Search

Acute lower respiratory tract infections pose a significant health challenge, affecting over 15% of Saudi Arabia’s population annually. Research over the past three decades highlights that viruses are the primary cause of these infections. The epidemiology of respiratory viruses in Saudi Arabia is influenced by the movement of foreign workers and the mass gatherings of Muslims in Mecca during the Hajj and Umrah seasons. This study exploring the impact of different vaccination rates across various age groups during the 2019/2020 influenza season in Saudi Arabia. We used a two-phased multi-objective optimization approach with a Controlled Elitism Non-Dominated Sorting Genetic Algorithm applied to an age-structured SEIR model (Susceptible-Exposed–Infectious-Recovered) to design vaccination policies tailored to age-specific infection rates. The research divides the population into five age categories—infants and children, adolescents, young adults, middle-aged adults, and seniors—across five regions: Eastern, Central, Northern, Western, and Southern Saudi Arabia. The study compares three vaccination strategies: no vaccination, national vaccination, and a two-phased multi-objective optimization approach under various pandemic scenarios. It evaluates the outcomes regarding infection rates, mortality, and morbidity. The findings highlight how demographic factors significantly influence disease transmission and vaccine distribution. The optimized vaccination strategy outperformed the national vaccination approach, leading to more significant reductions in infection rates (by 0.01% to 0.45%), morbidity (by 0.01% to 0.04%), and mortality (by 0.001% to 0.013%).

Managing the integration of Electronic Toll Collection (ETC) and Manual Toll Collection (MTC) systems is complex, especially during peak hours due to high traffic volumes. While ETC is more efficient, an inappropriate proportion of ETC lanes can increase congestion, making optimal ETC lane selection crucial. This paper proposes a novel framework combining deep learning and multi-objective optimization to improve toll plaza efficiency. A Gated Recurrent Unit (GRU) model with Optuna hyperparameter tuning predicts average queue lengths for various ETC lane proportions. The predictions guide the identification of the ideal ETC lane configuration to minimize queue lengths and balance traffic flow between MTC and ETC lanes. The framework dynamically adjusts ETC lane proportions based on real-time traffic data to ensure optimal lane allocations during peak hours. A multi-objective optimization function is applied to balance key objectives such as minimizing queue length, reducing queue time, lowering operational costs, and optimizing ETC utilization. Simulations with real-world data from high-traffic toll plazas demonstrate the framework’s effectiveness, reducing queue lengths by up to 95.03% during peak hours, decreasing operational costs by 28.72%, and improving overall toll plaza performance. The results are visualized with graphs showing predicted average queue lengths across different ETC lane proportions, aiding stakeholders in understanding the relationship between ETC adoption and congestion. The framework provides insights into operational costs and resource allocation, enhancing financial and operational planning. Validated through extensive simulations with real-time data, this research advances empirical studies with thorough model validation and introduces innovative visualization techniques for toll plaza management.

Dynamic pricing is a critical strategy in e-commerce, enabling merchants to optimize sales profit while adapting to varying market conditions. However, existing approaches often fall short in balancing commercial objectives with the preservation of cultural heritage, particularly in niche markets like Hanfu apparel. To address this challenge, we developed a dynamic pricing simulation environment based on a Markov Decision Process (MDP) and introduced a novel multi-objective hybrid particle swarm optimization algorithm combined with Deep Deterministic Policy Gradient (DDPG), referred to as MOHPSO-DDPG. By applying principal component analysis (PCA) to consumer preference data and constructing utility functions and Logit choice models, we accurately simulated consumer behavior. MOHPSO-DDPG, Multi-Objective Particle Swarm Optimization (MOPSO), and Multi-Objective Hybrid Particle Swarm Optimization (MOHPSO) were each deployed to interact with the environment to explore the Pareto front of pricing decisions. Experimental results demonstrate that MOHPSO-DDPG significantly outperforms other algorithms in terms of solution diversity and convergence efficiency. After 3,000 iterations, its Generation Distance (GD) reached 0.023 and the Diversity Metric $\Delta $ was 0.594, whereas GD and Diversity Metric $\Delta $ values remained larger for MOPSO and MOHPSO. Moreover, MOHPSO-DDPG continued to maintain a leading position in later iterations, underscoring its superiority in identifying comprehensive and near-optimal Pareto fronts. These findings validate that MOHPSO-DDPG provides an efficient multi-objective dynamic pricing decision-making framework for the Hanfu market, effectively balancing profit maximization with the demands of cultural heritage preservation.

Edge computing has emerged as a promising solution for delivering services that demand low latency, high bandwidth, and stringent privacy requirements in numerous data- and compute-intensive applications, such as those in Smart Cities. Heterogeneity in edge computing resources and diverse application requirements demand adaptive optimization techniques, such as service placement, to conform to changing conditions. A service placement model must optimize the selection of edge nodes for deploying and executing services, thereby improving application QoS and maximizing resource utilization. Numerous optimization techniques for adaptive service placement problem have been proposed in the recent past. However, in most cases, the results have been evaluated in limited scenarios. This paper presents a comprehensive comparative study evaluating representative optimization algorithms applied to the problem of dynamic service placement across various application scenarios. The study covers nature-inspired approaches, including both meta-heuristics and reinforcement learning. Our experimental findings offer valuable insights into the strengths and weaknesses of the selected nature-inspired algorithms for service placement optimization, evaluated for applications with different QoS requirements. In our analysis, the Genetic Algorithm shows superior performance in achieving lower average distance and the average number of servers selected. Particle Swarm Optimization excels in minimizing average waiting time and placement decision time. The Artificial Bee Colony maintains low average latency, whereas the RL Proximal Policy Optimization demonstrates superior performance in terms of balancing the utilization of network resources.

Salient-pole synchronous generator with rectifier load has been widely used in aviation power system for its advantages of high DC power supply quality and strong fault tolerance. Damper winding directly affects the steady-state and transient performance of salient-pole synchronous generator. To study the impact of damper winding, a prediction and optimization method based on the combination of neural network approximation model and multi-objective optimization is proposed in this paper. Taking a salient-pole synchronous generator with uncontrolled rectifier load as the research object, the OLHD is carried out with five parameters of the number, radius, slot pitch angle, slot width and slot height of damper windings as design variables. The sensitivity of different design variables to the commutated reactance, THD of AC side voltage, DC side output voltage and its ripple factor, output torque are analyzed. Based on the OLHD results, EBF neural network approximation model is established. With the goal of optimizing the commutated reactance, THD, ripple factor, and ensuring DC side voltage and torque, the AMGA is used to optimize the design of the damper winding structure. Finally, finite element simulation is conducted to verify the feasibility and accuracy of the method.

Fuel cell systems (FCSs) are based on a number of components whose electrochemical and physical interactions during operation are not fully understood either by empirical experience or by extensive modeling. The related FCS complexity limits technological applications and success. To exploit the potential of FCS-based deployments, this paper presents a complete and consistent toolchain that enables their model-based understanding and multi-objective optimization (MOO) regarding design and operation. The added value of the generic process is demonstrated by a MOO of an automotive FCS power train with fuel consumption and fuel cell aging as conflicting targets. For this purpose, an open-source object-oriented FCS model was extended to include cell degradation effects at stack level and power train components at system level. The resulting model covers the essential interactions at system level and enables the MOO of both the parametric design of the system components and their operational management. Highly efficient algorithms are used for MOO in view of the model’s considerable computing effort. In the context of the selected target criteria, this enables a comprehensive analysis of the technological potential in the form of target variable conflicts as well as the sensitivities of these target variables in their dependence on design and operational degrees of freedom. The results, which can be transferred to other conflicting target variables and their simultaneous optimization, demonstrate the fundamental superiority of a Pareto-optimal system design and operational management over previous development methods.

