Emergency material list generation method based on disaster evolution and historical case matching

Abstract

The application discloses an emergency material list generation method based on disaster evolution and historical case matching, and belongs to the technical field of emergency management decision-making; the method comprises the following steps: acquiring meteorological monitoring and forecasting data of a target disaster according to a current disaster category, and constructing a meteorological index set X; constructing a data set according to historical disaster cases, training a BP neural network model, predicting a current disaster grade according to the set X and the trained BP neural network model; searching for a most similar case matched according to the current disaster type and the disaster grade, and acquiring an emergency material demand list S; constructing a time sequence index set Y according to environmental information of the current disaster; constructing a time sequence sample data set D2 by using the historical disaster cases, and training an LSTM model; predicting an emergency material list required by the current disaster by using the set Y and the trained LSTM model; and dynamically adjusting the list S according to the list to obtain a final emergency material list.

Description

Technical Field

[0001] This invention belongs to the field of emergency management decision-making technology, specifically involving a method for generating emergency supplies lists based on disaster evolution and historical case matching. Background Technology

[0002] In recent years, extreme weather events have become more frequent, significantly increasing the demands for accuracy and timeliness in emergency response and disaster relief. Traditional emergency relief material allocation relies on manual experience, lacking scientific rigor and timeliness, and is prone to material redundancy or shortages. Meteorological disaster data and historical disaster cases are not fully integrated and utilized, making it difficult to dynamically adapt to the needs of different disaster scenarios. The lack of automated tools leads to low emergency response efficiency, especially in complex disaster environments where it is difficult to quickly generate accurate material lists. Therefore, researching dynamic material demand forecasting and list generation technologies based on intelligent matching of real-time disaster evolution and historical cases has become a key direction urgently needed to improve the intelligence level and response efficiency of emergency management.

[0003] Currently, some scholars have proposed methods for predicting and generating dynamic material demand lists, among which the technical solutions that are closer to this invention include: Patent application number 202310123648.7 discloses a method and apparatus for generating an emergency support plan. The method determines a support list for a public emergency based on basic information about the public emergency and generates an emergency support plan for the public emergency. However, this method requires manual input of the type and level of the event and cannot automatically predict based on meteorological disaster data. The literature (Cheng Mei. Research on earthquake emergency material demand prediction. Institute of Disaster Prevention Science and Technology, 2025.) proposes to predict the disaster-affected population by integrating particle swarm optimization algorithm and support vector machine, and to construct an earthquake emergency material demand prediction model by combining safety stock theory. This method only generates material demand based on the disaster-affected population, without fully combining disaster environmental information and historical case data, and fails to utilize the similarity patterns of historical cases to optimize real-time decision-making. The literature (Han Mengyao. Research on Emergency Material Allocation Based on Demand-Driven in the Context of Floods. Shanxi University of Finance and Economics, 2024.) constructs a dynamic demand prediction model for the affected population based on gray metabolism-Markov chain, considering the dynamic evolution of the disaster situation. Based on the prediction of the affected population, it also constructs a dynamic demand prediction model for emergency materials by combining the safety stock theory. However, it fails to fully integrate environmental information and historical case data for material demand prediction and only focuses on material prediction for floods. The literature (Wang Haixiao, Jiang Hua. Research on earthquake emergency material demand prediction based on BP neural network. Logistics Technology, 2025, 8:56-60.) predicts the affected population by constructing a prediction model based on BP neural network, and further estimates the emergency material demand by combining the relationship between the number of affected people and the material demand. This method cannot automatically predict the disaster level, requires human input, and can only predict the material demand for earthquake disaster types.

[0004] In summary, the current dynamic material demand forecasting methods have the following shortcomings: (1) They only target the forecasting of emergency material demand for a single type of disaster; (2) They require human assessment of the disaster level and cannot be automatically predicted; (3) They do not fully integrate meteorological environmental information and historical case data to optimize real-time decision-making by utilizing historical case patterns. Summary of the Invention

[0005] To address the shortcomings of existing technologies, the present invention aims to provide a method for generating emergency supplies lists based on disaster evolution and historical case matching, thereby solving the problems in existing technologies.

