Evolution Strategy Research Articles

Droplet Microfluidics for High-Throughput Screening and Directed Evolution of Biomolecules.

Directed evolution is a powerful technique for creating biomolecules such as proteins and nucleic acids with tailor-made properties for therapeutic and industrial applications by mimicking the natural evolution processes in the laboratory. Droplet microfluidics improved classical directed evolution by enabling time-consuming and laborious steps in this iterative process to be performed within monodispersed droplets in a highly controlled and automated manner. Droplet microfluidic chips can generate, manipulate, and sort individual droplets at kilohertz rates in a user-defined microchannel geometry, allowing new strategies for high-throughput screening and evolution of biomolecules. In this review, we discuss directed evolution studies in which droplet-based microfluidic systems were used to screen and improve the functional properties of biomolecules. We provide a systematic overview of basic on-chip fluidic operations, including reagent mixing by merging continuous fluid streams and droplet pairs, reagent addition by picoinjection, droplet generation, droplet incubation in delay lines, chambers and hydrodynamic traps, and droplet sorting techniques. Various microfluidic strategies for directed evolution using single and multiple emulsions and biomimetic materials (giant lipid vesicles, microgels, and microcapsules) are highlighted. Completely cell-free microfluidic-assisted in vitro compartmentalization methods that eliminate the need to clone DNA into cells after each round of mutagenesis are also presented.

A new genus and species of nudibranch-mimicking Syllidae (Annelida, Polychaeta)

Nudibranch mollusks, which are well-known for their vivid warning coloration and effective defenses, are mimicked by diverse invertebrates to deter predation through both Müllerian and Batesian strategies. Despite extensive documentation across different taxa, mimickers have not been detected among annelids, including polychaetes, until now. This study described a new genus and species of polychaete living on Dendronephthya octocorals in Vietnam and Japan. Belonging to Syllidae, it exhibits unique morphological adaptations such as a low number of body segments, simple chaetae concealed within the parapodia and large and fusiform antennae and cirri. Moreover, these appendages are vividly colored, featuring an internal dark red area with numerous terminal white spots and bright yellow tips, effectively contributing to mimicking the appearance of a nudibranch. This discovery not only documents the first known instance of such mimicry among annelids, but also expands our understanding of evolutionary adaptation and ecological strategies in marine invertebrates.

The rising SARS-CoV-2 JN.1 variant: evolution, infectivity, immune escape, and response strategies.

The JN.1 variant of COVID-19 has emerged as the dominant strain worldwide since the end of 2023. As a subclade of the BA.2.86 variant, JN.1 harbors a unique combination of mutations inherited from the BA.2.86 lineage, notably featuring the novel L455S mutation within its receptor-binding motif. This mutation has been linked to increased transmissibility and enhanced immune evasion capabilities. During the rise of JN.1, evidence of resistance to various monoclonal antibodies and reduced cross-neutralization effects of the XBB.1.5 vaccine have been observed. Although the public health threat posed by the JN.1 variant appears relatively low, concerns persist regarding its evolutionary trajectory under immune pressure. This review provides a comprehensive overview of the evolving JN.1 variant, highlighting the need for continuous monitoring and investigation of new variants that could lead to widespread infection. It assesses the efficacy of current vaccines and therapeutics against emerging variants, particularly focusing on immunocompromised populations. Additionally, this review summarizes potential vaccine advancements and clinical treatments for COVID-19, offering insights to optimize prevention and treatment strategies. This review thoroughly evaluates the JN.1 variant's impact on public health and its implications for future vaccine and therapeutic development, contributing to ongoing efforts to mitigate the risk of virus transmission and disease severity.

