTC). These contracts are designed to ensure traceability, accountability, and efficient resource allocation among donors, educational institutions, and administrators. The mathematical framework of EduDonateBlock determines the optimal level of blockchain transparency. This minimizes the Total Expected Cost (TEC) of smart-contract operations. Numerical analysis identifies an optimal transparency level of 87.16% on-chain integration. This finding underscores the economic trade-off between transaction costs and the benefits of automation, operational efficiency, and reduced fraud risk. The proposed framework achieves a campaign success probability of 89.45% and an institutional payoff of Rs. 11,335.99. Furthermore, executing smart contracts requires 0.0044 ETH, and the average latency remains at 6.25 s. The simulation results show that EduDonateBlock offers a more efficient, reliable, and transparent solution for decentralized educational crowdfunding and socially impactful digital supply chains.]]></description>
      </item><item>
        <guid isPermaLink="true">https://www.frontiersin.org/articles/10.3389/fams.2026.1773581</guid>
        <link>https://www.frontiersin.org/articles/10.3389/fams.2026.1773581</link>
        <title><![CDATA[Ant Colony Optimization applied to the Traveling Santa Problem]]></title>
        <pubdate>2026-07-23T00:00:00Z</pubdate>
        <category>Brief Research Report</category>
        <author>Elliot Fisher</author><author>Robin Smith</author>
        <description><![CDATA[The hypothetical global delivery schedule of Santa Claus must follow strict rolling night-time windows that vary with the Earth's rotation and obey an energy budget that depends on payload size and cruising speed. Given the vast number of permutations of city-to-city legs, different payload drop-offs at each location, and the necessity of delivering gifts during the hours of darkness, to minimize energy expenditure, a heuristic-based optimization was applied. To design this schedule, the Traveling-Santa Ant-Colony Optimization (TSaP–ACO) framework was developed. This heuristic framework constructs potential routes via a population of artificial ants that iteratively extend partial paths. Ants make their decisions much like they do in nature, following pheromones left by other ants, but with a degree of permitted exploration. This approach: (i) embeds local darkness feasibility directly into the pheromone heuristic, (ii) seeks to minimize aerodynamic work via a shrinking sleigh cross-sectional area depending on current payload, (iii) uses a low-cost “rogue-ant” reversal to capture direction-sensitive time-zones, and (iv) tunes leg-specific cruise speeds on the fly. On benchmark sets of 15 and 30 capital cities, the TSaP–ACO eliminates all daylight violations and reduces total work by up to 10% compared to a distance-only Ant-Colony Optimization (ACO). In a 40-capital-city stress test, it cuts energy use by 89% and shortens tour length by 60%. Population-first routing emerges naturally from work minimization (50% served by leg 11 of 40). These results demonstrate that rolling-window, energy-aware ACO has potential applications in more realistic global delivery scenarios.]]></description>
      </item><item>
        <guid isPermaLink="true">https://www.frontiersin.org/articles/10.3389/fams.2026.1842479</guid>
        <link>https://www.frontiersin.org/articles/10.3389/fams.2026.1842479</link>
        <title><![CDATA[Geodesic dominated coloring in certain product graphs]]></title>
        <pubdate>2026-07-13T00:00:00Z</pubdate>
        <category>Original Research</category>
        <author>M. Paruvatha Vathana</author><author>R. Jayagopal</author>
        <description><![CDATA[A geodesic dominated coloring of a graph G is a proper coloring in which each color class is dominated by at least one geodesic. The minimum number of colors required for such a coloring is the geodesic dominated chromatic number, denoted by χdomg(G). In this paper, we determine the geodesic dominated chromatic number for several families of graphs constructed through standard product operations, including the lexicographic product, the vertex corona, and the Hamiltonian edge corona. We further establish general bounds and characterize the influence of structural properties of these product graphs on χdomg(G).]]></description>
      </item><item>
        <guid isPermaLink="true">https://www.frontiersin.org/articles/10.3389/fams.2026.1741977</guid>
        <link>https://www.frontiersin.org/articles/10.3389/fams.2026.1741977</link>
        <title><![CDATA[Heuristics for the One-Dimensional Bin Packing Problem with Time Windows]]></title>
        <pubdate>2026-06-29T00:00:00Z</pubdate>
        <category>Original Research</category>
