Fuzzing is a critical technique for uncovering vulnerabilities in software libraries. However, current approaches often struggle with cross-language compatibility and integration with diverse fuzzing tools. We propose <monospace xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">EXo-Muta</monospace>, a novel universal library fuzzing framework based on exogenous mutation. We use the term ‘exogenous’ to describe this new mutation process because it operates externally to the fuzzer's core engine, executing within the fuzz driver as an independent component, unlike traditional endogenous mutations tightly integrated within the fuzzer itself. By decoupling the mutation process from specific fuzzers, <monospace xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">EXo-Muta</monospace> achieves unprecedented adaptability across diverse programming languages and fuzzing tools. It leverages static analysis to extract structured data representations and applies language-independent mutation operators at the code level. This design enables seamless integration with various existing fuzzers, enhancing their performance regardless of the target language. We further utilize large language models (LLM) for efficient cross-language data conversion. In experiments, we evaluated <monospace xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">EXo-Muta</monospace> on 20 real-world libraries across C++, Python, Java, and JavaScript, integrating it with multiple stateof- the-art fuzzers, such as AFL++ and libFuzzer, and languagespecific tools, such as Atheris and Jazzer. Results show significant improvements in code coverage across different fuzzers and languages, with up to 58% more edges discovered in C++ projects when integrated with libFuzzer, and consistent outperformance in other scenarios (27% on Python, 9% on Java, 6% on JavaScript). <monospace xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">EXo-Muta</monospace> represents a significant advancement in fuzzing technology, offering a universal, language-agnostic approach that substantially improves code coverage across diverse programming languages and fuzzing tools, thereby expanding the reach and effectiveness of library API testing.

American fuzzy lop (AFL), as a representative tool for fuzzing, is capable of uncovering security vulnerabilities in industrial systems. It suffers from consuming a large amount of computational resources during the mutation. To improve the performance of AFL, researchers adopt algorithms, such as particle swarm optimization and long short-term memory, to optimize mutation operator selection. However, challenges persist in these approaches integrated with AFL, including optimization model complexity, insufficient accuracy, and poor generalization scalability. To address these issues, the article proposes a new fuzzer called <sc xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">DPRFuzz</small> to optimize AFL’s mutation phases. First, in the deterministic mutation strategy mutation phase, deep Q network and trust region policy optimization are leveraged to precisely generate effective mutated samples through perceiving mutation process in a relatively short time. Then, to boost the efficiency of the Havoc random mutation phase, we improve the Thompson sampling algorithm based on a multiagent strategy to generate an overall optimal mutation strategy chain. Finally, the approach is tested on eight programs, such as <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">readelf</i>, <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">tcpdump,</i> and <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">nm</i>, and the advantages of <sc xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">DPRFuzz</small> are analyzed. Most importantly, the experiment reveals results that <sc xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">DPRFuzz</small> achieves better fuzzing performance compared to the traditional and other AFL-based fuzzers, such as AFL, AFL++, AFLSmart, etc. On average, <sc xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">DPRFuzz</small> achieves an improvement in code coverage of over 10%, along with a significant increase in the number of crashes.

Topology optimization (TO) emerges as a pivotal technique in engineering design, which aims at optimizing material distribution within given design spaces. TO often presents large-scale combinatorial optimization problems, which pose challenges to gradient methods in handling discrete variables and evolutionary algorithms in handling high-dimensional search spaces. This paper proposes a dynamic and heterogeneous representation of TO solutions, and designs novel crossover and mutation operators for searching for TO solutions. The representation features low granularity at early stages to explore optimal regions and high granularity at later stages to exploit optimal solutions, and the operators support evolutionary algorithms in effectively evolving populations using the representation. Experimental results verify that the proposed method significantly outperforms the state-of-the-art gradient methods and evolutionary algorithms across various TO tasks.