Research on interregional oil cooperation-sanctions with evolutionary game

PurposeThe game strategies differ when different regions participate in the oil game. Under what circumstances will different participants choose cooperation or sanction strategies? This is the core issue of this paper.Design/methodology/approachRegarding the current and future game behavior between different regions in the oil trade, this paper constructs an evolutionary game model between two regions to explore the possibility of sanctions strategies between the two sides in different situations.FindingsThe research finds: (1) When the benefits of in-depth cooperation between the two regions are greater, both sides tend to adopt cooperative strategies. (2) When the trade conflict losses between the two regions are smaller, both sides adopt sanctions strategies. (3) When a strong region trades with a weak region, if the former adopts a sanctions strategy, the net profits are greater than the benefits of in-depth cooperation between the two regions. If the latter adopts a sanctions strategy, the net profits are less than the trade conflict losses between the two regions. There will be the strong region adopting a sanctions strategy and the weak region adopting a non-sanctions strategy. At this time, the latter should reasonably balance the immediate and future interests and give up some current interests in exchange for in-depth cooperation between the two regions. Otherwise, it will fall into the situation of unilateral sanctions by the strong against the weak.Originality/valueThere is no paper in the existing literature that uses the evolutionary game method to analyze the oil game problem between the two regions. This paper constructs a two-party evolutionary game model composed of crude oil importers and crude oil exporters and, based on this, analyzes the evolutionary stability between the two regions under sanctions and cooperation strategies, which enriches the energy research field.

Competitive Game Model and Evolutionary Strategy Analysis of Green Power and Thermal Power Generation

Competitive Game Model and Evolutionary Strategy Analysis of Green Power and Thermal Power Generation

Heterogeneous Evolution of Breast Cancer Cells-An Endogenous Molecular-Cellular Network Study.

Breast cancer heterogeneity presents a significant challenge in clinical therapy, such as over-treatment and drug resistance. These challenges are largely due to its obscure normal epithelial origins, evolutionary stability, and transitions on the cancer subtypes. This study aims to elucidate the cellular emergence and maintenance of heterogeneous breast cancer via quantitative bio-process modeling, with potential benefit to therapeutic strategies for the disease. An endogenous molecular-cellular hypothesis posits that both pathological and physiological states are phenotypes evolved from and shaped by interactions among a number of conserved modules and cellular factors within a biological network. We hereby developed a model of core endogenous network for breast cancer in accordance with the theory, quantifying its intrinsic dynamic properties with dynamic modeling. The model spontaneously generates cell states that align with molecular classifications at both the molecular and modular level, replicating four widely recognized molecular subtypes of the cancer and validating against data extracted from the TCGA database. Further analysis shows that topologically, a singular progression gateway from normal breast cells to cancerous states is identified as the Luminal A-type breast cancer. Activated positive feedback loops are found to stabilize cellular states, while negative feedback loops facilitate state transitions. Overall, more routes are revealed on the cellular transition between stable states, and a traceable count explains the origin of breast cancer heterogeneity. Ultimately, the research intended to strength the search for therapeutic targets.

Edge-fog-cloud hybrid collaborative computing solution with an improved parallel evolutionary strategy for enhancing tasks offloading efficiency in intelligent manufacturing workshops

Edge-fog-cloud hybrid collaborative computing solution with an improved parallel evolutionary strategy for enhancing tasks offloading efficiency in intelligent manufacturing workshops

Novel optimal sensor placement method towards the high-precision digital twin for complex curved structures

The complex shape of the structure and the new needs for high-precision in digital twin modeling pose challenges for sensor placement optimization. A novel optimal sensor placement towards the high-precision digital twin (OSP-HDT) method is proposed for complex curved structures. It comprises three key aspects. Firstly, leveraging the spatial dimensionality reduction method, the complex curved surface is simplified into a planar representation. Subsequently, candidate sensor placement points can be easily identified by dividing the background mesh in the plane and screening them within the curved surface. These candidate points are then binary encoded to facilitate the subsequent optimization. Secondly, the method collects result data from the finite element model, treating it as virtual sensor data. Using this data, a surrogate model is constructed and then the objective function is formulated based on both the global and local critical areas precision of the surrogate model. Thirdly, the sensor placement optimization model is constructed, followed by optimization design using the efficient multi-objective covariance matrix adaptive evolutionary strategy. Through the steps above, the optimal sensor placement can be identified. To validate the proposed OSP-HDT method, an experiment is conducted on an S-shaped variable cross-section stiffened shell, with the construction of the corresponding digital twin. Compared to the uniform placement with an equivalent number of sensors, the OSP-HDT method demonstrated a significant 9.0% improvement in global precision and a remarkable 62.1% enhancement in local precision of critical areas. Furthermore, when compared to the random sensor placement strategies, the OSP-HDT method exhibited a 20.5% increase in global precision, together with a 44.2% increase in the local precision. Notably, even when compared to the full sensor placement, the OSP-HDT method can maintain comparable local precision, while significantly reducing the number of sensors by 77.6%. The above comparison indicates that the proposed OSP-HDT method can build a digital twin model with higher global and local precision for complex structures.

