Urban park risk toughness assessment method and system based on probabilistic reasoning

Abstract

The invention provides an urban park risk toughness assessment method and system based on probabilistic reasoning, and the method comprises the steps: carrying out the recognition of risk factors of a park target region, and building a risk toughness assessment index system suitable for a park; constructing a park risk toughness assessment Bayesian network model according to the risk toughness assessment index system and the dependency relationship of the influence factors in the system; determining prior information of the conditional probability of a root node in the Bayesian network by adopting a fuzzy Bayesian method, and initializing model parameters to obtain the prior probability of the root node of the Bayesian network model; the non-root node conditional probability of the Bayesian network model is calculated, and the risk transfer relation between the nodes is quantified; and based on the actual situation of the park, performing reasoning calculation on the target node and the related father node to obtain a park risk toughness evaluation result. The method can scientifically evaluate the toughness characteristics of different areas, especially parks, in a city, and provides a decision basis for improving the toughness of the city.

Description

Technical Field

[0001] The present invention relates to the field of urban planning and management technology, and in particular to a method and system for assessing urban park risk resilience based on probabilistic reasoning, as well as a corresponding computer terminal and computer-readable storage medium. Background Art

[0002] In recent years, with the acceleration of global urbanization and the frequent occurrence of various emergencies, enhancing the resilience of different urban regions has become a hot topic of widespread concern. Regional resilience refers to the ability of a specific geographic area (such as a park, commercial district, residential area, or community) to maintain its critical functions, recover rapidly, and adapt to new environments when faced with external shocks such as natural disasters, social conflicts, or public health crises. In complex urban systems, different regions exhibit significant differences in disaster risk and resilience due to differences in function, population density, and infrastructure. However, regional resilience is influenced by multiple dimensions, uncertainty, and dynamic changes, and traditional evaluation methods often struggle to fully characterize these characteristics. In recent years, Bayesian network-based modeling methods have gradually become an important tool for quantifying regional resilience due to their ability to handle causal relationships and uncertainty in complex systems. However, the use of Bayesian network modeling methods for park resilience assessment often suffers from the complexity of the conditional probability quantification process and the difficulty in identifying key variables.

[0003] After searching, we found:

[0004] The Chinese invention patent application, "Multi-level Urban Community Resilience Assessment Index System and Evaluation Method and System," with publication number CN115983703A, includes the following steps: first, based on the principles of independence and representativeness, a preliminary selection of resilience assessment indicators is conducted. A Delphi method is then used to screen and construct an evaluation system consisting of primary, secondary, and tertiary indicators. A step-by-step weighted approach is then used to calculate the scores for each level of indicators. The three-level indicators are scored in turn and aggregated to obtain the scores for the secondary and primary indicators, generating distribution maps for social and environmental resilience and institutional and management resilience. Finally, based on the scoring results, weak links are identified and targeted improvement strategies are formulated. This method integrates multiple indicators through expert scoring, weighted averaging, and Bayesian theory. The resulting evaluation model fails to establish hierarchical dependencies or causal influence paths between indicators, lacks structured logical deduction capabilities, and is unable to implement bottom-up or top-down path reasoning and failure path identification in complex systems, hindering the identification and precise intervention of high-risk nodes.

[0005] How to scientifically evaluate the resilience characteristics of different urban areas, especially industrial parks, to provide a basis for decision-making on improving urban resilience remains an urgent challenge in this field. Currently, no descriptions or reports of technologies similar to the present invention have been found, nor has any similar information been collected domestically or internationally. Summary of the Invention

[0006] In response to the above-mentioned deficiencies in the prior art, the present invention provides a method and system for assessing urban park risk resilience based on probabilistic reasoning, and also provides a corresponding computer terminal and computer-readable storage medium.

[0007] According to one aspect of the present invention, a method for assessing urban park risk resilience based on probabilistic reasoning is provided, comprising:

[0008] Identify risk factors in the park's target areas and establish a risk resilience assessment indicator system applicable to the park;

[0009] Based on the risk resilience assessment indicator system and the dependency relationship of the influencing factors in the system, a Bayesian network model for park risk resilience assessment is constructed;

[0010] A fuzzy Bayesian method is used to determine the prior information of the conditional probability of the root node in the Bayesian network, and the model parameters are initialized to obtain the prior probability of the root node of the Bayesian network model;

[0011] Calculating the conditional probabilities of non-root nodes in the Bayesian network model and quantifying the risk transfer relationship between nodes;

[0012] Based on the actual situation of the park, the target node and its related parent nodes are inferred and calculated to obtain the park risk resilience assessment results.

[0013] According to another aspect of the present invention, a system for assessing urban park risk resilience based on probabilistic reasoning is provided, comprising:

[0014] A resilience indicator system construction module, which is used to identify risk factors in the park's target areas and establish a risk resilience assessment indicator system applicable to the park;

[0015] A Bayesian network construction module, which is used to construct a Bayesian network model for park risk resilience assessment based on the risk resilience assessment indicator system and the dependency relationship of influencing factors in the system;

[0016] A node conditional probability calculation module is used to determine the prior information of the conditional probability of the root node in the Bayesian network using a fuzzy Bayesian method, initialize the model parameters, and obtain the prior probability of the root node of the Bayesian network model; calculate the conditional probability of non-root nodes in the Bayesian network model, and quantify the risk transfer relationship between nodes;

[0017] The risk change assessment module performs inference calculations on the target node and its related parent nodes based on the actual situation of the park to obtain the park risk resilience assessment results.