Fleet conversion using electric buses stipulates a comprehensive optimization of the fleet mix, size, and charging station decisions while minimizing the total cost of ownership and carbon dioxide emissions. Therefore, this study focuses on the optimal Electric Bus Fleet Conversion Planning in the urban transportation sector to reduce emissions under different incentives, carbon tax, retrofit bus usage, charging station type, and scheduling scenarios. It proposes a novel multi-objective mathematical model, a unique contribution of this research, to provide decision aid through this conversion process. The model minimizes the total cost of ownership and emissions to meet the Paris Agreement’s targets. A case study based on the Metrobus fleet of Istanbul was performed to exemplify the usage of the model. This is the first study to simultaneously assume all the critical aspects of the fleet transition problem: electric bus purchasing, retrofit bus usage, charging infrastructure development, carbon taxes, incentives, and charge scheduling. This study’s findings can help decision-makers develop optimal fleet conversion plans that balance economic and environmental objectives under different conditions. The findings imply that the fleet transition using new electric buses incurs very high capital costs, but retrofitted buses can make this transition more economically viable, offering potential economic benefits. Charge scheduling affects the sizing of and investment in charging stations. The government should provide bus purchasing and charging station incentives to support this conversion process. Besides, renewable resource usage in electricity generation plays a pivotal role in carbon dioxide emission reduction.

Hybrid Electrical Vehicles (HEVs) have three magnetic field sources: armature winding currents; permanent magnets; and field winding current. To initiate the investigation, the magnetic field distributions due to those sources are obtained. First, Maxwell’s equations need to be solved. Then, the machine is analyzed and the operational indices such as unbalanced magnetic force (UMF) in x- and y direction and instantaneous torque are calculated by using the magnetic flux density distribution. Next, various magnetization patterns, which are presented based on Fourier series, are considered to investigate their influences on the performance of the machine. It is assumed the volume of the machine is determined thanks to thermal considerations. As a result, it is constant so it is maximized average torque instead of torque per volume ratio. After that, the average and ripple of the instantaneous torque are expressed in terms of pole arc to pole pitch ratio by using curve fitting. Finally, the optimal value of average torque and torque ripple are also computed by multi-objective optimization algorithms.

This paper presents a multi-objective model predictive control-based energy management system. The proposed scheme design aims to perform the power sharing among available sources, minimizing the diesel engine fuel consumption, and regulating the battery state-of-charge, while considering power source constraints. The proposed MPC-based EMS is designed as a high-level controller whose control actions are the references for the low-level controllers for the diesel generator set. A Pareto front analysis is appealed to get a better understanding of the conflicts between the MPC’s objectives. The MPC scheme is designed following a weighting sum scalarization, where Pareto front analysis is crucial for the tuning of the MPC weights, resulting in the proposal of a weight schedule to achieve the desired trade-off performance. Moreover, six typical agricultural working cycles have been considered, where, as expected, a characteristic Pareto front is obtained for each cycle power profile.

Technology-driven startups often experience fluctuations in business volume and their delivery peak exceeds the conventional functional department’s human resource staffing. Resource allocation is crucial for ensuring timely completion of project deliverables within resource constraints. This study tackles the scheduling challenge under the condition of composite skills, with the goal of efficiently allocating diverse human resources while adhering to constraints such as external labor coordination and skill compatibility. Initially, an integer programming model was formulated to minimize the overall project duration and equalize the workload among employees. Subsequently, an enhanced fast and elitist non-dominated sorting genetic algorithm (NSGA-II) was developed by integrating heuristic-based population initialization and adaptive genetic strategies. A case study was then developed based on the R program of technology-driven company S. The allocation model was solved using both the classic and improved versions of the algorithm. The outcomes indicate that the fitness of the final generation significantly exceeds that of the initial generation, and the refined algorithm outperforms the conventional one in reducing the project duration and enhancing balance, with the potential to reduce project duration by 9% and enhance the balance by 40%. Cluster analysis and statistical methods were applied to extract three pivotal traits of the optimal allocation scheme, providing a scientific reference and decision-making foundation for managerial resource-allocation strategies.

Advanced airborne equipment liquid cooling system presents the development trend of multiple pumps in parallel, centralised heat dissipation and high efficiency circulation, but this type of system has the characteristics of many variable parameters and strong non-linearity, which makes the system operation and control become extraordinarily complex, posing a serious challenge to the system performance. Although some progress has been made in the current optimization of liquid cooling system performance, there are still gaps such as insufficient consideration of the multi-objective optimization problem and limited optimization effect. To fill this research gap, this study takes minimum specific energy consumption, minimum flow deviation and minimum impeller load as the optimization objectives, establishment of multi-objective optimization model for liquid-cooled power unit, the optimal solution set of the Pareto model is solved using the multi-objective grey wolf optimization algorithm (MOGWO), and the optimal solution set is substituted for the optimal solution, and the evaluation indexes are designed in combination with the actual working conditions to obtain the optimal schedule. By example analysis, the optimized system under variable operating conditions, the impeller load and specific energy consumption decreased by 16.88% and 3.64% respectively, while the rise in flow deviation was controlled within 0.76%. The newly established optimisation model effectively mitigates potential safety hazards during system operation. Finally, by combining all the indicators, three scheduling options are provided under variable operating conditions, focusing on reliability and economy, overall balance and high-precision flow, respectively, in order to adapt to the diversified requirements of different operating environments.

In cloud computing, multiple workflow scheduling is important to optimize resource allocation and utilization for concurrent execution of diverse workflows across different applications. While previous research has focused on clustering-based resource allocation to reduce communication overheads by grouping tasks, it often overlooks the significance of task execution ordering, limiting overall performance optimization. To address this limitation, we propose two genetic-based approaches, considering task and cluster-level characteristics, to introduce novel ordering techniques for multi-workflow scheduling under cluster-based resource allocation. By comparing two granularity ordering methods, we offer valuable insights for efficient task management in multi-workflow environments. Our experiments demonstrate that the proposed approaches, especially the task granularity-based ordering method, outperform existing primary clustering methods, particularly for scenarios involving a large number of workflows or highly parallel workflows.

With the expansion of communication technologies and their impact on trade patterns, e-commerce strategies have also undergone significant changes. Adapting to these changes necessitates the utilization of artificial intelligence techniques to automate a wide range of processes and further develop e-commerce. This article employs a combination of machine learning techniques and multi-objective optimization to enhance the supply chain performance in Cross-Border E-Commerce (CBEC). To achieve this, a framework for intelligent CBEC based on Internet of Things (IoT) technology is proposed. By deploying machine learning models within this framework, efforts are made to improve supply chain performance through demand volume prediction. The predictive model used in the proposed method is an ensemble system based on Adaptive Neuro-Fuzzy Inference System (ANFIS), which employs weighted averaging to predict demand volume for each retail unit. The configuration of this prediction model is done at two levels, utilizing Particle Swarm Optimization. At the first level, the hyperparameters of each ANFIS model are optimized, and at the second level, the weight values of each learning component are optimized using this algorithm. The performance of this predictive model in enhancing the CBEC supply chain structure is evaluated using real-world data. Based on the results, the proposed predictive model achieves an average absolute error of 2.54 in demand volume prediction, showcasing a minimum reduction of 8.58% compared to previous research. Moreover, the improvement in supply chain performance through this model will lead to reduced delays and increased efficiency in CBEC, demonstrating the effectiveness of the proposed model.