Optimal path planning problems for mobile robots have been particularly actively studied in the last decade. The goal is to find an optimal or near-optimal path from a starting terminal to one or more terminals in an environment with various obstacles, in terms of minimizing robot travel time, distance traveled, energy costs, or other optimization criteria. In this paper, we propose a hybrid algorithm combining a probabilistic roadmap algorithm (PRM) and an adapted genetic algorithm (AGA) to solve a path planning problem with one or more independent objectives. The robot's path length is used as an optimization criterion. Compared with existing approaches used in genetic algorithms (GAs), the proposed approach has two main differences. The first is the environment representation, which relies on image processing and morphological operations, which has proven to be a more efficient method than methods based on cellular representation. In particular, the proposed method eliminates the need to find a trade-off between accuracy and speed of processing geometric information. The second is a new tactic for creating an initial population of the genetic algorithm to accelerate convergence in the presence of multiple objectives. By leveraging the capabilities of a probabilistic roadmap algorithm. Another key feature of the algorithm's implementation is the appropriate (for the domain under study) selection of numerical parameters that determine the characteristics of all stages of the evolutionary strategy, including the time required to complete each stage. This applies in particular to the parameters of the mutation operator and the elite strategy. The proposed algorithm was tested on two real-world maps with varying levels of complexity. Its effectiveness was confirmed by comparison with path planning results for test maps obtained using a standard genetic algorithm and an ant colony optimization algorithm. Experimental results demonstrate that the hybrid algorithm expands the capabilities of a conventional genetic algorithm and finds rational path variants with the best objective function value for single and multiple objectives in significantly less time than other traditional GA implementations.

ABSTRACTThis study investigates the enzymatic synthesis of amoxicillin, focusing on its kinetic properties and their influence on antibiotic production in a batch‐operated enzymatic reactor. The reaction is catalyzed by penicillin G acylase (PGA, E.C.3.5.1.11), which is immobilized on glyoxyl‐agarose. The reaction involves the p‐hydroxyphenylglycyne methyl ester and 6‐aminopenicillanic acid (6‐APA) for amoxicillin formation. Under kinetic control, parallel hydrolytic pathways lead to product loss. Two kinetic models were evaluated: one based on Michaelis–Menten kinetics and another incorporating reaction and equilibrium constants for the process steps. Parameter estimation for the models was performed at different concentrations using two mathematical approaches: the Markov chain Monte Carlo (MCMC) method, rooted in Bayesian statistics and characterized as nondeterministic, and genetic algorithm, an evolutionary computation method incorporating crossover, mutation, and selection operators. The relative root mean squared error (rRMSE) was selected as the metric for evaluating the predictive performance of the models. MCMC presented the best results for low ester concentrations, with rRMSE values ranging from 1.48% to 6.10% for the Michaelis–Menten‐based model. The mathematical model was validated using data from an enzymatic reactor operating in semi‐batch mode, demonstrating a satisfactory capacity to predict the system's dynamic behavior under this operational condition.

ABSTRACT Cloud computing system task scheduling optimization has garnered substantial attention because it directly influences resource utilization, service quality, and system energy consumption. To address the demands of cloud computing environments, an enhanced seahorse optimizer using alpha balance factor and t‐distribution mutation is proposed. Firstly, the alpha balance adaptive mechanism was proposed. The original alpha balance factor exhibits a relatively smooth downward trend. To overcome this limitation, the standard normal distribution random numbers are considered to add, which introduces greater volatility and enhances the algorithm's ability to jump out of the local optima. Secondly, a multi‐dimensional t‐distribution mutation operator was designed, taking into account both exploration and exploitation. This promotes the dynamic exploration and exploitation of seahorse individuals during the movement behavior stage, enhances population diversity, reduces the possibility of aggregation at local points, and accelerates the convergence speed of the algorithm. The optimal variant HTSHOt2 was selected by validation on the CEC‐2022. Finally, HTSHOt2 was applied to small‐scale and large‐scale cloud computing task scheduling scenarios for the first time, achieving coordinated optimization of task completion time, resource utilization rate, and energy consumption. The results show that the total cost of HTSHOt2 in small‐scale scenarios ranges from 2.28E‐01 to 2.62E‐01, reducing by 12.37% to 24.25%. The total cost range in large‐scale scenarios is between 2.75E‐01 and 2.87E‐01, reducing by 4.33% to 7.72%. In addition, HTSHOt2 spends very little on price cost, load cost, and time cost, all of which are far superior to other comparison algorithms. This indicates that HTSHOt2 can effectively solve the problem of task scheduling optimization in cloud computing systems and has powerful performance.

