Method, system and equipment for evaluating rule base based on analytic hierarchy process and medium

Abstract

The invention belongs to the technical field of rule evaluation, and particularly relates to a method, a system, equipment and a medium for evaluating a rule base based on an analytic hierarchy process, and the method comprises the steps: specifying an evaluation index, building a hierarchical index structure according to the evaluation index, and obtaining an index weight through carrying out the consistency and normalization processing on each layer of index; an n-order membership matrix is established through indexes of each layer, a complexity evaluation result of the rule base is obtained through combination of weights and the membership matrixes, then reliability evaluation of the rule base is obtained through soft error analysis, evaluation score tables of 12 levels are specified, and the reliability of the rule base is evaluated through combination of the complexity evaluation and the reliability evaluation and comparison of the evaluation score tables. And obtaining the score of the rule base.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of rule evaluation, and particularly relates to a method, system, device and medium for evaluating a rule base based on an analytic hierarchy process. BACKGROUND

[0002] In recent years, with the promotion of big data, new algorithms and supercomputing, artificial intelligence is changing and even subverting every industry it touches. Scientific and efficient operational planning is crucial for future information warfare. As an important part of contemporary strategic technology, intelligent combat aircraft have become an important direction of development with autonomous decision-making and combat capabilities. Intelligent air combat refers to a form of warfare that uses artificial intelligence (AI) technology and unmanned aerial vehicles (UAVs) and other unmanned systems to enhance the combat capabilities and efficiency of aircraft. The emergence of intelligent air combat has brought about a huge change to the traditional air combat pattern and combat methods, with breakthrough development potential and wide application prospects.

[0003] Traditional air combat mainly relies on the skills and experience of pilots to make decisions and operate, which has good results to some extent, but still has many challenges and limitations. With the rapid development of artificial intelligence technology, using unmanned aerial vehicle systems and artificial intelligence technology for air combat has a broader development prospect. Intelligent air combat can greatly improve the combat capabilities and adaptability of aircraft, enabling them to perform various tasks more flexibly, quickly and intelligently.

[0004] The single-combat operational combination maneuver rule module aims to realize the conversion between various working modes of combat aircraft through a finite state machine; through a behavior tree, combat aircraft have corresponding behavior rules in different states such as attack, defense and coordination. An autonomous decision-making model based on the behavior tree model is required to be established, a system training rule base needs to be established, and information such as attack, defense and coordination is read from the rule base and the decision result, i.e., the task, is fed back to the system training platform. The training platform realizes individual actions inside the combat aircraft through an autonomous decision-making model inside the model. The decision rule is composed of a condition space and a command space. This project takes the input information required for decision-making (i.e., information obtained from the combat simulation deduction system) as the condition space of the rule, and the output information of the decision-making, i.e., the task, as the command space of the rule.

[0005] The prior art with publication number CN104376400A discloses a risk assessment method based on fuzzy matrix analytic hierarchy process, comprising: a plurality of evaluation indexes are generated by an expert group evaluating the hazard source to generate an evaluation index domain; a plurality of evaluation grades are generated by the expert group evaluating the hazard source to generate an evaluation grade domain, each evaluation grade corresponds to a grade fuzzy subset; the membership degrees of the hazard source to the grade fuzzy subsets are calculated according to the evaluation indexes, and a membership degree matrix is constructed; the weight vector of the evaluation indexes is calculated by referring to the analytic hierarchy process; the fuzzy comprehensive evaluation result vector of the hazard source is synthesized according to the membership degree matrix and the weight vector; the risk evaluation value is calculated according to the weight vector and the fuzzy comprehensive evaluation result vector; the risk value is calculated according to the risk evaluation value, and the risk situation of the hazard source is judged according to the risk matrix and the risk value.

