An artificial intelligence-based digital emergency auxiliary decision-making system

Abstract

The application relates to the technical field of decision support, in particular to a digital emergency decision support system based on artificial intelligence. The system comprises: an evaluation index calculation module for obtaining decision alternatives and real-time data; a scenario weight calculation module for calculating the scenario weight of each evaluation index; an inflow and outflow calculation module for converting two emergency plans into a priority probability distribution and calculating the priority inflow and outflow of each emergency plan; and a decision module for identifying and outputting the key evaluation index that contributes most to the net flow ranking value of the plan as decision information representing the core advantages or disadvantages of the plan. The scheme can improve the robustness and reliability of the ranking result, clearly reveal the core advantages and disadvantages of each alternative, and enhance the transparency and explainability of the decision process.

Description

Technical Field

[0001] This invention relates to the field of decision support technology, and in particular to a digital emergency decision support system based on artificial intelligence. Background Technology

[0002] In the field of emergency management, digital decision support systems play a crucial role in quickly and scientifically selecting the optimal solution from multiple pre-set emergency plans. These systems typically construct a multi-dimensional evaluation index system covering all aspects of emergency response, such as casualties, economic losses, response time, and resource consumption. By acquiring or evaluating the performance of each alternative emergency plan under these indicators, multi-attribute decision-making methods are used to rank and select the best option. Commonly used decision models include the Analytic Hierarchy Process (AHP), fuzzy comprehensive evaluation, and weighted summation models. These methods provide quantitative decision-making basis for emergency response, improving the efficiency and objectivity of decision-making to a certain extent.

[0003] However, most existing methods rely on pre-defined static indicator weights, which fail to reflect the dynamic evolution of emergency scenarios. In real emergencies, the relative importance of different evaluation indicators changes in real time as the situation develops, and fixed weights cannot accurately capture this dynamic prioritization. Moreover, the data used for emergency decision-making, whether from sensor monitoring or model simulations, is generally uncertain. Existing methods often simplify this uncertain data into definite single-point values ​​for calculation, ignoring the potential disturbances that uncertainty may cause to the ranking of contingency plans, thus reducing the robustness of the decision. Furthermore, many systems provide a "black box" ranking list, lacking in-depth explanation of the decision-making process and results. They cannot clearly tell decision-makers why a particular contingency plan ranks highly or low, or which specific indicators represent its key advantages or disadvantages, hindering decision-makers from building trust in the system's results and making informed decisions. Summary of the Invention

[0004] The purpose of this invention is to propose a digital emergency auxiliary decision-making system based on artificial intelligence to solve the problems of poor decision robustness, lack of transparency and poor interpretability in the decision-making process in the prior art; to this end, the present invention provides a solution in one aspect.

[0005] This invention provides a digital emergency auxiliary decision-making system based on artificial intelligence, comprising the following modules:

[0006] The evaluation index calculation module is used to acquire multiple emergency plans as decision-making alternatives, as well as real-time data of multi-dimensional evaluation indicators related to these emergency plans. Evaluation indicators include indicator values ​​and indicator uncertainties. The scenario weight calculation module is used to calculate the scenario weight of each evaluation indicator based on a pre-defined directed acyclic graph (DAG) of the relationships between the evaluation indicators. The scenario weight of any evaluation indicator is determined by its own baseline weight and the indicator value of its direct parent node indicator in the DAG. The inflow / outflow calculation module uses a non-linear preference function coupled with the global emergency situation level parameter to convert the numerical differences between the two emergency plans under various evaluation indicators, combined with indicator uncertainties, into a priority probability distribution for any two emergency plans. Based on the priority probability distribution and scenario weights... The system calculates the priority inflow and priority outflow for each emergency plan, where the priority inflow and priority outflow are random variables in the form of probability distributions. A decision module is used to calculate the net flow ranking value of each emergency plan based on statistical moment analysis of the probability distributions of the priority inflow and priority outflow, and to generate a ranking based on the net flow ranking value. For any emergency plan, the system identifies and outputs the key evaluation indicator that contributes the most to the net flow ranking value of the plan. For any emergency plan and evaluation indicator, the system calculates the contribution of the evaluation indicator to the net flow ranking value of the emergency plan, identifying the evaluation indicator with the largest positive contribution as the core advantage indicator of the plan, and the evaluation indicator with the smallest negative contribution as the core disadvantage indicator of the plan, serving as decision information characterizing the core advantages or disadvantages of the plan.

