Adaptive Constraint Research Articles

Learning Critically: Selective Self-Distillation in Federated Learning on Non-IID Data

Federated learning (FL) enables multiple clients to collaboratively train a global model while keeping local data decentralized. Data heterogeneity (non-IID) across clients has imposed significant challenges to FL, which makes local models re-optimize towards their own local optima and forget the global knowledge, resulting in performance degradation and convergence slowdown. Many existing works have attempted to address the non-IID issue by adding an extra global-model-based regularizing item to the local training but without an adaption scheme, which is not efficient enough to achieve high performance with deep learning models. In this paper, we propose a Selective Self-Distillation method for Federated learning (FedSSD), which imposes adaptive constraints on the local updates by self-distilling the global model’s knowledge and selectively weighting it by evaluating the credibility at both the class and sample level. The convergence guarantee of FedSSD is theoretically analyzed and extensive experiments are conducted on three public benchmark datasets, which demonstrates that FedSSD achieves better generalization and robustness in fewer communication rounds, compared with other state-of-the-art FL methods.

Self‐Determination Skills in Ageing Women With Intellectual Disabilities

ABSTRACTBackgroundWomen with intellectual disabilities confront distinct challenges as they age, compounded by cognitive and adaptive constraints. These challenges encompass heightened health risks, barriers to social support and susceptibility to violence, underscoring the critical role of self‐determination in enhancing their quality of life.MethodThis quantitative study utilised a scale based on the Functional Model of Self‐Determination to assess self‐determination levels in 218 Chilean women aged 45 to 80 with intellectual disabilities. The research also explored potential personal and environmental factors influencing these levels.FindingsEmployed participants and those in supportive family environments demonstrated higher self‐determination levels. Moreover, women without mental health issues exhibited advanced skills in self‐regulation, empowerment and self‐realisation. Significant disparities were noted based on varying levels of intellectual disability, types of support received and socioeconomic status, highlighting the necessity for tailored and diversified support systems.ConclusionDeveloping personalised support programs that address individual and contextual differences is crucial. Despite limitations, this study emphasises the urgency for inclusive public policies aimed at eliminating socioeconomic and cultural barriers, thereby fostering environments conducive to community engagement and autonomous decision‐making.

Phase-only hologram optimization with adaptive constraint strategy for high-quality optical reconstruction

Phase-only hologram optimization with adaptive constraint strategy for high-quality optical reconstruction

Adaptive Robust Control of Nonlinear Constrained Stirred‐Tank Reactors via Self‐Organizing Fuzzy Neural Network

ABSTRACTThis paper presents an adaptive robust constraint controller for a continuous stirred tank reactor (CSTR) system based on a self‐organizing fuzzy neural network (SOFNN). Due to the high complexity of the chemical reactions, the CSTR system contains many strong nonlinearities and uncertainties. This is the first time to introduce the SOFNN into the adaptive controller design of the CSTR system, which improves the adaptability to dynamic system changes through the adjustment of the fuzzy network structure. Meanwhile, the time‐varying integral barrier Lyapunov functions (TVIBLFs) are employed to ensure the dimensionless reactant concentration and mixture temperature within a reasonable scope, which can improve the stability and safety of the CSTR system. Based on Lyapunov stability analysis, all the signals in a closed‐loop system are ultimately bounded. Simulation results substantiate the efficacy of the proposed control scheme.

Adaptive robust constraint-following control method for improving carbody hunting stability of high-speed trains

The low-frequency carbody hunting motion of high-speed trains (HSTs) frequently occurs due to the complex operating environment and deterioration of wheel–rail contact conditions. It not only affects ride comfort but also poses a risk of derailment safety accidents for HSTs. In such scenarios, traditional passive suspension systems struggle to ensure satisfactory dynamic performance of HSTs. However, active suspension systems can offer an effective solution. To enhance the low-frequency carbody hunting stability and environmental adaptability of HSTs, this study proposes a novel adaptive robust displacement inequality constraint following control (AICFC) method for active suspension utilising the beneficial nonlinearity inspired by bio-inspired structures (BIS). The effectiveness of the proposed AICFC-BIS method is validated through numerical simulations comparing it with other control strategies. Simulation results show that the AICFC-BIS method can effectively suppress low-frequency carbody hunting motion and improve the dynamic performance of HSTs. Compared with conventional passive suspensions and other active suspensions, HSTs equipped with AICFC-BIS active suspension exhibit significantly reduced operational safety risk and improved ride comfort indices while being energy efficient.

