Recent Advances in Random Matrices for Mathematical ModelingView this Special Issue
Research on University Information Management Based on Nonlinear Matrix Organizational Structure
The construction of data and automation in the field of teaching materials management in colleges and universities is an important part of teaching management in colleges and universities, and it is also a specific and heavy task. Therefore, improving the function of the college textbook management system and giving full play to the informatization service role of the college information management system are of great significance to improve the management efficiency of college textbooks and the quality of college teaching management. The research done still needs to be further deepened and still needs to be gradually improved in many aspects. For example, in order to make the application of the decision-making method proposed in this paper more convenient, it is necessary to further develop the supporting decision-making analysis software, which are practical problems that need to be improved and solved in the future. Aiming at the problems of traditional college textbook management mode, such as large manual workload, low work efficiency, unsound query and statistical functions, and postreview perimeter, this paper studies and designs a set of college information management system based on nonlinear matrix organizational structure. Based on the comprehensive integration methodology of nonlinear complex system theory, reductionism, and holism, this paper innovatively proposes a new analysis structure for complex problem evaluation and a value system for solution evaluation. The non-linear ANP decision-making method (collectively referred to as the new method) is proposed in this study. The new method judges the system state layer by layer from bottom to top based on the hierarchical overall judgment thinking, effectively revealing the behavior characteristics such as the emergence and mutation of complex systems. Compared with the traditional ANP decision-making method, the new method is more scientific and reliable than the traditional ANP decision-making method. This paper designs and develops a college information management system based on nonlinear matrix. Through the implementation test, it can be found that the system is helpful to improve the management level of college textbooks, reduce the irrelevant workload in the management of textbooks, and use data management for the decision-making of college textbooks.
As the country attaches great importance to popularizing the informatization of higher education, school education is also developing in a better direction, gradually moving towards intelligence and informatization. This has also promoted the trend of education competition gradually becoming white-hot. Under the strong competition, not only the management and service ability of the school’s educational affairs, the comprehensive ability of teachers, and so on can be improved but also the school’s management efficiency and service efficiency. The main label of the current school education is competition . For admissions, colleges and universities hope to achieve a gradual increase in the number of students to a certain extent and then the scale of majors, disciplines, and courses will be enriched and the management of textbooks in colleges and universities will naturally also improve. A new situation will be ushered in. However, at present, although some colleges and universities are leading the teaching quality, the management of teaching materials has not kept up. With the diversification of learning modes, the teaching materials of colleges and universities are no longer a single form . Colleges and universities should first pay more attention to the management of teaching materials, and second, complete the management of teaching materials through rich technical means. According to incomplete statistical data, many colleges and universities in my country are now under the pressure of resource management when managing textbooks . With the enrichment of teaching forms, the enrichment of teaching knowledge, the improvement of teaching systems and other factors every year, the management of teaching materials in colleges and universities has changed in an order of magnitude. The lack of intelligent guidance for textbook management has always restricted the progress of textbook management in colleges and universities.