        <author>Huibing Cheng</author><author>Hong He</author><author>Shanshui Zheng</author><author>Jian Tu</author><author>Dengheng Zheng</author>
        <description><![CDATA[This paper investigates the one-dimensional bin packing problem with time windows (1DBPP-TW), a practical variant of the bin packing problem widely applied in logistics. Existing relevant studies mainly focus on bin packing problems with variable-sized bins under time constraints, while 1DBPP-TW with homogeneous bins and a shared time window requirement for items in one bin has not been fully explored. Given the NP-completeness of the bin packing problem, exact algorithms are not applicable to large-scale instances, so this paper first establishes a mathematical model for 1DBPP-TW and generates two benchmark datasets including optimal-solution-known instances and random instances for verification, then proposes two efficient algorithms: the Greedy on Time Range (GTR) heuristic which ensures real-time response and obtains reliable solutions within 0.001 seconds for large instances, and the Iterative Local Search (ILS) metaheuristic equipped with three neighborhood operators to improve convergence rate and reduce the number of used bins by an average of 7.3% compared with GTR. Comparative experiments with the CPLEX solver within a 3600-second time limit show that the two proposed algorithms have higher computational efficiency and equivalent or better solution quality for medium and large instances. In practice, these algorithms resolve logistics industry problems including low loading efficiency and time window conflicts, and theoretically, this research advances the combinatorial optimization theory of time-constrained bin packing variants and provides standard benchmarks for follow-up studies. This study is limited to homogeneous bins and synthetic test instances, and future work will expand to multi-dimensional packing, dynamic time window constraints and practical verification with real industrial data.]]></description>
      </item><item>
        <guid isPermaLink="true">https://www.frontiersin.org/articles/10.3389/fams.2026.1838903</guid>
        <link>https://www.frontiersin.org/articles/10.3389/fams.2026.1838903</link>
        <title><![CDATA[EMRSO: enhancing the modified rat swarm optimizer via hybrid memory strategies]]></title>
        <pubdate>2026-06-17T00:00:00Z</pubdate>
        <category>Original Research</category>
        <author>Noor K. Hamid</author><author>Azad A. Ameen</author><author>Aso M. Aladdin</author>
        <description><![CDATA[Metaheuristic algorithms play a crucial role in solving complex engineering optimization problems. However, the Modified Rat Swarm Optimizer (MRSO) often suffers from premature convergence and limited exploration capability. To address these limitations, this paper proposes an Enhanced Modified Rat Swarm Optimizer (EMRSO), which incorporates a hybrid memory mechanism and a spiral-based search strategy to improve the balance between exploration and exploitation. The performance of EMRSO was evaluated using 23 classical benchmark functions, the CEC-C06 2019 test functions, and six constrained engineering design problems. The results indicate that EMRSO provides competitive and stable performance across a wide range of optimization scenarios. In particular, it demonstrates improved robustness and the ability to avoid premature convergence in multimodal and complex search spaces, where maintaining diversity is essential. However, the performance improvements are not uniform across all problem types. In simpler or highly structured landscapes, classical algorithms such as GWO occasionally achieve higher precision, while some recent methods (e.g., DSWO and LEO) show superior performance in specific cases. These observations highlight the problem-dependent nature of optimization algorithms. Statistical analyses confirm that the proposed modifications enhance solution stability and consistency in many cases. Overall, EMRSO can be considered a robust and competitive optimization approach, particularly suitable for complex and multimodal problems, while still presenting limitations in exploitation-dominant scenarios.]]></description>
      </item><item>
        <guid isPermaLink="true">https://www.frontiersin.org/articles/10.3389/fams.2026.1824146</guid>
        <link>https://www.frontiersin.org/articles/10.3389/fams.2026.1824146</link>
        <title><![CDATA[A randomized block Krylov method for tensor train approximation]]></title>
        <pubdate>2026-05-26T00:00:00Z</pubdate>
        <category>Original Research</category>
        <author>Gaohang Yu</author><author>Jinhong Feng</author><author>Zhongming Chen</author><author>Xiaohao Cai</author><author>Liqun Qi</author>