[0006] The objective of this invention can be achieved through the following technical solutions: The method for generating an emergency supplies list based on disaster evolution and historical case matching includes the following steps: S1, acquire meteorological monitoring and forecast data of the target disaster for the current disaster category, and construct a set of meteorological indicators X including the quantitative values ​​of meteorological elements; S2, construct disaster level prediction sample dataset D1 based on historical disaster cases. S3, using the disaster level prediction sample dataset D1 to train a BP neural network model; S4: Input the quantified values ​​of meteorological elements in the meteorological index set X into the trained BP neural network model to predict the current disaster level L; S5, based on the current disaster type and disaster level L, search the historical disaster case database for the most similar case and obtain the actual emergency material demand list S for the similar case; S6, construct a time series index set Y of key influencing factors of emergency material demand based on the current environmental information of the disaster, including the quantitative values ​​of key influencing factors; S7, using historical disaster cases to construct an emergency supplies list to generate a time-series sample dataset D2; S8, use the time series sample dataset D2 to train the LSTM model; S9 inputs the quantified values ​​of key influencing factors in the time-series index set Y into the trained LSTM model to predict the emergency supplies list needed for the current disaster. ; S10, based on the predicted emergency supplies list The actual emergency supplies demand list S of similar cases is dynamically adjusted to obtain the final emergency supplies list.

[0007] Furthermore, the disaster level prediction sample dataset D 1={( X j , L j ) | j =1,2,…, J},in: J Indicates the number of historical disasters. j Indicates the index number of historical disasters. X j ={( x i , v ji ) | i =1,2,…, I} indicates the first j A set of meteorological indicators corresponding to historical disasters L j Indicates the first j The disaster level value corresponding to each historical disaster i Indicates the meteorological element index number. I It represents the total quantity of meteorological elements.

[0008] Furthermore, when training the BP neural network model, a set of meteorological indicators is used. X j Quantitative values ​​of all meteorological elements v ji For input.

[0009] Furthermore, a list of actual emergency supplies needed for similar cases. S ,in: A The types and quantities of supplies. m a For the first a Name of the material n a For the first a The required quantity of these supplies; The steps for searching the historical disaster case database to find the most similar case include: S51, Select cases of the same type as the current disaster from historical disaster cases; S52, For each selected historical disaster case, calculate the disaster level similarity between it and the current disaster case. : in, L max This represents the highest disaster level value that has occurred across all historical cases. L min The minimum disaster level value that has appeared in all historical cases. For the first j The disaster level value corresponding to each historical disaster The current disaster level is predicted by a BP neural network model based on the meteorological index set X. S53, select the historical disaster case with the highest disaster level similarity value as the most similar case.

[0010] Furthermore, the time-series sample dataset D 2={( Y j , S j ) | j =1,2,…, J},in: J Indicates the number of historical disasters. j Indicates the index number of disasters over the years. Y j ={( y k , ) | k =1,2,…, K} indicates the first j A historical disaster T A set of time-series indicators of key influencing factors at consecutive time points. K This indicates the total number of key factors influencing the demand for emergency supplies. k This indicates the index number of key factors influencing emergency supplies demand. y k Indicates the first k Identifiers of key influencing factors Indicates the first j The first historical disaster k Key influencing factors T A set of time series indicators at consecutive moments; S j Indicates the first j A list of actual emergency supplies corresponding to each historical disaster.

[0011] Furthermore, when training the LSTM model, the quantified numerical time series matrix of key influencing factors is used. For input.

[0012] Furthermore, the final emergency supplies list , For the first a The final quantity of the required materials; Calculate the first a The formula for the final quantity of the required materials is: in, and To integrate weights, satisfy , For the first a The minimum quantity of such materials For the first a The upper limit of the quantity of such materials For the predicted emergency supplies list The Middle a The quantity of each type of material is predicted to be needed.

