Research on Network Resources Integration of University English Blended Learning Model Based on Multi-Objective Optimization

The development of educational technology has led to teaching no longer a single traditional classroom instruction but a variety of learning modes. Hybrid teaching, is through the combination of different modes of instruction to provide the most effective teaching experience [1-3]. Through the continuous intervention of the Internet in the field of education, it gradually breaks the limitations of students’ learning time and space. Essentially, blended learning is defined as a distance learning approach that uses technologies such as the Internet, video conferencing, and email in conjunction with traditional education or training [4-7]. It aims to combine the best of classroom teaching with the best of online teaching. Even if the blended learning model is defined as a combination of classroom teaching and online learning, it can be seen that the essence of the blended learning model is to break the single mode of teaching and improve the quality of teaching and learning by using different teaching methods.

The blended learning mode is a product of the development of the times and the progress of education, which can not only stimulate students’ learning motivation but also meet the teaching goals of today’s society [8-9]. The application of blended learning mode in college English writing teaching with the help of online resources is of great significance, which is mainly reflected in the following two aspects: on the one hand, theoretically speaking, the blended learning mode is a learning mode combining “online + classroom”, which can to a certain extent improve students’ interest in learning and their English proficiency [10-12]. In this teaching mode, teachers should play the roles of guidance, encouragement, and supervision to gradually change students’ passive attitude toward English learning. On the other hand, under the background of the era of education informatization, network data has been widely used in the field of education. Its advantages are reflected in the promotion of real-time communication between teachers and students and the improvement of students’ learning initiative and enthusiasm, which is of great significance in improving students’ English proficiency [13-15]. Of course, the effective integration of network resources into the university English teaching mode to promote the improvement of college students’ English proficiency and to effectively avoid students putting the cart before the horse because of excessive indulgence in the network is worthy of in-depth study.

With the rapid development of information technology, the problem of network resource allocation has become more and more important. Network resource allocation involves many objective functions, such as network bandwidth utilization, transmission delay, data throughput, etc [16-17]. The emergence of multi-objective optimization methods provides strong support for the solution of this problem. The basic idea of multi-objective optimization methods is to optimize multiple objective functions simultaneously [18-19]. These objective functions are often contradictory and incompatible. The application of the multi-objective optimization method involves many fields, such as machine learning, supply chain management, intelligent control, and so on [20-21]. In network resource allocation, its application is particularly important.

Relying on the multi-objective optimization technology to realize the efficient integration of network resources under the blended learning mode of university English, a multi-objective optimization model of grouping papers is constructed, and the resources are fully integrated and appropriately applied in English teaching. Set the knowledge points of the English exam and construct the multi-objective optimization function of the intelligent group of the English exam to optimize the quality of the group paper. In order to ensure the reliability and validity of the test paper, we formulate several principles of test paper organization, and at the same time, we formulate the constraints of paper organization, such as the score points corresponding to the English knowledge points, the difficulty distribution and the structure of the test paper. Using a genetic algorithm to complete the model of this paper, determine the code and initial population of chromosomes, design the fitness function of the genetic algorithm, select chromosomes through the set selection probability, crossover, and mutation operations on chromosomes, and finally complete the English paper organization. Simulation experiments are carried out on the standard genetic algorithm and the improved genetic algorithm of this paper’s model, respectively by using a simulated standard test bank to verify the performance of this paper’s model for paper grouping. Carry out English teaching practice, set up experimental class and control classes, and analyze the effective assistance of this paper’s model for English teaching.

2

Multi-objective optimization model for grouping papers based on the integration of network resources

Traditionally, an English course is a syllabus and a textbook, and the process of course implementation is that teachers teach the content of the textbook according to the course plan [22]. Modern multi-objective optimization technology solves the problems of recording, storing, transmitting, displaying, and processing large amounts of information in the integration of network resources and provides a brand-new concept and technology for the preparation of English group volumes, the main goal in the integration of English teaching resources. In the implementation of the course, teachers can integrate and make full use of the network resources already available in the course according to their teaching needs and organize more flexible and efficient English exam papers. In this paper, we will integrate and make full use of the online English teaching resources under the blended learning mode of university English to realize the efficient grouping of English exams and avoid idle waste or repeated development of online teaching resources.