For the management of teaching materials of different levels and types, many colleges and universities have no construction due to the pressure of their own student management and teaching management and allow the free play of teaching material management, which is linked together, adding to the multiple pressures on management, to a large extent, it has caused an increase in the workload, which is not conducive to colleges and universities to give full play to their advantages and make rapid progress. However, with the help of electronic information management, the auxiliary function of teaching material management can be realized. Hand over the work beyond human resources to electronic information engineering, realize the automation and digitization of college teaching, take the lead in strengthening the importance of teaching material management, and take the lead in the first part of college management [4–7]. It is not difficult to find that at present, the management of teaching materials is extremely important for many colleges and universities . Therefore, in order to actively challenge the problems in the management of teaching materials, all colleges and universities should take active actions and strive to implement the strategies of teaching materials management, so that they can maximize their effectiveness and serve teachers and students better . Only in this way can the quality be improved. Increase efficiency, so as to promote the improvement of the quality of college education in a subtle way. Aiming at the shortcomings of existing methods that use numerical information and lack of semantic interpretation, a probabilistic matrix factorization recommendation method based on topic model neighbor selection is proposed [10–12]. The method first obtains the user’s interest distribution and item’s attribute distribution through the topic model and finds approximate users and items through the interest distribution and attribute distribution respectively and then decomposes the adjacent matrix and the probability matrix to obtain the user’s feature vector and recommends . In this paper, through the topic model, semantic matching can be effectively carried out to find users with similar interests and items with similar attributes. LDA is a document topic probability generation model, which is a three-layer Bayesian probability model, including document, topic, and word three-layer structure. Correspondingly, in the recommender system, users, potential interests of users and labels marked by users are regarded as a three-layer structure; items, potential attributes of items and labels marked on items are regarded as a three-layer structure . Each user (item) can be represented as a multinomial distribution of a set of interests (attributes), and each interest (attribute) can be represented as a multinomial distribution of a set of labels . Use LDA to obtain the user’s interest distribution and item’s attribute distribution respectively, and then calculate the similarity between the interest distribution and the attribute distribution to select users with similar interests and items with similar attributes .
Aiming at the independent loop system inside ANP, the theoretical viewpoint of constructing the influence matrix of scheme set on target set by interval estimation method is firstly proposed, and based on the technical core idea of DEA relative efficiency evaluation, an ANP supermatrix for this kind of system is proposed. Then, the original ANP method system is completely abandoned, and a new ANP decision-making method is proposed based on the construction of a new analysis structure, which can ideally reflect the nonlinear relationship between factors in complex system levels. In order to help colleges and universities successfully implement knowledge management strategies, it is of great practical significance to establish a practical, reasonable, and operable comprehensive evaluation system for knowledge management capabilities. Unfortunately, the existing traditional AHP evaluation methods and fuzzy comprehensive evaluation methods all ignore the interdependence between system factors. In order to overcome the aforementioned defects of the existing methods, this paper uses the nonlinear ANP inner loop dependency hierarchical system scheme ranking method to comprehensively evaluate the knowledge management ability of enterprises. The application process of the method shows that the nonlinear ANP decision-making method proposed by the author is feasible and has strong practical operability. The evaluation results obtained using this method not only provide an important basis for enterprise knowledge management workers to monitor the current situation of knowledge management and improve knowledge management work, but also provide a new idea for helping them carry out knowledge management work.
2. Nonlinear System Characteristics of Decision-Making Methods
2.1. Nonlinear System Characteristics
There may be a large number of nonlinear influence relationships among the factors within the factor set. Unlike AHP, which assumes that the factors within the hierarchy are independent of each other, ANP takes into account the interdependence among the factors within the factor set, which fully reflects that there may be complex nonlinear relationships among the factors within the factor set [17–19]. Of course, Professor Saaty tries to reveal this complex dependency through the traditional pairwise comparison method, but from his description of the internal dependency mechanism, the dependency mechanism given by this analysis method is chaotic. For example, for a relatively simple factor set internal loop dependence mechanism, set the internal loop dependence relationship of the target set as “target depends on target , target depends on target , target depends on target , if , and assume that the absolute importance of under as the control standard is and , then the pairwise comparison matrix between factors and under as the control standard can be represented as matrix Z.
There may also be complex nonlinear interaction relationships between different factor sets of ANP. It should be emphasized that the second ANP structure division method has been rarely (almost never) adopted by other scholars since it was proposed. Even Professor Saaty himself did not adopt this division structure in his latest ANP achievement monograph, so in the following analysis, the author still uses the traditional first structure division method to study the ANP decision-making method.
The selected question should first be identified with five indicators and represented by a five-dimensional vector (question number, difficulty, question type, section, and test score), denoted as (). If n is used to represent the total number of questions in the paper, then a paper can be represented by a matrix of :
This is an objective morphological matrix for problem-solving which meets the following constraints.