[0006] Therefore, for various rule bases established, there are different problems in storage and management of the rule bases, and there is no effective standard for evaluating the behaviors of the rule bases in the prior art, that is, there is a lack of rule base complexity evaluation, reliability evaluation and comprehensive evaluation based on the two. SUMMARY

[0007] To solve the above problems, the present application provides a method, system, device and medium for evaluating rule bases based on analytic hierarchy process, which realizes the evaluation of the complexity of the rule bases based on the evaluation method of fuzzy analytic hierarchy process, realizes the evaluation of the reliability of the rule bases based on the evaluation method of soft error analysis, and finally realizes the comprehensive evaluation of the rule bases by combining the complexity evaluation and the reliability evaluation.

[0008] To achieve the above purpose, the present application adopts the following technical solutions: The first aspect of the present application provides a method for evaluating rule bases based on analytic hierarchy process, comprising the following steps: Step S1: according to the actual situation of rule base complexity evaluation, evaluation indexes are formulated, and the acquisition method of the corresponding indexes is formulated. Step S2: according to the evaluation requirements, the hierarchical index structure of the problem to be analyzed is constructed, and a judgment matrix is constructed to obtain the index weight.

[0009] Step S3: according to the evaluation indexes, a membership degree relationship matrix is constructed, and the membership degree values of each index are calculated.

[0010] Step S4: the evaluation of the complexity of the rule bases is completed by combining the index weight and the membership degree values.

[0011] Step S5: the reliability of the rule bases is evaluated based on the soft error analysis method.

[0012] Step S6, the rule base comprehensive evaluation is carried out by combining rule complexity evaluation analysis and rule reliability analysis.

[0013] In the second aspect of the present application, a system for evaluating a rule base based on an analytic hierarchy process is provided, comprising a building evaluation index module, a building hierarchical index structure module, a fuzzy analytic hierarchy process module, a soft error analysis module, and a comprehensive evaluation module. The building evaluation index module formulates corresponding evaluation indexes by starting from the actual situation of the rule base, and establishes evaluation indexes of the rule base from three dimensions of time complexity, code complexity, and structure complexity of the rule base. The building hierarchical index structure module further refines indexes to be evaluated for each index based on the evaluation indexes, and further constructs a more complete hierarchical index structure to comprehensively evaluate the complexity of the rule base. The fuzzy analytic hierarchy process module comprises a weight calculation submodule and a membership degree matrix calculation submodule. The weight calculation submodule constructs a judgment matrix by constructing the evaluation index structure, obtains index weights by normalizing the judgment matrix, and determines whether the obtained index weights are correct by consistency checking of the judgment matrix. The membership degree matrix calculation submodule obtains a membership degree relationship matrix through mutual influence relationships between indexes, obtains a comprehensive contribution degree vector by combining the membership degree matrix with the index weights, and obtains standard weights by normalizing the contribution degree vector. The rule base comprehensive complexity evaluation is obtained by multiplying the index weights and the standard weights and then summing them. The soft error analysis module obtains reliability according to the total number of running and the number of success by running the rule base multiple times and recording the number of success.

[0014] The comprehensive evaluation module determines the evaluation result of the rule base by combining the rule base comprehensive complexity evaluation and the reliability.

[0015] In the third aspect of the present application, an electronic device is provided, comprising a memory 102, a processor 101, a display module 103, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of any one of the aforementioned methods for evaluating a rule base based on an analytic hierarchy process.

[0016] In the fourth aspect of the present application, a readable storage medium is provided, which stores a computer program, wherein the computer program is executable on a processor to implement the steps of any one of the aforementioned methods for evaluating a rule base based on an analytic hierarchy process.

