Combinatorial optimization underpins applications in artificial intelligence, logistics, and network design, yet classical techniques such as greedy search and dynamic programming struggle to balance efficiency and solution quality at scale. We present a probabilistic framework that embeds true random number generators based on spin-transfer-torque magnetic tunnel junctions into a greedy solver. Intrinsic stochastic switching enables configurable random number distributions, which we use to inject controlled randomness via a temperature parameter that interpolates between deterministic and stochastic choices, balancing exploration and exploitation. Applied to the traveling salesman problem, the framework yields high-quality tours and outperforms simulated annealing and genetic algorithms in solution quality and convergence speed. In larger instances with up to 70 cities, it maintains its advantage, reaching near-optimal solutions with fewer iterations and reduced computational cost. These results show that hardware true randomness with tunable statistics can improve heuristic search and motivate integrated, energy-efficient probabilistic hardware for scalable optimization.

Intelligent nanoparticle design: Unlocking the potential of AI for transformative drug delivery

Image interpretation and generation method integrating block sentinels and AAM in intelligent art design

Research on intelligent customization design technology for nuclear power unit systems

Intelligent fatigue-resistant design of crack-stop hole under arbitrary loading mode

A hybrid Bayesian model updating and non-dominated sorting genetic algorithm framework for intelligent mix design of steel fiber reinforced concrete

Conceptualization and validation of an intelligent digital twin design framework for supply chain risk management

Intelligent pattern design using 3D modelling technology for urban sculpture designing

Multi-objective intelligent optimization design method for mix proportions of hydraulic asphalt concrete facings

Intelligent design framework for finned solar air heaters: A synergy between PSO/GA-tuned MLPNN and multi-objective crystal structure algorithm (MOCryStAl)

The Construction of an Artificial Intelligence Model for the Intelligent Design of Ring Erbium-Doped Fiber Lasers

Intelligent design of broadband metasurface based on spectrum prediction neural network and transition simulated annealing algorithm

Multi-objective collaborative optimization of an LPP coaxial staged combustor: A data-driven intelligent design framework

From innate power to intelligent design: The evolution of NK cell-based cancer immunotherapy.

Electroactive hydrogel bone adhesives:from intelligent design to biomedical applications

This paper examines the developmental trajectory of age-friendly design from early barrier-free principles to contemporary intelligent ageing environments. Drawing upon international scholarship, ageing-policy developments, and advances in interactive technologies, the study identifies a three-stage evolution consisting of accessibility-oriented design, usability and dignity-centered design, and intelligent empowerment design. It further analyzes four underlying mechanisms—physiological, cognitive, emotional, and technological—that shape older adults’ interaction with built and digital environments. The findings demonstrate that age-friendly design has shifted from compensatory modifications toward proactive, adaptive, and capability-enhancing systems. The study argues that future age-friendly environments must integrate human–technology co-adaptation, emotional well-being, and long-term autonomy as core objectives. This integrated framework provides theoretical grounding for developing intelligent, inclusive, and sustainable design strategies that support meaningful later-life experiences.

With the rapid advancement of artificial intelligence technology, modern ceramic art design is undergoing a digital and intelligent transformation. This paper proposes a design method for modern ceramic works based on a visual expression model that integrates deep learning and cognitive computing technologies to enhance both design efficiency and artistic expression. By incorporating deep feature extraction, attention mechanisms, and the Swin Transformer framework, the detailed representation of ceramic art images is optimized. Additionally, 3D image reconstruction technology is employed to facilitate the generation of ceramic works. Experimental results demonstrate that the proposed method achieves higher accuracy and better performance than existing approaches in ceramic image classification and reconstruction tasks. This research fosters the integration of ceramic art creation with artificial intelligence technology and advances the application of intelligent design.