The total marks of the examination papers are as follows:
Marks occupied by each question type:
If a13 = j, then cij = 1; if a13 > 1, then cij = 0. j represents the question number:
Fractions occupied by each chapter:
Regarding the difficulty of type i questions:
First of all, from the subject’s point of view, the partial independence test is to determine that the individual subject did not receive any assistance from external factors in the process of answering the test and answered the test based on his or her own true level, including copying others’ answers, finding relevant information, and using external communication devices.
This interaction relationship is reflected in that the influence of the lower-level factor set on the upper-level factor set is not necessarily independent, but there may be nonlinear relationships such as complementarity, substitution, and matching, which may produce complex system behavior characteristics such as emergence and mutation. This is also an important theoretical problem that the nonlinear ANP decision-making method needs to solve.
2.2. Classification of Decision Structure Types
The structure of a system depends on the dependencies between factors (sets of factors) in the system. There are two kinds of dependencies in the ANP system: functional dependencies and structural dependencies. Functional dependencies refer to the quantitative dominance between factor sets and between factors in a system. Structural dependence refers to the set of factors and the structural relationship between factors in the system. Because structural dominance affects quantitative methods for determining functional dependencies, it is critical to the analysis and design of systems . Structural dependencies can be divided into two categories, namely external dependencies (dependencies between sets of factors) and internal dependencies (that is, dependencies between factors within a set of factors). External dependencies are divided into two categories, one is the hierarchical dominance relationship, and the other is the cyclic dominance relationship. As shown in Figure 1, according to the classification of external dependencies, the ANP system structure can be divided into a hierarchical system structure and a cyclic system structure accordingly. Hierarchical system structure can be further divided into simple hierarchical system structure and hierarchical system structure with feedback. According to whether the internal factors of the hierarchy (factor set) are independent, the simple hierarchical system structure can be divided into the simplest AHP structure and HSID system structure. The AHP structure is the simplest form of system structure. If the interdependence of the internal factors of each level is considered on the basis of the AHP structure, then this structure is called the HSID system structure. If a hierarchical dominance relationship may exist between the levels of a system, there may be a cyclic dominance relationship, and at the same time, dependencies within the hierarchy are allowed, then this type of system structure is called an HSOF system structure. Similar to the hierarchical system structure, the cyclic system structure can be divided into CSII system structure and CSID system structure according to whether the hierarchy is independent of each other.
2.3. System Structure Division
The difference between the Suparchy structure and the AHP system structure is that it implies the overall goal of the system, and there is a feedback dominance relationship between the top two layers; the Intarchy system structure not only has the overall goal, but also has feedback between the two middle layers . Dominance relationship; Sinarchy system structure is similar to Intarchy system structure, it differs from Intarchy structure only in that the two layers with dominance relationship are located at the bottom, not in the middle part of the system structure; Hiernet system structure is the most general network structure, in which there are not only dependencies between factors within a factor set (this dependency relationship is shown in Figure 2), but also a feedback relationship between factor sets. However, considering that the traditional division method is too rough in the division of the internal dependencies of the factor set, the author subdivides the internal dependencies into typical circular dependencies (that is, the internal factors of the factor set are shaped such as A affects B, B affects C, C It also affects the dependency of A) and multidependency (that is, there are multiple factors (greater than or equal to 2) in the factor set that jointly affect the role of a factor). Based on the aforementioned understanding, here we divide the most commonly used HSID system structure into the inner cyclic dependency hierarchical system structure (HSICD-Hierarchy System with Inner Circular Dependence) and the inner multidependent hierarchical system structure (HSIMD-Hierarchy System with Inner Multielement Dependence). Therefore, the following four typical ANP system structure problems will be studied in depth.