[0017] The beneficial effects of the present application are: the innovative fusion of fuzzy analytic hierarchy process and soft error analysis technology, the construction of complexity-reliability dual-dimensional evaluation system, the effective breakthrough of the limitation of single evaluation dimension of traditional rule base, the processing of the fuzziness of index importance in complexity evaluation by fuzzy analytic hierarchy process, the combination of hierarchical structure and membership degree relationship matrix, the conversion of subjective experience into rigorous quantitative results, the improvement of the scientificity and objectivity of complexity evaluation, and the reliance on soft error analysis to accurately detect the soft error risk in the operation of the rule base, the completion of the short board of reliability evaluation, and the realization of the comprehensive characterization of the static structure and dynamic operation state of the rule base. Through the complexity and reliability structure, a standard evaluation table is developed, which provides a standard for the evaluation of the rule base. BRIEF DESCRIPTION OF DRAWINGS

[0018] In order to more clearly illustrate the technical solutions of the present application, the drawings needed to be used in the description will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and all other drawings obtained by those skilled in the art without creative labor on the premise of the drawings are within the scope of the present application.

[0019] Figure 1 The flowchart of the method of the present application is shown in the figure. Figure 2 The system structure diagram of the present application is shown in the figure. Figure 3 The equipment structure diagram of the present application is shown in the figure.

[0020] Among them, 101-processor, 102-memory, 103-display module. DETAILED DESCRIPTION

[0021] In order to make the application purpose, features and advantages of the present application more obvious and easy to understand, the technical solutions in the embodiments of the present application will be described clearly and completely in the following with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of the present application.

[0022] Unless otherwise defined, all technical and scientific terms used in the present application have the same meaning as understood by those skilled in the art to which the present application belongs. The terms used in the specification of the present application are only for the purpose of describing the specific embodiments of the present application, not intended to limit the present application.

[0023] Embodiment 1, a method for evaluating rule base based on analytic hierarchy process, as shown in the figure, comprising the following steps: Figure 1 ​Step S1, according to the actual situation of the rule base complexity evaluation, the evaluation index is made, and the corresponding index acquisition method is made; the specific steps are as follows: Step S11, the rule base is composed of state machine and behavior tree, according to the complexity of the rule base of state machine and behavior tree, the evaluation index is made, including time complexity, code complexity and structure complexity; Step S12, time complexity is a relative measure of the running time of an algorithm, the running time of an algorithm is roughly equal to the time required to perform a simple operation (assignment, comparison, calculation, turn, return, input and output) and the number of simple operations in the algorithm. The time consumed by an algorithm is theoretically impossible to calculate, and must be tested on the machine. But we can't and don't need to test every algorithm on the machine, we just need to know which algorithm takes more time and which algorithm takes less time, the time an algorithm takes is proportional to the number of statements in the algorithm, the number of statement execution in an algorithm is called time frequency, denoted as T(n), n is called the size of the problem. Acquisition method: place the start and end statements at the front and tail of the interpretable model code respectively, calculate the time of each epoch of the model, take multiple data, and complete the acquisition of time parameter; When n changes continuously, the time frequency T(n) also changes continuously. But sometimes we want to know what the law is when it changes. For this purpose, we introduce the concept of time complexity. If there is a certain auxiliary function f(n), such that when n tends to infinity, lim(T(n) / f(n)) is not equal to zero, then f(n) is the same order function of T(n), finally use O to represent the time performance of the algorithm, denoted as T(n)=O(f(n)), called O(f(n)) as the time complexity of the algorithm, simply speaking, it is to keep the highest power of the number of times and remove the coefficient. For example, T(n)=2n 2 +n+1=O(n 2 ); Step S13, code complexity is space complexity, the calculation formula is S(n)=O(f(n)), where n is the size of the problem, f(n) is the function of the storage space occupied by the statement about n, S14, according to the table structure of the rule base, define the number of nodes as N, the number of behavior tree layers as L, the number of branches as B, and the structure complexity as SC, then the calculation formula of structure complexity is obtained: .