[0013] An emergency supplies inventory generation system based on disaster evolution and historical case matching includes: Meteorological Indicator Set Construction Module: Acquire meteorological monitoring and forecast data of the target disaster for the current disaster category, and construct a meteorological indicator set X that includes quantitative values ​​of meteorological elements; Disaster Level Prediction Sample Dataset Construction Module: Constructs a disaster level prediction sample dataset D1 based on historical disaster cases. BP Neural Network Model Training Module: Trains a BP neural network model using the disaster level prediction sample dataset D1; Disaster Level Prediction Module: Inputs the quantified values ​​of meteorological elements in the meteorological index set X into the trained BP neural network model to predict the current disaster level L; Similar Case List Acquisition Module: Based on the current disaster type and disaster level L, search the historical disaster case database for the most similar case and obtain the actual emergency material demand list S for that similar case; Time-series indicator set construction module: Based on the current disaster environmental information, construct a time-series indicator set Y of key influencing factors of emergency material demand, including the quantitative values ​​of key influencing factors; Time-series sample dataset construction module: The time-series sample dataset D2 is generated by constructing an emergency supplies list using historical disaster cases; LSTM model training module: Trains the LSTM model using the time series sample dataset D2; Inventory Prediction Module: Inputs the quantified values ​​of key influencing factors in the time-series indicator set Y into the trained LSTM model to predict the emergency supplies inventory required for the current disaster. ; Final Inventory Acquisition Module: Emergency Supplies Inventory Based on Forecasts The actual emergency supplies demand list S of similar cases is dynamically adjusted to obtain the final emergency supplies list.

[0014] A computer storage medium storing a readable program that, when run, instructs a computing device to perform the emergency supplies list generation method described above, based on disaster evolution and historical case matching.

[0015] An electronic device includes: a processor, a memory, a communication interface, and a communication bus, wherein the processor, the memory, and the communication interface communicate with each other through the communication bus; The memory is used to store at least one executable instruction, which causes the processor to perform operations corresponding to the emergency supplies list generation method based on disaster evolution and historical case matching described above.

[0016] The beneficial effects of this invention are: 1. By quantitatively modeling meteorological monitoring and forecasting data and using BP neural networks to predict disaster levels, it is possible to achieve automated and objective determination of the current disaster level, avoiding the subjectivity and uncertainty brought about by human experience judgment.

[0017] 2. By matching similar cases in the historical disaster case database based on disaster type and predicted disaster level, it is possible to quickly obtain emergency material needs from real historical cases, reducing the time cost of generating emergency material lists.

[0018] 3. By introducing a set of time-series indicators of key influencing factors of emergency material demand and constructing a time-series sample dataset, it is possible to characterize the dynamic changes in material demand during the development of a disaster and improve the timeliness of emergency material demand forecasting.

[0019] 4. By dynamically adjusting and integrating the actual emergency material demand list obtained based on historical cases with the prediction results of the LSTM model, the emergency material list can be adaptively corrected between historical experience and real-time disaster conditions, thereby improving the rationality of material allocation. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 This is a diagram of the BP neural network model for disaster level prediction constructed in Example 2; Figure 2This is a diagram of the Long Short-Term Memory (LSTM) neural network model constructed in Example 2 for generating an emergency supplies list. Detailed Implementation

[0022] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0023] Example 1 The method for generating an emergency supplies list based on disaster evolution and historical case matching includes the following steps: Step 1: Obtain meteorological monitoring and forecast data for the target disaster based on the current disaster category, and construct a set of meteorological indicators. X ={( x i , v i ) | i =1,2,…, I},in I This represents the total quantity of meteorological elements; the specific value depends on the type of disaster. i Indicates the meteorological element index number. x i Indicates the first i Identifiers for meteorological elements, v i Indicates the first i Quantitative values ​​of meteorological elements; Step 2: Construct a disaster severity prediction sample dataset using historical disaster cases. D 1={( X j , L j ) | j =1,2,…, J},in J Indicates the number of historical disasters. j Indicates the index number of historical disasters. X j ={( x i , v ji ) | i =1,2,…, I} indicates the first j A set of meteorological indicators corresponding to historical disasters L j Indicates the first jThe disaster level value corresponding to each historical disaster; Step 3: Disaster level prediction sample dataset obtained in Step 2 D 1. Train a BP neural network model for disaster severity prediction, denoted as . Model 1, Model The input layer neurons of layer 1 are a set of meteorological indicators. X j Quantitative values ​​of all meteorological elements v ji , Model The output layer neurons of layer 1 represent the disaster level. L j ; The BP neural network model includes: input layer neurons, hidden layer neurons, and output layer neurons. These neurons are interconnected through weighted connections to form a multi-layer feedforward neural network structure. The input layer unit is used to receive a set of meteorological indicators. X j Quantitative values ​​of various meteorological elements in China v ji Each input layer neuron corresponds to a quantitative index of a meteorological element. Hidden layer neurons are connected to the input layer neurons via weighted connections, and perform linear weighted summation and nonlinear activation function transformation on the quantitative meteorological element values ​​transmitted from the input layer. This is used to extract nonlinear correlation features between meteorological elements, realizing the mapping representation from multidimensional meteorological features to a disaster evolution feature space. Output layer units are connected to the hidden layer neurons, used to perform weighted fusion calculations on the hidden layer output features, and output the disaster level prediction result through an activation function. L j This enables the output of disaster level classifications.

