This study presents a comprehensive approach to solving an advanced Multi-Robot Task Allocation (MRTA) problem, in which heterogeneous agents, initially stationed at different depots, are required to perform a set of spatially distributed tasks. A key challenge lies in determining which agent performs each task and defining the order in which each agent executes the assigned tasks, while optimizing travel costs and respecting energy constraints. To address this problem, we propose an approach that combines the Greedy Randomized Adaptive Search Procedure (GRASP) with 2-Opt local search. A comparative analysis was conducted against a Mixed Integer Linear Programming (MILP) solver and standard GRASP variants. The results demonstrate that the GRASP + 2-Opt approach strikes a favorable balance between optimality and execution time. This work provides practical insights for applications such as industrial inspection and environmental monitoring, where autonomous multi-robot coordination and energy constraints are essential for sustained and reliable operation.

This study presents a comprehensive approach to solving an advanced Multi-Robot Task Allocation (MRTA) problem, in which heterogeneous agents, initially stationed at different depots, are required to perform a set of spatially distributed tasks. A key challenge lies in determining which agent performs each task and defining the order in which each agent executes the assigned tasks, while optimizing travel costs and respecting energy constraints. To address this problem, we propose an approach that combines the Greedy Randomized Adaptive Search Procedure (GRASP) with 2-Opt local search. A comparative analysis was conducted against a Mixed Integer Linear Programming (MILP) solver and standard GRASP variants. The results demonstrate that the GRASP + 2-Opt approach strikes a favorable balance between optimality and execution time. This work provides practical insights for applications such as industrial inspection and environmental monitoring, where autonomous multi-robot coordination and energy constraints are essential for sustained and reliable operation.

Wireless power transfer (WPT) based on inductive coupling has gained increasing interest in scientific research, particularly in the field of implantable medical devices (IMDs). The efficiency of a WPT system primarily depends on the operating frequency, the quality factors (Q1 and Q2) of the coils, and significantly on the coupling coefficient (k) between the two coils forming the link. This efficiency can be enhanced by optimizing the geometric parameters of the coils or adjusting the operating frequency. In this paper, we characterize the coil parameters constituting the inductive link to demonstrate the impact of their variation on the energy transfer efficiency. Furthermore, we address the design considerations of the coils for wireless power transfer applications in IMDs.

This paper presents a compact multi-physical model for lithium-ion batteries used in electric vehicles (EVs). The proposed framework combines an equivalent circuit for electrical behavior, a lumped thermal model, and a semi-empirical ageing model separating calendar and cycling effects. The model is parameterized and validated using experimental data from Kokam NMC-based High Energy (HE) and High Power (HP) cells. Results show high accuracy in voltage, temperature, and degradation prediction, with relative errors below 2%. The model structure offers a good balance between accuracy and computational cost, making it suitable for EV performance evaluation, ageing prediction, and battery management system integration.

Outpatient clinics are integral to healthcare, offering vital services without the need for hospital admission. However, appointment scheduling in these settings remains challenging due to uncertainties such as patient no-shows and variable service times. This study proposes a data-driven approach to minimize physician idle time and patient waiting time by analyzing an eight-year (2016-2023) dataset from primary and specialized care for American veterans. Adopting the Cross Industry Standard Process for Data Mining (CRISP-DM) methodology, four predictive models: Random Forest, Extreme Gradient Boosting (XGBoost), Adaptive Boosting (AdaBoost), and Artificial Neural Networks (ANN), were developed for both classification (noshow) and regression (service time) tasks. The ANN model demonstrated superior predictive performance in both domains. The key predictors of no-shows included waiting time, type of care and care provider, while type of care, care provider, and veteran ZIP code were the most influential in forecasting service time. These findings highlight the potential of machine learning to improve appointment scheduling in outpatient clinics.