2.1

Multi-objective optimization for intelligent grouping of English exams

Set English Exam Knowledge Point Attribute φ₁(y) to: 1 $φ_{1} (y) = \frac{1}{m} \sum_{j = 1}^{m} B_{j}$

where the total number of English test questions for the group paper task is m, the proportion of knowledge points covered by the jnd English test question is B_j, $B_{j} = \frac{S_{j}}{S_{totol}} B_{j}$ , S_j is the number of knowledge points covered by the jth English test question, and S_total is the total number of knowledge points.

The difficulty attribute can describe the difficulty of English test questions φ₂(y): 2 $φ_{2} (y) = \frac{1}{m} \sum_{j = 1}^{m} \exp [- π {(y_{j} - ϕ)}^{2}]$ 3 $Q_{1} = 1 - \frac{{\bar{Y}}_{K} + {\bar{Y}}_{Z}}{2 U}$ 4 $Q_{2} = 1 - \frac{G}{P}$

Where Q₁, Q₂ is the difficulty of the subjective and objective English test questions respectively, the average value of the scores of the high group and low group is ${\bar{Y}}_{K}, {\bar{Y}}_{Z}$ . The total score of this English test question and the number of people who answered this question correctly is U, G in that order, the number of people who participated in the English test is P. The difficulty value set by the organizer of the paper is ϕ, the value of ϕ is larger, the difficulty of the English test question is high, and y_j indicates the difficulty level of the jth test question.

The Distinctiveness attribute determines the level of scoring for English questions. The differentiation attribute is set to φ₃(y): 5 $φ_{3} (y) = \frac{1}{m} \sum_{j = 1}^{m} \exp [- π {(y_{j} - γ)}^{2}]$ 6 $y_{j} = \frac{{\bar{Y}}_{K} - {\bar{Y}}_{Z}}{U}$

Among them is γ group paper, where people set the differentiation value. The larger its value, the English test score high and low gap is more significant.

The time expectation property can describe how much time a set of English exams and the completion time of the English exam is closer to the time set by the group paper. Then, the time quality of the intelligent grouping of English exams is better. Test paper time expectation attribute φ₄(y) specific settings are: 7 $φ_{4} (y) = {\begin{matrix} \sum_{j = 1}^{m} \frac{H_{j}}{H_{t o t a l}}, \sum_{j = 1}^{m} \frac{H_{j}}{H_{t o t a l}} \leq 1 \\ 0, \sum_{j = 1}^{m} \frac{H_{j}}{H_{t o t a l}} > 1 \end{matrix}$

Where the expected time taken by the jst English test question is H_j, H_j value is set by the organizer, and H_total is the overall test time. In order to optimize the quality of the group, the model constructed needs to be centered on the above attributes, and the multi-objective optimization function for constructing the English exam smart group is: 8 $φ (y) = \max (\sum_{j = 1}^{4} ϖ_{j} φ_{j} (y))$

where the weight of each subfunction is ϖ_j.

2.2

Principles of examination paper organization and development of constraints

2.2.1

Principles for the organization of examination papers

In order to ensure the credibility and validity of the test paper, the process of organizing the paper should abide by the following rules. 1)

The questions should try to cover all the knowledge points, and in the same paper, repeated testing of the same knowledge points should be avoided.

2)

According to the type of questions and the difficulty of the knowledge points and other parameters make the appropriate distribution of marks.

3)

The question paper should not be too difficult or too easy to ensure normal differentiation of the question paper.

4)

The question paper should have a total mark limit and a time limit for answering the questions.

5)

The same question should not be repeated in the question paper.