2.4. Construction of University Information Management System
The university information management system designed in this paper adopts the browser/server mode architecture, and users only need to log in the browser to access the system. In addition, the university information management system based on nonlinear matrix realizes the design purpose of multilevel user access and professional management. System development should focus on providing professional software management performance, effectively improving the efficiency of college textbook information management, and optimizing all levels. The user provides a good operation interface for the software and processing flow of the university information management system based on nonlinear matrix, as shown in Figure 3, which lays a good foundation for the upgrading of the system platform in the future. The functional structure model of the university information management system based on nonlinear matrix covers five subsystems, including: (1) Book fee management subsystem, (2) Inbound and outbound management subsystem, (3) Comprehensive query subsystem, (4) Mobile app management subsystem system, and (5) System management subsystem. It should be noted that the traditional ANP does not recognize the existence of this nonlinear mechanism relationship but only reveals the intrinsic relationship between them by constructing the weighted matrix between the factor sets, so it is impossible to analyze the relationship from the theoretical analysis method. It reflects the nonlinear mechanism connection between factor sets.
2.4.1. User Layer
The so-called business user layer is the interface component displayed by the business system to the outside world, which can realize the interaction between the user and the application program. As far as the users of the system are concerned, the user layer can provide a good use page, and as far as the business system is concerned, it can provide standard interfaces with external business systems and realize extended services. The JBPM-based university information management system chooses the architectural design of the combination of B/S and HTML5. Some features can be directly displayed on the front end of the page, and can also be transplanted. Users only need to log in to the client without the help of other software. User layer system can be divided into three layers: presentation layer, control layer and interface layer. Different layers correspond to different calls and can also coordinate some of the requests sent back from the operation page by different users. Among them: the display layer uses advanced technology based on HTML5 and its scripting language, which can mobilize asynchronous data, and then directly call the business logic of the control layer, so as to ensure that this layer is more humanized and intelligent; the main function room of the control layer receives requests from different users, so as to call business logic methods at all levels; the interface layer provides convenience for third-party access systems to enter the system services, provides channels for business systems to connect with other systems, and can also provide necessary services for other related systems support.
2.4.2. Application Layer
The application layer can realize various management requirements for the business system customized by the component layer. As for the college information management system based on nonlinear matrix, it is necessary to manage the book fee, in-out and out-of-stock situation, and also realize the mobile app management and system management and management and comprehensive query functions. It can also provide interfaces between various business systems and the outside world and provide technical support for data management communication between business departments, business systems, and departments and systems in colleges and universities.
2.4.3. Model Layer
The model layer can provide no less than one interface to the outside world, and its component content includes multiple different classes, and different classes work together to achieve specific functions. The component layer belongs to the collection of components in the system. The users of the component layer located between the business management system and the corresponding operating system can provide environment support for the operation and development of the business system, and realize the data interaction between data, hardware platform and software. This support is not only reflected in business management applications but also provides convenient conditions for the connection and data between other business system modules in colleges and universities, enabling data sharing. In addition, the model layer can also provide support for the implementation of basic functions of the business system, such as basic personnel management, authority management, and specific business logic management, and can also implement core business logic based on interface services, providing access to the audit service bus. The components of the model layer are independent, and the system realizes the data connection between the components through the standardized interface provided by the independent components, and then completes the related business.
2.5. Construction of University Information Management System
A series of tests are carried out for the main functions of the system, user experience, system performance, and the use environment of the system. The main working principle of black-box testing is to set the system as a black box that cannot be observed by the outside world. In the environment of black-box testing because the core code of the program cannot be disclosed, developers will the test items, and the external functions of the program are tested, and the rationality of the system is checked for bugs and memory leaks. Black-box testing is used to find a series of serious operating errors and whether the overall operating results of the program conform to user needs, whether the system interface is beautiful, whether the user behavior design conforms to ergonomic and scientific perspectives, and whether improvements are needed, and so on, have significant effects. For some relatively complex attachments, in addition to a certain number of other classes, it also includes library files, interfaces, and configuration files that support business systems and may also include and apply some other components to provide JBPM-based university information management systems.