Adaptive constraint-guided surrogate enhanced evolutionary algorithm for horizontal well placement optimization in oil reservoir

In the face of escalating global energy demands, this study introduces an Adaptive Constraint-Guided Surrogate Enhanced Evolutionary Algorithm (ACG-EBS) for optimizing horizontal well placements in oil reservoirs. Addressing the complex challenge of maximizing oil production, the ACG-EBS integrates geological, engineering, and economic considerations into a novel optimization framework. This algorithm stands out for its adept navigation through a complex and discrete decision space of horizontal well placements, an area where traditional methods often encounter challenges. Key innovations include the Adaptive Constraint Initialization Mechanism (ACIM) and the Evolutionary Constraint-Tailored Candidate Refinement strategy (ECTCR), which collectively elevate the feasibility of candidate solutions. An enhanced balance strategy harmonizes comprehensive and niche surrogate models, optimizing the balance between exploration and exploitation. Through testing on both two-dimensional and three-dimensional reservoir models, the ACG-EBS has proven highly effective in identifying optimal well placements that align with field deployment realities and maximize economic returns. This contribution significantly supports the ongoing evolution of oilfield development optimization, showcasing the algorithm's potential to enhance oil production and economic outcomes.

Social robot navigation through constrained optimization: A comprehensive study of uncertainty-based objectives and constraints in the simulated and real world

This paper provides an empirical evaluation, in both simulation and real scenarios, of the social navigation problem when considering human motion prediction and its stochastic effects. To this end, we study several different optimization criteria and constraints related to the uncertainty of predicting pedestrians’ motion, embedded into the Model Predictive Control (MPC) scheme.The main research question of this work is the following: what are the most important uncertainty-based criteria for the social MPC both in simulated and real-world environments? In order to achieve a solid answer to this question, we extend the results previously obtained from our work (Akhtyamov et al., 2023) in the simulated environments and provide a real-world setting that mimics similar conditions, for a fair comparison of the qualitative and quantitative results.The main conclusions supported by both of the evaluation environments are the advantages of using adaptive constraints as a clear undisputed enhancement and the problems raised when considering uncertainty-aware criteria. We hope this paper is of interest to the community for deciding and designing uncertainty-aware approaches for social robot navigation.

Adaptive support constraint based on pixel difference distribution for twin-image removal in in-line holography

A tight support domain can enhance the reconstruction effect of phase retrieval algorithms using a support domain constraint. Hence, an adaptive support method based on pixel difference distribution to calculate a more accurate support domain is proposed in this paper. The support domain is continuously updated at each iteration of the phase retrieval algorithm, resulting in a support domain that closely corresponds to the shape and size of the object. Simulations are conducted to investigate the reconstruction effect of complex-valued objects as well as the influence of varying the binarization threshold and the difference window size on the convergence performance in the absence of noise. The changing dynamics of the support domain during the iterative process are also showcased. The experiments demonstrate that the adaptive support method can improve the reconstruction image quality and image contrast under actual conditions. Both simulations and experiments affirm the feasibility of this approach.