2.2.2

Test paper development constraints

The test paper metrics involved in this paper are based on the following conditions. It is assumed that B is a matrix of test set attributes in the form shown in (9): 9 $B = [\begin{matrix} b_{11} & b_{12} & b_{13} & b_{14} \\ b_{21} & b_{22} & b_{23} & b_{24} \\ \dots & \dots & \dots & \dots \\ b_{m 1} & b_{m 2} & b_{m 3} & b_{m 4} \end{matrix}]$ 1)

Knowledge Points Corresponding to Score Points Here, set I₁ is used to represent the key knowledge, and set I₂ is used to represent the secondary knowledge, and the corresponding score ranges are [L₁, u₁] and [L₂, u₂]. The constraints on the scores can then be expressed as Eq. (10). 10 ${\begin{array}{l} L_{1} \leq \sum C_{2 i} b_{i 2} \leq u_{1} \\ L_{2} \leq \sum C_{3 i} b_{i 2} \leq u_{2} \end{array}$

This constraint specifies the score requirements that should be met for different levels of knowledge. The composition of the focus and sub-focus knowledge sets and the specific range of score values are determined by the proposing instructor. 2)

Difficulty distribution: the expectation of the difficulty distribution for the different levels should conform to a normal distribution.

3)

The structure of the examination paper, the total marks of the examination paper, and the question types of the examination paper together constitute the structure of the examination paper. (1)

Total marks of the paper

The total score of the examination paper can be calculated by formula (11) in general. In the process of organizing the paper, the total marks of the paper will be given by the teacher of the question paper: 11 $K = \sum_{i = 1}^{\infty} b_{i 2}$

In the above formula, the total number of questions in the test is represented by m and the mark for question i is represented by b_i2.

(2)

Test paper questions

The demand used to set the corresponding question types that should comply with the total mark constraint is expressed by formula (12): 12 $Q_{t} = \sum_{i = 1}^{m} C_{4 i} \cdot b_{i 2}$

Where, Q_t denotes the score of the tnd question type, and the value of this variable is given by the proposing teacher.

(3)

Coverage of knowledge points

The method of calculating the coverage of knowledge points is expressed in equation (13): 13 $R = \frac{The knowledge already included in the test set}{The sum total of knowledge that the test should contain} \geq r$

Where R denotes the coverage of knowledge points, the maximum value of R is equal to 1 when the knowledge points contained in the test set cover the entire syllabus, and generally, r is taken to be more than 0.8.

2.3

Optimal solution solving for group volume based on improved genetic algorithm

In this paper, the biological evolution process is simulated by designing the test questions, test papers and constraints as genes and chromosomes in the genetic algorithm [23]. In the operation of the genetic algorithm through the process of population initialization, chromosome coding, crossover operation, mutation operation, and other processes to complete the process of population diversification and optimization and ultimately achieve the survival of the fittest. 1)

Chromosome coding

In the application of genetic algorithms, the coding of chromosomes should be solved first. Binary coding as a commonly used coding method, with 1 indicates the selection of test questions, 0 indicates not selected. The coding method is simple and clear, easy to color crossover and mutation operation afterwards. In this paper, this coding method will not be chosen and another real number coding method will be used.

The real number coding method directly uses the question number of the test to indicate the selected gene in the chromosome. The question number that appears is the selected test. The length of the encoding is the set of question numbers of the selected questions, and this encoding reduces the length of the encoding string significantly. As a result, the encoding and decoding time will be reduced, and the efficiency of paper organization will be improved. In this paper, this coding method is used when the genes are arranged in the order of question types so that the successive genes are composed of test questions of the same question type.

2)

Determination of initial population

In the traditional genetic algorithm, the population is generated by random form. In this paper, in order to improve the efficiency of the system of grouping papers, for the initial population set certain initial constraints. In this paper, the initial population (i.e., the test paper) needs to meet the conditions of the total score, the total number of test questions, the total number of questions, and the number of questions in each question type. This can effectively reduce the late chromosome crossover and the number of iterations of the mutation, which is conducive to overcoming the early convergence problem of the genetic algorithm.

3)

Fitness function determination

The good ring of the fitness function design directly affects the iterative execution of the algorithm [24]. When the individual has the appropriate fitness value can be considered that the individual has reached the optimization. Usually, the fitness function is directly related to the objective function. The higher the fitness, the closer the individual is to the target paper. The setting of the fitness is directly related to the objective function. The formula for the fitness function of the genetic algorithm in this paper is (14): 14 $f = 1 / (1 + \sum_{i = 1}^{3} r_{i} | e_{i} |)$

4)

Elite retention strategy

In order to avoid the destruction of excellent individuals in the process of evolution, we adopt the elite retention strategy for some individuals with the highest adaptability in the current population. That is to say, in the current population, some individuals with the highest adaptability do not participate in the selection, crossover, mutation, and other operations of the population. These individuals are used to replace the individuals with the lowest fitness values in the population after selection, crossover, and mutation. Therefore, it is guaranteed that the overall fitness value of the new population is higher than that of the previous population.