[0007] Preferably, the emergency plan is retrieved from the plan database, and the real-time data is obtained through IoT sensors, on-site information reporting systems, or simulation models.

[0008] Preferably, the index value is the expected value, and the index uncertainty is the standard deviation.

[0009] Preferably, the directed acyclic graph is pre-constructed by experts to characterize the causal or influence relationships between evaluation indicators.

[0010] Preferably, the calculation of the scenario weights for each evaluation indicator includes: for each evaluation indicator... Context weights The calculation formula is: ;in, Evaluation indicators Context weights Evaluation indicators The baseline weights, Parent node indicator Real-time indicator values, Parent node indicator The preset maximum value, Parent node indicator Evaluation indicators The preset influence coefficient.

[0011] Preferably, the nonlinear preference function is an sigmoid function in the form of a Gaussian cumulative distribution function.

[0012] Preferably, the step of converting the numerical differences between the two emergency plans under each evaluation indicator, combined with the uncertainty of the indicator, into a priority probability distribution includes: for any evaluation indicator Emergency Plan Compared to emergency plans priority Modeled as a Gaussian distributed random variable, the mean of the Gaussian distribution is the two emergency plans in the evaluation indicators. Difference in index values Standard deviation Through formula Calculate, where, Let be the standard deviation of the Gaussian distribution. Emergency response plan In evaluation indicators The following indicator values, Emergency response plan In evaluation indicators The following indicator values, Emergency response plan In evaluation indicators The uncertainty of the index, Emergency response plan In evaluation indicators The uncertainty of the index, This refers to the parameters for the overall emergency situation level.

[0013] Preferably, the calculation of the priority inflow and priority outflow for each emergency plan includes: for each emergency plan The priority inflow of the plan and priority outflow Let these be random variables calculated using the following formulas respectively: ; ;in, To remove Any emergency plan other than Priority inflow volume as per the contingency plan The priority outflow volume in the contingency plan, Iterate through all evaluation metrics, Evaluation indicators Context weights Evaluation indicators Emergency Response Plan Compared to emergency plans The priority of as a random variable Evaluation indicators Emergency Response Plan Compared to emergency plans The priority of as a random variable.

[0014] Preferably, the step of calculating the net flow ranking value of each emergency plan based on the statistical moment analysis of the probability distributions of the priority inflow and priority outflow includes: calculating the net flow ranking value of each emergency plan. Net flow random variable Extract the net flow random variable. Mathematical expectation and standard deviation ; through formula Calculate the net flow ranking value ;in, Sort by net flow rate. Let net flow be a random variable. Let the expected value of the net flow random variable be . Let the standard deviation be the net flow rate random variable. This is the preset risk aversion coefficient.

[0015] Preferably, the formula for calculating the contribution of the evaluation index to the net flow ranking value of the emergency plan is as follows: ;in Evaluation indicators Context weights Contingency Plan In evaluation indicators The following indicator values, Contingency Plan In evaluation indicators The following indicator values, Emergency response plan In evaluation indicators The uncertainty of the index, Emergency response plan In evaluation indicators The uncertainty of the index, This is a parameter for the overall emergency situation level. To remove Any emergency plan other than To be related to the risk aversion coefficient Related preset contribution penalty coefficients.