[0018] According to a third aspect of the present invention, a computer terminal is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the terminal can be used to execute the method described above in the present invention, or to execute the system described above in the present invention.

[0019] According to a fourth aspect of the present invention, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, it can be used to execute the method described above in the present invention, or to run the system described above in the present invention.

[0020] Due to the adoption of the above technical solution, the present invention has at least one of the following beneficial effects compared with the prior art:

[0021] The urban park risk resilience assessment method and system based on probabilistic reasoning provided by the present invention quantifies and evaluates the resilience performance of a region when facing various disturbance events by introducing Bayesian networks and probabilistic reasoning, providing decision makers with scientific resilience assessment data, optimizing risk management and emergency response strategies, thereby improving the resilience and recovery capabilities of the region when facing emergencies, and promoting urban sustainable development and disaster prevention and control.

[0022] The urban park risk resilience assessment method and system based on probabilistic reasoning provided by the present invention overcome the problems that traditional resilience assessment methods have difficulty in dealing with multi-factor coupling, fuzzy causal relationships between indicators, and uncertain information. It explicitly represents the causal dependency relationship between indicators, supports reasoning and analysis under incomplete information, and more realistically portrays the system dynamic behavior and response process of the park under disaster impact.

[0023] The urban park risk resilience assessment method and system based on probabilistic reasoning provided by the present invention have good mathematical stability and can extract weights without changing the consistency of the judgment matrix, ensuring the rationality and transparency of weight distribution, and laying a solid foundation for subsequent probability quantification.

[0024] The urban park risk resilience assessment method and system based on probabilistic reasoning provided by the present invention avoid the problem of exponential growth of CPT with the increase of the number of parent nodes in traditional Bayesian networks, and improve the scalability of the model.

[0025] The urban park risk resilience assessment method and system based on probabilistic reasoning provided by the present invention can simulate the response changes of the system under different disturbance scenarios through Bayesian reasoning, forward deduce the resilience level of the results, and identify high-risk inducement paths through reverse reasoning, providing data support for pre-emptive intervention and risk control. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Other features, objects and advantages of the present invention will become more apparent upon reading the detailed description of non-limiting embodiments with reference to the following drawings:

[0027] Figure 1 This is a workflow diagram of a method for assessing urban park risk resilience based on probabilistic reasoning in a preferred embodiment of the present invention.

[0028] Figure 2 This is a schematic diagram of the component modules of an urban park risk resilience assessment system based on probabilistic reasoning in a preferred embodiment of the present invention.

[0029] Figure 3 This is a workflow diagram of an urban park risk resilience assessment method based on probabilistic reasoning in a specific application example of the present invention.

[0030] Figure 4 Schematic diagram of a Bayesian network in a specific application example of the present invention.

[0031] Figure 5 This is a single-node impact analysis diagram in a specific application example of the present invention.

[0032] Figure 6 The figure is a schematic diagram of the calculation results of the Bayesian network reverse reasoning in a specific application example of the present invention.

[0033] Figure 7 The figure is a schematic diagram of the calculation results of the Bayesian network sensitivity analysis in a specific application example of the present invention.

[0034] Figure 8 Schematic diagram of sensitivity analysis calculation results of R node in a specific application example of the present invention. DETAILED DESCRIPTION

[0035] The following is a detailed description of an embodiment of the present invention. This embodiment is implemented based on the technical solution of the present invention, and provides a detailed implementation method and specific operation process. It should be noted that those skilled in the art may make various modifications and improvements without departing from the scope of the present invention, and these modifications and improvements fall within the scope of protection of the present invention.

[0036] Existing urban park risk resilience assessment technologies fail to establish hierarchical dependencies or causal influence paths between indicators, lack structured logical reasoning capabilities, and cannot achieve bottom-up or top-down path reasoning and failure path identification in complex systems, which is not conducive to the identification and precise intervention of high-risk nodes.

[0037] In response to the above problems, an embodiment of the present invention provides an urban park risk resilience assessment method based on probabilistic reasoning. The method fully integrates structural modeling, expert knowledge, probabilistic reasoning and sensitivity analysis to construct a systematic and scalable urban park resilience assessment technology. It has strong practicality and decision-making support value, and is suitable for resilience analysis and improvement path identification in complex system scenarios with incomplete multi-source data or dominated by expert knowledge. It can scientifically evaluate the resilience characteristics of different areas in the city, especially parks, and provide a decision-making basis for improving urban resilience.