Hyperspectral band selection has become a key focus in hyperspectral image processing as it reduces the spectral redundancy and computational overhead, thereby improving classification performance. However, optimal band selection remains challenging due to its combinatorial nature. Although numerous metaheuristic algorithms have been introduced in recent years to address this problem, achieving an effective balance between exploration and exploitation continues to pose a major challenge. This paper proposes a novel approach that combines a parameter-free binary Jaya algorithm with a mutation operator to enhance exploration and maintain solution diversity within the search space. We employ Opposition-based Learning (OBL) for population initialization and Quasi-Reflection reinitialization strategy to add diversity whenever fitness stagnation occurs. To simultaneously improve classification performance and band reduction we adopt weighted sum multi-objective fitness function that minimizes redundancy and enhances model generalization. Our proposed method is evaluated using three benchmark datasets, namely Indian Pines, Pavia University, and Salinas. Experimental results demonstrate that the pro-posed method outperforms recent metaheuristic-based band selection techniques. Its superior performance makes it well suited for various HSI applications.

In complex dynamic traffic environments, existing path planning methods are often limited by insufficient convergence efficiency, poor solution stability, and uneven load distribution. This paper proposes an optimization strategy based on a genetic algorithm (GA). This strategy incorporates an adaptive mutation operator to adjust mutation probabilities during evolution to maintain population diversity dynamically. A partial matching crossover operator is used to improve the recombination efficiency of high-quality gene segments. Furthermore, three metrics—travel time, path length, and path balance—are simultaneously incorporated into the fitness function to achieve optimization under multi-dimensional constraints. The experimental results demonstrate that the average travel time of the optimized path is 32.5 s, which significantly improves the traffic efficiency compared with the Dijkstra algorithm (42 s). The path length is 850 meters, which is better than the 860 meters of the A* (A-star) algorithm, and the path complexity is significantly lower than that of other algorithms. The results demonstrate the effectiveness and applicability of the proposed optimization strategy in improving convergence speed, path stability, and traffic flow balance.

This paper proposes a computational and optimization model for determining the effective shielding duration of enemy missiles by smoke countermeasure deployment from unmanned aerial vehicles (UAVs), focusing on the application of a multi-decision variable nonlinear optimization method based on the Differential Evolution (DE) algorithm. First, by analyzing the flat trajectory of smoke countermeasures and the detonation point location, combined with missile trajectory equations and line-of-sight distance criteria, the effective shielding duration is derived. Second, with maximizing shielding duration as the objective, parameters such as heading angle, flight speed, smoke grenade deployment time, and detonation delay are optimized. The DE algorithm performs a global search in continuous variable space through mutation, crossover, and selection operations. Finally, the approach was extended to scenarios involving a single UAV deploying three smoke grenades and a three-UAV deployment. In both cases, the DE algorithm optimized multiple decision variables to derive optimal masking strategies. This model enables precise quantification and optimization of smoke interference effects. The DE algorithm ensures the acquisition of globally optimal solutions, adapts to multi-scenario parameter optimization, and effectively enhances the effectiveness of countermeasures against enemy missiles.

This study examines the efficacy of Genetic Algorithms (GAs) in portfolio optimization by contrasting their results with those of more conventional techniques such as Mean-Variance Optimization (MVO) and Modern Portfolio Theory (MPT). The objective is to compare the ability of GAs to optimize risk-adjusted returns while ensuring diversification and flexibility in changing financial environments. A GA-based portfolio selection model is built and tested using ten years of actual financial data, employing selection, crossover, and mutation operations to attain maximum asset allocation. Its performance is evaluated against key financial metrics such as the Sharpe Ratio, portfolio return, and risk (standard deviation), in comparison with MPT, Simulated Annealing (SA), and Particle Swarm Optimization (PSO). The results indicate that GAs consistently outperform traditional methods in risk-adjusted returns, yield higher Sharpe Ratios, promote better diversification, and result in lower portfolio volatility. Sensitivity analysis demonstrates that GAs are robust across different parameter values. The pragmatic implications of this study are relevant for portfolio managers and algorithmic traders seeking improved risk-return trade-offs. Additionally, the research contributes to the literature by highlighting the superiority of GAs over traditional approaches and provides directions for future research into hybrid AI-driven portfolio optimization techniques.

Although the Dung Beetle Optimizer (DBO) is a promising new metaheuristic for global optimization, it often struggles with premature convergence and lacks the necessary precision when applied to complex optimization challenges. Therefore, we developed the Multi-Strategy Improved Dung Beetle Optimizer (MIDBO), an algorithm that incorporates several new strategies to enhance the performance of the standard DBO. The algorithm enhances initial population diversity by improving the distribution uniformity of the Circle chaotic map and combining it with a dynamic opposition-based learning strategy for initialization. A nonlinear oscillating balance factor and an improved foraging strategy are introduced to achieve a dynamic equilibrium between the algorithm’s global search and local refinement, thereby accelerating convergence. A multi-population differential co-evolutionary mechanism is designed, wherein the population is partitioned into three categories according to fitness, with each category using a unique mutation operator to execute targeted searches and avoid local optima. A comparative study against multiple metaheuristics on the CEC2017 and CEC2022 benchmarks was performed to comprehensively evaluate MIDBO’s performance. The practical effectiveness of the MIDBO algorithm was validated by applying it to three practical engineering challenges. The results demonstrate that MIDBO significantly outperformed the other algorithms, a success attributed to its superior optimization performance.