[0024] Step 4: Combine the meteorological index set constructed in Step 1 X Quantitative values ​​of all meteorological elements v i As Model The input layer neurons and output layer neurons predict the current disaster level. L ; Step 5: Based on the current disaster type and disaster level L Search the historical disaster case database for the most similar case and obtain the actual emergency supplies requirement list for that similar case. S ,in A The types and quantities of supplies. m a For the first a Name of the material n a For the firsta The required quantity of these supplies; The steps to search the historical disaster case database for the most similar case are as follows: S51, Select cases of the same type as the current disaster from historical disaster cases; S52, For each historical disaster case selected in S51, calculate the disaster level similarity between it and the current disaster case according to formula (1). : (1) in, L max This represents the highest disaster level value that has occurred across all historical cases. L min The minimum disaster level value that has appeared in all historical cases. For the first j The disaster level value corresponding to each historical disaster The current disaster level is predicted by a BP neural network model based on the meteorological index set X. S53: Select the historical disaster case with the highest disaster level similarity value as the most similar case to match.

[0025] Step 6: Construct a time-series index set of key influencing factors of emergency material demand based on current disaster environmental information. Y ={( y k , ) | k =1,2,…, K},in K This represents the total number of key factors influencing emergency supplies demand; the specific value depends on the type of disaster. k This indicates the index number of key factors influencing emergency supplies demand. y k Indicates the first k Identifiers of key influencing factors Indicates the first k The continuous influence of key factors in the evolution of disasters T The quantized value at each moment; Step 7: Construct a time-series sample dataset for generating emergency supplies lists using historical disaster case studies. D 2={( Y j , S j )| j =1,2,…, J},in J Indicates the number of historical disasters. jIndicates the index number of disasters over the years. Y j ={( y k , ) | k =1,2,…, K} indicates the first j A historical disaster T A set of time-series indicators of key influencing factors at consecutive time points. S j Indicates the first j A list of actual emergency supplies corresponding to each historical disaster; Step 8: The time-series sample dataset constructed in step 7 D 2. Train a Long Short-Term Memory (LSTM) neural network model for generating an emergency supplies list, denoted as . Model 2, Model The input for 2 is the time series matrix of the quantified numerical values ​​of the key influencing factors. , Model The output of 2 is an emergency supplies list. S j ; The Long Short-Term Memory (LSTM) neural network model includes: input layer units, LSTM layers, fully connected layers, and output layer units. Among them: The input layer unit receives a time-series matrix of quantified values ​​for key influencing factors. This time-series matrix consists of multiple time steps, each corresponding to a quantified value of a key influencing factor under the current disaster environment. The input layer unit inputs the quantified values ​​of the key influencing factors at each time step into a Long Short-Term Memory (LSTM) network layer in chronological order. The LSM layer is used to model the temporal features of the input time-series matrix of quantified values ​​for key influencing factors. A fully connected layer is connected to the output of the LSM layer and is used to perform feature mapping and dimension transformation on the hidden states output by the LSM layer, mapping the temporal features into an output feature vector that matches the structure of the emergency supplies list. The output layer unit outputs the emergency supplies list. S j .