[0038] Specifically, if Figure 1 As shown, the urban park risk resilience assessment method based on probabilistic reasoning provided in this embodiment may include:

[0039] S1, identify risk factors in the park’s target area and establish a risk resilience assessment indicator system applicable to the park;

[0040] S2, based on the risk resilience assessment indicator system and the dependency relationship of the influencing factors in the system, a Bayesian network model for park risk resilience assessment is constructed;

[0041] S3, using the fuzzy Bayesian method to determine the prior information of the conditional probability of the root node in the Bayesian network, initialize the model parameters, and obtain the prior probability of the root node of the Bayesian network model;

[0042] S4, calculates the conditional probability of non-root nodes in the Bayesian network model and quantifies the risk transfer relationship between nodes;

[0043] S5, based on the actual situation of the park, inference calculation is performed on the target node and its related parent nodes to obtain the park risk resilience assessment results.

[0044] In some preferred embodiments, the above S1 may further include:

[0045] Risk factors were identified for the park's target areas, including natural, engineering, social, and institutional factors. This led to the construction of a risk resilience assessment index system encompassing five dimensions: engineering resilience, organizational resilience, social resilience, economic resilience, and institutional resilience. Furthermore, utilizing an expert system, literature review, expert interviews, the Delphi method, and historical case studies, a systematic analysis of the various risk factors facing the park, including those related to nature, engineering, society, and institutions, was conducted. Based on this, and combining actual research results with multi-source data, a risk resilience assessment index system encompassing five dimensions: engineering resilience, organizational resilience, social resilience, economic resilience, and institutional resilience was constructed.

[0046] In some preferred embodiments, the above S2 may further include:

[0047] S21 constructs a directed acyclic graph structure of the risk resilience assessment indicator system based on the dependency relationship between the influencing factors in the system, forming a Bayesian network model; where:

[0048] Each node in the Bayesian network model represents a random variable, the edge represents the causal relationship between variables, and the direction of the edge represents the dependency order of conditional probability. Each node is associated with a conditional probability table, which is used to describe the probability distribution of the node under the condition that its parent node is known. By decomposing the joint probability distribution, the probability relationship between multiple random variables is structured as follows:

[0049]

[0050] Where P(R) represents the joint probability; n represents the number of nodes in the Bayesian network, x1, x2, ..., x n represents a random variable (i.e., a node) in the Bayesian network, x i represents the i-th variable, R(x i ) represents node x i The parent node set, P(x i ∣∣R(x i )) represents a given parent node R(x i ) condition, node x i The conditional probability distribution of .

[0051] In some preferred embodiments, the above S3 may further include:

[0052] S31, a triangular membership function is used to describe the fuzzy probability distribution of the risk variable; wherein the triangular membership function μ(x) is defined as follows:

[0053]

[0054] Where x is the evaluation value of the node in a specific situation; l, m, n represent the lower bound, peak value and upper bound of the triangular fuzzy number respectively, and l <m<n;

[0055] S32: Based on the expert system, obtain the expert survey results, and classify the expert survey results (including but not limited to: qualitative scores or quantitative valuations of dimensions such as risk event level, impact severity, and emergency response efficiency) into different levels of fuzzy language levels, where each fuzzy language level corresponds to a triangular fuzzy number, and each expert is assigned a different weight. Then:

[0056]

[0057] Where k is the number of experts, e j For each expert E j The weight, pj is the fuzzy evaluation value, and p′ is the weighted average fuzzy value:

[0058] S33, defuzzifies the weighted average fuzzy value and converts it into an accurate numerical value, which is used as the prior information of the conditional probability of the root node in the Bayesian network model, obtains the prior probability of each root node of the Bayesian network model for park risk resilience assessment, and completes the model parameter initialization process.

[0059] In the above steps, according to the correspondence between fuzzy language levels and triangular fuzzy numbers, the experts' evaluation results on the risk probability of the root nodes are matched, and the risk probability of each root node is quantified through fuzzy weighted calculation and fuzzy operation. The fuzzy evaluation results of the risk probability of each root node are expressed as triangular fuzzy numbers, and weighted average calculation is performed according to the expert weights to obtain the weighted mean statistical results of the risk probability of each root node. According to the scoring results, the weighted mean of the risk probability of each root node is processed with triangular fuzzy numbers to obtain the prior probability of each root node in the Bayesian network model of park risk resilience assessment.

[0060] In some preferred embodiments, the above S4 may further include:

[0061] S41, based on the conditional probability table in the Bayesian network model, the square root method is used in combination with the expert judgment matrix, and the Noisy-OR method is used to quantitatively analyze the conditional probability table to calculate the conditional probability of the non-root nodes of the Bayesian network model; where:

[0062] Based on the Noisy-OR method, let the parent node probability of each cause leading to the event be P i , the state of the parent node is X i , takes 0 or 1, then the Noisy-OR model of the probability of the event occurring P (event occurrence) is:

[0063]

[0064] Among them, P i The specific calculation method is:

[0065] The jth expert scores the influence degree of a parent node's child node as a j , the number of parent nodes of the child node is m, then the expert scoring matrix A=[a1,a2,…,a m ], for each score matrix, the square root method is used to synthesize expert opinions to obtain the parent node probability P i for:

[0066]

[0067] The obtained P iSubstituting the above Noisy-OR model into the model, a non-root node conditional probability table is generated, providing a data basis for the inference analysis of the Bayesian network model for park risk resilience assessment.