HighlightsWhat are the main findings?The proposed tSSA-Informer model achieves 90.85% diagnostic accuracy under strong noise (SNR = −9 dB) and 61.32% accuracy with only 10% labeled samples, outperforming SSA-Informer, GA-Informer, and other comparative models.The tSSA algorithm optimizes Informer hyperparameters 7.74% faster than the original SSA, and the integrated data preprocessing (IQR + CWT + KPCA) effectively enhances the representation of non-stationary vibration features.What is the implication of the main finding?The model’s anti-noise and small-sample adaptability precisely addresses the “multi-interference, scarce samples” dilemma of pumping station unit fault diagnosis in practical engineering.The tSSA-optimized Informer framework provides a referable paradigm for hyperparameter tuning and time-series feature mining in mechanical fault diagnosis fields.To address the problems of noise sensitivity, insufficient modeling of long-term time-series dependence, and high cost of labeled data in the fault diagnosis of pumping station units, an intelligent diagnosis method integrating the improved Sparrow Search Algorithm (tSSA) and Informer model is proposed in this study. Firstly, an adaptive t-distribution strategy is introduced into the Sparrow Search Algorithm to dynamically adjust the degree of freedom parameters of the mutation operator, balance global search and local development capabilities, avoid the algorithm converging to the origin, and enhance optimization accuracy, with time complexity consistent with the original SSA. Secondly, by combining the sparse self-attention and self-attention distillation mechanisms of Informer, the model’s ability to extract key features of long sequences is optimized, and its hyperparameters are adaptively optimized via tSSA. Experiments were conducted based on 12 types of fault vibration data acquired from pumping station units. Outliers were removed using the interquartile range (IQR) method, and dimensionality reduction was achieved through kernel principal component analysis (KPCA). The results indicate that the average diagnostic accuracy of tSSA-Informer under noise-free conditions reaches 98.73%, which is significantly higher than that of models such as SSA-Informer and GA-Informer; under noise interference of SNR = −1 dB, it still maintains an accuracy of 87.47%, outperforming comparative methods like 1D-DCTN; when the labeled sample size is reduced to 10%, its accuracy is 61.32%, which is more than 40% higher than that of traditional models. These results verify the robustness and practicality of the proposed method in strong-noise and small-sample scenarios. This study provides an efficient solution for the intelligent fault diagnosis of complex industrial equipment.

The Grouping Genetic Algorithm with Controlled Gene Transmission (GGA-CGT) is among the most effective algorithms for solving the one-dimensional Bin Packing Problem (1D-BPP), a classical NP-hard combinatorial optimisation problem with numerous industrial and logistical applications. This study aims to identify the characteristics that enable a mutation operator to perform better within this algorithm by implementing five state-of-the-art mutation operators: Elimination, Merge &amp; Split, Swap, Insertion, and Item Elimination. Performance was evaluated in terms of the number of optimal solutions obtained. Our findings indicate that the GGA-CGT performs best with the least disruptive operators and that both the gene selection strategy and the item selection strategy can enhance the performance of mutation operators. These findings were validated by redesigning and improving a state-of-the-art item-oriented operator, achieving a 26% improvement over the best baseline version of the same operator.