[0024] Step S2, according to the evaluation requirements, build the hierarchical index structure of the problem to be analyzed, and construct the judgment matrix to get the index weight; the specific steps are as follows: Step S21, constructing a hierarchical index structure according to the analyzed problem; the hierarchical index structure is: the first level influence factor set is {time complexity, space complexity, structure complexity}, the second level influence factor set is {O(n), decision time, problem size n, n storage space function f(n), code line number, node number, layer number, branch number}; Step S22, constructing a judgment matrix A according to the constructed hierarchical index structure; Step S23, calculating the maximum eigenvalue of the judgment matrix A and the characteristic vector corresponding to the maximum eigenvalue, and normalizing the characteristic vector; Step S24, performing consistency check on the judgment matrix A: when checking, first calculating the consistency index CI, the public formula is:

[0025] where λ max is the maximum eigenvalue, n is the order of the judgment matrix; then looking up the corresponding average random consistency index RI according to the matrix order, and finally calculating the consistency ratio CR, the calculation formula is:

[0026] If CR<0.1, it is determined that the judgment matrix A satisfies the given consistency standard, and the normalized characteristic vector obtained from S23 is used as the index weight; otherwise, if CR≥0.1, it is determined that the judgment matrix A does not satisfy the given consistency standard, then A needs to be modified, and the processes of S22 and S23 are repeated until A satisfies the given consistency standard.

[0027] Step S3, constructing a membership degree relationship matrix according to the evaluation index, and calculating the membership degree value of each index; the specific steps are as follows: Step S31, using the quantitative scale 0.1-0.9 scale method to perform roof beam description on the influence factors a i and a j to obtain the membership degree value r ij , i=1, 2,..., n, j=1, 2,..., n; Step S32, constructing an n-order membership degree matrix R according to the membership degree value, which is expressed as follows:

[0028] The actual meaning of the element r ij is: r ij represents the importance relationship of the influence factor a i to the influence factor a j , and r ij is the membership degree value of the "much more important than" fuzzy relationship.

[0029] Step S4, the rule base complexity is evaluated by combining the index weight and the membership value; the specific steps are as follows: Step S41, the weight vector W=(w1, w2, …, wn) is established according to the index weight of each layer index. n The weight vector is combined with the membership matrix to form a weighted comprehensive matrix B, B represents the comprehensive contribution degree vector of each factor in the rule base complexity:

[0030] The comprehensive contribution degree Bi of each factor in the comprehensive matrix in the rule base complexity is calculated by the following formula:

[0031] Step S42, the normalized operation is performed on the comprehensive contribution degree vector B to obtain the feature vector as the standard weight B' after de-fuzzification:

[0032] Step S43, the weight vector W=(w1, w2, …, wn) and the standard vector are used to calculate the complexity comprehensive evaluation value S, and the calculation formula is: n .

[0033] Step S5, the rule base reliability is evaluated based on the soft error analysis method; the specific steps are as follows: Step S51, the number of running state machines and behavior trees is set to m, which is determined by combining the actual application scene of the rule base and the evaluation accuracy requirement, and usually needs to cover enough running periods to capture potential occasional soft errors; the running state machine and the behavior tree model simulate various operation processes and logical judgment processes of the rule base in the actual running environment, during which a series of behaviors such as rule calling, condition matching, result outputting are automatically executed.

[0034] Step S52, keep running until m times of complete running are completed, even if program crashes, logical errors, abnormal results and other situations occur during running, it will also be recorded as a complete running and continue to execute the next one, in order to fully capture the influence of different types of soft errors on running, and record the number of error-free running n; S53, according to the number of error-free running n, the rule base reliability C is obtained, and the calculation formula is as follows: C=n / m.

[0035] Step S6, the rule base comprehensive evaluation is performed by combining the rule complexity evaluation analysis and the rule reliability analysis; the specific steps are as follows: ​Step S61, the rule reliability is divided into three levels, wherein: very reliable level 1: 0.8≤C≤1; more reliable level 2: 0.5≤C<0.8; unreliable level 3: 0≤C<0.5; Step S62, the rule complexity is divided into four levels: very complex level 1: 0.75<S≤1; more complex level 2: 0.5<S≤0.75; general level 3: 0.25<S≤0.5; not complex level 4: 0≤S≤0.25 S63, the levels of the two evaluations are integrated to obtain an evaluation score table:

[0036] By comparing with the values in the evaluation score table, the comprehensive evaluation of the rule base is obtained.