Black-box testing corresponds to functional testing, which uses observation and other methods to test the completed functional modules in the system to test whether these functional modules are consistent with the actual needs of customers. In the process of black-box testing, first of all, it is necessary to clarify the software functions and determine the system functions. If it is unnecessary to analyze the time structure of the software system, it is not necessary to conduct in-depth analysis. The purpose is to reduce the irrelevant impact of the system content structure on the system testing, thereby improving the efficiency of system testing. In addition, the use of black-box testing does not require high technical requirements for testers. The test focuses on the interface and function of the software product, the hardware function of the software operation, the external equipment and the software interface information. Compared with black-box testing, white-box testers must fully understand the internal structure of software products, combine the design environment, design conditions and other content and sequence structure, transfer branch, loop structure and other information to mine, and correct system loopholes and errors, and also actionable technical measures need to be proposed for the problem. In general, black-box testing can also be called functional testing because its source code is not visible, so the name is derived from its testing principle. Generally speaking, when testers conduct black-box testing, they usually test each functional module repeatedly and repeatedly to learn in detail whether the relevant operations of each functional module can be used accurately. Combined with white-box testing, they can also give the internal code, framework and function variables are all encapsulated and debugged, so that the black box with unclear internal structure is used as the main object of system testing, and the programming principle and internal structure can be detached in a targeted manner. Black box testing also has a relatively direct effect on the UI interface and system functions in the testing system.
3.1. System Management Implementation
Operation result of the university information management system based on nonlinear matrix is shown in Figure 4. The system management process is as follows: the user enters the system management information through the browser, the client sends back the query information to the background, the background realizes data processing according to the program logic, and the database and the background information interact with each other; the foreground refers to the business logic to display the system management results. In this way, the original five basic structures of ANP are expanded to the existing six basic structures. As the AHP system structure, CSII system structure, HSICD, and HSIMD system structure are the most typical system structures of the aforementioned six structures (in other words, the aforementioned four typical types of system structures have been studied clearly, other types of ANP structures exist).
In the process of system testing, testers need to first consider the user’s needs and the main norms of operating behavior, and on the basis of mastering technical disciplines such as ergonomics, to ensure that users can use the system quickly and efficiently. Before testing, testers need to master the main problems of common system testing, and make certain records of common problems, so as to ensure that common problems are discovered and avoided in the process of system testing, so as to reduce the situation that common problems affect the system testing work. Large content, such as the advantage is that it can prevent detailed problems, reduce potential system risks caused by detailed problems, and reduce the occurrence of small-probability events. Testers also need to simulate the needs and habits of as many users as possible, use diverse user roles and permissions to experience system functions, and refine the content of system testing.
3.2. Comparison of Expected Results and Actual Results
This article uses a combination of black-box testing and white-box testing. Therefore, it is necessary to combine the two testing methods to test multiple user roles of the system. Through testing, we can understand whether there are functional deficiencies between each user role and user behavior. The white-box test mainly conducts clear functional tests for each user role and its related user functions. Therefore, it is necessary to log in the user, understand the functional permissions of different user roles, and summarize their respective permissions to facilitate the white box. When testing, do a test check. For black-box testing, the user’s role also needs to be developed by different testers, so the login test operation is the necessary first step for the two testing methods. Through the user test, you can understand the various functional permissions of the user behavior, and also through the understanding of the relevant information, as the data basis for the system test. To test a software product is to regard the complete software system including the operator as a whole, and test whether there is any discrepancy with reference to the system specification requirements. System testing requires the participation and assistance of users to continuously improve and revise. For the software testing of university information management system based on nonlinear matrix, first, black-box testing is used to clarify the error range and then white-box testing is used for correction and diagnosis. Through the combination of the two test methods, it can be ensured that the system achieves the expected functional goals. For the user’s client, by simulating user behavior, the system is tested in black box and white box. The test case is as follows: the first step is the user login test, and the test result is shown in Figure 5. In the process of testing, the work content of the test should be carried out in steps from small to large. For small problems, they should be tested first and recorded in time. After the detailed problems are tested, they will be tested one by one.