        <description><![CDATA[Tensor train decomposition is a powerful tool to tackle high-dimensional large-scale tensor data and is not suffering from the curse of dimensionality. It relies on performing the singular value decomposition (SVD) of auxiliary unfolding matrices. Some randomized algorithms have been proposed to accelerate the SVD calculation of the unfolding matrix, but they are not suitable for noisy data. This paper proposes a randomized algorithm for low-rank tensor train approximation of large-scale tensors based on randomized block Krylov subspace iteration and provides theoretical guarantees. Extensive numerical trials conducted on synthetic as well as real-world tensor data showcase the outstanding efficacy of the suggested algorithm, particularly when dealing with noisy data.]]></description>
      </item><item>
        <guid isPermaLink="true">https://www.frontiersin.org/articles/10.3389/fams.2026.1700090</guid>
        <link>https://www.frontiersin.org/articles/10.3389/fams.2026.1700090</link>
        <title><![CDATA[Two-stage model predictive voltage control in distribution networks with high penetration of renewable generation]]></title>
        <pubdate>2026-05-19T00:00:00Z</pubdate>
        <category>Original Research</category>
        <author>Lei Chen</author><author>Jiapeng Xu</author><author>Ning Li</author><author>Zheng Tian</author><author>Naifu Yan</author><author>Qi Han</author><author>Jie Lou</author>
        <description><![CDATA[This study proposes a two-stage model predictive voltage control (TS-MPC) framework for distribution networks (DNs) with high penetration of renewable generation. The framework integrates traditional voltage control devices, such as on-load tap changers (OLTCs) and capacitor banks (CBs), with distributed generation (DG) active- and reactive-power control. In the upper stage, OLTC and CB operations are optimized on a slow timescale to minimize switching, whereas in the lower stage, DG outputs are dynamically adjusted on a fast timescale to address rapid voltage fluctuations. The proposed framework leverages both the advantages of model predictive control (MPC) and DG real/reactive power control to enhance voltage regulation performance. Simulation results on a modified 33-bus test system demonstrate the effectiveness of the proposed approach.]]></description>
      </item><item>
        <guid isPermaLink="true">https://www.frontiersin.org/articles/10.3389/fams.2026.1752154</guid>
        <link>https://www.frontiersin.org/articles/10.3389/fams.2026.1752154</link>
        <title><![CDATA[An EOQ-based inventory model for circular economy implementation in municipal solid waste management]]></title>
        <pubdate>2026-05-15T00:00:00Z</pubdate>
        <category>Original Research</category>
        <author>Vennila Sivalingam</author><author>Karthikeyan Kanagasabai</author>
        <description><![CDATA[IntroductionRapid urbanization has intensified municipal solid waste (MSW) management challenges in developing regions. Tamil Nadu exemplifies this trend, with MSW generation increasing from approximately 9,000 tons per day in 2005 to 17,843 TPD in 2024 and projected to reach 18,300 TPD by 2025.MethodsTo address this growing pressure, this study develops a circular economy-based Economic Order Quantity (EOQ) model to optimize windrow composting systems for sustainable fertilizer production. A circularity index θ (0 ≤ θ ≤ 1) is introduced to quantify material recovery and reuse and is explicitly linked to demand and unit profit through linear and nonlinear functional forms. The model integrates ordering, holding, shortage, and carbon emission-related costs to evaluate their combined effect on total profit under controllable emissions. Optimal solutions are derived using Lagrangian optimization and Karush–Kuhn–Tucker (KKT) conditions.ResultsNumerical results and sensitivity analysis demonstrate exceptional parameter stability, with ±0.1% variations in key parameters causing negligible changes in total profit, while optimal order quantity and green technology levels remain unchanged. Among the examined cases, logarithmic and logistic functional forms yield higher profits, with improvements exceeding 8–12% compared to linear structures.DiscussionIncreased circularity and green technology investment simultaneously enhance profitability and reduce landfill dependency and carbon emissions. Overall, the proposed model provides a practical decision-support tool for municipal authorities to strengthen 3R (Reduce-Reuse-Recycle) strategies, improve resource efficiency, and advance toward a sustainable circular MSW management system.]]></description>
      </item><item>
        <guid isPermaLink="true">https://www.frontiersin.org/articles/10.3389/fams.2026.1742828</guid>