C++ is a system-level programming language for modern software development, which supports multiple programming paradigms, including object-oriented, generic, and functional programming. The intrinsic complexity of these paradigms and their interactions grants C++ powerful expressiveness while posing significant challenges for compilers in correctly implementing its type system. A type system encompasses various aspects such as type inference, type checking, subtyping, type conversions, generics, scoping, and binding. However, systematic testing of the type systems of C++ compilers remains largely underexplored in existing studies. In this work, we present TyMut, the first approach specifically designed to test the C++ type system. TyMut is a mutation-based compiler fuzzer equipped with advanced type-driven mutation operators, carefully crafted to target intricate type-related features such as template generics, type conversions, and inheritance. Beyond differential testing, TyMut introduces enhanced test oracles through a must analysis that partially confirms the validity of generated programs. Specifically, mutation operators are classified into well-formed and not-well-formed: Programs generated by well-formed mutation operators are valid and must be accepted by compilers. Programs generated by not-well-formed operators are validated against a set of well-formedness rules. Any violation indicates the program is invalid and must be rejected. For programs that pass the rules but lack a definitive oracle, TyMut applies differential testing to identify behavioral inconsistencies across compilers. The testing campaign took about 32 hours to generate and test 250584 programs. The must analysis provides definite test oracles for nearly 80% of all generated programs. TyMut uncovered 102 bugs in the recent versions of GCC and Clang, with 56 confirmed as new bugs by compiler developers. Among the confirmed bugs, 26 of them cause compiler crashes, and more than 50% cause miscompilation. Additionally, 7 of them had remained hidden for over 20 years, 22 for over 10 years, and 39 for over 5 years. One long-standing bug discovered by TyMut was later confirmed as the root cause of a real-world issue in TensorFlow. Before submitting this paper, 13 bugs were fixed, most of which were fixed within 60 days. Notably, some unconfirmed bugs have led to in-depth discussions among developers. For instance, one bug led a compiler developer to submit a new issue to the C++ language standard, showing that we uncovered ambiguities in the language specification.

This article presents an innovative approach to optimizing the seismic modeling and analysis of high-rise buildings by automating the process with Python 3.13 and the ETABS 22.1.0 API. The process begins with the collection of information on the base building, a structure of seventeen regular levels, which includes data from structural elements, material properties, geometric configuration, and seismic and gravitational loads. These data are organized in an Excel file for further processing. From this information, a code is developed in Python that automates the structural modeling in ETABS through its API. This code defines the sections, materials, edge conditions, and loads and models the elements according to their coordinates. The resulting base model is used as a starting point to generate an optimal solution using a genetic algorithm. The genetic algorithm adjusts column and beam sections using an approach that includes crossover and controlled mutation operations. Each solution is evaluated by the maximum displacement of the structure, calculating the fitness as the inverse of this displacement, favoring solutions with less deformation. The process is repeated across generations, selecting and crossing the best solutions. Finally, the model that generates the smallest displacement is saved as the optimal solution. Once the optimal solution has been obtained, it is implemented a second code in Python is implemented to perform static and dynamic seismic analysis. The key results, such as displacements, drifts, internal and basal shear forces, are processed and verified in accordance with the Peruvian Technical Standard E.030. The automated model with API shows a significant improvement in accuracy and efficiency compared to traditional methods, highlighting an R2 = 0.995 in the static analysis, indicating an almost perfect fit, and an RMSE = 1.93261 × 10−5, reflecting a near-zero error. In the dynamic drift analysis, the automated model reaches an R2 = 0.9385 and an RMSE = 5.21742 × 10−5, demonstrating its high precision. As for the lead time, the model automated completed the process in 13.2 min, which means a 99.5% reduction in comparison with the traditional method, which takes 3 h. On the other hand, the genetic algorithm had a run time of 191 min due to its stochastic nature and iterative process. The performance of the genetic algorithm shows that although the improvement is significant between Generation 1 and Generation 2, is stabilized in the following generations, with a slight decrease in Generation 5, suggesting that the algorithm has reached its level has reached a point of convergence.

To address the challenges of increasing carbon dioxide (CO2) emissions and climate change caused by the growth of air traffic, accurate prediction of CO2 emissions in civil aviation has become crucial. This study proposes a CO2 emission prediction method based on an improved back propagation (BP) neural network, where the Improved Sparrow Search Algorithm (ISSA) is employed to optimize the hyperparameters of the BP neural network, thereby enhancing the prediction capability for CO2 emissions in civil aviation. To overcome the limitations of the traditional SSA, such as the tendency to fall into local optima during population initialization and the search process, this paper introduces Tent mapping for population initialization and incorporates adaptive t-distribution-based perturbation for individual position updates during the mutation operation, aiming to improve the algorithm’s global search ability and convergence performance. Subsequently, the ISSA algorithm is applied to optimize the weights and biases of the BP neural network, further constructing an ISSA-BP neural network-based prediction model for civil aviation CO2 emissions. Experimental results demonstrate that the improved BP neural network outperforms other comparative models in terms of prediction accuracy and error control, enabling accurate prediction of civil aviation CO2 emissions. This research provides a solid theoretical foundation for formulating precise energy-saving and emission-reduction strategies in civil aviation.