[0016] The beneficial effects of this invention are as follows: This invention can systematically handle the uncertainty prevalent in emergency decision-making information. By quantifying the superiority-inferiority relationships between plans into probability distributions and performing statistical moment analysis on the ranking indicators, the robustness and reliability of the ranking results are improved. Furthermore, by constructing correlations between evaluation indicators, the weights of the indicators can change according to the actual values ​​of the relevant indicators, overcoming the limitations of fixed weights and making the evaluation process more aligned with the inherent requirements of specific emergency scenarios. In addition to providing ranking results, this invention can also trace back and identify the key evaluation indicators that determine the ranking of each plan, clearly revealing the core advantages and disadvantages of each alternative, enhancing the transparency and interpretability of the decision-making process, thereby providing decision-makers with in-depth insights and assisting them in making more scientific and prudent decisions. Attached Figure Description

[0017] Figure 1 The schematic diagram illustrates the module flowchart of the AI-based digital emergency auxiliary decision-making system in this embodiment. Detailed Implementation

[0018] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

[0019] like Figure 1 As shown in this embodiment, a digital emergency auxiliary decision-making system based on artificial intelligence includes the following modules:

[0020] The evaluation index calculation module is used to acquire multiple emergency plans as decision-making alternatives, as well as real-time data of multi-dimensional evaluation indicators related to the multiple emergency plans. The evaluation indicators include index values ​​and index uncertainties.

[0021] Emergency plans are retrieved from a plan database, and real-time data is acquired through IoT sensors, on-site information reporting systems, or simulation models. In the evaluation indicators, the numerical values ​​are expected values, and the uncertainty is typically expressed as standard deviation or variance, together forming a Gaussian distribution or a similar probability distribution to describe the actual state of the indicator.

[0022] The scenario weight calculation module is used to calculate the scenario weight of each evaluation indicator based on the pre-set directed acyclic graph of the relationship between evaluation indicators. The scenario weight of any evaluation indicator is determined by its own baseline weight and the indicator value of the direct parent node indicator in the directed acyclic graph.

[0023] Directed acyclic graphs are pre-constructed by experts to characterize causal or influence relationships between indicators.

[0024] Calculate the scenario weights for each evaluation indicator, including: for each evaluation indicator Context weights The calculation formula is: ;

[0025] in, Evaluation indicators Context weights Evaluation indicators The baseline weights, Parent node indicator Real-time indicator values, Parent node indicator The preset maximum value, Parent node indicator Evaluation indicators The preset influence coefficient.

[0026] The importance of evaluation indicators is dynamically adjusted based on the real-time situation. The weight of each indicator in an emergency response is not static but changes with relevant environmental factors. For example, in responding to urban flooding disasters, the baseline weight of the evaluation indicator, personnel evacuation efficiency, changes accordingly. It can be set to 0.3. The importance of the indicator is directly affected by the current rainfall and upstream water level of the parent node indicator. When the values ​​of these two parent node indicators increase, it means that the risk is aggravated, and the evacuation of people becomes more urgent and important.

[0027] Assuming the current real-time rainfall is 50 mm per hour, with a preset maximum of 100 mm; and the upstream water level is 8 meters, with a preset maximum of 10 meters. The parent node indicator, current rainfall, has an impact coefficient of 0.1 on evacuation efficiency, while the upstream water level, due to its more direct threat, has an impact coefficient of 0.15. Based on the formula, the weight increment from current rainfall is calculated to be 0.05, and the weight increment from upstream water level is 0.12. The scenario weights for evacuation efficiency are... It is 0.47.

[0028] This result indicates that as the real-time risks of rainfall and water levels increase, the importance of the evacuation efficiency indicator has significantly increased from a baseline of 0.3 to 0.47 in the current scenario. This dynamic adjustment makes contingency plan assessments more closely reflect reality, ensuring that decisions prioritize the most critical response aspects.

