This paper presents an Expert Data-Induced Learning Control (E-DiLC) framework, formulated within the Adaptive Iterative Learning Control paradigm, to achieve high-precision tracking for uncertain nonlinear systems in repeatable processes. While building upon the dual-timescale adaptation structure, the proposed method mainly contributes by systematically integrating statistical expert knowledge to constrain the learning process. Specifically, an expert knowledge-based smooth soft projection is designed to confine parameter estimates within physically feasible intervals determined from expert-provided data (expected values and confidence bounds). By restricting the adaptive search to a high-confidence region, the E-DiLC framework enhances robustness and accelerates convergence compared to conventional unguided adaptation. The global stability is formally proven via Lyapunov analysis, establishing error convergence and signal boundedness. The framework's effectiveness is demonstrated on a robotic manipulator, showcasing its ability to synergise expert knowledge with online adaptation for superior control performance.

Permanent magnet synchronous motors (PMSMs) are utilized in demanding conditions and applications requiring precision and accuracy, such as servo systems. Especially at low speeds, the effects of cogging torque, current measurement and offset errors, improper controller gains, mechanical resonance, and torque fluctuations caused by load torque and flux result in fluctuations at various frequencies in the motor output speed. This study, motivated by two factors, proposes an extended state observer (ESO)-based multivariable fast response extremum-seeking (FESC) interval type-2 fuzzy PID (IT2FPID) controller to improve dynamic response and reduce speed ripple at low speeds in situations where all these negative factors could arise. This approach enables the real-time adaptation of parameters to counteract the decline in controller performance caused by the nonlinear characteristics of PMSMs and parameter fluctuations while also optimizing disturbance rejection in the speed response under varying operating conditions and existing speed ripple. The experimental results from the prototype setup validate that the proposed control mechanism is functional, valid, and precise in diminishing speed ripples during low-speed operations. The simulation and test outcomes of the control scheme show that speed noise at low speeds is reduced from 26% to 3% compared to traditional proportional-integral (PI) controller and supertwisting (STW) sliding mode controller (SMC) responses and that the scheme exhibits a 16–23% reduction in undershoot amplitude and faster recovery in the presence of load torque variations.

This paper introduces a novel cascaded softsign function-based PID (CSoft-PID) controller designed for precise pressure regulation in highly nonlinear shell-and-tube steam condenser systems. For the first time in the literature, the classical PID control structure is enhanced through a cascaded nonlinear transformation using the softsign function, which dynamically adjusts the controller input according to the magnitude of the error. This architecture allows for high sensitivity near the setpoint while gracefully limiting excessive control efforts during larger deviations, thereby improving stability and transient performance. To optimally tune the six parameters of the proposed controller, a new hybrid optimization algorithm, termed hGASO-PS, is proposed. This method synergistically integrates an adaptive gbest-guided atom search optimization (ASO) strategy with the precision of the pattern search (PS) technique, ensuring both effective global exploration and fine-tuned local exploitation. The controller parameters are optimized by minimizing the integral of time-weighted absolute error (ITAE), subject to a step change in the condenser pressure setpoint. Extensive simulations and statistical evaluations demonstrate the superiority of the proposed approach. The hGASO-PS-based CSoft-PID controller achieved the lowest ITAE value of 2.1608, with an average of 2.2746 across 30 runs. It also demonstrated the fastest settling time (12.51 s) and the lowest overshoot (1.98%) among all tested controllers. Comparisons with recent PI, FOPID, and cascaded PI-PDN controllers confirm the consistent outperformance of the proposed method in both transient response and control precision, making it a promising candidate for industrial condenser applications.

The widespread commercialization of wireless chargers for electric vehicles generally suffers from one main problem, which is the perfect alignment between the two coils, leading to a decrease in mutual inductance, which causes a drop in magnetic coupling and even a failure to transfer power. To address this persistent problem, this work proposes a comprehensive and integrated method for optimizing the coils and control architecture for reliable and safe battery charging. To address the challenges of a complex, nonlinear design space and the need for misalignment-tolerant geometries, we employ a memetic algorithm (MA) that hybridizes Particle Swarm Optimization (PSO) for broad global exploration with Mesh Adaptive Direct Search (MADS) for precise local refinement. This combination effectively avoids poor local solutions—a limitation of standalone PSO or GA approaches reported in recent studies—while efficiently converging to coil geometries that maintain strong magnetic coupling under misalignment. After the coils have been designed, electromagnetic validation is tested using finite element analysis (FEA), which allows the magnetic field distribution to be evaluated, as well as the coupling coefficient under different scenarios of misalignment and variation in the air gap between the ground side and the vehicle side. At the same time, a comprehensive control strategy for the primary side of the system has been developed. This control method ensures power management on the primary side, enabling system interoperability for charging multiple types of vehicles, as well as reducing vehicle weight for greater range. All this is combined with an innovative pulsed current charging method, chosen for its advantages in terms of thermal stability, ensuring safe and efficient recharging that is mindful of battery health. Simulation and experimental validation demonstrate that the proposed framework maintains stable wireless power transfer and achieves over 87% DC–DC efficiency under lateral misalignments up to 100 mm, fully complying with SAE J2954 alignment tolerance requirements.