[0068] In some preferred embodiments, the above S5 may further include:

[0069] S51. For the non-root node conditional probability table obtained in the above steps, based on the Bayesian reasoning mechanism, the park resilience of the target node is subjected to forward probability analysis and contribution decomposition through inference calculation, the key influence of each related parent node of the target node on the overall resilience level is identified, and its contribution to the comprehensive resilience of the park is quantified, thereby evaluating its weak links and strengths.

[0070] In some preferred embodiments, the above-mentioned inference calculation may further include:

[0071] S511, Single Node Impact Analysis: Assume that any key parent node in the model fails completely and observe the impact of the parent node on the resilience level of the parent node and the target node to evaluate the key bottleneck elements and vulnerabilities in the model;

[0072] S512, reverse reasoning analysis: Assuming the target node's resilience is in a failed state, backtrack and reason about the state changes and failure probabilities of each upstream node to assess potential high-risk impact paths and sensitive nodes. Specifically, assuming a node is in a failed state, calculate the state changes of its upstream nodes and the corresponding failure probabilities, thereby identifying high-risk paths and sensitive nodes.

[0073] S513, Sensitivity Analysis: Quantify the sensitivity of each indicator to the target node, used to evaluate the impact of key variables and direction, and assist in key interventions and resource allocation.

[0074] The probabilistic reasoning-based urban park risk resilience assessment method provided in the above-mentioned embodiment of the present invention constructs a hierarchical system containing multiple indicators (such as engineering resilience and organizational resilience), determines the network structure and weights through fuzzy hierarchical analysis and expert scoring, uses the Noisy-OR model to simplify conditional probability table modeling, and calculates the positive probability of nodes based on expert evaluation. Bayesian reasoning is then used to quantify the contribution of each indicator to comprehensive resilience, and single-node failure impact analysis and reverse path backtracking are conducted to identify key vulnerabilities and risk paths. Finally, sensitivity analysis and visualization tools are used to reveal core influencing factors and provide decision support for resilience improvement. This method integrates expert knowledge and probabilistic reasoning and is suitable for complex system resilience assessment.

[0075] Based on the same inventive concept, an embodiment of the present invention also provides an urban park risk resilience assessment system based on probabilistic reasoning.

[0076] Specifically, if Figure 2 As shown, the urban park risk resilience assessment system based on probabilistic reasoning provided in this embodiment may include:

[0077] A resilience indicator system construction module, which is used to identify risk factors in the park's target areas and establish a risk resilience assessment indicator system applicable to the park;

[0078] A Bayesian network construction module, which is used to construct a Bayesian network model for park risk resilience assessment based on the risk resilience assessment indicator system and the dependencies of the influencing factors in the system;

[0079] Node conditional probability calculation module, which is used to determine the prior information of the conditional probability of the root node in the Bayesian network using the fuzzy Bayesian method, initialize the model parameters, and obtain the prior probability of the root node of the Bayesian network model; calculate the conditional probability of non-root nodes in the Bayesian network model, and quantify the risk transfer relationship between nodes;

[0080] The risk change assessment module performs inference calculations on the target node and its related parent nodes based on the actual situation of the park to obtain the park risk resilience assessment results.

[0081] It should be noted that the steps in the method provided by the present invention can be implemented using corresponding modules, devices, units, etc. in the system. Those skilled in the art can refer to the technical solution of the method to implement the composition of the system, that is, the embodiments in the method can be understood as preferred examples of constructing the system, which will not be elaborated here.

[0082] The technical solution provided by the above embodiment of the present invention is further described in detail below with reference to a specific application example.

[0083] like Figure 3 As shown in FIG, the urban park risk resilience assessment method based on probabilistic reasoning involved in this specific application example includes the following steps:

[0084] Step 1: Select the area to be assessed and identify the risk factors in the specific area. The urban park resilience assessment index system is an important tool for measuring a city's ability to resist shocks, recover and adapt when facing shocks such as natural disasters and emergencies. Based on the five dimensions of engineering resilience, organizational resilience, social resilience, economic resilience and institutional resilience, the system constructs a systematic and comprehensive indicator framework from multiple levels such as infrastructure, personnel quality, park integration, economic resources and management mechanisms. The established indicator system is shown in Table 1:

[0085] Table 1 Urban park resilience assessment index system

[0086]

[0087]

[0088]

[0089] Step 2: Construct a resilience Bayesian network for the park according to the known risk resilience evaluation index system and the dependency relationships of the influencing factors in the system, as shown in Figure 4 shown;

[0090] Step 3: Use the fuzzy Bayesian method to determine the prior information of the conditional probabilities of the nodes in the Bayesian network, complete the parameter initialization process of the model, and obtain the prior probabilities of the root nodes of the resilience evaluation Bayesian network for the park.