[0029] The inflow and outflow calculation module is used to, for any two emergency plans, employ a nonlinear preference function coupled with the global emergency situation level parameter to convert the numerical differences between the two emergency plans under various evaluation indicators, combined with the uncertainty of the indicators, into a priority probability distribution; based on the priority probability distribution and the scenario weights, calculate the priority inflow and priority outflow of each emergency plan, wherein the priority inflow and priority outflow are random variables in the form of probability distributions.

[0030] In one embodiment, the nonlinear preference function is preferably a Gaussian preference function, which is essentially a cumulative distribution function. The global emergency situation level parameter is coupled to this function to adjust the sensitivity or severity of the decision.

[0031] The numerical differences between the two emergency response plans under various evaluation indicators, combined with the uncertainty of the indicators, are converted into a priority probability distribution, including: for any evaluation indicator Emergency Plan Compared to emergency plans priority Modeled as a Gaussian distributed random variable, the mean of the Gaussian distribution is the two emergency plans in the evaluation indicators. Difference in index values Standard deviation Through formula Calculate, where, Let be the standard deviation of the Gaussian distribution. Emergency response plan In evaluation indicators The following indicator values, Emergency response plan In evaluation indicators The following indicator values, Emergency response plan In evaluation indicators The uncertainty of the index, Emergency response plan In evaluation indicators The uncertainty of the index, This refers to the parameters for the overall emergency situation level.

[0032] The values ​​assessed in contingency plans are often predicted rather than exact, and therefore subject to uncertainty. For example, there might be two contingency plans for a chemical plant leak. and One of the evaluation indicators is the expected economic loss. (Contingency Plan) The estimated loss is 1 million yuan, but due to the use of new technology, the uncertainty is higher, resulting in a loss of 200,000 yuan. (Contingency plan) The estimated loss is 1.1 million yuan. The mature solution is used, and the uncertainty is low, amounting to 100,000 yuan.

[0033] Calculate the mean of the priority distribution, i.e., the contingency plan. The loss value minus the contingency plan The average loss value is -20, which means that without considering uncertainties, the contingency plan... Better than the contingency plan The reliability of this comparison is quantified by calculating the standard deviation. This assumes a global emergency situation level. The standard deviation is calculated as a contingency plan, and the level is 3 (medium). Uncertainty 20 squared plus contingency plan The uncertainty is 10 squared, and the standard deviation is approximately 7.5.

[0034] Therefore, contingency plan Compared to the contingency plan The priority of economic loss indicators is described as a Gaussian distribution with a mean of -20 and a standard deviation of 7.5. This probability distribution is more informative than a simple difference, indicating that although the contingency plan... The expected loss is lower, but the contingency plan The uncertainty of the outcome is also greater. Global emergency situation level. In the formula, using it as the denominator signifies a more severe situation. The larger the value, the lower the tolerance for uncertainty, the smaller the standard deviation, making decisions more dependent on the difference in expected values.

[0035] Calculate the priority inflow and priority outflow for each emergency plan, including: for each emergency plan The priority inflow of the plan and priority outflow Let these be random variables calculated using the following formulas respectively:

[0036] ;

[0037] ;

[0038] in, To remove Any emergency plan other than Priority inflow volume as per the contingency plan The priority outflow volume in the contingency plan, Iterate through all evaluation metrics, Evaluation indicators Context weights Evaluation indicators Emergency Response Plan Compared to emergency plans The priority of as a random variable Evaluation indicators Emergency Response Plan Compared to emergency plans The priority of as a random variable.

[0039] A comprehensive pairwise comparison of all contingency plans is conducted using a simulated round-robin format to summarize the overall strengths and weaknesses of each plan. Assume there are three contingency plans. , and And two evaluation indicators, namely, response time with a weight of 0.6 and resource consumption with a weight of 0.4. Calculation plan. Compared to the contingency plan and contingency plans The overall advantage is the priority inflow.