Supply-chain scheduling in modern production systems faces significant challenges due to the complexity of balancing cost efficiency, delivery lead times, and order variability—especially under stochastic demand and uncertain transportation conditions. Conventional metaheuristic approaches, such as the standard Non-dominated Sorting Genetic Algorithm II (NSGA-II), often struggle with premature convergence, limited population diversity, and rigid parameter settings that hinder adaptability in dynamic environments. To address these limitations, this study proposes a novel Hybrid Adaptive NSGA-II (HA-NSGA-II) framework, enhanced by two key innovations: (1) a Global-Intensity Mutation Operator (GIMO) that strengthens population diversity by dynamically adjusting mutation intensity across the entire solution space; and (2) an Adaptive Parameter Self-Tuning (APST) module based on reinforcement learning, which intelligently regulates crossover and mutation rates in response to evolutionary progress and environmental changes. A tri-objective optimization model is formulated to simultaneously minimize total supply chain cost, order quantity variance (mitigating the bullwhip effect), and customer delivery delays. The proposed HA-NSGA-II demonstrates superior convergence and diversity performance compared to traditional NSGA-II, offering a more robust and adaptive solution for complex, real-world supply chain scheduling under uncertainty. The model is tested on a simulated supply chain with 5 factories, 15 warehouses, and 50 customers facing random demand and Gaussian-distributed lead times. The new HA-NSGA-II is compared with traditional NSGA-II, SPEA2, and MOEA/D based on Hypervolume (HV), Inverted Generational Distance (IGD), Spread (Δ), Total Inventory Cost, Bullwhip Index (BI), and Robustness Index (RI). Results of 50 simulation runs demonstrate that HA-NSGA-II surpasses all the baseline algorithms with 12.9% enhancement in HV, 41.1% improvement in IGD, and 13.7% reduction in convergence time. In addition, it delivers a 14.6% supply chain cost reduction, a 20.4% bullwhip effect suppression, and 25% accelerated times of fulfillment. The research concludes that the HA-NSGA-II provides a scalable and effective scheduling approach applicable for real-time Industry 4.0 supply chains. Work in the future involves expanding the framework towards multi-period dynamic environments using IoT and blockchain-based visibility.

This study investigates the application of the Non-dominated Sorting Genetic Algorithm II (NSGA-II) for optimizing multiple conflicting objectives related to student academic performance. Using the Student Performance dataset from the UCI Machine Learning Repository, which contains demographic, behavioral, and academic information of 395 secondary school students, the research aimed to maximize final grades (G3), minimize absenteeism, and maximize study time. The study began with exploratory data analysis, which revealed wide variability in academic outcomes, low average absenteeism, and moderate study time, justifying the selection of these three objectives. NSGA-II was then implemented with a population of 100 individuals across 200 generations, employing crossover and mutation operators to generate Pareto-optimal solutions. The results demonstrated diverse non-dominated solutions, illustrating trade-offs between academic achievement, attendance, and study time. Absenteeism emerged as the most significant negative factor, while study time and school support were positively associated with better outcomes. Unlike conventional regression or classification methods that produce a single prediction, NSGA-II provided a spectrum of optimal alternatives, offering flexibility in policy and decision-making. These findings highlight the relevance of multi-objective optimization in education and emphasize the importance of integrating behavioral, social, and digital dimensions to design adaptive strategies for improving student performance.

A hyper-heuristic enhanced neuro-evolutionary algorithm with self-adaptive operators and various activation functions for classification problems.

Evolutionary computation (EC)-based neural architecture search (NAS) has achieved remarkable performance in the automatic design of neural architectures. However, the high computational cost associated with evaluating searched architectures poses a challenge for these methods, and a fixed form of learning rate (LR) schedule means greater information loss on diverse searched architectures. This paper introduces an efficient EC-based NAS method to solve these problems via an innovative meta-learning framework. Specifically, a meta-learning-rate (Meta-LR) scheme is used through pretraining to obtain a suitable LR schedule, which guides the training process with lower information loss when evaluating each individual. An adaptive surrogate model is designed through an adaptive threshold to select the potential architectures in a few epochs and then evaluate the potential architectures with complete epochs. Additionally, a periodic mutation operator is proposed to increase the diversity of the population, which enhances the generalizability and robustness. Experiments on CIFAR-10, CIFAR-100, and ImageNet1K datasets demonstrate that the proposed method achieves high performance comparable to that of many state-of-the-art peer methods, with lower computational cost and greater robustness.