Optimization of NOX emissions of a CRDI DIESEL engine using CMA-ES method

Engine calibration is the tuning of embedded parameters in the engine control unit (ECU) software to improve vehicle characteristics and meet legal requirements. Due to the stricter emission limits and rising customer expectations, current ECU software may include variables up to 30,000, which require very much time for engine calibration development. For this reason, automotive manufacturers continuously develop mathematical-based optimization methods to find optimum operating conditions for the engines. This study aimed to develop an online optimization algorithm to conduct automated dynamometer tests in the calibration development process. A modified covariance matrix adaptation (CMA) algorithm, which is an evolutionary strategy (ES) method belonging to meta-heuristic optimization, was integrated with an automation system for online calibration optimization. Some CMA method parameters such as step size and damping factor were initially revised to achieve the method to function efficiently in online engine calibration. Optimization was conducted at three different operating points of a 2-liter common rail direct injection (CRDI) diesel engine, where NOx emission mainly impacts the New European Driving Cycle (NEDC) results. The main injection timing, rail pressure, pilot injection quantity and timing, manifold pressure, and mass air flow were controlled in the optimization process. Optimization targets were determined according to the NOx-PM Pareto curve for each operating point. Covariance matrix adaptation was used to generate Pareto curves. Sixty-five measurements were taken for each operating point in the optimization process. Once optimization targets were determined, optimization occurred at each operating point. A total NOx emission reduction of 3.8% was obtained in the NEDC test, while fuel consumption and PM remained almost the same at steady-state operating points. The modified CMA-ES algorithm is expected to be an efficient method for online calibration optimization.

Coordination Mechanisms for Traceable Products in the Context of Live Streaming E-Commerce from the Perspective of Platform, Anchor, and Consumer

The meteoric rise of live streaming e-commerce, an emergent form of electronic commerce, is accompanied by frequent occurrences of “live streaming mishaps”. Live streaming’s visual nature offers a more varied and intuitive method for showcasing traceable products, thereby enhancing customer engagement and contracting customers. However, due to the diverging interests, conflicts and contradictions among the stakeholders in traceable products persist, introducing challenges and complexities to its management. We develop an evolutionary game-theoretical model involving three key entities: the platform, anchors (live stream hosts), and consumers. Through the establishment of a payment matrix, we analyze their interactive behaviors and equilibrium states. Additionally, we examine the impacts of key factors on the system and find that platforms and anchors play a pivotal “dual-core” role in traceable product management within the live streaming e-commerce landscape. Moreover, factors such as market attention, anchor characteristics, and consumer dissatisfaction emerge as critical determinants for effective traceable product management. This research contributes valuable insights into enhancing coordination of traceable products in live streaming e-commerce. It offers perspectives on the evolutionary dynamics and strategies for selecting traceable products by platforms, broadcasters, and consumers.

Trilateral Evolutionary Game Strategy for the Design Optimization of Engineering General Contracting Projects in the Chinese Context

The engineering general contracting mode is an advanced engineering transaction mode, and design optimization is one of the important driving forces for vigorously promoting the general contracting mode. The application proportion of the general contracting mode in infrastructure projects is not high, the number of successful projects is not large, and the implementation effect is not ideal. One of the main reasons is that the design optimization theory and practice of international standard general contracting projects cannot adapt to the general contracting projects in the Chinese context, making local general contracting projects face huge challenges such as low enthusiasm for design optimization from all parties and unsatisfactory design optimization effects. Therefore, under the premise of bounded rationality, when the owner adopts control methods of different intensities, an evolutionary game study on the selection of design optimization strategies between the design and construction parties is carried out, and stability control strategies are proposed through case experiments and simulations. The research results indicate the following: firstly, the design optimization of general contracting projects in the Chinese context is feasible, but it depends on the distribution ratio of benefits obtained from the design optimization. Compared with general civil construction general contracting projects, the design optimization allocation ratio of industrial construction general contracting projects is more significant; secondly, the mixed control method of strong control and weak control is the optimal choice for the owner of the general contracting project; and thirdly, there are multiple evolutionary stable points, and mechanism design or incentive measures should be used to guide owners to choose weak control strategies, while design and construction parties should choose their design optimization strategies. The research results provide a reference for owners to determine the proportion and scheme of design optimization allocation, and for construction parties to determine design optimization strategies.