        <link>https://www.frontiersin.org/articles/10.3389/fams.2026.1742828</link>
        <title><![CDATA[Fuzzy hyperheuristic optimization of a facilitated hub-and-spoke drone-enabled logistics network: a case study of Australia Post]]></title>
        <pubdate>2026-04-10T00:00:00Z</pubdate>
        <category>Original Research</category>
        <author>Kassem Danach</author><author>Samir Haddad</author><author>Wissam Khalil</author><author>Ziad El Balaa</author><author>Jinane Sayah</author>
        <description><![CDATA[IntroductionThe rapid growth of e-commerce has increased pressure on postal logistics networks, especially in remote regions.MethodsThis study proposes a fuzzy hyperheuristic genetic algorithm for optimizing a facilitated hub-and-spoke network with drone integration under uncertainty.ResultsThe proposed approach improves service robustness by 25–35% and expands drone coverage to 93.4% of remote demand, with only a modest cost increase (8–12%).DiscussionResults highlight the effectiveness of fuzzy optimization and adaptive hyperheuristics in designing resilient and cost-efficient postal logistics systems.]]></description>
      </item><item>
        <guid isPermaLink="true">https://www.frontiersin.org/articles/10.3389/fams.2026.1809903</guid>
        <link>https://www.frontiersin.org/articles/10.3389/fams.2026.1809903</link>
        <title><![CDATA[Low-rank tensor completion with fractal-inspired multi-scale energy regularization]]></title>
        <pubdate>2026-03-23T00:00:00Z</pubdate>
        <category>Original Research</category>
        <author>Chuling Wen</author><author>Weijie Liang</author><author>Chen Xu</author><author>Yuru Zou</author>
        <description><![CDATA[Low-rank tensor completion has become a fundamental tool for recovering high-dimensional data from incomplete observations. However, conventional methods rely primarily on algebraic low-rank priors and do not explicitly regulate how signal energy is distributed across scales. This study introduces a fractal-inspired multi-scale energy regularization that enforces approximate power-law scaling of tensor energy across resolution levels. The proposed formulation integrates scale-consistency constraints with tensor nuclear norm regularization in a unified framework. An inexact ADMM algorithm is developed to solve the resulting non-convex problem. Experimental results demonstrate consistent improvements in reconstruction accuracy across tensor sizes and observation ratios.]]></description>
      </item><item>
        <guid isPermaLink="true">https://www.frontiersin.org/articles/10.3389/fams.2026.1763637</guid>
        <link>https://www.frontiersin.org/articles/10.3389/fams.2026.1763637</link>
        <title><![CDATA[Advancing bearing fault detection through a modified metaheuristic optimization approach]]></title>
        <pubdate>2026-03-17T00:00:00Z</pubdate>
        <category>Original Research</category>
        <author>Lana A. Abullah</author><author>Chnoor M Rahman</author>
        <description><![CDATA[IntroductionDetecting bearing faults plays a vital role in industrial maintenance since discovering problems early can help avoid unexpected breakdowns and expensive production losses. Yet, spotting these faults in their initial stages is still difficult because vibration signals are often complex and change over time.MethodsIn this study, optimized Mel Frequency Cepstral Coefficients (MFCC) feature extraction approach enhanced through a modified FOX optimization algorithm. The enhancement focuses on fine-tuning MFCC hyperparameters to maximize the discriminative power of extracted features for fault detection tasks. The proposed Enhanced FOX (EFOX) algorithm integrates different random distribution method and improved exploration–exploitation balance, enabling more effective parameter optimization compared to conventional methods.ResultsExperimental evaluations were conducted using benchmark datasets, and the optimized MFCC features were compared against those obtained via standard MFCC settings and other metaheuristic optimization techniques. Results demonstrate that our approach consistently outperforms competing methods in terms of classification accuracy and the robustness of the proposed model was assessed by testing it on two distinct bearing’s datasets with different noise ratios including −3 dB and −6 dB.DiscussionThe analysis highlights the impact of each of hyperparameter’s of MFCC to bearing fault detection.]]></description>
      </item><item>
        <guid isPermaLink="true">https://www.frontiersin.org/articles/10.3389/fams.2026.1774262</guid>
        <link>https://www.frontiersin.org/articles/10.3389/fams.2026.1774262</link>
        <title><![CDATA[Flexibility-oriented robust optimization planning for electro-hydrogen energy storage in high-renewable grids]]></title>