[0026] Step 9: Combine the time-series indicators of key influencing factors of emergency material demand constructed in Step 6. Y Quantitative numerical sequences of all key influencing factors enter Model 2. Predict the list of emergency supplies needed for the current disaster through model reasoning. , For the first a The quantity of various materials required for forecasting; Step 10: Based on the predicted emergency supplies list Dynamically adjust the actual emergency supplies requirements list for similar cases S To obtain the final list of emergency supplies. , For the first a The final quantity of the required materials.

[0027] Based on the predicted emergency supplies list The process of dynamically adjusting the actual emergency supplies demand list S for similar cases is as follows: For each item on the emergency supplies list m a Calculate the corresponding final quantity according to formula (2). : (2) in, and To integrate weights, satisfy , For the first a The minimum quantity of such materials For the first a The upper limit of the quantity of such materials For the first a The quantity of each type of material is predicted to be needed.

[0028] Based on a similar inventive concept, embodiments of the present invention also provide a computer storage medium storing a readable program that, when run by a processor, can execute the above-described method for generating an emergency supplies list based on disaster evolution and historical case matching.

[0029] Based on a similar inventive concept, this invention provides an electronic device, including: a processor, a memory, a communication interface, and a communication bus, wherein the processor, the memory, and the communication interface communicate with each other through the communication bus; The memory is used to store at least one executable instruction, which causes the processor to perform the operation corresponding to the above-described emergency supplies list generation method based on disaster evolution and historical case matching.

[0030] Based on a similar inventive concept, embodiments of the present invention also provide a computer program product, including computer instructions, which instruct a computing device to perform the operations corresponding to the above-described emergency supplies list generation method based on disaster evolution and historical case matching.

[0031] Example 2 This embodiment uses a case study of a severe typhoon disaster in a coastal city to illustrate in detail the method for generating an emergency supplies list based on dynamic matching of disaster evolution and historical cases; it includes the following steps: Step 1: Obtain meteorological monitoring and forecast data for the target disaster based on the current disaster category, and construct a set of meteorological indicators. X ={( x i , v i ) | i =1,2,…, I},in I This represents the total quantity of meteorological elements; the specific value depends on the type of disaster. i Indicates the meteorological element index number. x i Indicates the first i Identifiers for meteorological elements, v i Indicates the first i The quantitative values ​​of meteorological elements, and the set of meteorological indicators constructed in this embodiment. X ={(Typhoon center wind speed, 48m / s), (24-hour cumulative rainfall, 160mm), (average wind force level, level 13), (minimum air pressure, 965hPa), (duration of sustained wind, 14h)}; Step 2: Construct a disaster severity prediction sample dataset using historical disaster cases. D 1={( X j , L j ) | j =1,2,…, J},in J Indicates the number of historical disasters. j Indicates the index number of historical disasters. X j ={( x i , v ji ) | i =1,2,…, I} indicates the first j A set of meteorological indicators corresponding to historical disasters L j Indicates the first j For each historical disaster, a corresponding disaster level value is defined. In this embodiment, 100 historical disaster cases related to typhoons are extracted from the historical database to construct... D 1={( X j , L j ) | j =1,2,…,100}, where the disaster level is 1,2,…,100. L j∈{1,2,3,4,5}, where 5 is the most severe; Step 3: Disaster level prediction sample dataset obtained in Step 2 D 1. Train a BP neural network model for disaster level prediction, denoted as Model 1, Model The input layer neurons of layer 1 are a set of meteorological indicators. X j Quantitative values ​​of all meteorological elements v ji , Model The output layer neurons of layer 1 represent the disaster level. L j In this embodiment, the constructed BP neural network model is a standard three-layer BP neural network, with an input layer containing 5 neurons, a hidden layer containing 16 neurons with ReLU activation, and an output layer containing 1 neuron, as shown below. Figure 1 As shown; Step 4: Combine the meteorological index set constructed in Step 1 X Quantitative values ​​of all meteorological elements v i As Model The input layer neurons and output layer neurons predict the current disaster level. L In this embodiment, the predicted current disaster level is 4; Step 5: Based on the current disaster type and disaster level L Search the historical disaster case database for the most similar case and obtain the actual emergency supplies requirement list for that similar case. S ,in A The types and quantities of supplies. m a For the first a Name of the material n a For the first a The required quantity of these materials, obtained in this embodiment S ={(Water, 2000), (Instant Noodles, 1500), (Tent, 300), ..., (Flashlight, 500)}; Among them, based on the current disaster type and disaster level L The steps to search the historical disaster case database for the most similar case are as follows: 5.1): Select cases with the same type as the current disaster from historical disaster cases. In this embodiment, a total of 40 cases with the same type were selected. 5.2): For each historical disaster case selected in step 5.1), the disaster level similarity between it and the current disaster case is calculated according to formula (1), where... L max The maximum disaster level value that occurred in all historical cases is 5 in this embodiment. L min The minimum disaster level value that has occurred in all historical cases is 1 in this embodiment; 5.3): Select the historical disaster case with the highest disaster level similarity value as the most similar case matched.