An evolutionary game model for indirect data sharing in manufacturing big data consortium

In the context of the explosive growth of manufacturing data, a manufacturing big data consortium is a collaborative organization aimed at promoting efficient decision-making and technological innovation. It is of vital significance to study how to effectively promote the data sharing among the members of manufacturing big data consortium. In this paper, we propose an evolutionary game model to investigate how to facilitate data sharing in the manufacturing big data consortium. Then, we analyze the synergistic revenue allocation method based on the indirect data sharing mode, and calculate the probability of the ideal event. We explore the influencing factors of data sharing in the manufacturing big data consortium by simulating Evolutionary Stable Strategies. Our findings reveal that the initial willingness of participants has a profound impact on their sharing behavior. Furthermore, the optimal interval revenue allocation coefficient can improve the probability of the manufacturing big data consortium to share data. Factors such as incentive revenue and social penalties positively influence the evolution towards active sharing strategies. In contrast, factors such as speculative revenue and data sharing costs inhibit the evolution towards active sharing strategies.

Deep Learning-Based Multifidelity Surrogate Modeling for High-Dimensional Reliability Prediction

Abstract Multifidelity surrogate modeling offers a cost-effective approach to reducing extensive evaluations of expensive physics-based simulations for reliability prediction. However, considering spatial uncertainties in multifidelity surrogate modeling remains extremely challenging due to the curse of dimensionality. To address this challenge, this paper introduces a deep learning-based multifidelity surrogate modeling approach that fuses multifidelity datasets for high-dimensional reliability analysis of complex structures. It first involves a heterogeneous dimension transformation approach to bridge the gap in terms of input format between the low-fidelity and high-fidelity domains. Then, an explainable deep convolutional dimension-reduction network (ConvDR) is proposed to effectively reduce the dimensionality of the structural reliability problems. To obtain a meaningful low-dimensional space, a new knowledge reasoning-based loss regularization mechanism is integrated with the covariance matrix adaptation evolution strategy (CMA-ES) to encourage an unbiased linear pattern in the latent space for reliability prediction. Then, the high-fidelity data can be utilized for bias modeling using Gaussian process (GP) regression. Finally, Monte Carlo simulation (MCS) is employed for the propagation of high-dimensional spatial uncertainties. Two structural examples are utilized to validate the effectiveness of the proposed method.

Sub-genomic RNAi-assisted strain evolution of filamentous fungi for enhanced protein production.

Genetic engineering at the genomic scale provides a rapid means to evolve microbes for desirable traits. However, in many filamentous fungi, such trials are daunted by low transformation efficiency. Differentially expressed genes under certain conditions may contain important regulatory factors. Accordingly, although manipulating these subsets of genes only can largely reduce the time and labor, engineering at such a sub-genomic level may also be able to improve the microbial performance. Herein, first using the industrially important cellulase-producing filamentous fungus Trichoderma reesei as a model organism, we constructed suppression subtractive hybridization (SSH) libraries enriched with differentially expressed genes under cellulase induction (MM-Avicel) and cellulase repression conditions (MM-Glucose). The libraries, in combination with RNA interference, enabled sub-genomic engineering of T. reesei for enhanced cellulase production. The ability of T. reesei to produce endoglucanase was improved by 2.8~3.3-fold. In addition, novel regulatory genes (tre49304, tre120391, and tre123541) were identified to affect cellulase expression in T. reesei. Iterative manipulation using the same strategy further increased the yield of endoglucanase activity to 75.6 U/mL, which was seven times as high as that of the wild type (10.8 U/mL). Moreover, using Humicola insolens as an example, such a sub-genomic RNAi-assisted strain evolution proved to be also useful in other industrially important filamentous fungi. H. insolens is a filamentous fungus commonly used to produce catalase, albeit with similarly low transformation efficiency and scarce knowledge underlying the regulation of catalase expression. By combining SSH and RNAi, a strain of H. insolens producing 28,500 ± 288 U/mL of catalase was obtained, which was 1.9 times as high as that of the parent strain.IMPORTANCEGenetic engineering at the genomic scale provides an unparalleled advantage in microbial strain improvement, which has previously been limited only to the organisms with high transformation efficiency such as Saccharomyces cerevisiae and Escherichia coli. Herein, using the filamentous fungus Trichoderma reesei as a model organism, we demonstrated that the advantage of suppression subtractive hybridization (SSH) to enrich differentially expressed genes and the convenience of RNA interference to manipulate a multitude of genes could be combined to overcome the inadequate transformation efficiency. With this sub-genomic evolution strategy, T. reesei could be iteratively engineered for higher cellulase production. Intriguingly, Humicola insolens, a fungus with even little knowledge in gene expression regulation, was also improved for catalase production. The same strategy may also be expanded to engineering other microorganisms for enhanced production of proteins, organic acids, and secondary metabolites.