This paper proposes a Long Short-Term Memory (LSTM) based approach for accurate State Of Charge (SOC) and State Of Health (SOH) estimation in lithium-ion batteries, which is critical for improving the safety and longevity of electric vehicles. The nonlinear dynamics of batteries, influenced by factors such as temperature, voltage, current, and SOC, poses significant challenges to traditional estimation methods. Using the Long-Term Degradation dataset, our LSTM model captures temporal dependencies and complex electrochemical interactions to predict SOC and SOH under varying operating conditions. The experimental results demonstrate robust performance, with mean squared errors as low as 5.3121×10−5for the estimation of SOC and 5.572×10−5for the estimation of SOH for different current profiles. The proposed framework provides a scalable solution for real-time battery management systems, reducing the reliance on manual feature extraction and enabling generalization across different battery technologies.

This paper proposes a novel Puma Optimizer-based PID control (PO-PID) strategy for quadruped robots and presents a comprehensive comparison against two established approaches: Harris Hawks Optimization-PID (HHO-PID) and Slime Mould Algorithm-PID (SMA-PID). While conventional PID tuning methods struggle with high-dimensional search spaces and local optima, the PO algorithm leverages adaptive exploration-exploitation mechanisms inspired by puma hunting behavior to achieve globally optimal PID gains. Detailed simulations evaluate the three controllers in terms of trajectory accuracy and control effort. The PO-PID controller achieved the lowest joint trajectory error and significantly outperformed HHO-PID and SMA-PID in Cartesian accuracy while maintaining a lower or comparable torque demand. Additionally, PO-PID demonstrated faster convergence and more consistent optimization, indicating its robustness and reliability. These findings validate the PO-PID controller as a powerful and efficient solution for enhancing quadruped robot locomotion.

Accurate real-time estimation of battery State of Charge (SOC) and State of Health (SOH) is essential for electric vehicle (EV) performance and safety. We propose a hybrid framework combining deep learning with physics insights using LFP cell data. Our core model is a series CNN-TCN-DNN network trained on raw voltage/current signals; replacing measured temperature with cumulative charge ($\int I d t$) as an input improves accuracy. The model is tested under input noise and bias to ensure robustness. For example, it consistently achieves SOC errors below 2% MAPE and more accurate SOH tracking, enabling more reliable range prediction.

Digital Twin (DT) technology has recently gained significant attention in both industry and academia. By improving efficiency, reducing costs, and enabling predictive maintenance, DT plays a crucial role in control, diagnostics, prediction, and optimization. However, at the definition level, the rapid growth of research has led to a variety of definitions and inconsistent terminology, underscoring the need for a clearer and more unified understanding of the DT concept. At the design stage, two major challenges remain. The first is to develop models that are both accurate and dynamic, capable of supporting optimization and real-time supervision, including control and diagnosis. The second lies in ensuring a reliable link to the physical system through online identification, which is particularly challenging due to the non-stationary nature of real-world processes. This study addresses the definition level and the first challenge at the design level by offering a comprehensive overview of the DT concept, including its definition, development phases, and an illustration of its design carried out in a two-tank system. The research consists of using the Bond Graph (BG) tool as a multi-physical graphical tool for designing a DT not only for dynamic modeling but also for online supervision, including diagnosis and recovery decisions. The results first show the effectiveness of BG in constructing the DT using 20Sim software for dynamic modeling and fault detection and isolation. This BG tool enables DT to monitor the state of the system and provides the health indicators necessary to reconfigure the system controller.

This paper investigates the implementation of feedback linearization for the control of a planar two-link robotic manipulator, with the objective of achieving precise and robust trajectory tracking. The nonlinear dynamic model of the manipulator is systematically derived using the Euler-Lagrange formalism, capturing the full complexity of its coupled dynamics. Based on this model, a nonlinear control law is designed to cancel the system’s intrinsic nonlinearities and impose a desired linear behavior.The proposed controller guarantees rapid convergence of tracking errors, stability of the closed-loop system, and robustness against modeling inaccuracies and dynamic interactions between the links. Through numerical simulations, the effectiveness of the approach is demonstrated, showing high tracking fidelity even in the presence of parameter uncertainties and external disturbances.