For the system, the black-box test is mainly carried out in the functional test, so for the book fee management function of the system, it is necessary to use the black-box test. Testing is important. As one of the main functions of the system, book fee management is connected to several other important modules of the system, such as student management, textbook management, and administrator management. The main function of the system is to verify whether the various interfaces of the system can be used normally. The third step of this test link is the book fee management test. The test table is shown in Figure 6:
School textbook management business. It needs to be considered that, as the system is fully used, there may also be some problems. Therefore, system testing must focus on performance testing. In addition, during the performance test, the tester first simulated 50 users and issued transaction operation requests at the same time, and the project success ratio was 11,324/11,325; then the tester simulated 100 users and issued transaction operation requests at the same time, and the project success ratio was 12,673/12,680. In general, the system runs smoothly, the response speed is ideal, the data are relatively safe and meets the user’s expectations.
3.3. System Security Detection
As an open environment, Big Data technology has been recognized for its advantages, but its security has also become a concern for users. Therefore, the information management system in colleges and universities considers security work as a key point in the construction process. The system first considers the security protection mechanism. By protecting all layers of security construction, the system’s protection function against intrusions is improved to ensure the safety and reliability of data; second, the system should establish a security review mechanism to review location data in the cloud environment, waybill data, and so on to authorize, to realize the tracking of data operation behavior, so as to ensure the safety and reliability of cloud data flow.
System security is mainly through antivirus software for intrusion detection, vulnerability scanning and assessment, and antivirus security measures. Improve system security and reliability. In this system, Huawei’s IDS is used for intrusion detection, OpenVAS is used for vulnerability scanning and assessment, and security protection is carried out from the whole system. The main function of the intrusion detection system is to monitor and control illegal intrusions, while the vulnerability scanning system uses professional software to check the vulnerable vulnerabilities in the system, reminding personnel to fix the vulnerabilities and carry out security protection in advance; the network security assessment system can dynamically evaluate the system risks, detection of network security level, so as to provide a guarantee for the safe operation of the system; antivirus software can strengthen the filtering of virus files, and can be regularly antivirus. The results are shown in Figure 7. The system adopts the method of client classification, and divides the client into the access portals of different users, so as to control the access content and authority of users. The access rights of fixed clients are set in advance, and other types of clients do not have access. For the carrier client and enterprise user client, where the functions overlap, you can set the user permission level to manage. This can be identified by adding a marker to the user account when the user is registered. When the user logs in, he can pass markers to distinguish permission levels, thereby controlling access. In addition, for the access of the same client but belonging to different enterprises, the system stores data according to the enterprise partition during data storage, and different enterprise data are stored in their respective areas, and when the enterprise registers users, the enterprise classification mark is also added, according to the mark. Set user access addresses to control user access to sensitive data. Through the previous test, it can be found that the system performance basically meets the user’s needs.
We have understood a series of processes of system implementation and presented results for the earlier requirements analysis and system design, which added help to the application of college textbook management system and promoted the development of JBPM technology. In the second half of the chapter, this paper describes the system testing, combined with the actual related work of the system testing, analyzes and summarizes, verifies the operability of the system, and provides later support for the later application and practice of the system.
4.1. Detection Sensitivity of Nonlinear Matrix Method
The research on CSII system points out that even if the traditional ANP method is feasible, it is not reasonable to use point estimates to reflect the relative preference of experts to the compared targets when constructing an important submatrix in the supermatrix, namely the IMAG matrix. The reason is that comparing the relative importance of each target factor based on the scheme set essentially allows experts to make logical judgments from the perspective of holism, while experts analyze the behavioral characteristics of complex social and economic systems from the perspective of holism. It is often necessary to rely on a lot of experience and intuition, so it is often only possible to give judgment results with high uncertainty, and only one point estimate cannot reflect the real opinions of experts on the evaluation problem, as shown in Figure 8. In order to solve the aforementioned problems, this paper takes CSII-type ANP system as the research object, and puts forward the theoretical viewpoint of constructing IMAG matrix by interval estimation method. The analysis of a specific example and the comparison with the ANP supermatrix construction method proposed by Saaty show that the improved supermatrix construction method is more scientific and reasonable than the traditional ANP method.