[0091] Obtain the experimental probabilities through the fuzzy set method and conduct fuzzy evaluation with the help of expert experience. When specifically implemented, the fuzzy probability distribution of the risk variables is described by the triangular membership function. The triangular membership function μ(x) is defined as follows:

[0092]

[0093] where l, m, and n represent the lower bound, peak value, and upper bound of the triangular fuzzy number respectively, and satisfy l < m < n. In order to more accurately quantify the possibility of risk factors, divide the expert survey results into different levels of fuzzy languages according to the actual situation. Each fuzzy language corresponds to a triangular fuzzy number, and different weights are assigned to different experts. Suppose there are k experts, and the weight of each expert E j is e j , and its fuzzy evaluation value is p j , then the weighted average fuzzy value p′ is expressed as:

[0094]

[0095] To conduct a quantitative evaluation of each risk factor, a fuzzy language level will be assigned to each risk factor. The risk levels are divided into seven levels: extremely low, low, relatively low, medium, relatively high, high, and extremely high. To quantify the fuzzy levels, triangular fuzzy numbers are used for representation. Each fuzzy level corresponds to a triangular fuzzy number. The fuzzy numbers are defined in Table 2. For each risk factor, experts provide corresponding fuzzy values according to their judgments.

[0096] Table 2 Fuzzy Languages and Corresponding Triangular Fuzzy Numbers

[0097]

[0098] To evaluate the resilience indicators of Shanghai Park X, five experts from various fields were invited to participate. These experts, all with extensive experience in park management and related fields, included park planning experts (E1), facility maintenance experts (E2), traffic organization experts (E3), emergency management experts (E4), and comprehensive management experts (E5). They conducted professional assessments of each node in the park resilience evaluation indicators and provided authoritative evidence for the prior probabilities of the nodes in the Bayesian network.

[0099] According to the needs of park resilience evaluation, the judgment matrix method is used to quantify the weight of each expert. Judgment matrix A=(a ij ) n×n Each element a in ij Indicates expert E i and E j The relative importance of. The weight B is calculated by geometric mean method. i , the formula is as follows:

[0100]

[0101] Among them, n is the total number of experts, a ij is the element in the judgment matrix.

[0102] Next, to B i Perform normalization to obtain the final weight C of each expert i :

[0103]

[0104] The calculated relative weight C i The results are shown in Table 3.

[0105] Table 3 Expert weight judgment matrix

[0106]

[0107]

[0108] Based on the correspondence between fuzzy language and triangular fuzzy numbers in Table 3, the expert evaluation results of the root node risk probability were matched. The risk probability of each root node was quantified through fuzzy weighted calculation and fuzzy operation. After expressing the fuzzy evaluation results of each root node risk probability as triangular fuzzy numbers, a weighted average calculation was performed based on the expert weights, ultimately resulting in the weighted mean statistical results of the risk probability of each root node, as shown in Table 4.

[0109] Table 4 Weighted mean statistics of triangular fuzzy numbers

[0110]

[0111]

[0112] According to the scoring results, the weighted mean of each root node is processed with triangular fuzzy numbers to obtain the prior probability of each root node of the Bayesian network of the park resilience index, as shown in Table 5.

[0113] Table 5 Prior probability of the root node of the Bayesian network for campus resilience assessment

[0114]

[0115] Step 4: Calculate the conditional probability of non-root nodes and construct a conditional probability table.

[0116] The conditional probability tables for non-root nodes in the Bayesian network used in the campus resilience assessment were quantitatively analyzed using the square root method, combined with an expert judgment matrix, and the Noisy-OR method. The Noisy-OR method is a classic and efficient Bayesian network probabilistic modeling tool suitable for describing the combined impact of multiple causes on a single outcome. Its fundamental assumption is that the impact of each conditional variable on the target node is independent of each other. This assumption effectively simplifies the computational complexity of the conditional probability table, allowing the probability distribution of the target variable to be derived even under multivariate conditions.

[0117] The Noisy-0R method assumes that the probability of each cause leading to the parent node of the event is P i , the state of the parent node (value is 0 or 1) is X i , then the probability of the event occurring is:

[0118]

[0119] In this step, taking the evaluation of public operating equipment reliability (M1) as an example, M1 is determined by five parent nodes (X1 to X5), which correspond to key characteristics such as equipment maintenance response speed, equipment utilization, equipment operation stability, equipment maintenance investment, and safety drill frequency. In the Noisy-OR method, the probability calculation formula for M1 is:

[0120]

[0121] Through the above formula, the independent contribution of each indicator can be accumulated one by one, and the comprehensive activation probability of M1 can be derived. Combined with the definition of fuzzy membership, the generated membership value can reflect the comprehensive impact of different indicator combinations on the reliability of the target node.

[0122] Step 5: Quantify the risk transfer relationship between nodes. Based on the actual situation of the park, perform inference calculations on the core nodes and their related parent nodes to evaluate the risk changes under different scenarios.

[0123] Based on the urban park risk conditional probability table obtained in the above steps, a Bayesian inference mechanism is used to conduct a positive probability analysis and contribution decomposition of the target node "park resilience," identifying the key influences of each parent node on the overall resilience level. For example, the positive probability of each primary and secondary indicator is calculated to quantify its contribution to the park's overall resilience, thereby assessing its weaknesses and strengths.