The rapid adoption of electric mining trucks in open-pit mining faces a significant scheduling challenge: effectively coordinating transportation tasks with loading/unloading and battery swapping. To address this, we model the integrated operation process as a flow shop with battery swapping, enabling the coordination of these interdependent processes. Given the problem's computational intractability, we develop a three-stage customised ALNS algorithm with proactive battery management, featuring a novel two-dimensional encoding scheme and problem-specific destroy-repair operators. Extensive experiments demonstrate the superior performance of the proposed ALNS over commercial solvers and benchmark metaheuristics. A case study shows the approach can reduce carbon emissions by 31.97% under average grid conditions, translating to an annual reduction of approximately 11.7 thousand tonnes of CO 2 . This study provides a practical decision-making tool for achieving sustainable and continuous mining production.

Online social networks have become crucial platforms for marketing, leveraging their vast user bases and capacity for rapid information dissemination. Influence maximization (IM) problem plays a central role in applications such as viral marketing and information propagation. Nevertheless, IM remains challenging due to the difficulty in balancing accuracy and efficiency with existing methods. To overcome these limitations, a discrete hybrid optimizer called PDNS-DODE (PageRank-based Diffusion Neighborhood Search and Discrete Osprey Optimization with Differential Evolution) was proposed. The algorithm incorporates an adaptive three-hop diffusion neighborhood search (PDNS) strategy based on PageRank centrality, which dynamically adjusts the search scope to enable broader global exploration and targeted local optimization. A PageRank-descending initialization strategy is introduced to improve population diversity. Furthermore, the Osprey Optimization Algorithm (OOA) and Differential Evolution (DE) are discretized with revised individual representation and update mechanisms for solving the influence maximization problem. Extensive experiments on seven real-world networks under both Independent Cascade and Linear Threshold models demonstrate the robustness and superiority of PDNS-DODE. It consistently outperforms seven state-of-the-art baselines, achieving statistically significant improvements (Wilcoxon test) and the highest overall ranking (Friedman test). These advancements are attained while maintaining competitive computational efficiency.

To address the issues of premature convergence and insufficient local search capability in genetic algorithms (GAs), this article proposes a hybrid optimization algorithm: the harmony search–genetic algorithm (HS-GA). By integrating the principles of GA and harmony search (HS), this algorithm significantly enhances global and local search capabilities through fine division of the GA population and adaptive optimization of crossover and mutation probabilities for each subgroup. It is suitable for truss structure optimization design with both discrete and continuous variables. Size optimization analyses based on classical truss examples demonstrate that, under structural constraints, HS-GA reduces structural weight by an average of 9.27% and 4.69% compared to GA and the improved genetic algorithm (IGA) (across three examples), respectively, exhibiting superior comprehensive performance.

Effective monitoring is vital for maintaining interconnected infrastructure systems, where components are prone to failure without proper servicing. However, designing inspection routes remains computationally difficult due to high complexity and inherent uncertainties of large-scale infrastructure systems. This paper investigates the deployment of multi-vehicle fleets, such as unmanned aerial vehicles (UAVs), to inspect spatially distributed components subject to uncertain travel times, inspection durations, and failure risks. Notably, the probability of component failure depends on inspection timing, creating decision-dependent (endogenous) uncertainty. We model this as a variant of a stochastic multi-vehicle routing problem and formulate a two-stage stochastic mixed-integer program based on finite samples. We propose a scenario decomposition framework that integrates column generation and random coloring techniques to accelerate subproblem resolution. We further provide theoretical analyses of the algorithm’s finite convergence and optimality guarantees under a user-specified probabilistic error tolerance. Numerical experiments on networks of varying topologies, including IEEE and EPANET systems, demonstrate the computational efficiency and effectiveness of our approaches. Across all large instances, our algorithm achieves an optimality gap below 4% and consistently outperforms state-of-the-art optimization solvers and the adaptive large neighborhood search as a heuristic benchmark.