This paper presents the design of a Terminal Sliding Mode Controller (TSMC) for trajectory tracking of an upper limb rehabilitation exoskeleton. Addressing dynamic uncertainties and disturbances inherent in human-robot interaction, the TSMC ensures finite-time convergence of tracking errors, providing fast, precise, and robust performance. The nonlinear dynamic model of the exoskeleton is derived using the Euler-Lagrange formulation, focusing on planar shoulder and elbow motions. Simulation results demonstrate the effectiveness of the proposed controller in significantly reducing tracking errors in joint angles and velocities, while effectively mitigating the chattering phenomenon. This work lays the foundation for robust control solutions tailored to the demanding requirements of functional robotic rehabilitation.

The increasing adoption of legged robots for applications such as search and rescue, environmental monitoring, and planetary exploration presents unique challenges in navigating diverse and complex terrains. Accurate terrain classification is crucial for adaptive locomotion and reliable performance, yet existing approaches often suffer from limited generalizability and accuracy due to the inherent variability of terrains and sensor noise. This study proposes a multi-terrain classification framework leveraging the HistGradient Boosting (HGB) machine learning technique to address these challenges. The system utilizes a force sensor and IMU data from a quadruped robot to extract meaningful features for robust classification. The HistGradient Boosting model achieved the highest accuracy of 0.9931, highlighting its superior classification performance compared to k-Nearest Neighbors (KNN), Support Vector Machine (SVM), and Random Forest (RF) with accuracy of 0.9410, 0.9878, and 0.9729, demonstrating its effectiveness in handling the complexities of multi-terrain environments.

Data analysis and classification can be affected by the availability of missing data in datasets. To deal with missing data, either deletion- or imputation-based methods are used that result in the reduction of data records or imputation of incorrect predicted value. Quality of imputed data can be significantly improved if missing values are generated accurately using machine learning algorithms. In this work, an analysis of machine learning-based algorithms for missing data imputation is performed. The K-nearest neighbors (KNN) and Sequential KNN (SKNN) algorithms are used to impute missing values in datasets using machine learning. Missing values handled using a statistical deletion approach (List-wise Deletion (LD)) and ML-based imputation methods (KNN and SKNN) are then tested and compared using different ML classifiers (Support Vector Machine and Decision Tree) to evaluate the effectiveness of imputed data. The used algorithms are compared in terms of accuracy, and results yielded that the ML-based imputation method (SKNN) outperforms the LD-based approach and KNN method in terms of the effectiveness of handling missing data in almost every dataset with both classification algorithms (SVM and DT).

A Smart-Grid (SG) represents an advanced electrical network designed for intelligent and efficient management across its entire infrastructure, facilitating seamless communication and coordination among its interconnected components, including IoT-enabled devices, and leveraging real-time data transfer mechanisms. This paper focuses on examining the local level within the SG, primarily tasked with monitoring energy consumption. To gain a comprehensive understanding of these concepts and the strategies employed for effective energy management, an energy management system has been devised. The primary objective of this system is twofold: first, to minimize the overall energy consumption within the SG, ensuring it remains within or below the received energy quantity; and second, to optimize the utility of appliances while ensuring they do not surpass the total energy capacity supplied by the Photovoltaic Panels (PVPs). To achieve this, we employ the Knapsack algorithm, wherein the locally produced energy serves as the algorithm's capacity, and the devices seeking energy consumption are treated as objects. Each device's weight, fixed consumption, and utility values are considered as attributes defining these objects, aiding in the algorithm's decision-making process. Through this approach, we aim to develop a robust energy management framework capable of efficiently allocating resources while maximizing overall utility within the SG ecosystem.