[0124] To enhance the depth and applicability of the system evaluation, the following three types of analysis are further conducted:

[0125] (1) Single node impact analysis

[0126] The single node impact analysis aims to quantify the independent impact of a single indicator on the resilience level of the parent node by setting the condition of complete failure of the single indicator, thereby revealing the importance of key factors in complex systems. Taking the reliability of public operation equipment (M1) as the target, the changes in the resilience level when its parent node fails are evaluated, such as Figure 5 and as shown in Table 6.

[0127] The failure of equipment maintenance response speed (X1) has the most significant impact on the M1 resilience level. Its failure causes the M1 resilience level to drop to 83, indicating that fast and efficient maintenance response is a core element in maintaining equipment reliability. This result reflects that in complex systems, time-sensitive tasks (such as maintenance response) have a higher weight for system stability. Similarly, the failure of equipment operation stability (X3) causes the M1 resilience level to drop to 89, further illustrating the fundamental supporting role of equipment daily operation performance in the reliability of public operation equipment. In contrast, the impact of single failures of equipment utilization (X2), equipment maintenance investment (X4), and safety drill frequency (X5) is relatively small, indicating that although these factors are important, their short-term failure will not have catastrophic consequences for overall reliability.

[0128] Table 6 Analysis of changes in single node resilience level

[0129]

[0130] (2) Reverse reasoning analysis

[0131] By performing reverse reasoning analysis on the Bayesian network model for park resilience assessment, we can intuitively observe the mutual influence relationship between the variables of each node and their contribution to the overall resilience level of the park. Figure 6 To explore the key factors of campus resilience failure, the target node campus resilience (R1) is set to be in a failure state (no, with a probability of 100%), and the state changes of other indicators and their impact on the overall resilience failure are analyzed through Bayesian network reasoning.

[0132] Based on the inference results of the first-level indicators, the failure probability of engineering resilience (T1) reached 58%, indicating that engineering resilience, as one of the core pillars of park resilience, is highly vulnerable in the context of park resilience failure. As a key component directly supporting the daily operations of the park, the high failure probability of engineering resilience indicates that the park's infrastructure (such as public operating equipment and emergency equipment) suffers from insufficient maintenance, degraded performance, or weak emergency response capabilities. These problems can significantly reduce the overall resilience of the park in extreme situations (such as disasters and emergencies). The high failure probability of engineering resilience also indicates that the daily operations of the park are highly dependent on the normal operation of the infrastructure. This high dependence requires risk mitigation through strengthening facility reliability and redundant design.

[0133] Among the secondary indicators, facility and system resilience (F1), personnel emergency preparedness (F4), and park adaptability (F7) warrant special attention. Facility and system resilience (F1) demonstrates high importance in the reasoning, indicating that infrastructure maintenance and management play a key role in the overall resilience of the park. Personnel emergency preparedness (F4) emphasizes the necessity of employees' emergency response capabilities for rapidly handling emergencies. Park adaptability (F7) reveals the indispensability of social collaboration in disaster response.

[0134] Among the three-level indicators, the reliability of public operation equipment (M1), the disaster response capability of management personnel (M9), and the disaster emergency response capability of the park (M1 14 Among them, M1 is directly related to the stability and security of infrastructure operation, M9 reflects the decision-making and coordination capabilities of management in emergency scenarios, and M 14 It is an important indicator to measure the park's emergency response mechanism and resource allocation efficiency.

[0135] Among the four-level indicators, equipment maintenance response speed (X1), safety drill frequency (X5), management personnel safety drill frequency (X 25 ), the completeness of the park's disaster emergency plan (X 40 ) are specific key influencing factors. Among them, X1 reflects the rapid response capability of the equipment in an emergency, X5 and X 25 They respectively emphasized the importance of safety awareness of infrastructure operation and management personnel; X 40 It reflects the degree of perfection of the park's anti-risk mechanism and is an important basis for enhancing the park's resilience.

[0136] (3) Sensitivity analysis

[0137] Conduct sensitivity analysis on the nodes, that is, find out which node factors in the Bayesian network structure model will have a greater impact on the target node "campus resilience (R)" when they change, so as to focus on these node factors during risk control. Figure 7 As shown in the figure, the color depth of the node reflects the sensitivity of the node.

[0138] The results show that among the primary indicators, engineering resilience (T1) and organizational resilience (T2) have high sensitivities, indicating that infrastructure maintenance response capabilities and the stability of emergency response organizations have a significant impact on overall resilience. Social resilience (T3) also has significant sensitivity, highlighting the central role of park participation and collaboration in park resilience. Among the secondary indicators, facility and system resilience (F1), personnel emergency preparedness (F4), and organizational maturity (F5) are particularly sensitive, requiring special attention to optimizing these factors.