[0037] The rule base comprehensive evaluation method based on fuzzy analytic hierarchy process and soft error analysis improves the problems existing in the traditional analytic hierarchy process by using the fuzzy analytic hierarchy process, the consistency test of the matrix of the method is easier than that of the analytic hierarchy process; the fuzzy comprehensive judgment method has a simple mathematical model; the object system with many evaluation factors can be evaluated; the fuzzy transformation of the incomplete determined information increases the reliability of the evaluation, and the rule base complexity is evaluated through the structural complexity, time complexity and space complexity, and the rule reliability is evaluated, finally, the rule base is comprehensively evaluated by integrating the rule complexity evaluation and the rule reliability evaluation, so the method has practicality.

[0038] In example 2, taking the rule base based on state machine and behavior tree in the intelligent air combat as an example, a rule base comprehensive evaluation method based on fuzzy analytic hierarchy process and soft error analysis is used to study the comprehensive evaluation of the rule base.

[0039] The complexity of the rule base is evaluated by formulating the evaluation indexes of time complexity, space complexity and structural complexity; the time complexity is obtained, the interpretable model test uses 100 epochs in this time, and the time of each neural network decision is calculated, as shown in the table:

[0040] Finally, the average value of the interpretable neural network in obtaining 100 decisions is 20.1453s. By the theoretical method, the T(n) of the interpretable neural network model is 7n 2 +31n+17, so O(n)=O(n 2 ). The space complexity is obtained by counting the number of code lines, the number of loops, the number of edges, the number of nodes, the number of disconnected parts of the graph, the number n1 of different operators appearing in the program, the number n2 of different operands appearing in the program, the total number N1 of operators, and the total number N2 of operands, to obtain three indexes of the model, i.e., the number of code lines, the loop complexity, and the Halstead volume (also known as program complexity); the structure complexity is obtained by counting the number of nodes at each layer of the behavior tree as shown in the table:

[0041] Based on the rule base representation, i.e., the rule base table structure, the number of branches is counted as 26. Now, the number of nodes is defined as N, the number of layers of the behavior tree is L, the number of branches is B, and the structure complexity is SC. The calculation formula of the structure complexity is obtained as SC = N / L + B The hierarchical index structure of the problem to be analyzed is constructed. The first-level influence factor set S1 is {time complexity, space complexity, structure complexity}, and the second-level influence factor set S2 is {O(n), decision time, problem size n, function f(n) of the storage space occupied by n, number of code lines, number of nodes, number of layers, and number of branches}. The weight of each index is calculated by the fuzzy analytic hierarchy process, and the weights of the indexes are {0.21808417, 0.58810136, 0.19381447}.

[0042] The membership degree relationship matrix of the indexes is established according to the mutual influence between the indexes. The weight vector is established by using the index weights. The comprehensive contribution degree vector is obtained by multiplying the membership degree relationship matrix and the weight vector. The standard weight is obtained by normalizing the comprehensive contribution degree vector. The value range of the standard weight is 0-1, and S 11 , S 12 , and S 13 , respectively. Finally, the calculation formula of the rule base complexity value S is obtained as.

[0043] S = 0.21808417S 11 + 0.58810136S 12 + 0.19381447S 13 The value of S is finally calculated as 0.68387.

[0044] The rule base reliability C is calculated as follows when the rule base is run m times, and the number of times of normal operation without errors is n times. C = n / m The closer the value of C is to 1.0, the higher the reliability of the interpretable model. Finally, after testing, the value of C was 98%.