The power-current relationship of a photovoltaic generator (GPV) is non-linear and contingent upon environmental factors. Nonetheless, achieving the highest possible power output from a GPV can only occur at a specific point along the characteristic curve. The development of Maximum Power Point Tracking (MPPT) techniques is fundamental to designing solar systems that optimize power generation. The Adaptive Fuzzy Neural Inference System (ANFIS) is one of the most effective ways to attain the maximum power point (MPP) in PV systems due to its prompt response time and minimal oscillations. Furthermore, sliding mode control (SMC) is a popular method for managing linear and nonlinear systems because of its robustness. The primary objective of this research is to introduce a novel approach that utilizes a combination of ANFIS and Sliding Mode Control (ANFIS-SMC) to safeguard the PV system against uncertain conditions and achieve the optimum power point. The simulation outcomes indicate that the ANFIS-SMC controller delivers a precise, swift, and resilient response, compared to other algorithms like perturb and observe (P&O).

The interest of research and automotive industries is concentrating progressively on the Electric Vehicles (EV) which are a global transportation development currently in order to achieve considerable carbon emission reductions. Electric batteries are the essential component of the EV and precise remaining useful life prediction is the main to ensure its reliability. As a result, the inside workings of these battery systems must be fully included. There is presently no precise model for predicting an EV battery's aging. This paper presents an intelligent method for estimating the State Of Charge (SOC) of the battery.

Artificial intelligence methods have aided the advancement of several disciplines of science and technology. Furthermore, they have had a significant influence on smart grid management. One of the most significant information for optimal management in Smart Grid is the ability to predict its parameters: electricity consumption and meteorological factors. It is mostly utilized to develop improved control ways for building appliances (such as lighting and heating/cooling systems). Several methodologies for load and weather data characterization prediction have recently been proposed. The work discussed in this paper is aimed at the development and the comparison of three artificial intelligence forecasting approaches used to manage the Smart Grid by integrating load and climate data predictions. To achieve this objective, we primarily looked on the predicting accuracy of some artificial intelligence methodologies: a neural network, a neuro-fuzzy and a deep learning prediction algorithms are applied and compared to forecast the smart grid parameters (temperature, solar radiation, wind speed and the energy consumption). The simulation results are checked based on real database of wind speed, temperature, solar radiation, and consumption data. The findings of the simulation give us an idea about the most appropriate and performant algorithm to use in this aim.

Short-term load forecasting is a topic of considerable interest as it is of major importance for specifying and managing power resources and needs. In the literature, Deep Neural Networks have been successfully recently applied in load forecasting using single modalities as an improvement to traditional Artificial Neural Networks (ANN). In this paper, the main objective is to tackle the load forecasting problem with the intention of enhancing the prediction performance by combining and testing different multi-modal deep learning approaches and architectures in order to process and relate information from multiple modalities. The benefits of using multi-modality instead of one modality, applied to the time series modeling problem of short-term load forecasting is therefore investigated. The models are trained and evaluated using the hourly temperature and electrical consumption in addition to auto-regressive variables as the first modality. The day type characteristic, such as: weekends, week days, bank holidays, religious holidays etc., may be considered as a second modality. The approach liability is tested by comparing the empirical results of different Deep architectures namely: Stacked Denoising Auto- Encoders (SDAEs) and Convolutional Neural Network (CNN), with multiple and single modalities where processing the same task of one day ahead load forecasting.

Renewable energy or alternative energy is extracted through renewable resources. These are considered as an alternative from conventional fossil fuel-based sources because conventional energy sources are depleting rapidly and raised concerns over increasing environmental impacts. Among many renewable sources, solar energy has a substantial part to meet the increased energy demand with reduced environmental effects. Solar irradiance and temperature are key factors upon which photovoltaic (PV) power generation depends but its optimum operating point gets affected by variation in the above-mentioned environmental factors. Finding the optimum operating point is a challenge due to the nonlinear solar behavior and varying nature of environmental conditions. To overcome these challenges, maximum power point (MPP) searching algorithms are exploited to get optimum power from the PV energy system. Maximum power point tracking (MPPT) behavior is different for various weather conditions, for instance, partial shading (PS), and uniform irradiance (UI) conditions. Numerous MPPT methods came to be used to find the optimum power. This work deals with the development of a novel technique for MPP finding of a PV system on the basis of the Water Cycle Algorithm (WCA) under PS conditions. It turns out to be good in terms of exploration and exploitation. Thus, it has the capability to avoid getting stuck in local minima (LM) and to find the global maxima (GM). The performance of the WCA technique is examined on four different types of P-V patterns for UI, PS and fast changing environmental conditions through MATLAB simulation and experimental setup. The findings of WCA are compared with the previous well-known soft computing methods such as PSO, ACS, DFO, and conventional method P&O to evaluate performance. The outcomes reveal that the WCA algorithm overtakes P&O from the perspective of robustness, accuracy, efficiency, and stability, as well as PSO in respect of converging speed and efficiency.