[0139] Further analysis of the third and fourth level indicators, emergency site reliability (M3), park emergency management capabilities (M 14 ), equipment maintenance response speed (X1) and management personnel emergency drill frequency (X 25 ) are the most sensitive factors affecting target nodes. Small changes in these factors can significantly reduce campus resilience. Therefore, it is necessary to strengthen routine equipment maintenance, enhance the campus's collaborative emergency response capabilities, and optimize management processes to mitigate potential risks.

[0140] Perform sensitivity analysis on R1 node, such as Figure 8 The tornado diagram shown in Figure 1 shows the calculated results, which intuitively demonstrates the degree of influence of each factor on the target node R1 and its range of variation. The length of the bar in the figure represents the sensitivity of each factor to the target node, with longer bars indicating higher sensitivity. Figure 8 The right side of the middle bar indicates positive changes, and the left side indicates negative changes. The specific range is shown by the upper and lower limits in the figure.

[0141] T2 and T1 have a significant impact on node R1. T2's state change (State = YES) has the most significant impact on R1 when combined with F4 = State 0 and F5 = State 0. The target node's value range changes from [0.729038, 0.767431], with a span of 0.038393. This indicates that T2's state plays a key role in decision-making and is a key focus for risk control.

[0142] F4 and F5 are more sensitive to the target node, especially when F4 = State 0, the value of the target node changes the most, reflecting the importance of organizational resilience-related factors in the overall resilience assessment of the park. 10 and M11 The change significantly affected the sensitivity of F5, which further illustrates that strengthening organizational management and personnel quality is crucial to improving the overall performance of the target node R1.

[0143] An embodiment of the present invention further provides a computer terminal comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the processor can be used to execute any one of the methods described in the foregoing embodiments of the present invention, or to execute any one of the systems described in the foregoing embodiments of the present invention.

[0144] Optionally, the memory is used to store programs; the memory may include volatile memory (English: volatile memory), such as random-access memory (English: random-access memory, abbreviated: RAM), such as static random-access memory (English: static random-access memory, abbreviated: SRAM), double data rate synchronous dynamic random access memory (English: Double Data Rate Synchronous Dynamic Random Access Memory, abbreviated: DDR SDRAM), etc.; the memory may also include non-volatile memory (English: non-volatile memory), such as flash memory (English: flash memory). The memory is used to store computer programs (such as applications, functional modules, etc. that implement the above-mentioned methods), computer instructions, etc., and the above-mentioned computer programs, computer instructions, etc. can be partitioned and stored in one or more memories. In addition, the above-mentioned computer programs, computer instructions, data, etc. can be called by the processor.

[0145] The processor is configured to execute the computer program stored in the memory to implement the various steps of the method or various modules of the system involved in the above embodiments. For details, please refer to the relevant descriptions in the above method and system embodiments.

[0146] The processor and memory can be independent structures or integrated structures. When the processor and memory are independent structures, the memory and processor can be coupled via a bus.

[0147] An embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, it can be used to execute any method of the above embodiments of the present invention, or to run any system of the above embodiments of the present invention.

[0148] Computer-readable media include computer storage media and communication media, wherein communication media include any medium that facilitates the transfer of computer programs from one location to another. The storage medium can be any available medium that can be accessed by a general-purpose or special-purpose computer. An exemplary storage medium is coupled to a processor so that the processor can read information from the storage medium and write information to the storage medium. Of course, the storage medium can also be an integral part of the processor. The processor and storage medium can be located in an ASIC. Alternatively, the ASIC can be located in a user device. Of course, the processor and storage medium can also exist as discrete components in a communication device.

[0149] The urban park risk resilience assessment method and system based on probabilistic reasoning provided in the above-mentioned embodiments of the present invention quantify and evaluate the resilience performance of a region when facing various disturbance events by introducing Bayesian networks and probabilistic reasoning, provide decision makers with scientific resilience assessment data, optimize risk management and emergency response strategies, thereby improving the resilience and recovery capabilities of the region in the face of emergencies, and promoting urban sustainable development and disaster prevention and control.

[0150] Matters not mentioned in the above embodiments of the present invention are well known in the art.

[0151] The above describes the specific embodiments of the present invention. It should be understood that the present invention is not limited to the above specific embodiments, and those skilled in the art may make various variations or modifications within the scope of the claims, which do not affect the essence of the present invention.

Claims

1. A method for assessing urban park risk resilience based on probabilistic reasoning, characterized by: include: Identify risk factors in the park's target areas and establish a risk resilience assessment indicator system applicable to the park; Based on the risk resilience assessment indicator system and the dependency relationship of the influencing factors in the system, a Bayesian network model for park risk resilience assessment is constructed; A fuzzy Bayesian method is used to determine the prior information of the conditional probability of the root node in the Bayesian network, and the model parameters are initialized to obtain the prior probability of the root node of the Bayesian network model; Calculating the conditional probabilities of non-root nodes in the Bayesian network model and quantifying the risk transfer relationship between nodes; Based on the actual situation of the park, the target node and its related parent nodes are inferred and calculated to obtain the park risk resilience assessment results.