Analysis of Shared Parking Game Model under Dynamic Parking Pricing

In recent years, the problem of "parking difficulty" has led to a large number of illegal parking incidents. The shared parking mode has been considered as an effective way to alleviate the conflict between parking supply and demand and urban traffic pressure. To address the issue of illegal parking and promote the development of shared parking, this paper constructs an evolutionary game model with shared platforms and motor vehicle drivers as the main entities. The study investigates the evolutionary stability strategy of the model, conducts sensitivity analysis on model parameters, and further analyzes the impact of highly sensitive parameters on the evolutionary paths of both players in the game. Finally, numerical simulations are performed on dynamic parking pricing standard. The research findings demonstrate that the sensitivity of discounts received by drivers from shared platforms and the additional revenue gained by the shared platform is higher than that of other parameters. Moderately increasing the penalty for illegal parking and the additional revenue of the shared platform can encourage drivers to choose legal parking and promote the development of shared parking. Under given parameterized and periodic parking pricing standard, finally, according to the particle swarm optimization algorithm, a set of relatively optimal parameter values is derived to enable the model to evolve rapidly into a stable state where drivers choose to park legally and the shared platform selects surrounding parking lots. It can effectively reduce the frequency of illegal parking.

Ship Autonomous Berthing Strategy Based on Improved Linear-Quadratic Regulator

There has been significant interest in the research field of ship automatic navigation, particularly in the area of autonomous berthing. To address the key challenges of path planning and control during ship berthing, we propose an enhanced Linear−Quadratic Regulator (LQR) control approach, reinforced by the Covariance Matrix Adaptation Evolution Strategy (CMA−ES), along with an adaptive berthing strategy decision model. This integrated framework encompasses ship motion control, path planning, and berthing strategy selection to facilitate adaptive and autonomous ship berthing. Initially, a dynamic mathematical model of ship motion is established, taking into account wind and current interference effects. Subsequently, an adaptive environment−aware berthing strategy model is introduced to enable automatic selection of berthing strategies based on spatial relationships between environmental factors and the berth. By utilizing the refined LQR method, autonomous motion control for ship berthing is achieved. To validate the effectiveness of our controller, comprehensive simulation analyses are conducted under varying operating conditions to encompass crucial factors such as large drift angle characteristics of ships, shallow water effects, and bank effects across seven diverse working conditions. The simulation results underscore the robustness of our proposed method in responding to environmental interference while demonstrating its capability to select appropriate berthing strategies based on varying operational scenarios.

Salinity acclimation of nitrifying microorganisms: Nitrification performance, microbial community, osmotic adaptation strategies

Wastewater with high salinity (> 1%) presents a significant challenge to conventional wastewater treatment, particularly for the nitrification process. However, the osmotic adaptation strategies of nitrifying microorganisms remain poorly understood. In this study, we examined the impacts of salinity on the ammonia and nitrite oxidation processes in wastewater. The biofilm samples without salinity acclimation (0 g NaCl/L), after 1% salinity acclimation (10 g NaCl/L), and after 3% salinity acclimation (30 g NaCl/L) were inoculated to conical flasks containing synthetic high-salt wastewater (30 g NaCl/L), respectively. The research findings indicate that, following the salinity acclimation of biofilm, the activity of ammonia oxidation surpassed that of nitrite oxidation. 16S rRNA gene amplicon analysis revealed a noteworthy increase in the abundance of Nitrosomonas (ammonia-oxidizing bacteria) and an unclassified ammonia-oxidizing archaeon within the Nitrososphaeraceae family. In contrast, Nitrospira (nitrite-oxidizing bacteria) exhibited a significant decrease (p < 0.01). Metagenomic analysis indicates certain strains, such as Nitrosomonas sp. PL2, Nitrosomonas mobilis PL3, and Nitrososphaeraceae gen. sp. PL5, possessed various genes related to Na+ efflux, K+ uptake, glutamate synthesis or transport. However, Nitrospira sp. PL6 and Nitrospira sp. PL7 lacked K+ uptake genes. This study elucidates the microbial mechanisms underlying the variations in nitrification observed before and after salinity acclimation of biofilm, which helps to develop microbial evolution strategies to remove nitrogen pollutants under high salinity conditions.