Peak-hour congestion in metro systems poses significant challenges to operational reliability and passenger experience. This study investigates a coordinated operational strategy that integrates passenger flow control, reservation-based entry, and skip-stop train operations to alleviate congestion in high-density metro corridors. A mathematical optimization model is formulated to jointly capture passenger demand, station crowding, and train capacity constraints, and is solved using an adaptive large neighborhood search algorithm. Numerical experiments based on a real-world metro line demonstrate that the proposed framework can effectively reduce passenger waiting time and improve the balance of passenger distribution across stations under peak-hour conditions. The results indicate that coordinating multiple operational measures yields better performance than applying individual strategies in isolation, highlighting the practical value of the proposed approach for congested metro systems.

The quality of food packaging plays a significant role in determining the shelf life of consumer products. This is a good preservative that extends the shelf life of food. In some cases, food packaging may contain some defects, such as blow holes, holes, burrs, shrink defects, mold material defects, metal defects, metal defects, etc., which may affect the purity of the food. To avoid these issues, in this paper, an efficient automatic detection system using a deep learning (DL) algorithm is proposed. The proposed approach consists of two stages namely, pre-processing and defect detection. Initially, the images are collected from the dataset. Median filters are used to remove noise from the images. Once the pre-processing has been completed, the pre-processed images are given to the classifier. Optimised fully connected CNNs are proposed in this paper for classification. With adaptive reptile search optimization algorithm (ARSA), full-connected layer weight parameters are optimized for enhanced performance. The proposed algorithm effectively detects the problem and correctly segments the affected region. The performance of the proposed approach is analysed based on accuracy, sensitivity, specificity, and precision.

Estimating the number of sources in challenging scenarios with low signal-to-noise ratio (SNR), limited snapshots, or colored noise is a fundamental problem in array signal processing. This letter presents a method that recasts source number estimation from a statistical task into a problem of geometric feature detection, achieved by exploiting the invariants embedded within the data’s covariance structure. The method formulates dual determinant-trace criteria that leverage principal angle geometry to yield a characteristic cliff-like signature at the true source number. This signature is then harnessed by a unified framework integrating adaptive search and intelligent fusion, with hyperparameters optimized by an attentional Bayesian neural network. Experiments validate the method’s superior effectiveness and robustness against challenging conditions such as low-to-medium SNR, limited snapshots, and colored noise, where traditional techniques often falter.

This study addresses the Unequal Area Facility Layout Problem (UA-FLP), a classic NP-hard optimization problem where traditional exact methods are impractical for large-scale instances. To overcome this challenge, a hybrid Particle Swarm Optimization (PSO) framework is proposed, which integrates adaptive local search strategies tailored to two widely used layout representations. The algorithm features a global learning phase for efficient exploration and a self-evolutionary phase for fine-grained local refinement, supported by a diverse set of neighbourhood operators. To further enhance adaptability, two operator selection mechanisms are introduced: a probabilistic variable neighbourhood descent and a Q-learning-based strategy that adjusts operator choices based on the search context. Experimental results demonstrate the proposed method's effectiveness, scalability, and robustness in solving complex, large-scale UA-FLP instances.

The optimization of resource allocation in a cloud computing environment is a problem that has been challenging due to heterogeneous tasks with varying resource requirements and different optimization objectives for execution time, energy consumption, service level agreements (SLAs) and so on. In this paper, a hybrid multi-objective optimization algorithm called AH-NSGAIII-VND is proposed for solving a multi-objective optimization problem in a cloud computing environment. The proposed algorithm integrates a global search process using a variant of a multi-objective evolutionary algorithm called Non-Dominated Sorting Genetic Algorithm III (NSGA-III) and a local search process using a variant of a local search algorithm called Variable Neighborhood Descent (VND). The problem of resource allocation in a cloud computing environment is formulated as a multi-objective optimization problem considering makespan, energy consumption, cost, service level agreements (SLAs) and resource utilization. Researchers compared the results of the PSO, GA, and GWO optimization methods to the baseline using CloudSim-based model conditions on the changing workload scale. Experimental results indicate that cloud supply governance efficiency is enhanced with the AH-NSGAIII-VND architecture. It achieves around 11.3%, 12.8%, and 28% lower costs than the baseline NSGA-III approach while increasing overall asset utilization by over 5 percentage points. Moreover, with increasing workload, the proposed model exhibits improved convergence behavior and scalability. These results confirm that global evolutionary optimization supported by adaptive local search successfully reinstates or enhances the efficiency of resource allocation.

In semiconductor manufacturing assembly systems, a master schedule spanning several weeks outlines the production timeline. However, unexpected events like random machine failures can disrupt this schedule, causing sequencing delays and time window violations. These disruptions often render the master schedule suboptimal or infeasible, resulting in productivity losses. Remaking a new schedule is time-consuming and may lead to greater losses than the initial disruption. To address this challenge, we propose a novel Adaptive Large Neighbourhood Search (ALNS) method for rapid rescheduling. This approach accounts for complex re-entrant flows and time window constraints. Unlike traditional methods using disjunctive graph representations, the proposed ALNS employs sequence representation to mitigate infeasibility caused by machine failures. We also introduce specialised destroy and repair operators tailored to this problem context. The objective is to minimise deviations from the original schedule, making it a practical solution for real-world applications. Experimental results demonstrate that our method significantly reduces delays and time window violations, effectively enhancing system resilience.