Proton Exchange Membrane Fuel Cells (PEMFCs) are a popular alternative to replace combustion engines, providing a clean pathway for green mobility and green energy. With their ability to deliver high power density at relatively low operating temperatures, PEMFCs are ideally suited for applications ranging from heavy-duty vehicles to grid balancing. However, optimizing their performance while extending stack lifetime requires a deep understanding of the complex electrochemical and transport phenomena that govern their operation.This study presents a comprehensive investigation of an industrial-scale PEMFC stack from Plug Power operating at high current densities—up to 1 A/cm²—across 27 different operating conditions. The operational condition design of experiments varied key parameters including current density, temperature, fuel ratio, cathode and anode humidity, and cathode and anode pressure, measuring each of the 20 cells within the stack simultaneously. The goal was to map how these variables interact and influence electrochemical behavior under realistic operating conditions.To achieve this, we employed in-operando electrochemical impedance spectroscopy (EIS)— using specialized instruments developed by Pulsenics—to continuously monitor internal electrochemical processes during fuel cell operation. Unlike conventional EIS, which is performed post-operation or at rest, in-operando EIS captures real-time impedance signatures, allowing direct observation of performance-limiting mechanisms as they occur. Time-resolved data revealed strong correlations between operating conditions and changes in cell impedance. These insights link measurable electrochemical parameters to overall stack performance, enabling the development of operational strategies to optimize efficiency and durability.This work demonstrates the value of in-operando EIS as a high-resolution tool for fuel cell diagnostics, enabling more intelligent design, operation, and management of PEMFC systems in industrial applications. By bridging the gap between lab-based measurements and real-world operation, this technique offers a powerful pathway to accelerate the commercialization and reliability of fuel cell technology.

Abstract Electrospun fiber mats, as a class of high-performance nonwoven materials, are widely applied in textiles, filtration, medical, and other fields. However, the precise three-dimensional characterization of their microstructure and quantification of volume fraction face challenges such as low resolution, poor computational efficiency, and reliance on expensive experimental imaging. This study aims to develop a computer modeling method independent of experiments, achieving high-precision reconstruction and performance prediction of fiber mats. Methodologically, by simulating the electrospinning deposition process, a parameterized deposition model is constructed, and a solvent-orientation coupled dynamic contact model is proposed, which integrates solvent residual concentration with von Mises orientation distribution and quantifies fiber cross-penetration behavior through adhesion offset equations. The main work includes developing an efficient voxelization algorithm that analyzes fiber-voxel interactions via multi-level detection (center point-corner point-ray penetration) and tolerance compensation mechanisms, enabling rapid calculation of volume fraction. Experimental results demonstrate that the error rate of this method is below 2%, and it remains robust in high fiber density scenarios. This model not only provides a high-precision tool for studying the relationship between microstructure and performance of electrospun materials but can also be extended to multi-process parameter optimization and multi-scale performance prediction, thereby promoting the intelligent design and application of nonwoven materials.

Abstract This study presents SupportGAN, a knowledge‐guided framework based on a two‐stage generative adversarial network, for preliminary conceptual plan‐view layout of corner‐ and cross‐supporting structures in foundation pits. Design drawings of support structures collected from authoritative design institutes were semantically processed and expanded through a structural knowledge‐guided augmentation (SKGA) approach. The SupportGAN model was then trained with multiple hyperparameter configurations to achieve optimal performance. Additionally, SupportGAN was evaluated and compared with two mainstream GAN models (pix2pix and pix2pixHD), demonstrating its superior capabilities. The design results generated by SupportGAN were evaluated using visual assessment and quantitative metrics. A feasibility assessment was also performed to confirm the economic and mechanical viability of the generated layouts: A pixel‐count proxy showed a 6.90% material gap versus engineer designs, and results of the finite element (FE) analysis on two cases indicated comparable structural performance (force difference ). The results indicate that SupportGAN's outputs exhibit significant similarities to those of expert engineers across various aspects, demonstrating its potential to aid designers in the preliminary conceptual layout design of corner‐ and cross‐supporting structures.