[0040] In order to calculate the contingency plan Priority inflows, assessment right The advantages. This includes... In terms of response time The priority is multiplied by the indicator's weight of 0.6, plus... In terms of resource consumption The priority is multiplied by a weight of 0.4. Evaluate in the same way. right The advantages, soon On two indicators The priorities are multiplied by their corresponding weights and then summed. right and right The weighted summation of advantages yields the contingency plan. The total priority inflow. Inflow is a random variable that integrates all comparison items and all evaluation indicators, representing the plan's... The overall degree to which it is preferred by all other contingency plans.

[0041] Correspondingly, contingency plans will also be calculated. The priority outflow is calculated. The process is similar to calculating the inflow, but it calculates the planned outflow. and contingency plans Compared to the contingency plan The sum of advantages. For example, it will summarize... right Weighted advantage and right The weighted advantage. This sum represents the contingency plan. It is less comprehensive than the overall scope of all other contingency plans. In this way, each contingency plan is assigned a priority inflow and a priority outflow, both of which comprehensively reflect its relative position within the entire set of contingency plans.

[0042] The decision-making module is used to calculate the net flow ranking value of each emergency plan based on the statistical moment analysis of the probability distribution of the priority inflow and priority outflow, and generate a ranking based on the net flow ranking value; for any emergency plan, it identifies and outputs the key evaluation index that contributes the most to the net flow ranking value of the plan, as decision information characterizing the core advantages or disadvantages of the plan.

[0043] Based on statistical moment analysis of the probability distributions of the priority inflow and priority outflow, the net flow ranking values ​​for each emergency response plan are calculated, including:

[0044] Calculate each emergency plan Net flow random variable ;

[0045] Extract the net flow random variable Mathematical expectation and standard deviation ;

[0046] Through formula Calculate the net flow ranking value ;

[0047] in, Sort by net flow rate. Let net flow be a random variable. Let the expected value of the net flow random variable be . Let the standard deviation be the net flow rate random variable. This is the preset risk aversion coefficient.

[0048] The final ranking is determined by comprehensively considering the expected performance and stability of the proposed plans. After obtaining the priority inflow and outflow amounts for each plan, the net flow of the plan is calculated, which is the inflow minus the outflow. This represents the overall relative advantage of a plan. Net flow is still a random variable with uncertainty. We further extract the expected value of the random variable, i.e., the average advantage value, and the standard deviation of the random variable, i.e., the volatility or risk of the advantage.

[0049] For example, after calculation, the contingency plan The expected net flow is 15, with a standard deviation of 6, indicating that the plan... The average performance was good, but the results were highly volatile, posing a certain risk. The contingency plan... The expected net flow is 12, with a standard deviation of only 2, indicating that the plan... Although the average performance was slightly inferior However, the results are very stable and reliable. The choice of which contingency plan to use depends on the decision-maker's risk preference, which is reflected in the risk aversion coefficient.

[0050] When the decision-making environment allows for a certain level of risk, a lower risk aversion coefficient can be set. For example, 0.5. At this time, the contingency plan... The sorting value is 12; contingency plan The sorting value is 11. In this case, the recommended plan is... However, in matters of great importance, decision-makers tend to be highly risk-averse; in such cases, a higher risk threshold can be set. Value, such as 2. Contingency Plan The sorting value becomes 3; contingency plan The sorting value becomes 8. At this time, due to the contingency plan... High risk and contingency plan Instead of imposing penalties, a more robust contingency plan B is recommended.

[0051] Identify and output the key evaluation indicators that contribute most to the net flow ranking value of the contingency plan, including: for any emergency plan and any evaluation index Calculate evaluation indicators for emergency response plans Contribution of net flow ranking value The calculation formula is:

[0052] ;

[0053] in Evaluation indicators Context weights Contingency Plan In evaluation indicators The following indicator values, Contingency Plan In evaluation indicators The following indicator values, Emergency response plan In evaluation indicators The uncertainty of the index, Emergency response plan In evaluation indicators The uncertainty of the index, This is a parameter for the overall emergency situation level. To remove Any emergency plan other than To be related to the risk aversion coefficient The relevant preset contribution penalty coefficient;

[0054] Will have the largest positive contribution The evaluation indicators were identified as the core advantages of the plan, and those with the smallest negative contribution values ​​were selected. The evaluation indicators were identified as the core weaknesses of the proposed plan.