5)

Selection

The purpose of selection is to increase the number of excellent individuals in the population so that the offspring chromosomes have high fitness values. In this paper, combined with the actual demand situation of the group volume, the roulette strategy is selected to carry out selection operations on the population. The specific operation process is as follows. (1)

Calculate the total fitness value of the population F. The formula is as follows: 15 $F = \sum_{i = 1}^{m} f_{i}$

where m is the number of individuals in the population.

(2)

Calculate the probability of an individual being selected with the following formula: 16 $p_{i} = f_{i} / F$

(3)

Construct a roulette wheel and calculate the cumulative probability: 17 $K_{i} = \sum_{i = 1}^{m} P_{i}$

(4)

Select individuals by roulette wheel to generate a random number between [0, 1] q. When q ≤ K₁, then select individual B₁. When K₁ < q ≤ K₂, then select individual U₂.

Repeat the above steps to generate N population individuals.

6)

Crossover operation

The individuals in the population are teamed up in two pairs. If the population is M, then M/2 pairs of individuals will be generated. Set the crossover probability of the genes in the chromosome (Pc in this paper) and specify the crossover conditions so that the crossover can be completed and a new chromosome can be generated. In this paper, the crossover probability of chromosomes adopts a dynamic adjustment strategy. The specific crossover probabilities are taken as follows: 18 $P c = {\begin{matrix} L_{1} & f < f_{a v g} \\ L_{2} + (L_{1} - L_{2}) (f_{\max} - f) / (f_{\max} - f_{\arg}) & f \geq f_{a v g} \end{matrix}$

Where f_max denotes the value of the largest fitness function in the population, f_avg denotes the value of the average fitness function of the population in the generation, and f denotes the value of the fitness function of the larger of the two individuals in the crossover operation. L₁ denotes the larger of the two individuals in the fitness function, L₁ = 0.9, L₂ denotes the smaller of the two individuals in the fitness function, and L₂ = 0.65. Since the situation of the population in each generation is constantly changing, the above equation can be used to dynamically describe the crossover probability between individuals in each generation of the population.

When f < f_avg, i.e., when the individual is less adapted, then more gene crossover operations are required to improve the individual. So, the probability of chromosome crossover is the larger value of L₁.

When f ≥ f_avg, i.e., when the individual adaptation value is high, then more genes need to be retained for that individual, so the crossover probability is smaller.

When f_max − f is close to 0, i.e., when the value of individual fitness is high and the genes are good, the crossover probability is smaller, and the value of Pc is L₂. When the value of individual fitness is very small, and the genes are poor, the crossover probability needs to be increased at this time, and the value of Pc is L₁. Through this operation, it helps to avoid the problem of precocious maturation of the individual, and to accelerate the convergence of the population by dynamically adjusting crossover arithmetic and other forms of stabilizing the evolution.

7)

Mutation

The mutation is an important way to produce new individuals, mainly by replacing genes at certain positions in the chromosome with other alleles according to probability P_m to produce new individuals. Commonly used mutation methods mainly include basic position mutation, insertion mutation, displacement mutation, and so on.