This issue of Interdisciplinary Methods and Approaches for Cybernetics and Systems Modeling includes a collection of extended versions of best presented papers in CoDIT 2020 conference. The aim con...

Safety and risks prevention in numerous industrial domains require a systematic monitoring of areas with high-risk level. The surveillance tasks can be assessed by smart sensor systems. This work is devoted to configure a patrol of mobile robots associated with sets of sensors in order to perform a given sequence of surveillance tasks at minimal cost. The environment to be monitored as the trajectories of the robots are modelled with Petri nets. The objective is reformulated as an initial marking optimization problem in the Petri net framework. A real case is considered with the port area in Le Havre City.

This paper deals with the problem of the optimal sizing in the PVP/Battery system. To achieve this aim, an ANFIS estimation algorithm has been developed in order to estimate a data base of instantaneous photovoltaic power. The estimated instantaneous PV power has been used in an optimal algorithm for sizing a PVP / Battery power station to supply a 1.5 Kw AC load.The simulation of the proposal sizing system has been implemented in Matlab. The results of the simulation give a good performance of our proposal sizing system.

This paper deals with the problem of estimating the Minimum Initial Marking (MIM) of Labeled Petri Nets (L-PN). By the observation of a sequence of labels, we determine the set of possible MIMs related to a given L-PN through an approach based on GRASP (Greedy Randomized Adaptive Search Procedure) inspired method – GMIM. The objective is to get the maximum of feasible MIMs by exploring the search space and giving best solutions for real time cyber systems in short time. We consider four basic assumptions during the reasoning: (i) the L-PN structure is known; (ii) for each transition of L-PN, a label is associated, (iii) the label sequence is known, and (iv) all transitions of L-PN are observable. We show the validity and efficiency of our approach by applying the proposed GMIM metaheuristic to two validation examples: Initialization of two parallel machines (example widely cited in literature) and resources allocation in a monitoring problem via mobile robot network.

This issue of New Methods and Approaches in Decision and Control of Intelligent and Cyber Systems contains a collection of extended versions of best presented papers in CoDIT 2019 conference and ad...

This work develops a technique for estimating the minimum initial marking (an initial marking with the minimum total token number) of a real-time system modelled by a P-Time labelled Petri net system under partial observation. Indeed, the set of events is partitioned into a set of observable events, which can be detected by an external agent and a set of unobservable ones. Furthermore, some observable events can be associated to the same observation, i.e., they produce the same output signal and are called undistinguishable. In addition, the obtained marking must be consistent with a given label occurrence vector (LOV). The proposed method is based on an iterative procedure combined with a schedulability analysis technique for particular behaviors of the studied system.

International audience

This paper is devoted to the problem of optimal investment portfolio design on the base of mathematical modeling tools and method of genetic algorithms. The purpose relates to investing the funds into financial assets such that certain requirements regarding the expected profits and possible losses would be reached. The main result is developing a state-space dynamical model of portfolio management and applying a genetic algorithm in order to obtain the optimal solution.

The modelling of the chemical reactor behavior is always difficult task due to the absence of more detailed knowledge about the considered chemical transformation. We treat in this context the Nickel-Metal Hydride (Ni-MH) battery system. In this paper, an improved fuzzy c-regression model is proposed in order to develop a Ni-MH battery model on which a modified distance is introduced in the objective function of fuzzy c-regression model algorithm in the purpose of taking into account the outliers. After that the obtained model is employed to estimate the Ni-MH battery's State Of Heath (SOH). The experimental results indicate that the proposed method can be ensured an acceptable accuracy of the SOH estimation for Ni-MH battery system.