When traditional ANP evaluates this kind of system scheme, it is difficult to construct a weighting matrix, and the connotation of the elements (weights) of the judgment matrix is not clearly defined, and there are two defects. In order to overcome the aforementioned defects, this paper presents a new method for sorting solutions with a stable system analysis structure by constructing a new system analysis structure, an extraction platform for expert judgment information, and a value system for proposal evaluation under the new analysis structure. The new method has the following three advantages: First, on the basis of dividing the state of the system elements, the new analysis structure adopts the hierarchical holism way of thinking to judge the system state from the bottom to the top, which better reflects the state of the system. The complex nonlinear relationships such as substitution and matching between different levels of the system and between different factors at the same level reveal the characteristics of the mutation and emergence of complex systems.
4.2. Hierarchical Analysis of Information Management
According to the traditional ANP decision-making method, the relative importance weights of the four dimensions of B, O, C, and R judged by experts are b = 329.0, o = 231.0, c = 256.0, r = 184.0. In order to compare and analyze the traditional ANP/BOCR scheme ranking method and the new method proposed in this paper, it may be assumed that the single network evaluation value of each scheme can be measured by the actual monetary value. To this end, this paper presents two different scenarios. Although the B, O, C, R monetary measures in these two scenarios are completely different, they both produce the same scheme ranking weights (partial scheme compound weights). The theoretical relationship with the monetary measure value of the scheme is explained here only through the calculation method of a specific data, for example: in the scenario 1, the calculation method of the compound weight of the income of scheme 1a is 1350/(1350 + 1926 + 1788 + 936) = 0.225. According to the total integrated expression based on the monetary measurement value, the ranking of the pros and cons of each scheme under Scenario 1 and Scenario 2 is calculated to be 2 and 341 and 2 and 314, respectively (as shown in Figure 9). It can be seen that, Scenario 1 and Scenario 2 reflect two different objective ranking mechanisms. In order to make this theoretical point of view applicable to the actual complex economic system evaluation and decision-making problem, based on the technical core idea of DEA relative efficiency evaluation (that is, selecting a set of favorable virtual weights to maximize the relative efficiency index of the decision-making unit), this paper proposes an improved method for supermatrix construction of CSII-type ANP system is presented.
The hierarchical holism is used to judge the system state from bottom to top layer by layer, which ideally reflects the complex nonlinear relationship between adjacent levels of factors, thereby revealing the characteristics of the system’s sudden change and emergence. Third, the concept of weight used is clear in connotation. Fourth, since the average ranking value of the scheme is in one-to-one correspondence with its corresponding scheme, adding a new scheme or deleting individual original schemes in this method will not change the ranking of the other schemes, thus realizing the scheme evaluation in essence. The results of comparative analysis of numerical examples show that the new method given in this paper has stronger sensitivity and scientific rationality than traditional ANP in terms of whether it can reflect the nonlinear value preference of decision makers. The decision-making method is scientific and reliable.
5. Strengths and Limitations
For the AHP system, the author believes that a few specific AHP schemes based on the fixed-weight system and reflecting the interaction mechanism of system elements may appear method models because they cannot reflect the essential characteristics of complex systems such as nonlinearity and emergence. In order to solve the fixed-weight problem of AHP and the failure of the scheme compound ranking method that may be caused by traditional AHP, this paper draws on the variable weight idea of the existing multiobjective variable weight decision method (that is, the idea that the factor weight changes with the change of the factor state value), according to the systematic analysis idea combining holism and reductionism and by constructing a new analysis structure for complex evaluation problems and a value system for scheme evaluation under the new analysis structure, a nonlinear AHP for scheme optimization is presented. In the nonlinear AHP decision-making method, the weight of the system factor does not only depend on the change of the value of the factor but also depends on the change of the combination of all factors at the same level, which is a combination variable weight. Compared with the traditional AHP, the nonlinear AHP decision-making method not only adopts a new analysis structure that is more conducive to reflecting the nonlinear and emergent characteristics of complex systems, and the extraction method of expert judgment information based on the new analysis structure, that is, the hierarchical overall judgment method and fundamentally solve the problem of reverse order of AHP scheme that has troubled experts and scholars.