The value of P_m needs to be careful. Too small is not conducive to the generation of new individuals in the population, and too large will make the algorithm become a random search algorithm. Usually, in order to protect the diversity of the population, improve the efficiency of population search, and prevent the search from falling into the local optimum, the value of the variation probability is controlled within a small range (usually between 0.001 and 0.1). In this paper, the variation probability is taken according to the following formula: 19 $P m = {\begin{matrix} K_{1} & f < f_{a v g} \\ K_{2} + (K_{1} - K_{2}) (f_{\max} - f) / (f_{\max} - f_{a v g}) & f \geq f_{a v g} \end{matrix}$

K₁ is taken as 0.1, K₂ is taken as 0.01, f is the adaptive degree of the current individual, f_max is the adaptive degree of the most individuals in the current population, and f_avg is the average adaptive degree of the current population. 8)

Algorithm termination conditions

During the operation of the algorithm, the population will change after each iteration. There may be individuals that satisfy the constraints that have already appeared, but the algorithm is still executing, or there may be many iterations after the subgeneration individual fitness value still can not reach the target fitness function value. Reasonable termination conditions of the algorithm are very important to improve the efficiency of grouping. The termination conditions in this paper are as follows. (1)

The number of iterations of the population plan reaches a certain value. The system can be set in the range of 100~200 generations, and the default value of this system is 100 generations.

(2)

The algorithm can be terminated when the ratio of the fitness value of the optimal individual to the fitness value required by the goal reaches 95%.

3

Simulation experiment of English group paper based on the integration of network resources

In order to verify the feasibility of this paper in the integration of network resources under the blended learning mode of university English, simulation experiments are used to test the performance of the multi-objective optimization model of paper grouping constructed in this paper in the grouping of English exams. In this paper, a simulated standard test bank is used to simulate and verify the standard genetic algorithm and the improved genetic algorithm of this paper’s model, respectively. The test bank consists of three sub-banks: multiple-choice questions, judgment questions, and subjective questions, each of which consists of 10,000 questions. The computer randomly generates all the attributes of the test questions in the question bank, and the attributes of the test questions are difficulty, differentiation, number of times of use, completion time, and final exposure time, as mentioned in this paper. The specific environment of this experiment is shown in Table 1.

Table 1.

Hardware environment

Hardware environment	Concrete configuration
Processor	Intel core i3 M350 @ 2.27GHz
Memory	Samsung DDR3 1067MHz 2G
Hard disk	WDC WD3200BEVT-08A23T1 320GB
Operating system	Microsoft Windows Server 2010
Compiler	Microsoft Visual C++ 6.0

By testing and comparing the standard genetic algorithm and the improved genetic algorithm proposed in this paper, the resulting population maximum fitness value situation and the population average fitness value situation are specifically shown in Fig. 1, and Fig. (a) and Fig. (b) are the comparison of the population maximum fitness value and the comparison of the population average fitness value, respectively.

For the results of the simulation verification and further comparison of the grouping quality and execution efficiency using different models and algorithms, the running results of different grouping models and algorithms are specifically shown in Table 2. Combined with the above figure and the data in the table, it can be seen that the use of a random sampling algorithm and backtracking trial method can not quantitatively describe the quality of grouping papers, the running time is too long, and there is the phenomenon of not being able to find a feasible solution. Using the model proposed in this paper, applying the standard genetic algorithm avoids the phenomenon of not finding feasible solutions and can assess the quality of the group paper. Still, the algorithm is very easy to enter the local optimal solution and enter the precocious convergence trap. The average fitness value rises slowly, and the quality of the group does not improve obviously, resulting in the overall evolution of slow. The improved algorithm proposed in this paper will overcome the defects of the standard genetic algorithm in the application of grouping, greatly improve the execution efficiency of the intelligent grouping task, and can quantitatively assess and control the quality of grouping. And there is basically no big difference in the running time between the improved algorithm and the standard genetic algorithm.

Table 2.

The results of the model and the algorithm

-	Maximum adaptive value	Maximum average adaptive value	Algorithm running time/ms	Finding the probability of a feasible solution
Random extraction algorithm (general model)	Indescribable	-	>1000	81%
Retrospective test method (general model)	Indescribable	-	>1000	97.00%
Standard genetic algorithm (Model of this article)	0.68	0.53	207	100%
Improved genetic algorithm (Model of this article)	0.8	0.71	200	100%

Using the model of this paper to generate five sets of optimal solution sets for the organizer to choose, the final group paper results are specifically shown in Table 3. The difficulty span of the group paper is 0.42~0.92, and the differentiation and quality of the group paper is not less than 0.8. and the group paper organizer can choose the group of the group paper according to the type and form of the examination so that the whole group paper task maximally meets the group paper organizer’s specific needs of English teaching.