[0032] Step 6: Construct a time-series index set of key influencing factors of emergency material demand based on current disaster environmental information. Y ={( y k , ) | k =1,2,…, K},in K This represents the total number of key factors influencing emergency supplies demand; the specific value depends on the type of disaster. k This indicates the index number of key factors influencing emergency supplies demand. y k Indicates the first k Identifiers of key influencing factors Indicates the first k The continuous influence of key factors in the evolution of disasters T The quantized values ​​at each moment, and the key influencing factors selected in this embodiment include {( y 1: Changes in typhoon wind speed; y 2: Cumulative rainfall trend; y 3: Urban drainage load rate; y 4: Distribution ratio of key population groups; y 5: Traffic disruption index}, collecting continuous data T =5 hours of monitoring data; Step 7: Construct a time-series sample dataset for generating emergency supplies lists using historical disaster case studies. D 2={( Y j , S j )| j =1,2,…, J},in J Indicates the number of historical disasters. j Indicates the index number of disasters over the years. Y j ={(y k , ) | k =1,2,…, K} indicates the first j A historical disaster T A set of time-series indicators of key influencing factors at consecutive time points. S j Indicates the first j A list of actual emergency supplies corresponding to each historical disaster; Step 8: The time-series sample dataset constructed in step 7 D 2. Train a long short-term memory neural network model for generating an emergency supplies list, denoted as . Model 2, Model The input for 2 is the time series matrix of the quantified numerical values ​​of the key influencing factors. , Model The output of 2 is an emergency supplies list. S j In this embodiment, the Long Short-Term Memory (LSTM) neural network model has an input dimension of 5x5, corresponding to the quantified values ​​of five key influencing factors of emergency material demand at five consecutive moments during the disaster evolution process. The LTM network layer contains 64 units, and the fully connected layer outputs an emergency material list vector as shown below. Figure 2 As shown; Step 9: Combine the time-series indicators of key influencing factors of emergency material demand constructed in Step 6. Y Quantitative numerical sequences of all key influencing factors enter Model 2. Predict the list of emergency supplies needed for the current disaster through model reasoning. , For the first a The quantity of the required materials is predicted in this embodiment. ={(Water, 2200), (Instant Noodles, 1700), (Tent, 350), ..., (Flashlight, 600)}; Step 10: Combine the predicted emergency supplies list Dynamically adjust the basic emergency supplies requirement list for similar cases S To obtain the final list of emergency supplies. , For the first a The final quantity of each type of material required, and the final emergency material list obtained in this embodiment. ={(water, 2120), (instant noodles, 1620), (tent, 330), ..., (flashlight, 560)}.

[0033] Among them, the emergency supplies list is based on the forecast. Dynamically adjust the emergency supplies demand list for similar cases S The process is as follows: For each item on the emergency supplies list m a The final quantity is calculated according to formula (2), where and To integrate weights, satisfy , For the first a The minimum quantity of such materials For the first a The maximum quantity of this type of material is specified in this embodiment. , .