This paper tackles the complex problem of optimizing resource configuration for microservice management in heterogeneous cloud environments. To address this challenge, an enhanced framework, the multi-objective microservice allocation (MOMA) algorithm, is developed to formulate the efficient resource management of cloud microservice resources as a constrained optimization problem, guided by resource utilization and network communication overhead, which are two important factors in microservice resource allocation. The proposed framework simplifies the deployment of cloud services and streamlines workload monitoring and analysis within a diverse cloud system. A comprehensive comparison is made between the effectiveness of the proposed algorithm and existing algorithms on real-world datasets, with a focus on resource balancing, network overhead, and network reliability. Experimental results reveal that the proposed algorithm significantly enhances resource utilization, reduces network transmission overhead, and improves reliability.

With an increase in the complexity of society, solving multi-objective optimization problems (MOPs) has become crucial. In this study, we introduced a novel method called “quadratic unconstrained binary optimization based on the weighted normal” for solving MOPs using Ising machines, such as quantum annealing and digital annealer (DA), in the field of combinatorial optimization. The proposed method applies the penalty-based boundary intersection method to Ising machines under a setting limited to linear objective functions and maximizes the speed and performance of the DA, which is a quadratic unconstrained binary optimization-specific solver. We demonstrated the effectiveness of the proposed method by solving a real-world problem with a nonconvex shaped Pareto front (component combination problem). The results suggested that the proposed method could handle both convex- and nonconvex-shaped Pareto fronts, expanding the potential applications of Ising machines to solving complex MOPs. This development could significantly enhance decision-making processes, particularly in achieving sustainable development goals.

Computational biology research faces a formidable challenge in the detection of complexes within protein-protein interaction (PPI) networks, critical for unraveling cellular processes, predicting functions of uncharacterized proteins, and diagnosing diseases. While evolutionary algorithms (EAs), particularly state-of-the-art methods, often partition PPI networks based on graph properties or biological semantics, their resilience to noisy or missing interactions remains an underexplored territory. In this paper, we propose a groundbreaking heuristic operator, termed “strong neighbor-node migration”, specifically designed to elevate solution quality during the evolutionary process of our proposed EA. Through the application of EAs, we systematically evaluate the robustness of three single-objective models and two multi-objective models dedicated to addressing the complex detection problem. Our comprehensive assessment spans three well-known PPI networks, including two Saccharomyces cerevisiae datasets and the Human Protein Reference Database. To challenge the models further, we generate artificial networks by introducing varying percentages of noise to the original PPI networks. The experimental results showcase the superiority of the multi-objective model that incorporates our novel heuristic operator, demonstrating enhanced prediction accuracy compared to state-of-the-art models. Encouragingly, we advocate for the expansion of this research to integrate biological information, such as gene ontology. We propose the development of an objective function and heuristic operator based on this biological data, aiming to advance protein complex detection.

COVID-19 is a major public health emergency, the Chinese government adopts pandemic prevention and control measures such as home quarantine in order to block the strong contagiousness of the virus. In order to solve the problem of distribution difficulties due to the limited transportation resources brought by the pandemic and to guarantee the demand of fresh agricultural products for the isolated residents, this study takes the emergency supply as the main objective, and designs three distribution modes, namely, single distribution center, multi-vehicle model based on the classification of temperature zones, and multi-distribution center for isolated communities of different sizes, with the objectives of minimizing the number of vehicles in use, minimizing the average response time, and minimizing the risk of viral infections. A multi-objective optimization model was constructed as the objectives. Then the genetic algorithm is improved by adding destruction operator, repair operator and greedy operation, and the effectiveness of the algorithm is verified through empirical research, and the advantages of the algorithm in convergence speed are compared and analyzed. Finally, different distribution methods are suggested for the number of communities of different sizes. This study can provide a scientific basis for government departments to formulate an optimized decision-making plan for the emergency supply and distribution of fresh agricultural products under public health emergencies.

The time-dependent green vehicle routing problem with time windows is a further deepening of the research on vehicle routing problems with time windows. Its simultaneous consideration of vehicle transportation time, carbon emissions, and customer satisfaction under time-dependent variables makes it more challenging to solve than traditional vehicle routing problems. This work proposes a multi-objective optimization algorithm that combines the learnable crossover strategy and the adaptive search strategy based on reinforcement learning to overcome the local optima, poor convergence, and reduced variety of solutions that plague the multi-objective optimization algorithms when solving this problem. The proposed approach solves the problem in two stages: In the first stage, a hybrid initialization strategy is used to generate initial solutions with high quality and diversity, and crossover strategies are used to further explore the solution space and improve convergence by learning the characteristics of pareto solutions. In the second stage, the adaptive search is designed and used for learning and searching in the later stage of the algorithm. The experimental results show better solution quality obtained by the proposed approach, and the effectiveness and superiority of the proposed approach over existing methods in terms of solution convergence and diversity are demonstrated through experimental comparisons.

This paper addresses the challenging problem of Unit Commitment (UC), which involves the optimal scheduling of power generation units while adhering to numerous network operational constraints called security-constrained UC (SCUC). SCUC problem aims to minimize costs subject to turning on economically efficient generators and turning off expensive ones. These operational constraints include load balancing, voltage level at buses, minimum up and down time requirements, spinning reserve, and ramp up and down constraints. The SCUC problem, subject to these operational constraints, is a complex mixed-integer nonlinear problem (MINLP). There has been a growing interest in using evolutionary algorithms (EAs) to tackle large-scale multi-objective MINLP problems in recent two decades. This paper introduces a novel approach to address the SCUC problem, which is further complicated by including network constraints. They are pioneering the integration of single and multi-objective EAs to solve the SCUC problem while incorporating AC network constraints through hybrid binary and real coded operators. The development of an ensemble algorithm that combines mixed real and binary coded operators, extended by a bidirectional coevolutionary algorithm to tackle multi-objective SCUC problems. The paper implements a new formulation based on three conflicting objective functions: cost of energy supplied, startup and shutdown costs of generators, energy loss, and voltage deviation to solve the SCUC problem. Implementing a new formulation also addresses the solution of single and multi-objective SCUC problems using a combination of proposed technical and economic objective functions. The proposed algorithm is rigorously tested on a 10-unit IEEE RTS system and a 6-unit IEEE 30-bus test system, both with and without security constraints, addressing week-ahead and day-ahead SCUC scenarios. Simulation results show that the proposed algorithm finds near-global optimal solutions compared to other state-of-the-art EAs. Additionally, the research demonstrates the effectiveness of the proposed search operator by integrating it with a multi-objective coevolutionary algorithm driven by both feasible and infeasible solutions, showcasing superior performance in solving multi-objective SCUC problems. These results are compared with various recently implemented Multi-Objective Evolutionary Algorithms (MOEAs), demonstrating the superiority of