        <pubdate>2026-02-23T00:00:00Z</pubdate>
        <category>Original Research</category>
        <author>Wang Yan</author>
        <description><![CDATA[The large-scale integration of renewable energy sources poses significant challenges to grid stability due to inherent intermittency and volatility. This paper presents a novel robust optimization framework for planning electro-hydrogen energy storage systems (EHESS) that differs from traditional capacity planning by explicitly incorporating flexibility margin indices. We develop a comprehensive electro-hydrogen coupling model that captures the coordinated operational characteristics of battery storage (short-term regulation) and hydrogen systems (long-term shifting). Unlike existing works that treat flexibility qualitatively, we introduce a quantified flexibility margin index to measure the supply-demand gap of ramping capabilities. We formulate a two-layer robust optimization model: the upper layer minimizes investment costs, while the lower layer minimizes operational and flexibility penalty costs under worst-case scenarios. Wasserstein distance-based uncertainty sets are employed to handle the distributional uncertainty of renewable output. Case simulations on a modified IEEE 33-node system validate that the proposed method effectively determines the optimal configuration, reducing total costs by 10.6% compared to baselines by mitigating high-cost flexibility violations.]]></description>
      </item><item>
        <guid isPermaLink="true">https://www.frontiersin.org/articles/10.3389/fams.2025.1698876</guid>
        <link>https://www.frontiersin.org/articles/10.3389/fams.2025.1698876</link>
        <title><![CDATA[A multiple stakeholder-based target-oriented robust optimization approach and its applications]]></title>
        <pubdate>2026-01-21T00:00:00Z</pubdate>
        <category>Methods</category>
        <author>Jivulter Mangubat</author><author>Celbert Himang</author><author>Marlon Solloso</author><author>Dexter Alit</author><author>Melanie Himang</author><author>Milcah Mangubat</author><author>Patrobinson Salumag</author><author>Miriam Bongo</author>
        <description><![CDATA[In real-world decision-making, multiple stakeholders often participate, each with diverse and sometimes conflicting interests, which may fall exclusively under the expertise of individual decision-makers. Existing multiple-criteria decision-making (MCDM) methods can accommodate multiple criteria but typically fail to reconcile conflicting stakeholder priorities into a satisficing solution. To address this gap, this paper proposes the multiple stakeholder-based target-oriented robust-optimization (MS-TORO) approach, which explicitly embeds stakeholder interests into an optimization framework that minimizes deviations among priorities. The implementation procedure involves eliciting ordinal stakeholder preferences, parameterizing trade-offs, and solving the optimization model to generate an aggregated solution. Three case studies demonstrate the applicability and viability of MS-TORO, showing that it effectively produces solutions that satisfy the performance targets defined by each decision-maker across all criteria.]]></description>
      </item><item>
        <guid isPermaLink="true">https://www.frontiersin.org/articles/10.3389/fams.2025.1764289</guid>
        <link>https://www.frontiersin.org/articles/10.3389/fams.2025.1764289</link>
        <title><![CDATA[Editorial: Optimization for low-rank data analysis: theory, algorithms and applications]]></title>
        <pubdate>2026-01-12T00:00:00Z</pubdate>
        <category>Editorial</category>
        <author>HanQin Cai</author><author>Dong Xia</author><author>Ernest Domanaanmwi Ganaa</author><author>Abiy Tasissa</author>
        <description></description>
      </item><item>
        <guid isPermaLink="true">https://www.frontiersin.org/articles/10.3389/fams.2025.1640044</guid>
        <link>https://www.frontiersin.org/articles/10.3389/fams.2025.1640044</link>
        <title><![CDATA[EESB-FDO: enhancing the fitness-dependent optimizer through a modified boundary handling mechanism]]></title>
        <pubdate>2025-10-17T00:00:00Z</pubdate>
        <category>Original Research</category>
        <author>Aram Kamal Faraj</author><author>Aso M. Aladdin</author><author>Azad A. Ameen</author>