6. Future Research
Although this paper explores the ANP decision method from the perspective of nonlinear complex systems and has achieved some valuable research results, because the understanding and research of complexity science and complex systems in academia are still immature. The research done still needs to be further deepened and still needs to be gradually improved in many aspects. For example, in order to make the application of the decision-making method proposed in this paper more convenient, it is necessary to further develop the supporting decision-making analysis software, which are practical problems that need to be improved and solved in the future.
This paper designs and develops a college information management system based on nonlinear matrix. Through the implementation test, it can be found that the system is helpful to improve the management level of college textbooks, reduce the irrelevant workload in the management of textbooks, and use data management for the decision-making of college textbooks. Provide reference and help. Specific to the design of college information management system based on nonlinear matrix, first, clarify the research background and research significance and have a comprehensive understanding of relevant research at home and abroad. Second, clarify the design ideas and functional requirements of the system, and focus on analyzing the book cost management, inbound and outbound management, comprehensive query management, mobile app management, and system management. Each module of the system is implemented and tested to ensure the safe and stable operation of the system. Through the aforementioned research steps, the author got the inspiration of technical research. After this process, the experience accumulated in the research also laid the foundation for the future research. Therefore, through the research of this paper, the author has mainly obtained the following important conclusions, which can be summarized as follows: (1) Using the object-focused design method and the UML modeling tool, the related design of the university information management system based on nonlinear matrix, class diagram, sequence diagram, analyze and design the relevant modules of the system; (2) according to the database design principles and standards, store its table structure information, optimize the database storage and retrieval functions; (3) according to the system business logic, describe the system; and (4) use black-box testing and white-box testing to test the system to ensure the safety and stability of the system.
The data used to support the findings of this study are included within the article.
Conflicts of Interest
The authors declare that they have no conflicts of interest or personal relationships that could have appeared to influence the work reported in this paper.
This work was supported by Hunan Polytechnic of Environment and Biology.
H. Zhang, Y. Liang, H. Su, and C. Liu, “Event-driven guaranteed cost control design for nonlinear systems with actuator faults via reinforcement learning algorithm,” IEEE Transactions on Systems, Man, and Cybernetics: Systems, vol. 50, no. 11, pp. 4135–4150, 2020.View at: Publisher Site | Google Scholar
J. Hu, H. Zhang, X. Yu, H. Liu, and D. Chen, “Design of sliding-mode-based control for nonlinear systems with mixed-delays and packet losses under uncertain missing probability,” IEEE Transactions on Systems, Man, and Cybernetics: Systems, vol. 51, no. 5, pp. 3217–3228, 2021.View at: Publisher Site | Google Scholar
L. Ma, N. Xu, X. Zhao, G. Zong, and X. Huo, “Small-gain technique-based adaptive neural output-feedback fault-tolerant control of switched nonlinear systems with ud,” IEEE Transactions on Systems, Man, and Cybernetics: Systems, vol. 51, no. 11, pp. 7051–7062, 2021.View at: Publisher Site | Google Scholar
L. Aristodemou and F. Tietze, “The state-of-the-art on Intellectual Property Analytics (IPA): a literature review on artificial intelligence, machine learning and deep learning methods for analysing intellectual property (IP) data,” World Patent Information, vol. 55, pp. 37–51, 2018.View at: Publisher Site | Google Scholar