Cloud Computing and the Internet of Things play pivotal roles in advancing the healthcare system through enhanced observation mechanisms. These mechanisms can be implemented using various algorithms, including the Sparrow Search Algorithm (SSA), Goal Programming Algorithm (GPA) and Reptile Search Algorithm (RSA) etc. Task scheduling is one of the major challenges in cloud computing, as solving this problem requires reducing costs while meeting deadlines. Efficient task scheduling is essential to optimize resource utilization and ensure timely completion of tasks. To overcome this challenge, this study presents an innovative algorithm named Adaptive Parameter Control Reptile Search Algorithm (APC-RSA) which is designed to optimize healthcare tasks scheduling by achieving a balance between time and cost as well as ensuring tasks completion within deadlines. The performance of APC-RSA is driven by dynamic parameter adjustment that balances exploration and exploitation during optimization. This study evaluates the effectiveness of APC-RSA demonstrating the significant improvements over existing algorithms like SSA, GPA and RSA. Experimental results indicate that APC-RSA achieves a minimized cost of 17,061.68 with a 99.83% task success rate. In comparison, RSA achieves a cost of 45,821.17 with a 99.70% success rate, SSA achieves a cost of 50,411.52 with a 99.67% success rate, and GPA incurs a cost of 138,448.73 with a 95.37% task success rate. The findings suggest that APC-RSA has the potential to significantly enhance cloud task scheduling in the healthcare sector, offering a cost-effective and reliable solution to improve global healthcare systems.

High-voltage circuit breakers (HVCBs) are critical switching devices whose mechanical reliability directly affects power system safety and operational continuity. Accurate fault diagnosis remains challenging due to nonlinear vibration characteristics and the sensitivity of support vector machines (SVMs) to hyperparameter selection. To address this issue, a multi-strategy improved dung beetle optimization–support vector machine (MIDBO–SVM) framework is proposed for vibration-based mechanical fault diagnosis. Frequency-domain features are extracted from vibration signals using the fast Fourier transform to characterize fault-related spectral variations. A multi-strategy improved dung beetle optimization (MIDBO) algorithm incorporating chaotic initialization, adaptive search regulation, and mutation enhancement is developed to improve population diversity, global exploration, and convergence stability. The optimized MIDBO is used to determine the penalty and kernel parameters of the SVM, constructing a robust and well-generalized diagnostic model. Experimental results show that MIDBO–SVM achieves a diagnostic accuracy of 96.67%, outperforming conventional SVM (86.25%) and random forest (89.17%). The proposed method also demonstrates faster convergence and maintains accuracy above 86% under imbalanced sample conditions, confirming its robustness and generalization capability. These advantages contribute to more reliable mechanical condition assessment and improved maintenance decision support for HVCBs.

Electric cargo (e-cargo) bikes offer a promising, sustainable alternative for last-mile logistics. However, rider fatigue remains a critical, yet often overlooked, constraint, impacting both operational performance and rider well-being. This study introduces and formulates a Chance-Constrained Heterogeneous and Multi-Trip Vehicle Routing Problem (CC-HMVRP) that accounts for rider fatigue, incorporating load mass, environmental conditions, and rider characteristics into delivery planning. A mixed-integer linear programming (MILP) formulation and a modified adaptive large neighborhood search (ALNS) solution method are proposed to handle larger instances. We examine how wind speed and temperature influence battery and rider energy levels. Numerical results show a 27.2% reduction in total energy consumption compared to deterministic models, while riders retain 62.8% more available time before reaching fatigue. These findings enhance sustainability, improve efficiency, and support rider well-being, offering new insights for optimizing last-mile delivery operations under real-world constraints. • Introduces a chance-constrained programming model for e-cargo bike routing with rider fatigue limits. • Develops an Adaptive Large Neighborhood Search (ALNS) algorithm for large-scale optimization. • Demonstrates a 27.2% reduction in total energy consumption while improving rider endurance.

Slack-aware scheduling of AGVs in just-in-time matrix manufacturing systems via adaptive large neighborhood search

An electric commuter bus battery-swapping and charging–discharging schedules collaborative optimization strategy based on the collaborative operation of electric buses and renewable energy stations

A two-stage adaptive variable neighborhood search approach for electric vehicle routing problem with fuzzy demand