PPP time transfer using an adaptive clock constraint model

ABSTRACT In the processing of Precise Point Positioning (PPP) data, the receiver clock is approached with nearly infinite uncertainty, rendering it difficult to fulfill the requirements of high-precision time frequency applications. Therefore, a receiver clock model is essential. In this study, we first analyze the “over-constraint” problem in the existing clock model and subsequently propose a new clock model, called the Adaptive Clock Constraint (ACC) model, which relies on a sliding window to update covariance and frequency characteristics parameters in real-time. To verify the robustness of the ACC model, three experiments were conducted, and the results show that the ACC model not only is suitable for different types of atomic clock stations but also has superior frequency stability and time transfer precision in contrast to the BIPM PPP, the IGS products and White Noise (WN) model results. Using the optical fiber results as a time reference, the STD of the time difference between the ACC model and optical fiber results is 0.13 ns and the frequency stability is 1.28 × 10−16 on average for one week, representing improvements of more than 10% and 15% compared with the BIPM PPP results.

Adaptive Asymptotic Tracking Funnel Constraint Control for High-Power Nonlinear Multi-Agent Systems With Non-Affine Nonlinear Faults and Fixed-Time Convergence

Adaptive Asymptotic Tracking Funnel Constraint Control for High-Power Nonlinear Multi-Agent Systems With Non-Affine Nonlinear Faults and Fixed-Time Convergence

RSDM: Research of Rotation and Small Object Detection Models for Remote Sensing Images

Abstract Instance objects captured in aerial imagery display a range of orientations and exhibit significant scaling differences in their dimensions. Recent researches mainly address the above problems in terms of label assignment strategies and extended backbone spatial receptive fields, however, the former may still introduce lower quality samples, while the latter usually introduces considerable background noise. In this paper, the proposed Multiscale Convolutional Kernels Combined with Bi-Level Routing Attention for backbone Feature Extraction Network (MKBANet) employs multi-scale convolutional kernels to extract feature information at different scales, capture local context information, and combine with a Bi-Level Routing Attention (BRA) to capture remote context information and reduce computational complexity. In addition, an adaptive constrained oriented point set strategy (ACOPSS) is proposed. The adaptive point set constraint strategy is introduced in training, and then the quality metric function is utilized to select the top k high-quality directed point set samples in training, and finally the outliers are penalized. 81.23% mAP on DOTA dataset, the efficacy of the introduced approach is demonstrated.

Dynamic noise self-recovery ECM clustering algorithm with adaptive spatial constraints for image segmentation

Dynamic noise self-recovery ECM clustering algorithm with adaptive spatial constraints for image segmentation

Adaptive Constraint Relaxation-Based Evolutionary Algorithm for Constrained Multi-Objective Optimization

The key problem to solving constrained multi-objective optimization problems (CMOPs) is how to achieve a balance between objectives and constraints. Unfortunately, most existing methods for CMOPs still cannot achieve the above balance. To this end, this paper proposes an adaptive constraint relaxation-based evolutionary algorithm (ACREA) for CMOPs. ACREA adaptively relaxes the constraints according to the iteration information of population, whose purpose is to induce infeasible solutions to transform into feasible ones and thus improve the ability to explore the unknown regions. Completely ignoring constraints can cause the population to waste significant resources searching for infeasible solutions, while excessively satisfying constraints can trap the population in local optima. Therefore, balancing constraints and objectives is a crucial approach to improving algorithm performance. By appropriately relaxing the constraints, it induces infeasible solutions to be transformed into feasible ones, thus obtaining more information from infeasible solutions. At the same time, it also establishes an archive for the storage and update of solutions. In the archive update process, a diversity-based ranking is proposed to improve the convergence speed of the algorithm. In the selection process of the mating pool, common density selection metrics are incorporated to enable the algorithm to obtain higher-quality solutions. The experimental results show that the proposed ACREA algorithm not only achieved the best Inverse Generation Distance (IGD) value in 54.6% of the 44 benchmark test problems and the best Hyper Volume (HV) value in 50% of them, but also obtained the best results in seven out of nine real-world problems. Clearly, CP-TSEA outperforms its competitors.