2. The urban park risk resilience assessment method based on probabilistic reasoning according to claim 1 is characterized in that: Risk factors in the target area of ​​the park are identified, including natural factors, engineering factors, social factors and institutional factors, and a risk resilience assessment index system with five dimensions including engineering resilience, organizational resilience, social resilience, economic resilience and institutional resilience is constructed.

3. The urban park risk resilience assessment method based on probabilistic reasoning according to claim 1 is characterized in that: Based on the risk resilience assessment indicator system and the dependency relationships among the influencing factors in the system, a Bayesian network model of park resilience is constructed, including: Each node in the Bayesian network model represents a variable, the edge represents the causal relationship between variables, and the direction of the edge represents the dependency order of conditional probability; each node is associated with a conditional probability table, which is used to describe the probability distribution of the node under the condition that its parent node is known; by decomposing the joint probability distribution, the probability relationship between multiple random variables is structured as follows: Where P(R) represents the joint probability; n represents the number of nodes in the Bayesian network, x1, x2, ..., x n Represents a node in the Bayesian network, is a random variable, x i represents the i-th variable, R(x i ) represents node x i The parent node set, P(x i ∣∣R(x i )) represents a given parent node R(x i ) condition, node x i The conditional probability distribution of .

4. The urban park risk resilience assessment method based on probabilistic reasoning according to claim 1 is characterized in that: The fuzzy Bayesian method is used to determine the prior information of the conditional probability of the nodes in the Bayesian network and initialize the model parameters, including: The triangular membership function is used to describe the fuzzy probability distribution of the node; wherein the triangular membership function μ(x) is defined as follows: Where x is the evaluation value of the node in a specific situation; l, m, n represent the lower bound, peak value and upper bound of the triangular fuzzy number respectively, and l <m<n; The expert survey results are obtained based on the expert system and divided into different levels of fuzzy language levels, where each fuzzy language level corresponds to a triangular fuzzy number. The specific fuzzy value of each expert is: Where k is the number of experts, e j For each expert E j The weight, p j is the fuzzy evaluation value, p′ is the weighted average fuzzy value; The weighted average fuzzy value is defuzzified and converted into an accurate numerical value as the prior information of the conditional probability of the root node in the Bayesian network model, and the prior probability of each root node of the Bayesian network model for park risk resilience assessment is obtained to complete the model parameter initialization.

5. The urban park risk resilience assessment method based on probabilistic reasoning according to claim 1 is characterized in that: Calculating the conditional probability of non-root nodes of the Bayesian network model includes: Based on the Noisy-OR method, let the parent node probability of each parent node variable event be P i , the state of the parent node is X i , then the Noisy-OR model of the probability of the event occurring P(event occurrence) is: Where a j is the score given by the jth expert on the influence of a parent node’s child nodes, and m is the number of parent nodes of the child nodes; The obtained P i Substitute the Noisy-OR model to generate a non-root node conditional probability table, providing a mathematical basis for the inference analysis of the Bayesian network model for park risk resilience assessment.

6. The urban park risk resilience assessment method based on probabilistic reasoning according to claim 1 is characterized in that: Based on the actual situation of the park, the target node and its related parent nodes are inferred and calculated to obtain the park risk resilience assessment results, including: Based on the non-root node conditional probability table, the park resilience of the target node is analyzed through inference calculation, the impact of each of the target node's related parent nodes on the overall resilience level is identified, and the contribution of the target node to the comprehensive resilience of the park is quantified, thereby obtaining the park risk resilience assessment result; The inference calculation includes: Single-node impact analysis: Assume that any parent node in the model fails completely and observe the impact of the parent node on the resilience level of the parent node and the target node. This is used to evaluate the key bottleneck elements and vulnerable nodes in the Bayesian network model for campus risk resilience assessment. Backward reasoning analysis: Assuming the target node's resilience is in a failed state, Bayesian network reverse reasoning technology is used to back-infer the state changes and failure probabilities of each upstream node to assess potential high-risk impact paths and sensitive nodes. Sensitivity analysis: Quantify the sensitivity of each child node to the target node, which is used to evaluate the impact magnitude and direction of key nodes.

7. An urban park risk resilience assessment system based on probabilistic reasoning, characterized by: include: A resilience indicator system construction module, which is used to identify risk factors in the park's target areas and establish a risk resilience assessment indicator system applicable to the park; A Bayesian network construction module, which is used to construct a Bayesian network model for park risk resilience assessment based on the risk resilience assessment indicator system and the dependency relationship of influencing factors in the system; A node conditional probability calculation module is used to determine the prior information of the conditional probability of the root node in the Bayesian network using a fuzzy Bayesian method, initialize the model parameters, and obtain the prior probability of the root node of the Bayesian network model; Calculating the conditional probabilities of non-root nodes in the Bayesian network model and quantifying the risk transfer relationship between nodes; The risk change assessment module performs inference calculations on the target node and its related parent nodes based on the actual situation of the park to obtain the park risk resilience assessment results.

8. A computer terminal comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When executing the computer program, the processor can be used to perform the method according to any one of claims 1 to 6, or run the system according to claim 7.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, it can be used to perform the method according to any one of claims 1 to 6, or to run the system according to claim 7.

Images