In the process of daily product development and design, optimizing the shape and external dimensions of the vehicle chassis truss structure while considering both weight and reliability indicators often involves interdisciplinary collaboration and inefficient communication, leading to repetitive mechanical labor and low efficiency. This is widely regarded as a difficult and challenging task. In order to improve the design quality of the chassis truss structure and enhance work efficiency, this paper proposes a novel lightweight multi-objective optimization design system for vehicle chassis trusses based on the ANFIS-SHAMODE-IWOA model. Firstly, an adaptive spiral search strategy is introduced based on the multi-objective hybrid metaheuristic algorithm (SHAMODE) which is based on successful history. Then, a new SHAMODE-IWOA algorithm is proposed. In order to estimate the reliability level of the chassis truss structure under different design parameter combinations, a new ANFIS-SHAMODE-IWOA model is constructed by using the proposed SHAMODE-IWOA algorithm to learn the ANFIS model. Finally, in order to obtain the optimal design parameter combination, multi-objective optimization based on minimum design quality and optimal reliability measurement functions is studied using the SHAMODE-IWOA algorithm. Experimental results show that SHAMODE-IWOA has leading global optimization capability on CEC2017 test set benchmark functions. Compared with other intelligent models, the ANFIS-SHAMODE-IWOA model has better performance in reliability coefficient estimation. At the same time, the SHAMODE-IWOA algorithm obtains a more optimal design parameter combination for the chassis truss, with improvements of 9.5%, 12.5%, and 15% in Mass, Bending, and Torsion indicators, respectively. Ultimately, the proposed ANFIS-SHAMODE-IWOA multi-objective optimization design system, as a novel intelligent model, can be used to evaluate the reliability of chassis truss structures, improve development and design efficiency, and obtain the best design parameter combination, which is beneficial to improving the level of green intelligent manufacturing design.

Traditionally, projects have been scheduled with a primary focus on optimizing economic bene-fits, often at the expense of environmental considerations. However, in today’s context, there is an increasing societal demand for organizations to incorporate sustainability practices into their project management policies. This research aims to address this pressing issue by developing a comprehensive mathematical model that takes into account two crucial objectives: project net present value (NPV) and greenhouse gas (GHG) emissions. This model introduces carbon emission constraints on a per-period basis, alongside the usual precedence and resource constraints. the problem represented by this model is proven to be NP-hard, which adds complexity to its solution. To tackle this challenge, two multi-objective metaheuristic algorithms, specifically the Modified Grey Wolf Optimizer (MGWO) and the Strength Pareto Evolutionary Algorithm 2 (SPEA2), have been employed to solve the model. Extensive comparative analyses of these me-ta heuristics across various multi-objective performance measures indicate that MGWO consistently outperforms SPEA2 in most scenarios. To illustrate the practical applicability of the pro-posed model, a real-world case study has been utilized. The results of this case study offer compelling insights, showing that project NPV can be significantly improved with only a marginal increase in GHG emissions. This finding highlights the model’s potential to provide pragmatic solutions and valuable trade-off options for project planners and practitioners striving to balance economic objectives with environmental sustainability.

This study delves deeply into the traffic light intersection control issue for connected autonomous vehicles (CAVs). First, we employed the vehicle dynamics simulation software, CarSim, to model the vehicle and utilized the intelligent driving simulation software, PreScan, to establish a road environment model. Subsequently, we designed a rule-based traffic light crossing controller (RB-TLCC) for CAVs in MATLAB/Simulink and implemented a co-simulation using CarSim and PreScan. Furthermore, we conducted driver-in-the-loop studies using the Logitech G27 driving simulator kit and compared the experimental results with those from the RB-TLCC. Our findings indicate that RB-TLCC enhances the regenerative braking energy by 23.5%, improves traffic efficiency by 17.9%, and increases driving smoothness by 50.7%. Additionally, based on Markov Chain theory, we proposed a multi-objective optimization model (MO-OM) for CAVs traffic light intersection crossing using the Proximal Policy Optimization algorithm (PPO). A comparison was conducted under two different operating conditions between RB-TLCC and dynamic programming algorithms (DP). The results indicated that the MO-OM proposed in this study exhibited the best comprehensive control performance among the three control methods. While adhering to traffic regulations, it enhances the recuperation of braking energy, efficiency of vehicle passage, and smoothness of travel at traffic light intersections. This study offers effective methodologies for enhancing the performance of CAVs in traffic light intersection control and holds significant reference value for the advancement of future intelligent transportation systems (ITS).

Owing to the high speed of the interior permanent magnet synchronous motor (IPMSM), the high-frequency howling noise of electromagnetic vibration is disruptive to passengers. The harmonics of the electromagnetic force serve as the primary source of vibration noise. In this work, a non-uniform air gap with a microgroove surface of the motor was proposed for optimizing the harmonics of the electromagnetic force. The research was based on a developing project of a 150 kW IPMSM for a battery electric vehicle (BEV). To find the optimal microgroove surface of the rotor, six variables were defined for the microgrooves through 9900 calculations using the Ansys Workbench. By finite element analysis (FEA) comparisons, the non-uniform air gap with microgroove surface significantly reduced the no-load back Electromotive Force (EMF) harmonics, the total distortion rate (THD) from 2.57% to 1.52%, and the torque ripple from 10.84% to 4.06%. A motor prototype with a non-uniform air gap was manufactured to prove the accuracy of the simulation. Under the whole open throttle (WOT) condition, the vibration acceleration of the 24th and 48th order was approximately 10 dB lower than the target vibration noise limit.

In light of the escalating global concerns surrounding climate change, the significance of sustainable development in the realm of logistics cannot be overstated. This study undertakes the imperative task of devising strategies aimed at mitigating carbon emissions, reducing logistics costs, minimizing transportation time, and enhancing customer satisfaction. The research delves into the intricacies of an optimization model tailored for a specific iteration of the Location-Routing Problem (LRP), namely the Multi-Objective Multi-Period Low-Carbon Location-Routing Problem (MMLCLRP). This variant of the LRP takes into meticulous consideration several crucial parameters, such as the overall logistics cost, the arrival times of demand points, and carbon emissions. These factors are pivotal in determining both the optimal location for depots and the programming of routes within a multi-period planning horizon. The proposed model guarantees the long-term sustainability of logistics operations while flexibly adapting location routing decisions for each period in response to evolving market demands. To tackle the inherent complexity of this problem, an improved version of the Non-dominated Sorting Genetic Algorithm (NSGA-II) was employed. This approach integrates a pioneering similarity distance metric to quantify the resemblance between potential solutions. Additionally, a crowding clustering strategy was implemented to enhance the diversity within the NSGA-II. Empirical results illustrate the capability of the proposed optimization model in effectively harmonizing various objectives, encompassing economic, efficiency, and environmental aspects within the logistics domain. Additionally, the enhanced algorithm exhibits notable advantages in addressing the complexities inherent in the optimization model.

In response to the challenge of three-dimensional path planning for unmanned aerial vehicles (UAVs), this paper introduces a Multi-Strategy Plant Growth Path Planning (MSPGPP) algorithm. Initially, a threat sphere and obstacle avoidance cone model are constructed to ensure safety in avoidance and enhance algorithm accuracy and convergence velocity, while B-spline curves are introduced to improve the quality of the path. Next, the concept of threat level is incorporated along with UAV maneuverability to establish a multi-objective optimization function, enabling the algorithm to plan optimal paths. Finally, an algorithm framework is developed, and comparative simulation results of the improved algorithm against the original algorithm and other optimization algorithms are presented in a three-dimensional environment. The results demonstrate that compared to the original algorithm, the improved algorithm better balances obstacle avoidance safety and UAV maneuverability, planning safer and more effective paths. Additionally, it achieves higher accuracy and faster convergence. Compared to other optimization algorithms, the improved algorithm has better optimization capabilities, higher computational efficiency, and superior path quality.