This paper addresses an improved Genetic Algorithm (GA) combined with local search technique to solve the dynamic job shop scheduling problem (DJSSP) with new job arrivals and change in processing time. The objective function is the minimization of the makespan known to be one of the performance criterion used to optimize manufacturing system requirements. To enhance the scheduling process, a rescheduling strategy is used to solve dynamic disturbances. Various problems including the number of jobs, the number of machines and the number of new job arrivals are compared with a collection of state of the art Dispatching Rules(DRs) and other metrics. Obtained results are satisfactory for rescheduling of new job arrivals, change in processing time and makespan minimization.

Up to now, the majority of researches on scheduling assume the difficulty of scheduling the job shop manufacturing system, especially the Dynamic Job Shop Scheduling Problem (DJSSP) which is the main purpose of our contribution. Looking to minimize the makespan value (Cmax), setup times and precedence constraints are considered. During the did work, we have proposed an inspired Genetic Algorithm(GA) approach based on the genetic operators to solve the DJSSP. The main task of the proposed method is to show how efficiently the system will schedule dynamically the new jobs. Eventually, the DJSSP based GA approach is proven to be successfully solved through experimental results. Hence, this later make it possible to generate minimal makespan values compared to the well known priority dispatching rules.

We develop an optimization algorithm in order to improve the nonlinear optimal robust controller for diesel engine which equipped by Exhaust Gas Recirculation and Variable Geometry Turbocharger systems. For this goal we present a new way to search the optimal shape of control Lyapunov function which permits to calculate the parameters to achieve stabilization and optimal control criteria. The idea is to verify different form of Lyapunov functions in order to desire better performance.

This paper studies the problem of the parameter identification based on fuzzy c-regression models for nonlinear systems. The novel procedure combines the possibilistic c-means procedure with fuzzy c-regression models (FCRM) in order to reduce the effects of noisy data. In comparison to the existing algorithms in the literature, the proposed method utilizes a generalized objective function that reduces the errors of partitioning data sets contaminated by noise and as a consequence an accurate model is obtained. The results of this study demonstrate the effectiveness of proposed method compared with other extended versions of FCRM algorithm.

This paper deals with the Job-shop scheduling problem. We propose to solve this problem by exploiting the Particle Swarm Optimization Global Velocity (PSOVG) algorithm. The PSOVG by its nature focus on the global optimum within a given set of solutions. In this paper a solution is PSO particle, it consists in a possible scheduling solution for the given problem. The PSO-VG-JSSP is a PSO-VGO with a constraints control policy embedded in, allowing to detect and remove non admissible solutions. In this paper the particle representing a non admissible solution is simply removed and replaced by a new random particle.

There is a growing interest in intelligent maintenance in which supervision and its main component - the diagnosis - occupy a fundamental place leaning more towards dynamic monitoring applications and therefore to prognosis. In this sense, the use of artificial intelligence techniques has been under much research. The ability to monitor without modeling the system in an environment or if its subject to modifications and permanent reconfigurations, is of obvious interest for manufacturers. In this paper, we plan to improve the performance of a chemical reactor. We wish to add changes to the operation of the reactor in order to make it more secure by: a Adding a chain that measures temperature, alerting the operator if the threshold is exceeded and requesting the operator to open a valve (or PLC) to open a deluge system that lowers the temperature.

This paper seeks to establish within a formal framework for modeling the robust control laws of manufacturing systems to temporal and non-temporal constraints. The goal is to conclude the type from the production system robustness. A methodology of robust control construction generating the margins of passive and active robustness is elaborated. The redundancy of the robustness of the elementary parameters between passive and active is used. Indeed, local models are developed; each of them modeling resource robustness. Then, a synchronized assembly of these models enabled us to constitute our final model. The Intervals Constrained Petri Nets (ICPN) tool is used. To this goal, a whole of definitions, lemmas and theorems are committed and affirmed per applicable examples.