[0045] The obtained complexity value S and reliability C are mapped to the evaluation score table. The example used in this invention has a score of 10 according to the evaluation table. Therefore, this invention can effectively conduct a comprehensive evaluation of the rule base and give a conclusion.

[0046] Example 3, as Figure 2 As shown, a system for evaluating a rule base based on the analytic hierarchy process includes a module for constructing evaluation indicators, a module for constructing hierarchical indicator structures, a fuzzy analytic hierarchy process module, a soft error analysis module, and a comprehensive evaluation module. An evaluation index module is constructed. Based on the actual situation of the rule base, corresponding evaluation indexes are formulated. The evaluation indexes of the rule base are mainly established from three dimensions: time complexity, code complexity, and structural complexity. Constructing a hierarchical indicator structure module: Based on the evaluation indicators, further refine the indicators that each indicator needs to be evaluated, thereby constructing a more complete hierarchical indicator structure and more comprehensively evaluating the complexity of the rule base. The fuzzy hierarchical analysis module includes a weight calculation submodule and a membership matrix calculation submodule; The weight calculation submodule constructs a judgment matrix based on the evaluation index structure, normalizes the judgment matrix to obtain the index weights, and performs a consistency check on the judgment matrix to determine whether the obtained index weights are correct. The membership matrix calculation submodule obtains the membership relationship matrix through the mutual influence relationship between various indicators, obtains the comprehensive contribution vector by combining the membership matrix with the indicator weights, and obtains the standard weights by normalizing the contribution vector. The overall complexity evaluation of the rule base is obtained by multiplying the indicator weights and standard weights and then summing them. The soft error analysis module runs the rule base multiple times and records the number of successful runs. The reliability is obtained based on the total number of runs and the number of successful runs.

[0047] The comprehensive evaluation module determines the evaluation result of the rule base by combining the comprehensive complexity evaluation and reliability evaluation of the rule base.

[0048] Example 4, as Figure 3 As shown, a computer device includes a processor 101, a memory 102, a display module 103, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the steps in the method for evaluating a rule base based on the hierarchical analysis method described in Embodiment 1.

[0049] Embodiment 5. A computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements the steps of the method of evaluating a rule base based on analytic hierarchy process according to embodiment 1.

[0050] Those skilled in the art will appreciate that embodiments of the application can be supplied as a method, a system, or a computer program product. Accordingly, the application can be embodied in the form of hardware embodiments, software embodiments, or embodiments combining software and hardware aspects. Furthermore, the application can be embodied in the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, magnetic disk storage and optical storage) having computer-usable program code embodied therein.

[0051] The present application is described in reference to the flowchart and / or block diagrams of the method, apparatus (system) and computer program product according to embodiments of the application. It should be understood that each flow and / or block in the flowchart and / or block diagrams, and combinations of flows and / or blocks in the flowchart and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart and / or block diagrams block or blocks. Figure 1 one or more flows and / or blocks Figure 1 means for carrying out the function specified by the flow or flows and / or block or blocks.

[0052] These computer program instructions can also be stored in a computer- readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instructions which implement the flowchart and / or block diagrams block or blocks. Figure 1 one or more flows and / or blocks Figure 1 means for carrying out the function specified by the flow or flows and / or block or blocks.

[0053] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the flowchart and / or block diagrams block or blocks. Figure 1 one or more flows and / or blocks Figure 1 means for carrying out the function specified by the flow or flows and / or block or blocks.

[0054] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by a computer program instructing relevant hardware, and the program can be stored in a computer readable storage medium. When the program is executed, the program can include the processes of the above-mentioned embodiment methods. The storage medium can be a magnetic disc, an optical disc, a Read-Only Memory (ROM) or a Random Access Memory (RAM) and the like.