An outer rotor double permanent magnet (PM) excited flux switching generator is designed and optimized in this paper for direct drive wind turbine applications. This generator consists of two sets of PMs: ferrite PMs embedded in the stator yoke and neodymium PMs sandwiched between the rotor segments. In this regard, the main justification for employing ferrite PMs in the stator yoke is that the risk of demagnetization of ferrite PMs at high temperatures is lower than that of neodymium PMs (the temperature of the machine’s stationary parts is higher than that of its rotating parts). For the design of the machine, the Taguchi design of experiments is deployed, while a decision-making algorithm based on the technique for order of preference by similarity to the ideal solution is used to solve the contradiction that results from the Taguchi design of experiments in the multi-objective design optimization process. During the multi-objective design optimization steps, simultaneously maximizing the no-load phase voltage and minimizing the cogging torque and total harmonic distortion of the no-load phase voltage are defined as the objective functions. The optimally designed machine is prototyped and subsequently subjected to experimental validation to verify the predictions in satisfying the objective functions.

The fog computing paradigm was introduced to overcome challenges that cannot be addressed by conventional cloud computing, such as the lower response latency for real-time applications. Task scheduling in fog environments sets forth more complexity using novel objectives beyond scheduling in the cloud. In this study, a task scheduling model with five common objectives and two latency metrics is presented. We propose a latency aware multi-objective multi-rank scheduling algorithm, LAMOMRank, for fog computing. The performance of our algorithm was compared with that of three well known multi-objective scheduling algorithms, Non-dominated Sorting Genetic Algorithm (NSGA-II), Strength Pareto Evolutionary Algorithm (SPEA2) and Multi-objective Heterogeneous Earliest Finish Time (MOHEFT) algorithm, using three multi-objective metrics and two latency addressing metrics. We populate workload sets using Pegasus workflows and the DeFog benchmark to be distributed over two fog clusters generated with various Amazon Web Services instances. The empirical results validate the significance of our algorithm for better latency fronts including the response latency and task delivery time without performance degradation on multi-objective metrics.

This research presents a multi-channel phase shift inductor-inductor-capacitor (LLC) resonant converter with a wide input and output voltage range that has been specially optimized for fuel cell applications. The worst-case minimum stack voltage and the battery voltage range were used to determine the optimal parameters of the LLC converter. The voltage gain of the converter operating at a constant power was considered, and an algorithm for such a calculation is proposed. Current balancing is implemented using channel temperature measurements, circumventing the need for expensive current sensors. A natural self-balancing mechanism is investigated for converters operating at high frequencies exceeding the series resonance frequency of a resonant converter. Efficiency optimization recommendations for the transformer turn ratio of the converter are illustrated. Multi-object optimization is implemented to achieve an optimized design so that both the minimum and maximum voltage gains can be realized; simultaneously, the transformer turn ratio is kept at a minimum to maintain high efficiency. An experimental prototype of a four-channel converter is implemented and verified at a 5-kW power level. Active content and the C language code files for converter calculations are also provided in this paper.

This paper aims to address the time-dependent multi-depot vehicle routing problem with time windows (MDVRPTW) in urban cold chain logistics under a dynamic road network. The study considers the impact of carbon emissions and traffic congestion on urban cold chain logistics distribution activities. It proposes a cross-period road segment travel time calculation method, constructs a multi-objective optimization model that minimizes total costs encompassing comprehensive transportation costs, carbon emission costs, time penalty costs, cargo damage costs, and refrigeration costs. An adaptive large neighborhood search ant colony optimization algorithm (ALNS_ACO) is designed, which combines the exploration capability of ant colony optimization algorithm (ACO) with the local search capability of adaptive large neighborhood search algorithm (ALNS) to optimize and solve the model. Finally, the model is optimized and solved through simulation using six sets of C-type, R-type, and RC-type instances from the Solomon test database. The results indicate that: 1) The planned routes can reasonably avoid peak congestion periods in the morning and evening. Compared to the single-center scenario, the multi-center approach achieves superior solutions in terms of the total cost, carbon emissions, and total travel time in urban cold chain logistics distribution; 2) The exacerbation of traffic congestion leads to increased costs, and different optimization objectives have a significant impact on the model solutions; 3) Finally, through multidimensional comparisons of simulation performance with ACO and GA algorithms, the effectiveness of the proposed optimization algorithm is validated.

As a non-invasive neuromodulation technique, transcranial magnetic stimulation (TMS) has shown great potential in scientific research and clinical application. Using TMS to stimulate the deep brain is important for enhancing the therapeutic effects of mental disorders and explaining the causes of mental disorders. However, the intracranial induced electric field (E-field) generated by conventional magnetic stimulation coils dissipates severely with increasing stimulation depth and the focalized stimulation is limited to superficial areas. In this paper, we first propose a novel multi-quadrant spatial array (MQS array) based on curved-shaped coils. Each coil in the MQS array is tangent to the head and bent away from the human head, which is conducive to reducing the accumulation of non-longitudinal induced E-field components and improving the stimulation focalization in the deep brain. Then, we propose a new spatial magnetic array optimization method based on the BP-NSGA-II algorithm. The predictive models of deep brain transcranial magnetic stimulation characteristics of the spatial array are obtained by the BP neural network and the multi-objective optimization of the stimulation currents applied in the spatial array is performed with the NSGA-II algorithm. Results show that the MQS array can produce an obvious focusing area at 10 cm below the scalp which satisfies the depth requirement of deep brain transcranial magnetic stimulation. Under the same constraint of Joule loss, the optimized MQS array can enhance intracranial stimulation intensity by 131%, increase the longitudinal attenuation ratio to 4.5 times, and reduce the focusing area by 76% compared to the conventional planar magnetic stimulation array. The proposed MQS array has significant advantages in deep brain focalized stimulation, and the spatial array optimization method described in this study may provide a valuable reference for coil or array optimization processes in other application contexts.

As a renewable energy solution for remote marine environments, marine raft microgrid clusters differ from terrestrial multi-microgrid systems and traditional single-island microgrids. In the absence of large-scale grid support, these marine raft microgrids must maintain the stability and economic efficiency of power supply within a collaborative multi-microgrid context. To address this, a multi-objective optimization approach for energy scheduling is proposed. This study initially constructs a microgrid cluster system model and introduces two economic objective functions. These functions consider both inter-microgrid power scheduling and the economic benefits of power procurement. By applying the Non-dominated Sorting Genetic Algorithm II (NSGA-II) and the Constraint-based Multi-objective Evolutionary Algorithm based on Decomposition (CMOEA/D) to solve the objective functions, the results indicate that the CMOEA/D algorithm demonstrates high efficiency and accuracy in pursuing economically optimal solutions. Compared to the NSGA-II algorithm, CMOEA/D outperforms in terms of the quality of optimal solutions and iteration time, thereby enhancing the economic benefits of the microgrid cluster and validating the effectiveness of the proposed model. This research provides significant theoretical and practical guidance for energy management in remote marine environments, showcasing its profound theoretical significance and application value.