[0034] (2) Example 3 This embodiment proposes an emergency supplies list generation system based on disaster evolution and historical case matching, specifically including: Meteorological Indicator Set Construction Module: Acquire meteorological monitoring and forecast data of the target disaster for the current disaster category, and construct a meteorological indicator set X that includes quantitative values ​​of meteorological elements; Disaster Level Prediction Sample Dataset Construction Module: Constructs a disaster level prediction sample dataset D1 based on historical disaster cases. BP Neural Network Model Training Module: Trains a BP neural network model using the disaster level prediction sample dataset D1; Disaster Level Prediction Module: Inputs the quantified values ​​of meteorological elements in the meteorological index set X into the trained BP neural network model to predict the current disaster level L; Similar Case List Acquisition Module: Based on the current disaster type and disaster level L, search the historical disaster case database for the most similar case and obtain the actual emergency material demand list S for that similar case; Time-series indicator set construction module: Based on the current disaster environmental information, construct a time-series indicator set Y of key influencing factors of emergency material demand, including the quantitative values ​​of key influencing factors; Time-series sample dataset construction module: The time-series sample dataset D2 is generated by constructing an emergency supplies list using historical disaster cases; LSTM model training module: Trains the LSTM model using the time series sample dataset D2; Inventory Prediction Module: Inputs the quantified values ​​of key influencing factors in the time-series indicator set Y into the trained LSTM model to predict the emergency supplies inventory required for the current disaster. ; Final Inventory Acquisition Module: Emergency Supplies Inventory Based on Forecasts The actual emergency supplies demand list S for similar cases is dynamically adjusted to obtain the final emergency supplies list.

[0035] The methods of the present invention can be implemented in hardware, firmware, or as software or computer code that can be stored in a recording medium (such as a CD-ROM, RAM, floppy disk, hard disk, or magneto-optical disk), or as computer code originally stored on a remote recording medium or a non-transitory machine-readable medium and subsequently stored on a local recording medium, downloaded via a network. Thus, the methods described herein can be processed by software stored on a recording medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware (such as an ASIC or FPGA). It is understood that the computer, processor, microprocessor controller, or programmable hardware includes storage components (e.g., RAM, ROM, flash memory, etc.) capable of storing or receiving software or computer code that, when accessed and executed by the computer, processor, or hardware, implements the methods described herein. Furthermore, when a general-purpose computer accesses the code used to implement the methods shown herein, the execution of the code transforms the general-purpose computer into a dedicated computer for performing the methods shown herein.

[0036] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the claimed invention.

Claims

1. A method for generating an emergency supplies list based on disaster evolution and historical case matching, characterized in that, Includes the following steps: S1, acquire meteorological monitoring and forecast data of the target disaster for the current disaster category, and construct a set of meteorological indicators X including the quantitative values ​​of meteorological elements; S2, construct disaster level prediction sample dataset D1 based on historical disaster cases. S3, using the disaster level prediction sample dataset D1 to train a BP neural network model; S4: Input the quantified values ​​of meteorological elements in the meteorological index set X into the trained BP neural network model to predict the current disaster level L; S5, based on the current disaster type and disaster level L, search the historical disaster case database for the most similar case and obtain the actual emergency material demand list S for the similar case; S6, construct a time series index set Y of key influencing factors of emergency material demand based on the current environmental information of the disaster, including the quantitative values ​​of key influencing factors; S7, using historical disaster cases to construct an emergency supplies list to generate a time-series sample dataset D2; S8, use the time series sample dataset D2 to train the LSTM model; S9 inputs the quantified values ​​of key influencing factors in the time-series index set Y into the trained LSTM model to predict the emergency supplies list needed for the current disaster. ; S10, based on the predicted emergency supplies list The actual emergency supplies demand list S of similar cases is dynamically adjusted to obtain the final emergency supplies list.

2. The emergency supplies list generation method based on disaster evolution and historical case matching according to claim 1, characterized in that, The disaster level prediction sample dataset D 1={( X j , L j ) | j =1,2,…, J },in: J Indicates the number of historical disasters. j Indicates the index number of historical disasters. X j ={( x i , v ji ) | i =1,2,…, I } indicates the first j A set of meteorological indicators corresponding to historical disasters L j Indicates the first j The disaster level value corresponding to each historical disaster i Indicates the meteorological element index number. I It represents the total quantity of meteorological elements.

3. The emergency supplies list generation method based on disaster evolution and historical case matching according to claim 2, characterized in that, When training the BP neural network model, a set of meteorological indicators is used. X j Quantitative values ​​of all meteorological elements v ji For input.