        <description><![CDATA[The fitness-dependent optimizer (FDO) has recently gained attention as an effective metaheuristic for solving different optimization problems. However, it faces limitations in exploitation and convergence speed. To overcome these challenges, this study introduces two enhanced variants: enhancing exploitation through stochastic boundary for FDO (EESB-FDO) and enhancing exploitation through boundary carving for FDO (EEBC-FDO). In addition, the ELFS strategy is proposed to constrain Levy flight steps, ensuring more stable exploration. Experimental results show that these modifications significantly improve the performance of FDO compared to the original version. To evaluate the performance of the EESB-FDO and EEBC-FDO, three primary categories of benchmark test functions were utilized: classical, CEC 2019, and CEC 2022. The assessment was further supported by the application of statistical analysis methods to ensure a comprehensive and rigorous performance evaluation. The performance of the proposed EESB-FDO and EEBC-FDO algorithms was evaluated through comparative analysis with several existing FDO modifications, as well as with other well-established metaheuristic algorithms, including the Arithmetic Optimization Algorithm (AOA), the Learner Performance-Based Behavior Algorithm (LPB), the Whale Optimization Algorithm (WOA), and the Fox-inspired Optimization Algorithm (FOX). The statistical analysis indicated that both EESB-FDO and EEBC-FDO exhibit better performance compared to the aforementioned algorithms. Furthermore, a final evaluation involved applying EESB-FDO and EEBC-FDO to four real-world optimization problems: the gear train design problem, the three-bar truss problem, the pathological igg fraction in the nervous system, and the integrated cyber-physical attack on a manufacturing system. The results demonstrate that both proposed variants significantly outperform both the FDO and the modified fitness-dependent optimizer (MFDO) in solving these complex problems.]]></description>
      </item><item>
        <guid isPermaLink="true">https://www.frontiersin.org/articles/10.3389/fams.2025.1594873</guid>
        <link>https://www.frontiersin.org/articles/10.3389/fams.2025.1594873</link>
        <title><![CDATA[Plug-and-play low-rank tensor completion and reconstruction algorithms with improved applicability of tensor decompositions]]></title>
        <pubdate>2025-09-17T00:00:00Z</pubdate>
        <category>Original Research</category>
        <author>Manabu Mukai</author><author>Hidekata Hontani</author><author>Tatsuya Yokota</author>
        <description><![CDATA[In this paper, we propose a new unified optimization algorithm for general tensor completion and reconstruction problems, which is formulated as an inverse problem for low-rank tensors in general linear observation models. The proposed algorithm supports at least three basic loss functions (ℓ2 loss, ℓ1 loss, and generalized KL divergence) and various TD models (CP, Tucker, TT, TR decompositions, non-negative matrix/tensor factorizations, and other constrained TD models). We derive the optimization algorithm based on a hierarchical combination of the alternating direction method of multipliers (ADMM) and majorization-minimization (MM). We show that the proposed algorithm can solve a wide range of applications and can be easily extended to any established TD model in a plug-and-play manner.]]></description>
      </item><item>
        <guid isPermaLink="true">https://www.frontiersin.org/articles/10.3389/fams.2025.1628652</guid>
        <link>https://www.frontiersin.org/articles/10.3389/fams.2025.1628652</link>
        <title><![CDATA[Accounting data anomaly detection and prediction based on self-supervised learning]]></title>
        <pubdate>2025-09-02T00:00:00Z</pubdate>
        <category>Original Research</category>
        <author>Yingying Zhang</author><author>Bingbing Duan</author>
        <description><![CDATA[This study proposes a Hierarchical Fusion Self-Supervised Learning (HFSL) framework to address the challenge of scarce labeled data in accounting anomaly detection, integrating domain knowledge with advanced deep learning techniques. Based on financial data from Chinese listed companies in the CSMAR database spanning 2000–2020, this framework integrates temporal contrastive learning, a dual-channel LSTM autoencoder structure, and financial domain knowledge to construct a three-tier cascaded detection system. Empirical research demonstrates that the HFSL framework achieves a precision of 0.836, recall of 0.805, and F1 score of 0.820 in accounting anomaly detection, significantly outperforming traditional methods. In terms of practical metrics, the framework attains an early detection rate of 0.726 while maintaining a false alarm rate of just 0.068, providing technical support for early risk warning. Financial feature contribution analysis reveals that core indicators such as Return on Assets (ROA), Return on Equity (ROE), and their interaction effects play crucial roles in anomaly identification. Through analysis of 2,150 samples in the test set, the study identifies five typical financial fraud patterns (revenue inflation 38.6%, expense concealment 21.7%, asset overvaluation 17.4%, liability understatement 15.2%, and composite manipulation 7.1%) and their temporal evolution characteristics. The research also finds that financial anomalies typically exhibit three evolutionary patterns: progressive deterioration (64%), sudden anomalies (22%), or cyclical fluctuations (15%), providing empirical evidence for regulatory practice. This study applies self-supervised learning to accounting anomaly detection, not only solving the detection challenges in unlabeled data scenarios but also providing effective tools for financial supervision and risk management.]]></description>