Table 3.

Group results

Alternative solution	difficulty	Differentiating	Usage frequency	Completion time	Final exposure time	Overall quality
1	0.92	0.84	0.73	0.82	0.62	0.82
2	0.77	0.8	0.65	0.93	0.97	0.8
3	0.82	0.84	0.74	0.88	0.89	0.8
4	0.61	0.91	0.63	0.85	0.77	0.82
5	0.42	0.81	0.71	0.81	0.8	0.8

4

English teaching practice based on group paper multi-objective optimization

4.1

Experimental design

4.1.1

Experimental hypotheses

The English group paper based on the integration of network resources can provide students with high-quality English learning test questions, improve students’ academic performance and precise learning ability, and have a certain role in promoting the learning improvement of students of different levels.

4.1.2

Experimental variables

1)

Experimental independent variables

The multi-objective optimization model of group papers is based on the integration of online resources constructed in this study. In the experimental class, the case of the lecture and assessment class designed according to the model was used for teaching, while the control class continued to use traditional teaching.

2)

Experimental Dependent Variables

Students’ academic performance in English is obtained through English paper-and-pencil tests.

3)

Experimental Control Variables

Teaching progress and assessment content, knowledgeability and teaching experience of the instructor, time of instruction, and students’ knowledge level.

4.1.3

Experimental subjects

In this study, the first-year class 1 of the administrative management major of XH University was selected as the experimental subject and the class 2 of the same platform in the same grade with it as the control class. In the process of university freshmen admission and class placement, the two classes have similar overall scores in English in the college entrance examination, and the number of students is also comparable (50 students), and the same English teacher serves as the English teacher of the two classes at the same time, which has the conditions for educational experimentation. The experimental class will adopt the multi-objective optimization model constructed in this paper to assist English teaching, while the control class will still maintain the traditional teaching method. The experiment will last for a total of 15 weeks.

4.1.4

Experimental tools

A total of two sets of examination papers, the pre-test, and the post-test, were needed as experimental tools for this experiment. Among them, the pre-test paper is an experimental pre-test exam prepared according to the content of the teaching progress of the experimental and control classes, and the post-test is the end-of-semester exam, and the papers used in the two exams are provided by an authoritative organization specializing in organizing papers. It is generally recognized that a difficulty index of between 0.3 and 0.7 is appropriate for the test questions, and the average difficulty of the whole paper should ideally be around 0.5. A differentiation index higher than 0.3 indicates that the test paper possesses a good degree of differentiation. The difficulty and differentiation of the three tests in this experiment are shown in Table 4. As can be seen from the table, the difficulty and differentiation values of the two test papers are in a reasonable range, basically in line with the development of the student’s English knowledge level, which can be used to test the students’ English learning level in this teaching stage, and the test results can be used for the experiments of this study.

Table 4.

Test difficulty and distinction

-	Pretest	Posttest
Difficulty	0.45	0.42
Differentiating	0.36	0.39

4.2

Analysis of experimental results

Before and after the experiment, the English test scores of the students in the experimental class and the control class were counted separately, and the data content is specifically shown in Table 5. It can be seen that in the pre-test scores, the average score of the control class is slightly higher than that of the experimental class by 1.21 points, and the significance P-value is 0.721, which is greater than 0.05, indicating that the difference between the pre-test scores of the experimental class and the control class is not significant and that the two classes are not on a par in terms of their English learning level before the experiment formally begins. The average post-test score of the experimental class is 77.26, which is 9.4 points higher than that of the control class. The average academic performance of the students in the experimental class is better than that of the control class, with a significant p-value of 0.016, which is less than 0.05. The post-test scores of the students in the experimental class and the control class show a significant difference. This indicates that the English group paper based on the integration of network resources can provide students with high-quality English learning test questions, and the impact on the academic performance of students in the experimental class is significant.

Table 5.