[0055] After calculating the ranking value of each contingency plan, the ranking value is broken down to analyze the specific contribution of each evaluation indicator. This contribution calculation takes into account the weight of the evaluation indicators, the average performance difference of the contingency plan compared with all other contingency plans on the evaluation indicators, and the uncertainty of this performance difference.

[0056] A forest fire emergency plan For example, it is related to the contingency plan. and Joint assessment. Calculate the impact of each evaluation indicator on the contingency plan one by one. The contribution of ranking values. For example, for fire extinguishing efficiency indicators, the plan... The expected efficiency is much higher than and Furthermore, the certainty of the assessment value is very high. Calculations show that the evaluation indicators are effective for the contingency plan. The contribution is a large positive value, such as 25. As for the required logistical support indicators, the contingency plan... The requirements are extremely stringent and the uncertainty is very high, leading to the contingency plan being... The contribution is a large negative value, such as -20.

[0057] After calculating the contribution of all indicators, they were ranked. Firefighting efficiency, with a contribution of 25 (the largest positive value), was identified as a priority indicator. The core advantage indicators. This clearly tells decision-makers which contingency plan to choose. The biggest reason is the contingency plan. Excellent firefighting efficiency. Meanwhile, the required logistical support, with a contribution level of -20 being the minimum negative value, is identified as part of the contingency plan. This is a core weakness indicator. This serves as a warning to decision-makers that if they want to implement contingency plans... We must focus on resolving contingency plans. Huge logistical problems.

[0058] As can be seen from the above description of the embodiments, those skilled in the art can clearly understand that the present invention can be implemented by means of software plus necessary general-purpose hardware platforms. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in various embodiments or some parts of the embodiments of the present invention.

[0059] In the description of this specification, "multiple" means at least two, such as two, three or more, etc., unless otherwise expressly and specifically defined.

[0060] While various embodiments of the invention have been shown and described in this specification, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Many modifications, alterations, and alternatives will occur to those skilled in the art without departing from the spirit and essence of the invention.

Claims

1. An artificial intelligence-based digital emergency assistant decision-making system, characterized in that, The method comprises the following modules: An evaluation index calculation module is configured to obtain a plurality of emergency plans as decision alternatives and real-time data of multi-dimensional evaluation indexes related to the plurality of emergency plans, the evaluation indexes including index values and index uncertainties; The scenario weight calculation module is configured to calculate scenario weights of the evaluation indexes based on a preset correlation directed acyclic graph among the evaluation indexes, wherein the scenario weight of any evaluation index is determined by a reference weight of the evaluation index itself and an index value of a direct parent node index in the directed acyclic graph; and for the evaluation indexes , a calculation formula of the scenario weight is as follows: ； Wherein, is the evaluation index scenario weight, is the evaluation index baseline weight, is the real-time index value of the parent node index , is the preset maximum value of the parent node index , is the preset influence coefficient of the parent node index on the evaluation index . An inflow and outflow calculation module is configured to, for any two emergency plans, use a nonlinear preference function coupled with a global emergency situation level parameter to convert a difference between the two emergency plans in values under each evaluation index into a priority probability distribution in combination with the index uncertainties; and calculate a priority inflow and a priority outflow of each emergency plan based on the priority probability distribution and a scenario weight, wherein the priority inflow and the priority outflow are random variables in the form of probability distribution; A decision module is configured to, based on statistical moment analysis on the probability distribution of the priority inflow and the priority outflow, calculate a net flow ranking value of each emergency plan and generate a ranking according to the net flow ranking value; for any emergency plan, identify and output a key evaluation index that contributes most to the net flow ranking value of the plan; for any emergency plan and evaluation index, calculate a contribution degree of the evaluation index to the net flow ranking value of the emergency plan, identify an evaluation index with the largest positive contribution degree as a core advantage index of the plan, and identify an evaluation index with the smallest negative contribution degree as a core disadvantage index of the plan as decision information representing a core advantage or disadvantage of the plan. The method for calculating the net flow ranking value of each emergency plan comprises: calculating a net flow random variable for each contingency ;​ Extracted net flow random variable mathematical expectation and standard deviation ; The net flow ranking value is calculated by the formula ; is a preset risk aversion coefficient;​ When the decision environment allows for some risk, set a lower risk aversion coefficient In a high-stakes decision, the decision maker is highly risk-averse, in which case set a higher risk aversion coefficient .