Reconfigurable inverse design of phononic crystal sensor based on a deep learning accelerated evolution strategy

Phononic crystals offer valuable sensing capabilities due to their high sensitivity to changes in sound velocity of analytes. In this work, a reconfigurable inverse design approach of phononic crystal sensors is achieved using a deep learning accelerated evolution strategy. The training data is acquired through finite element method (FEM). Two multilayer perceptrons (MLP) are constructed and trained to predict the center frequency and bandwidth of a passband in the dispersion relation. Utilizing a two-step training approach enables rapid accuracy enhancements and swift reconstruction of network targets from NaCl to KCl solutions. The trained networks accelerate the optimization process, yielding a phononic crystal with good detection ability. Compared to FEM, invoking the trained networks can reduce optimization time by a factor of 105. The optimized and initial structures are both fabricated and experimentally tested. The robust linear relation between the resonant peak and the solution concentration indicates significant sensing value. The experimental results are in good agreement with the FEM simulations. In the detection of NaCl solution, the optimized phononic crystal sensor has a sensitivity increase of 375% and a Q-factor of 999%. Our research demonstrates that the data-driven deep learning network is a very powerful tool for the design and optimization of phononic crystal devices.

Integrating a dimensional perturbation module into exponential distribution optimizer for solving optimization problems

Exponential distribution optimizer (EDO) is a recently proposed optimization technique that is based on the exponential probability distribution model. As most swarm intelligence algorithms (SIAs), EDO is very good at handling common optimization problems. However, when addressing multimodal and center-bias problems, it exhibits limitations like imbalanced global and local search capabilities and poor solution accuracy. Building upon this, we introduce an improved version of exponential distribution optimizer that incorporates a dimensional perturbation module (DPM) called EDO_DPM. By testing the impact of dimensional perturbation, the K-dimensional evolution strategy is adopted in our approach. Concurrently, to modulate the usage frequency of DPM, an adjustment parameter, named evolutionary state factor (ESF), is designed contingent on the population’s evolutionary state. Moreover, to alleviate the deficiency of non full-dimensional perturbation, a bounce out operation is embedded in the algorithm. The efficacy of EDO_DPM has been substantiated through testing on 20 different types of functions and CEC2017 benchmark suite. Comparative analyses with the state-of-the-art algorithms have demonstrated a marked enhancement in EDO_DPM’s capability to manage multimodal problems and its proficiency in resolving center-bias problems. Meanwhile, it can achieve excellent results on CEC2017 by comparing excellent improved variants. Furthermore, EDO_DPM is applied to the problem of parameter estimation for photovoltaic (PV) generation systems. Comparison results show that EDO_DPM has good application capability.

A Student Performance Prediction Model Based on Hierarchical Belief Rule Base with Interpretability

Predicting student performance in the future is a crucial behavior prediction problem in education. By predicting student performance, educational experts can provide individualized instruction, optimize the allocation of resources, and develop educational strategies. If the prediction results are unreliable, it is difficult to earn the trust of educational experts. Therefore, prediction methods need to satisfy the requirement of interpretability. For this reason, the prediction model is constructed in this paper using belief rule base (BRB). BRB not only combines expert knowledge, but also has good interpretability. There are two problems in applying BRB to student performance prediction: first, in the modeling process, the system is too complex due to the large number of indicators involved. Secondly, the interpretability of the model can be compromised during the optimization process. To overcome these challenges, this paper introduces a hierarchical belief rule base with interpretability (HBRB-I) for student performance prediction. First, it analyzes how the HBRB-I model achieves interpretability. Then, an attribute grouping method is proposed to construct a hierarchical structure by reasonably organizing the indicators, so as to effectively reduce the complexity of the model. Finally, an objective function considering interpretability is designed and the projected covariance matrix adaptive evolution strategy (P-CMA-ES) optimization algorithm is improved. The aim is to ensure that the model remains interpretable after optimization. By conducting experiments on the student performance dataset, it is demonstrated that the proposed model performs well in terms of both accuracy and interpretability.