4. The emergency supplies list generation method based on disaster evolution and historical case matching according to claim 1, characterized in that, List of actual emergency supplies needed in similar cases S ,in: A For the types and quantities of supplies, m a For the first a Name of the material n a For the first a The required quantity of these supplies; The steps for searching the historical disaster case database to find the most similar case include: S51, Select cases of the same type as the current disaster from historical disaster cases; S52, For each selected historical disaster case, calculate the disaster level similarity between it and the current disaster case. : in, L max This represents the highest disaster level value that has occurred across all historical cases. L min The minimum disaster level value that has appeared in all historical cases. For the first j The disaster level value corresponding to each historical disaster The current disaster level is predicted by a BP neural network model based on the meteorological index set X. S53, select the historical disaster case with the highest disaster level similarity value as the most similar case.

5. The emergency supplies list generation method based on disaster evolution and historical case matching according to claim 1, characterized in that, The time-series sample dataset D 2={( Y j , S j ) | j =1,2,…, J },in: J Indicates the number of historical disasters. j Indicates the index number of disasters over the years. Y j ={( y k , ) | k =1,2,…, K } indicates the first j A historical disaster T A set of time-series indicators of key influencing factors at consecutive time points. K This indicates the total number of key factors influencing the demand for emergency supplies. k This indicates the index number of key factors influencing emergency supplies demand. y k Indicates the first k Identifiers of key influencing factors Indicates the first j The first historical disaster k Key influencing factors T A set of time series indicators at consecutive moments; S j Indicates the first j A list of actual emergency supplies corresponding to each historical disaster.

6. The emergency supplies list generation method based on disaster evolution and historical case matching according to claim 5, characterized in that, When training the LSTM model, the quantified numerical time series matrix of key influencing factors is used. For input.

7. The emergency supplies list generation method based on disaster evolution and historical case matching according to claim 4, characterized in that, The final emergency supplies list , For the first a The final quantity of the required materials; Calculate the first a The formula for the final quantity of the required materials is: in, and To integrate weights, satisfy , For the first a The minimum quantity of such materials For the first a The upper limit of the quantity of such materials For the predicted emergency supplies list The Middle a The quantity of each type of material is predicted to be needed.

8. An emergency supplies list generation system based on disaster evolution and historical case matching, characterized in that, include: Meteorological Indicator Set Construction Module: Acquire meteorological monitoring and forecast data of the target disaster for the current disaster category, and construct a meteorological indicator set X that includes quantitative values ​​of meteorological elements; Disaster Level Prediction Sample Dataset Construction Module: Constructs a disaster level prediction sample dataset D1 based on historical disaster cases. BP Neural Network Model Training Module: Trains a BP neural network model using the disaster level prediction sample dataset D1; Disaster Level Prediction Module: Inputs the quantified values ​​of meteorological elements in the meteorological index set X into the trained BP neural network model to predict the current disaster level L; Similar Case List Acquisition Module: Based on the current disaster type and disaster level L, search the historical disaster case database for the most similar case and obtain the actual emergency material demand list S for that similar case; Time-series indicator set construction module: Based on the current disaster environmental information, construct a time-series indicator set Y of key influencing factors of emergency material demand, including the quantitative values ​​of key influencing factors; Time-series sample dataset construction module: The time-series sample dataset D2 is generated by constructing an emergency supplies list using historical disaster cases; LSTM model training module: Trains the LSTM model using the time series sample dataset D2; Inventory Prediction Module: Inputs the quantified values ​​of key influencing factors in the time-series indicator set Y into the trained LSTM model to predict the emergency supplies inventory required for the current disaster. ; Final Inventory Acquisition Module: Emergency Supplies Inventory Based on Forecasts The actual emergency supplies demand list S of similar cases is dynamically adjusted to obtain the final emergency supplies list.

9. A computer storage medium storing a readable program, characterized in that, When the program runs, it can instruct the computing device to execute the emergency supplies list generation method based on disaster evolution and historical case matching as described in any one of claims 1-7.

10. An electronic device, characterized in that, include: The processor, memory, communication interface, and communication bus are provided, wherein the processor, memory, and communication interface communicate with each other via the communication bus. The memory is used to store at least one executable instruction, which causes the processor to perform the operation corresponding to the emergency supplies list generation method based on disaster evolution and historical case matching as described in any one of claims 1-7.

Images