[0055] The above description of disclosed embodiments enables one of ordinary skill in the art to make or use the application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein can be applied to other embodiments without departing from the spirit or scope of the application. Thus, the present application is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for evaluating a rule base based on the Analytic Hierarchy Process (AHP), characterized in that: Includes the following steps: Step S1: Develop evaluation indicators based on the actual situation of rule base complexity assessment, and determine the corresponding methods for obtaining these indicators. Step S2: Construct a hierarchical index structure for the problem to be analyzed based on the evaluation requirements, and construct a judgment matrix to obtain the index weights; Step S3: Construct a membership relation matrix based on the evaluation indicators and calculate the membership value of each indicator; Step S4: Combine the weights of each indicator and their membership values ​​to complete the evaluation of the complexity of the rule base; Step S5: Evaluate the reliability of the rule base based on soft error analysis methods; Step S6: Conduct a comprehensive evaluation of the rule base by combining rule complexity assessment and rule reliability analysis.

2. The method for evaluating a rule base based on the analytic hierarchy process according to claim 1, characterized in that: The specific steps for S1 are as follows: Step S11: The rule base consists of a state machine and a behavior tree. Based on the complexity of the rule base of the state machine and behavior tree, evaluation indicators are formulated, including time complexity, code complexity, and structural complexity. Step S12: Time complexity is a relative measure of the running time of an algorithm. The time spent by an algorithm is directly proportional to the number of times the statements in the algorithm are executed. The number of times the statements in an algorithm are executed is called the time frequency, denoted as T(n), where n is called the problem size. The method to obtain the time is: place the start and end statements at the beginning and end of the interpretable model code, respectively, calculate the time of each epoch of the model, take multiple data, and complete the acquisition of the time parameter. If there is an auxiliary function f(n) such that when n approaches infinity, the value of lim(T(n) / f(n)) is a non-zero constant, then f(n) is said to be a function of the same order of magnitude as T(n). Finally, the time performance of the algorithm is represented by O, denoted as T(n) = O(f(n)), and O(f(n)) is called the time complexity of the algorithm. Step S13, the code complexity is the same as the space complexity, calculated by the formula: S(n) = O(f(n)), where n is the problem size and f(n) is a function of the storage space occupied by the statement in terms of n; Step S14: Based on the table structure of the rule base, define the number of nodes as N, the number of bean tree layers as L, the number of branches as B, and the structural complexity as SC. Then, the formula for calculating the structural complexity is: 。 3. The method for evaluating a rule base based on the analytic hierarchy process according to claim 2, characterized in that: The specific steps for S2 are as follows: Step S21: Construct a hierarchical index structure based on the analyzed problem; Step S22: Construct judgment matrix A based on the constructed hierarchical index structure; Step S23: Calculate the largest eigenvalue and the corresponding eigenvector of the judgment matrix A, and normalize the eigenvector; Step S24, perform a consistency check on the judgment matrix A: During the check, first calculate the consistency index CI using the following formula: Where λ max Let n be the largest eigenvalue and n be the order of the judgment matrix; Next, find the corresponding average random consistency index RI based on the matrix order, and finally calculate the consistency ratio CR. The calculation formula is as follows: If CR < 0.1, then the judgment matrix A is determined to meet the given consistency criterion, and the normalized eigenvector obtained by S23 is used as the index weight; otherwise, if CR ≥ 0.1, then the judgment matrix A is determined to not meet the given consistency criterion, and A needs to be modified, and the process of S22 and S23 is repeated until A meets the given consistency criterion.