Time-efficient model predictive control for autonomous tugs with adaptive input constraints

The marine autonomous surface ship (MASS) has gained significant attention because of its wide application in waterborne transportation in recent years. The trajectory control of the MASS is crucial in ensuring safety, particularly in narrow navigation areas and urgent encounter situations. Model predictive control (MPC) is well known as a sufficient method to solve the tracking problem considering the actuator saturation with model uncertainties and environmental disturbances. However, due to the constraints on computing resources and hardware limits, the controller may fail to find a feasible solution for MPC within a reasonable amount of time, especially when the system models are coupled and complex. To improve the solution efficiency, this paper introduces adaptive constraints to reduce search space for optimization, i.e., the current tracking point’s speed is transformed into input constraints in the MPC formulation to decrease the computation burden, as well as reduce mechanical wear on the thrusters with smaller amplitudes of control actions. Simulations are carried out to evaluate the effectiveness of the proposed method with different MPC forms.

Robust adaptive asymmetric state constraint control of brush DC motor systems driving a one-link robot manipulator: A feasibility-condition-free method

In this paper, we investigate the tracking control problem for a class of brush direct current (DC) motor systems driving a one-link robot manipulator subject to asymmetric full-state constraints. By constructing a state-dependent nonlinear transformation function (NTF), we present an adaptive robust dynamic surface control (DSC) strategy that can directly address both symmetric and asymmetric state constraints, so that there is no need to convert the problem of state constraint into the constraints on tracking errors as necessitated by the Barrier Lyapunov Function (BLF)-based existing works. Furthermore, by employing the first-order filter and constructing a new coordinate transformation, the demanding feasibility condition imposed on the BLF methods is removed, allowing the designer more freedom to select design parameters. Moreover, it is worth mentioning that under the proposed nonlinear transformation function the extra condition on the constraining function is not required. The effectiveness of the proposed control is verified via the Simulation results.

Adaptive center constraint for joint release rate estimation and model correction: Multi-scenario validation against wind tunnel experiments

Release rate estimation is crucial for the consequence assessment and emergency decision-making in nuclear accidents. However, inevitable model biases can lead to significant deviations. This study proposes an Adaptive Center Constraint for joint release rate estimation and model correction (ACC joint) for improved robustness and accuracy. It uses a tailored cost function to determine the optimal center constraint, which can automatically adapt to different cases. It was validated against four wind tunnel experiments, which simulated complex dispersion scenarios with densely built-up and highly heterogeneous terrains. The ACC joint method was compared with the Tikhonov and joint correction. The results indicate that the proposed method significantly improves the accuracy of release rate estimation. Compared to the joint correction method, the mean relative error is reduced by 38.6% and 31.4% in the all-measurement and independent validation, respectively. Furthermore, sensitivity analysis reveals that the ACC joint method provides lower mean relative error with different numbers of measurements and shows ultimate stability in all scenarios. It also suggests that measurement sites should be positioned in downwind high-concentration areas and the foot of mountainous areas for reliable estimation. The results from different cost functions verify the scalability of the proposed method, providing potential applications to other complex scenarios.

Coevolutionary multitasking for constrained multiobjective optimization

Addressing the challenges of constrained multiobjective optimization problems (CMOPs) with evolutionary algorithms requires balancing constraint satisfaction and optimization objectives. Coevolutionary multitasking (CEMT) offers a promising strategy by leveraging synergies from distinct, complementary tasks. The primary challenge in CEMT frameworks is constructing suitable auxiliary tasks that effectively complement the main CMOP task. In this paper, we propose an adaptive CEMT framework (ACEMT), which customizes two adaptive auxiliary tasks to enhance CMOP-solving efficiency. The first auxiliary task dynamically narrows constraint boundaries, facilitating exploration in regions with smaller feasible spaces. The second task focuses specifically on individual constraints, continuously adapting to expedite convergence and uncover optimal regions. In solving the main CMOP task, this dual-auxiliary-task strategy not only improves search thoroughness but also clarifies the balance between constraints and objectives. Concretely, ACEMT incorporates an adaptive constraint relaxation technique for the first auxiliary task and a specialized constraint selection strategy for the second. These innovations foster effective knowledge transfer and task synergy, addressing the key challenge of auxiliary task construction in CEMT frameworks. Extensive experiments on three benchmark suites and real-world applications demonstrate ACEMT’s superior performance compared to state-of-the-art constrained evolutionary algorithms. ACEMT sets a new standard in CMOP-solving by strategically constructing and adapting auxiliary tasks, representing a significant advancement in this research direction.