The intelligent reflecting surface (IRS) supports communication systems well, especially in physical layer security for the cooperative power domain non-orthogonal multiple access (PDNOMA). In this work, we investigate the secrecy performance of PDNOMA with the assistance of the IRS and a multiple-relaying network in the presence of an eavesdropper. Three selection strategies are considered at the relaying network to boost the system’s performance: the first method is based on the best relay selection, the second on the max-min concept, and the third on harmonious characteristics. Moreover, the phase shift of IRS element and power allocation for each NOMA user can be controlled to improve the secrecy quality and reduce the influence of an eavesdropper (E). Besides, applying the technique of transmitting artificial noise (AN) from the source is also considered in this paper to interfere with the signal at E. Furthermore, in this paper, we determine two primary metrics to evaluate the secrecy performance of our proposed system: the worst secrecy capacity and secrecy energy efficiency. The balance of these two metrics needs to be assured to improve the secrecy performance. Thus, in this paper, we consider the multi-object problem and propose the genetic algorithm-based approach, a non-dominated sorting genetic algorithm with three procedures (NSGA-II), to solve this problem. Then, to highlight the proposed algorithm’s outstanding performance, we compare it with other algorithms, Reference point based NSGA-II (R-NSGA-II) and the exhaustive search (ES). Additionally, the impacts of critical system parameters are investigated for both cases as IRS and none-IRS assistance comprises three relaying selection techniques, the number of IRS elements, the strength of AN signal, the distances of source-relay link, relay-IRS link, and IRS-Eavesdropper link. Finally, the summaries of these archived results show the benefits of our proposed model in different cases of the deployment of the IRS and without the IRS.

Ensuring the well-being of patients is paramount for the effective functioning of healthcare systems. This research focuses on developing a comprehensive framework aimed at improving hospital systems with a primary emphasis on enhancing patient satisfaction, specifically addressing Length of Stay (LOS) and physician assignment issues. To achieve this objective, a multi-faceted approach was employed. The data collection involved a combination of surveys, observations, and interviews, which enabled a thorough understanding of the hospital environment. A unique formula was derived to compute the satisfaction scores from the gathered survey data. Subsequently, a simulation model was developed. The research applies two distinct multi-objective optimization approaches, that is, a weighted max-min for fuzzy multi-objective optimization and the Torabi-Hassini (TH) approach to balance the importance of various objectives. The application of a multi-criteria decision guideline further streamlines the decision-making process. The outcomes of this study provide suggestions for tailored improvements for hospitals of different sizes. For large hospitals, this study indicates enhancements that positively impact both LOS and physician assignment satisfaction. Medium and small-sized hospitals are presented with two distinct choices: improvements geared toward LOS satisfaction or those focused on physician assignment satisfaction. While this study is based on a case analysis of the Ophthalmology Department at Thammasat University Hospital in Thailand, its implications extend beyond this specific department. This study offers an adaptable framework that can be applied to other hospital departments, promoting an integrated approach to healthcare system enhancement.

In the rapidly evolving landscape of Beyond 5G (B5G) networks, Cell-Free Massive Multiple-Input Multiple-Output (CFmMIMO) systems have emerged as a pivotal solution to address the escalating demands for seamless connectivity and low-latency communication. However, the effective management of radio resources, particularly in the uplink direction, presents formidable challenges in optimizing system performance while ensuring equitable treatment for all users. Existing uplink power control methods often grapple with the intrinsic complexity of the optimization problems they encounter. To mitigate these challenges, this paper introduces a novel approach termed the metaheuristics-based uplink power control scheme tailored explicitly for user-centric CFmMIMO systems. Leveraging advanced metaheuristic optimization techniques, our scheme navigates the intricacies of uplink power control by prioritizing three primary objectives: maximizing the minimum user Spectral Efficiency (SE), maximizing the sum SE, and striking a balance between these two objectives. Through rigorous exploration of potential solutions, our proposed scheme transcends the limitations of conventional methods, offering near-optimal solutions for uplink power control. Numerical simulations corroborate the efficacy of our approach, showcasing enhanced fairness among users and substantial gains in SE compared to traditional methods. This research represents a significant advancement in the practical implementation of user-centric CFmMIMO systems within B5G networks, presenting promising solutions to address the evolving requirements of future wireless communication.

Induction heating is an efficient alternative to fuel combustion in thermal applications. The design of the coil arrangement is a complex problem, constrained by electromagnetic and thermal dynamics. This article proposes a multi-objective optimization methodology for the geometric design of winding distribution on both the radial and the height arrangement for coils in induction heating systems around generic axisymmetric convex workpieces. A novel parametrization using the curvilinear abscissa allows for an efficient formulation for the windings distribution coordinates and avoids the use of non-linear geometric constraints in the formulation. Coil configuration problems with many degrees of freedom are solved as black-box optimization problems using the Set Membership Global Optimization method in conjunction with a finite element-based simulator. The proposed methodology is verified over several electromagnetic finite element analysis-based case studies with increasing complexity. The methodology is first tested over two target power density profiles obtained by the finite element forward problem and set over the surface of an axisymmetric convex cookware. Then, the methodology is verified with four arbitrary non-smooth power density profiles. The results show that the proposed method is suitable for solving simultaneous design optimization problems in induction heating systems, improving the optimization complexity capability.

Caching plays an important role in reducing latency and increasing the overall performance of fog computing systems. With the rise of the Internet of Things (IoT) and edge computing, fog computing has become an essential paradigm for addressing the challenges related to latency-sensitive and bandwidth-intensive applications. The integration of Information-Centric Networking (ICN) and Fog Computing (ICN-Fog) has emerged as a promising solution for meeting the demands of low-latency and high-throughput applications in rapidly growing IoT devices. A step was taken by ICN-Fog to reduce latency and achieve higher data communication and information gathering for fog computing. Using Artificial Intelligence (AI) methods, the Firefly Optimization Technique was introduced as an effective optimization algorithm to enhance the caching technique. In this study, we aim to improve several performance metrics, such as the cache hit ratio, internal link load, and average query duration. To achieve these enhancements, we propose a unique solution inspired by the Firefly Optimization Technique. This technique applies to the ICN-Fog caching model, which was utilized to intelligently determine cache placement and network topology adaptations. CloudSim was employed to execute a simulation of the proposed strategy. The results of the experiments suggest that the Multi-Objective Firefly Algorithm (MOFA) outperforms the compared algorithms in terms of both efficiency and effectiveness in identifying the optimal caching technique.

This paper proposes a novel methodology based on the Genetic Programming (GP) to derive behavioral models describing the transient evolution of the terminal voltage of a battery. These models analytically relate the battery voltage to its state of charge, charge/discharge rate, and temperature. Compared to the popular equivalent circuit-based models, one of the main advantages is the significant reduction of the effort to produce the experimental dataset required to identify the model parameters. The GP generates a family of optimal “candidate” analytical models, each associated with suitable metrics that quantify performance indicators like simplicity and accuracy. The methodology is applied to describe the transient discharge phase of a Lithium Iron Phosphate (LiFePO4 or LFP) battery under realistic operating conditions, considering the state-of-charge between 20% and 80%, discharge rates comprised between 0.25C and 1C, and temperature ranging from 5°C to 35°C. The GP provides different solutions that can be chosen by imposing the desired trade-off between accuracy and simplicity. Two models are selected and validated against experimental results. The chosen models guarantee a quite low level of the relative root mean square error (maximum 0.31% and 0.22%, respectively) over the range of analysis.

In recent decades, multi-objective evolutionary algorithms (MOEAs) have been evaluated on artificial test problems with unrealistic characteristics, leading to uncertain conclusions about their efficacy in real-world applications. To address this issue, a few benchmark test suites comprising real-world problems have been proposed for MOEAs, encompassing numerous multi-objective problems and a select few many-objective problems. Given the distinct challenges posed by many-objective optimization problems (MaOPs) and their inherent difficulty, developing a test suite that includes real-world problems with many conflicting objectives is crucial. Hence, in this paper, we propose a comprehensive test suite for benchmarking the many-objective real-world complex problems. This test suite consists of 11 problems collected from different disciplines of engineering. Furthermore, we comprehensively analyzed the problems in our newly proposed test suite, employing eight state-of-the-art algorithms rooted in various fundamental principles specifically designed to address MaOPs. The experimental findings highlight the strong performance of indicator-based, weight-vector-based decomposition, Pareto-dominance-based, and hybrid MOEAs on the proposed test suite. In contrast, reference-vector-based decomposition approaches, Pareto front shape estimation-based methods, and multi-evolution approaches exhibit relatively weaker performance.