      </item><item>
        <guid isPermaLink="true">https://www.frontiersin.org/articles/10.3389/fams.2025.1589033</guid>
        <link>https://www.frontiersin.org/articles/10.3389/fams.2025.1589033</link>
        <title><![CDATA[A sparse tensor generator with efficient feature extraction]]></title>
        <pubdate>2025-07-30T00:00:00Z</pubdate>
        <category>Original Research</category>
        <author>Tugba Torun</author><author>Ameer Taweel</author><author>Didem Unat</author>
        <description><![CDATA[Sparse tensor operations are increasingly important in diverse applications such as social networks, deep learning, diagnosis, crime, and review analysis. However, a major obstacle in sparse tensor research is the lack of large-scale sparse tensor datasets. Another challenge lies in analyzing sparse tensor features, which are essential not only for understanding the nonzero pattern but also for selecting the most suitable storage format, decomposition algorithm, and reordering methods. However, due to the large size of real-world tensors, even extracting these features can be computationally expensive without careful optimization. To address these limitations, we have developed a smart sparse tensor generator that replicates key characteristics of real sparse tensors. Additionally, we propose efficient methods for extracting a comprehensive set of sparse tensor features. The effectiveness of our generator is validated through the quality of extracted features and the performance of decomposition on the generated tensors. Both the sparse tensor feature extractor and the tensor generator are open source with all the artifacts available at https://github.com/sparcityeu/FeaTensor and https://github.com/sparcityeu/GenTensor, respectively.]]></description>
      </item><item>
        <guid isPermaLink="true">https://www.frontiersin.org/articles/10.3389/fams.2025.1629658</guid>
        <link>https://www.frontiersin.org/articles/10.3389/fams.2025.1629658</link>
        <title><![CDATA[Corrigendum: Expectation-maximization alternating least squares for tensor network logistic regression]]></title>
        <pubdate>2025-07-01T00:00:00Z</pubdate>
        <category>Correction</category>
        <author>Naoya Yamauchi</author><author>Hidekata Hontani</author><author>Tatsuya Yokota</author>
        <description></description>
      </item><item>
        <guid isPermaLink="true">https://www.frontiersin.org/articles/10.3389/fams.2025.1593680</guid>
        <link>https://www.frontiersin.org/articles/10.3389/fams.2025.1593680</link>
        <title><![CDATA[Expectation-maximization alternating least squares for tensor network logistic regression]]></title>
        <pubdate>2025-05-14T00:00:00Z</pubdate>
        <category>Original Research</category>
        <author>Naoya Yamauchi</author><author>Hidekata Hontani</author><author>Tatsuya Yokota</author>
        <description><![CDATA[In recent years, a learning method for classifiers using tensor networks (TNs) has attracted attention. When constructing a classification function for high-dimensional data using a basis function model, a huge number of basis functions and coefficients are generally required, but the TN model makes it possible to avoid the curse of dimensionality by representing the huge coefficients using TNs. However, there is a problem with TN learning, namely the gradient vanishing, and learning using the gradient method cannot be performed efficiently. In this study, we propose a novel optimization algorithm for learning TN classifiers by using alternating least square (ALS) algorithm. Unlike conventional gradient-based methods, which suffer from vanishing gradients and inefficient training, our proposed approach can effectively minimize squared loss and logistic loss. To make ALS applicable to logistic regression, we introduce an auxiliary function derived from Pólya-Gamma augmentation, allowing logistic loss to be minimized as a weighted squared loss. We apply the proposed method to the MNIST classification task and discuss the effectiveness of the proposed method.]]></description>
      </item>
      </channel>
    </rss>