4. The method for evaluating a rule base based on the analytic hierarchy process according to claim 3, characterized in that: The specific steps for S3 are as follows: Step S31: Using the quantity scale method of 0.1-0.9, analyze the influencing factor a. i Influencing factor a j The membership value r is obtained by describing the top beam. ij ,i=1,2,...,n,j=1,2,...,n; Step S32: Construct an n-order membership matrix R based on the membership values, as follows: 。 5. The method for evaluating a rule base based on the analytic hierarchy process according to claim 4, characterized in that: The specific steps for S4 are as follows: Step S41: Establish a weight vector W = (w1, w2, ..., w3) based on the weights of the indicators at each level. n The weight vector is combined with the membership matrix to form a weighted composite matrix B, where B represents the overall contribution vector of each factor to the complexity of the rule base. The combined contribution B of each factor in the rule base complexity in the comprehensive matrix i The calculation formula is: Step S42: Normalize the comprehensive contribution vector B to obtain the feature vector, which is then used as the standard weight B′ after defuzzification. Step S43, using the weight vector W=(w1, w2, ..., w n The complexity evaluation value S is obtained by calculating the complexity of the standard vector. The calculation formula is as follows: 。 6. The method for evaluating a rule base based on the analytic hierarchy process according to claim 5, characterized in that: The specific steps for S5 are as follows: Step S51: Set the number of times the state machine and behavior tree are run to be m, and run the state machine and behavior tree models; Step S52: Continuously run until the program has completed m times. Program crashes or errors are also considered as completing one run. Record the number of times the program runs without errors, n. S53, based on the number of times n runs without errors, the reliability C of the rule base is derived, and the calculation formula is as follows: C=n / m.

7. The method for evaluating a rule base based on the analytic hierarchy process according to claim 6, characterized in that: The specific steps for S6 are as follows: Step S61: Divide the reliability of the rule into 3 levels, where: very reliable level 1: 0.8≤C≤1; relatively reliable level 2: 0.5≤C<0.8; unreliable level 3: 0≤C<0.5; Step S62: Divide the rule complexity into four levels: Very complex level 1: 0.75 < S ≤ 1; Moderately complex level 2: 0.5 < S ≤ 0.75; Average level 3: 0.25 < S ≤ 0.5; Uncomplicated level 4: 0 ≤ S ≤ 0.25 Step S63: Combine the two evaluation levels to obtain the assessment score table: The overall evaluation of the rule base is obtained by comparing the values ​​with those in the evaluation score table.

8. A system for evaluating a rule base based on the analytic hierarchy process (AHP), characterized in that: This includes a module for constructing evaluation indicators, a module for constructing hierarchical indicator structures, a module for fuzzy hierarchical analysis, a module for soft error analysis, and a comprehensive evaluation module; An evaluation index module is constructed. Based on the actual situation of the rule base, corresponding evaluation indexes are formulated. The evaluation indexes of the rule base are mainly established from three dimensions: time complexity, code complexity, and structural complexity. Constructing a hierarchical indicator structure module: Based on the evaluation indicators, further refine the indicators that each indicator needs to be evaluated, thereby constructing a more complete hierarchical indicator structure and more comprehensively evaluating the complexity of the rule base. The fuzzy hierarchical analysis module includes a weight calculation submodule and a membership matrix calculation submodule; The weight calculation submodule constructs a judgment matrix based on the evaluation index structure, normalizes the judgment matrix to obtain the index weights, and performs a consistency check on the judgment matrix to determine whether the obtained index weights are correct. The membership matrix calculation submodule obtains the membership relationship matrix through the mutual influence relationship between various indicators, obtains the comprehensive contribution vector by combining the membership matrix with the indicator weights, and obtains the standard weights by normalizing the contribution vector. The overall complexity evaluation of the rule base is obtained by multiplying the indicator weights and standard weights and then summing them. The soft error analysis module runs the rule base multiple times and records the number of successful runs. The reliability is obtained based on the total number of runs and the number of successful runs. The comprehensive evaluation module determines the evaluation result of the rule base by combining the comprehensive complexity evaluation and reliability evaluation of the rule base.

9. An electronic device comprising a memory (102), a processor (101), a display module (103), and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps of the method for evaluating a rule base based on the analytic hierarchy process as described in any one of claims 1 to 7.

10. A readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method for evaluating a rule base based on the analytic hierarchy process as described in any one of claims 1 to 7.

Images