2. The artificial intelligence-based digitalized emergency assistant decision-making system according to claim 1, characterized in that, The emergency plans are retrieved from a plan library, and the real-time data are obtained through Internet of Things sensors, a field information reporting system or a simulation model.

3. The artificial intelligence-based digitalized emergency assistant decision-making system according to claim 1, characterized in that, The index values are expected values, and the index uncertainties are standard deviations.

4. The artificial intelligence-based digitalized emergency assistant decision-making system according to claim 1, characterized in that, The directed acyclic graph is pre-constructed by experts and is used to represent causal or influence relationships between the evaluation indexes.

5. The artificial intelligence-based digitalized emergency assistant decision-making system according to claim 1, wherein, The nonlinear preference function is an S-shaped function in the form of a Gaussian cumulative distribution function.

6. The artificial intelligence-based digitalized emergency assistant decision-making system according to claim 1, wherein, The method for converting the difference between the two emergency plans in values under each evaluation index into a priority probability distribution in combination with the index uncertainties comprises: For any evaluation index Emergency Plan Compared to emergency plans priority Modeled as a Gaussian distributed random variable, the mean of the Gaussian distribution is the two emergency plans in the evaluation indicators. Difference in index values Standard deviation Through formula Calculate, where, Let be the standard deviation of the Gaussian distribution. Emergency response plan In evaluation indicators The following indicator values, Emergency response plan In evaluation indicators The following indicator values, Emergency response plan In evaluation indicators The uncertainty of the index, Emergency response plan In evaluation indicators The uncertainty of the index, This refers to the parameters for the overall emergency situation level.

7. The artificial intelligence-based digital emergency assistant decision-making system according to claim 1 or 6, characterized in that, The method for calculating the priority inflow and the priority outflow of each emergency plan comprises: For emergency plans , the priority inflow and outflow of the plan and are random variables calculated by the following formulas, respectively: ； ； wherein, is any emergency plan other than is a priority inflow amount for the plan, is a priority outflow amount for the plan, is a priority inflow amount for the plan, traverses all evaluation indices, is a scenario weight for the evaluation index is a scenario weight for the evaluation index is a scenario weight for the evaluation index is a priority degree for the emergency plan is a priority degree for the emergency plan is a priority degree for the emergency plan is a priority degree for the emergency plan is a priority degree for the emergency plan is a priority degree for the emergency plan is a priority degree for the emergency plan 8. The artificial intelligence-based digitalized emergency assistant decision-making system according to claim 1, wherein, The calculation formula of the contribution degree of the evaluation index to the net flow ranking value of the emergency plan is: ； wherein, is a scenario weight of the evaluation index , is a preplan , is an index value of the evaluation index under the preplan , is an index value of the evaluation index under the preplan , is an index uncertainty of the evaluation index under the preplan , is an index uncertainty of the evaluation index under the preplan , is a global emergency situation grade parameter, is any emergency preplan except , is a preset contribution penalty coefficient related to the risk aversion coefficient

Images