The impact of elevation and prediction of climate change on an ultra high-elevation ectotherm.

Climate change may affect the survival and reproduction of ectotherms. The toad-headed lizard Phrynocephalus theobaldi, which holds the distinction of occupying the highest elevation among all reptile species on Earth, with an elevational range from 3600 to 5000 m, represents an ideal model for studying the adaptations to climatic changes across elevational gradients. Here, we used mechanistic and hybrid species distribution models (HSDM) together with characteristic measurements of thermal biology (CTmax, CTmin, and Tsel) to simulate and compare the distribution and activity periods of the lizard across elevations in response to climate change. NicheMapR simulations using only climate factors predicted that all populations will be negatively impacted by climate change (+3°C) by suffering a reduced distribution. However, the impact was clearly reduced in simulations that accounted for thermal physiological traits. Longer activity periods were predicted for all populations during climate change. The suitable distribution is predicted to change slightly, with an increase anticipated for both high and low elevation populations. However, the forecast indicates a more pronounced increase in suitable habitats for populations at higher elevations (>4200 m) compared to those at lower elevations (<4200 m). This study underscores the key influence of climate change on population establishment and stresses the importance of physiological traits in distribution simulation for future studies to understand the potential constraints in animal adaptation to extreme high environments.

General output constraints control of interconnected nonlinear systems with sensor faults

This paper delves into the problem of generalized output-constrained control for interconnected nonlinear systems, where sensor faults are modeled as time-varying unknown functions. First, a barrier function related to an output transformation function is proposed, which can meet the diverse constraint requirements by adjusting some parameters and functions. Second, the sensor fault is described as an unknown time-varying function, and its difficulties to the controller are successfully overcome with the help of adaptive control technology. Then, an output related function is introduced into the controller to handle the interconnected functions without approximators, while retaining the uncertainties in the interconnected functions. Based on the dynamic surface control technology, an adaptive general output constraint fault-tolerant controller is proposed. With the aid of the Lyapunov stability theory, it rigorously proves that the output satisfies the pre-given constraint requirements, and all signals of the closed-loop systems are bounded. Finally, the numerical simulation is presented to illustrate the validity of the proposed algorithm.

Optimizing Room Temperature Control: A Comparative Analysis of Deep Reinforcement Learning and Contextual Bayesian Optimization Methods

Building temperature control presents significant modeling challenges due to the complexity of internal systems and varied usage patterns. Consequently, black box modeling has become a prevalent approach in this domain. This technique broadly encompasses deep learning and Bayesian optimization methods, each offering unique mechanisms for deriving optimal control parameters. Deep learning methods iteratively refine control parameters using extensive datasets, whereas Bayesian optimization employs Gaussian process regression and acquisition functions to pinpoint optimal solutions. This paper conducts a comparative analysis of two innovative methods within these branches: the application of Deep Reinforcement Learning (DRL) for integrated control of household temperature and electric vehicle charging, and the Primal-Dual Contextual Bayesian Optimization (PDCBO) method for adaptive thermal control in buildings. The study reveals that the DRL approach, integrated with electric vehicle charging systems, excels in responsiveness and reducing energy consumption. In contrast, the PDCBO method enhances control by allowing for the artificial definition and continuous optimization of constraints on energy usage and comfort during the operation, facilitated by algorithmic design. By synthesizing the strengths of both methods and considering electric vehicle integration and tailored adaptive control constraints, enhanced outcomes are achievable. This paper proposes that specific control strategies be developed for different building types to address diverse real-world requirements effectively.