English test results

Test	Class	Mean	Standard deviation	T	P
Pretest	Experimental class	65.11	15.268	0.392	0.721
Pretest	Control class	66.32	15.174	0.392	0.721
Posttest	Experimental class	67.86	13.739	-3.306	0.016
Posttest	Control class	77.26	14.956	-3.306	0.016

In order to accurately and dynamically grasp the magnitude of the change in students’ learning level during the experiment, in addition to the two scores of the pre-test and post-test, this experiment collects the scores of a unified examination in the middle of the pre-test and post-test as the mid-test. The test difficulty and differentiation of the middle test are 0.45 and 0.35, respectively, which are in a reasonable interval. First of all, the results of the students in the control class in the pre-test, mid-test, and post-test were statistically analyzed, and the specific analysis results are shown in Figure 2. As can be seen from the figure, the number of students in the control class in the 110-119 points and 100-109 points score interval does not change significantly, 80-89 points and 70-79 points score interval have increased, and the number of people accounted for a larger proportion, 69 points and below the number of low score interval decreased significantly. It can be seen that the traditional teaching mode also has a certain promotion effect on the learning of students at the level of the lower score interval, but it does not have much effect on students in the higher score interval.

The pre-test, mid-test, and post-test scores of the students in the experimental class were counted, and the statistical results are specifically shown in Figure 3. It can be seen that the number of students in both the high score ranges of 120-150 and 110-119 increased in the experimental class, while the number of students in both the low score ranges of 70-79 and 69 and below decreased. In the mid-test scores, the number of students in the 80-89 score range reached a maximum of 16, while in the post-test scores, the number of students in the 90-99 score range reached a maximum of 18, suggesting that there was a more substantial increase in the performance of intermediate students between the mid-test and post-test timeframes. It can be seen that after five months of teaching experiments, the number of students in the high-score section increased, the number of low-score students decreased, and the number of students in the upper middle level reached a peak, which can indicate that the multi-objective optimization model based on the integration of network resources can realize the efficient integration of English online teaching resources, formulate high-quality English test papers, etc., provide useful help for college English teaching, and improve students at all learning levels.

5

Conclusion

In this paper, we form a multi-objective optimization model of grouping papers based on the integration of network resources, which efficiently integrates the network resources to achieve high-quality English grouping papers under the blended learning mode of university English. The performance of the model is tested through simulation experiments. The improved genetic algorithm used in this paper’s model for solving the optimal solution of paper organization, with the maximum adaptation value of the population and the average adaptation value of the population being higher than the standard genetic algorithm, overcomes the defects of the standard genetic algorithm in paper organization and greatly improves the efficiency of the execution of the paper organization task. In the final grouping results produced, the grouping difficulty span is rich, ranging from 0.42 to 0.92, and the differentiation and quality of the grouping is not less than 0.8, which can greatly satisfy the English teaching needs of the groupers. In order to further verify the utility of the model of this paper in college English teaching, English teaching practice was carried out. The posttest scores of the students in the experimental class were higher than those of the control class by 9.4 points, showing a significant difference (P=0.016<0.05). In terms of the score intervals of students’ pre-test, mid-test, and post-test scores, the number of students in the experimental class in the two high score intervals of 120-150 and 110-119 in the mid-test and post-test increased, while the number of students in the two low score intervals of 70-79 and 69 and below decreased, and the number of students of intermediate or higher level in the post-test reached a peak. Overall, the model in this paper can realize the efficient integration of English online teaching resources, assist university English teaching by developing high-quality English test papers, and promote the improvement of students’ English learning levels.

Lingua:: Inglese

Frequenza di pubblicazione:: 1 volte all'anno
Argomenti della rivista:: Scienze biologiche, Scienze della vita, altro, Matematica, Matematica applicata, Matematica generale, Fisica, Fisica, altro

Feed RSS della rivista

Research on Network Resources Integration of University English Blended Learning Model Based on Multi-Objective Optimization

Ling Sun

Pubblicato online: 21 mar 2025

Ricevuto: 14 ott 2024

Accettato: 11 feb 2025

DOI: https://doi.org/10.2478/amns-2025-0575

Parole chiaveMulti-objective optimization, Network resource integration, Objective optimization function, Genetic algorithm, English grouping paper

© 2025 Ling Sun, published by Sciendo

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Parole chiave
Multi-objective optimization, Network resource integration, Objective optimization function, Genetic algorithm, English grouping paper