Multiobjective evolutionary algorithms are highly effective in solving multiobjective optimization problems (MOPs). The selection strategy, involving mating and environmental selection, is crucial in shaping these algorithms. However, when applied to many-objective optimization (MaOPs) with more than three objectives, existing methods face challenges due to reduced selection pressure and issues in maintaining diversity, making them less efficient. To address these challenges, we present a novel approach in this paper: the clustering-aided grid-based one-to-one selection-driven evolutionary algorithm (ClGrMOEA), designed to handle both MOPs and MaOPs effectively. The proposed ClGrMOEA introduces a hybrid approach that combines clustering-based mating selection with grid-based one-to-one environmental selection to balance convergence and diversity in MOPs and MaOPs. The mating selection employs K-means clustering to partition the objective space and utilizes a convergence indicator based on Euclidean distance to select promising solutions for offspring generation. The environmental selection combines Pareto dominance with grid-based one-to-one selection, using grid coordinate point distance to select promising solutions. An external archive based on Pareto dominance and crowding distance preserves elite individuals. Extensive experiments are conducted on 19 benchmark problems and 16 real-world problems to validate the superior performance of ClGrMOEA compared to seven state-of-the-art algorithms. The experimental results demonstrate that ClGrMOEA significantly outperforms these benchmark algorithms.

With the rapid growth of the mobile computing techniques, a wide variety of mobile edge computing (MEC) applications have emerged recently, aiming to provide computationally intensive and delay-sensitive network services. Through MEC, various complex tasks of mobile devices can be offloaded to the edge of network system for execution by edge servers, which greatly reduces the local computing burden. However, how to effectively allocate computational and communication resources in edge-cloud remains a challenging task, especially when multiple mobile users and edge servers are involved. In this paper, we propose a decomposition-based multi-objective optimization algorithm based on estimation-of-distribution models (MOEA/D-EoD) to deal with the task offloading and resource allocation problem in MEC. Especially, considering the features of multi-user and multi-server cloud-edge-end collaboration wireless MEC system, we construct a joint optimization model of task offloading and resource allocation, where limited communication and computational resource constraints are considered. To deal with the optimization model, we design an efficient decomposition-based algorithm, which incorporates two novel estimation-of-distribution models to deal with discrete and continuous decision variables of the problem. Experimental results obtained from benchmark test suites DTLZ and ZDT demonstrate that the proposed method exhibits significantly superior performance compared to other comparative algorithms. The proposed model and algorithm are simulated and tested on different test instances, and experimental results show the effectiveness and efficiency of our proposed method.

The production planning problem in the Pinus radiata sawmill industry revolves around determining how to cut a set of logs of different diameters to obtain pieces with a rectangular base, typically of the same length as the original log. The primary objective is often to maximize the volumetric yield or the ratio between the total volume of the produced pieces and the available volume of the logs. Given the scarcity of timber forests, small and medium-sized (SME) sawmills must optimize their operations to maximize the volumetric yield of logs, usually procured from third parties and may vary in quality and dimensions. In this context, the absence of decision support tools directly contributes to inefficient raw material utilization, consequently impacting the business’s profit. This study introduces a multi-objective mixed-integer linear programming model incorporating log availability and product demand as input parameters. The objective functions aim to minimize the total volumetric loss of utilized logs and the surplus quantity associated with products exceeding demand. The model integrates cutting patterns pre-determined for each log diameter as an additional input. The Cutlog© software was used to identify all the optimal cutting patterns. The $\varepsilon $ -constraint method, implemented in the CPLEX© solver, was employed to solve the model. The model was validated using representative instances designed to emulate the challenges faced by SME sawmills. Real industry data, surveys, and commercial records from SME sawmills in southern Chile were utilized. The results confirm the effectiveness of the proposed model in addressing the multi-objective challenges encountered by these businesses. The model successfully identifies multiple solutions on the Pareto frontier, offering valuable insights for decision-making.

Many machine learning algorithms excel at handling problems with conflicting objectives. Multi-Objective Optimization (MOO) algorithms play a crucial role in this process by enabling them to navigate these trade-offs effectively. This capability is essential for solving complex problems across diverse scientific and engineering domains, where achieving optimal solutions often requires balancing multiple objectives. One of these MOO algorithms Multi-Objective Particle Swarm Optimization (MOPSO) extends it to handle problems with multiple objectives simultaneously, but like many swarm-based algorithms, MOPSO can suffer from premature convergence or local optima solutions. Therefore, this article introduces a novel Multi-Exemplar Particle Swarm Optimization with Local Awareness (MEPSOLA) as a potent solution. The algorithm presents a developed multi-objective-aware criterion for multi-exemplar selection, adeptly balancing exploration and exploitation to avoid local optima and enhance performance across multiple objectives. It also introduces a conditional and periodic Tabu search tailored specifically for exemplar selection enhancement, improving both exploration and exploitation capabilities and avoiding premature convergence. Additionally, our method employs an improved initialization phase using equal sampling for each decision variable to ensure a comprehensive exploration of the entire solution space. A comprehensive assessment utilizing standard mathematical functions such as Fonseca-Fleming (FON), Kursawe (KUR), ZDT1, ZDT2, ZDT3, and ZDT6, and a comparison with state-of-the-art benchmarks in the field such as the Multi-Objective Evolutionary Algorithm (MOEA), Non-Dominated Sorting Genetic Algorithm (NSGA-II), and NSGA-III, validate the efficiency of MEPSOLA. Notably, MEPSOLA’s solutions outperform other benchmarks in key metrics across the majority of mathematical problems, for instance in set coverage, where our method dominates other methods’ solutions by 99.22%, 69%, and 93.58 %, respectively, highlighting its enhanced capability in optimizing capability within complex multi-objective optimization scenarios.

The core idea of Multi-Task Learning (MTL) is to develop neural networks with a shared feature extraction backbone and multiple prediction heads, each capable of inferring a different task simultaneously. Parameters in the backbone contribute to all tasks while those in the prediction heads contribute to only one or fewer tasks. Challenges arise when multiple tasks compete for resource. Existing methods focus on resource competition in shared parameters and proposed explanatory factors of task conflicts, task dominance, and gradient stability. However the fundamental nature of MTL is still understudied. In this paper, instead of following the existing methodology research directions, we carry out large-scale empirical study and provide deeper insight on understanding MTL. In particular, instead of focusing on resource competition in the shared parameters in the backbone, we shift our attention to resource competition in the backbone output, which is the embedded representation that is shared by all prediction heads. We show that the existing explanatory factors display weak causal relationship with model performance. We propose a novel measurement, which we term Feature Disentanglement, and show that understanding MTL problems from the perspective of how the shared representation is leveraged by different prediction heads, is a more faithful and reliable way than that from the perspective of how supervision signals from different tasks are interfering in the shared parameters. Additionally, it has been a commonly employed technique to replace gradients w.r.t. shared parameters with gradients w.r.t. shared representation for reduced computation. We conduct a comprehensive study and show that unless a theoretical analysis could be developed, there is not general guarantee that this fast approximation technique would work in practice.

This paper addresses the critical challenge of effective resource allocation in emergency logistics distribution, particularly during disaster scenarios where demand is uncertain and traditional deterministic optimization methods are inadequate. We propose a novel multi-objective chance-constrained model that minimizes transportation deviation, creating a computationally feasible deterministic equivalent model using uncertainty theory. Furthermore, this paper introduces a hybrid of the Estimation of Distribution Algorithm and the Multi-Objective Snow Goose Algorithm (EDA-MOSGA), designed to efficiently navigate the solution space and identify near-optimal solutions. The EDA-MOSGA enhances Pareto front diversity with its adaptive population adjustment and specialized operator model, marking a significant advancement over existing methods. Validated through the Zitzler-Deb-Thiele (ZDT) and Deb-Thiele-Laumanns-Zitzler (DTLZ) multi-objective test suites, and a case study during the Coronavirus Disease 2020 (COVID) pandemic in Chengdu, the EDA-MOSGA demonstrates exceptional performance in real-world applications. The results of this study lay the foundation for the future integration of artificial intelligence technology to improve the scalability of logistics distribution in different emergency situations.

In real-world multi-objective optimization, dealing with many objectives and a large number of solutions is a common challenge that complicates data visualization and analysis. This study aims to simplify decision-making by analyzing tools to better explore Pareto optimal solutions in many-objective scenarios, integrating clustering, filtering, and ranking with existing graphics techniques. The dynamic combination of these tools should reduce complexity and highlight significant patterns in the data set, allowing decision-makers to tailor the visualization to their specific needs and preferences. Central to the approach presented in this work is the innovative application of shape-based clustering to organize the solution set and the use of this clustering to define distinct types of filters. Additionally, ranking methods originally proposed to enhance search in many-objective evolutionary algorithms are used here to identify the best solutions based on predefined criteria in combination with other techniques. The efficacy of the proposed integrated approach was evaluated using an application developed with this aim and considering a five-objective problem as a case study. The analysis suggests that using these combined strategies aids interactive visual exploration, effectively reducing solution volume and improving data understanding, potentially facilitating decision-making tasks.

The development and the upgrade of railway networks is one of the strategies to reach decarbonization targets in the transportation field, thanks to the considerably lower energy consumption of electric trains with respect to other vehicles, typically fossil fuel powered. The design process of electric railway power systems is complex, requiring advanced simulation tools. The paper proposes a novel methodology for the design of the electrical power system of railway tracks, using genetic optimization. For this purpose, the authors developed ROAR, a flexible simulation and optimization software that generates optimized railway power system designs, helping engineers find the most efficient design solutions from a technical and economic feasibility perspective. After validating the simulation engine and comparing it with well-established software, the proposed method was applied to an operational electrified railway line in Italy to assess the effectiveness of the optimization algorithm. The results demonstrate excellent convergence properties, finding a different infrastructure design that achieves the same electrical performance, reducing costs with respect to the existing design.

Most credit scoring datasets suffer from imbalanced class distribution, with most borrowers repaying their loans and only a few defaulting. This can affect the accuracy of traditional machine learning algorithms. To address the imbalance issue, we propose a new approach that combines borderline conditional generative adversarial networks, ensemble learning, and multi-objective optimization techniques. Our method involves employing an autoencoder to extract relevant features from the dataset, followed by generative adversarial networks (GAN)-based oversampling method to balance the dataset. We utilize a multi-objective evolutionary algorithm to optimize the selection of balanced data subsets, which are then used to train multiple classifiers. The predictions from these classifiers are combined using a stacking method, enhancing overall accuracy. Our experiments indicate that our proposed model MOCGAN, effectively handles class imbalance, achieving a mean AUC of 90.12% and a mean F1-score of 88.25%, thereby significantly enhancing the reliability of credit scoring models.

In bioinformatics, medical diagnosis models might be significantly impacted by high-dimensional data generated by high-throughput technologies. This data includes redundant or irrelevant genes, making it challenging to identify the relevant genes from such high-dimensional data. Therefore, an effective feature selection (FS) technique is crucial to mitigate dimensionality, thereby enhancing the performance and accuracy of medical diagnosis. The Cuckoo Search Algorithm (CSA) has proven effective in gene selection, demonstrating prowess in exploitation, exploration, and convergence. However, most of the current CSA-based FS techniques deal with gene selection problems as a single objective rather than adopting a multi-objective mechanism. This article proposes the Multi-Objective Binary Cuckoo Search Algorithm (MOBCSA) for gene selection. MOBCSA extends the standard CSA by incorporating multiple objectives, including accuracy of classification and number of selected genes. MOBCSA utilizes an S-shaped transfer function for transforming the algorithm’s search space from a continuous to a binary search space. MOBCSA integrates two components: an external archive to save the pareto optimal solutions attained during the search process, and an adaptive crowding distance updating mechanism integrated into the archive to maintain diversity and increase the coverage of optimal solutions. To assess MOBCSA’s performance, evaluation experiments were conducted on six benchmark biomedical datasets using three different classifiers. Then, the obtained experimental results were compared against four multi-objective-based state-of-the art FS methods. The findings prove that MOBCSA surpasses the other methods in both accuracy of classification and number of selected genes, where it has obtained an average accuracy ranging from 92.79% to 98.42% and an average number of selected genes ranging from 15.67 to 27.88 for different classifiers and datasets.

This article introduces BetaMemo models, a set of advanced probabilistic forecasting models aimed at predicting the hourly power output of photovoltaic plants. By employing a semiparametric approach based on beta distributions and deterministic models, BetaMemo offers detailed forecasts, including point forecasts, variance, quantiles, uncertainty measures, and probabilities of power generation falling within specific intervals or exceeding predefined thresholds. BetaMemo models rely on input data derived from weather forecasts generated by a Numerical Weather Prediction model coupled with variables pertaining to solar positioning in the forthcoming hours. Eleven BetaMemo models were created, each using a unique combination of explanatory variables. These variables include data related to the location of the plant and spatiotemporal variables from weather forecasts across a broad area surrounding the plant. The models were validated using a real-life case study of a photovoltaic plant in Portugal, including comparisons of their performance with benchmark forecasting models. The results demonstrate the superior performance of the BetaMemo models, surpassing those of benchmark models in terms of forecasting accuracy. The BetaMemo model that integrates the most extensive set of spatiotemporal explanatory variables provides notably better forecasting results than simpler versions of the model that rely exclusively on the local plant information. This model improves the continuous ranked probability score by 13.89% and the reliability index by 45.66% compared to those obtained from a quantile random forest model using the same explanatory variables. The findings highlight the potential of BetaMemo models to enhance decision-making processes related to photovoltaic power bidding in electricity markets.

Aiming at the problems of punctuality, parking accuracy and energy saving of urban rail train operation, an intelligent control method for automatic train operation (ATO) based on deep Q network (DQN) is proposed. The train dynamics model is established under the condition of satisfying the safety principle and various constraints of automatic driving of urban rail train. Considering the transformation rules and sequences of working conditions between train stations, the agent in the DQN algorithm is used as the train controller to adjust the train automatic driving strategy in real time according to the train operating state and operating environment, and optimizes the generation of the train automatic driving curve. Taking the Beijing Yizhuang Subway line as an example, the simulation test results show that the DQN urban rail train control method reduces energy consumption by 12.32% compared with the traditional train PID control method, and improves the running punctuality and parking accuracy; at the same time, the DQN train automatically driving control method can adjust the train running state in real time and dynamically, and has good adaptability and robustness to the change of train running environment parameters.