Disaster relief plan generation method and device based on disaster characteristics

Abstract

The invention relates to the technical field of emergency rescue and disaster relief plan auxiliary recommendation, in particular to a disaster relief plan generation method and device based on disaster characteristics. According to the plan recommendation model, current disaster information is matched with a self-defined plan and a historical disaster plan to obtain a first matching probability and a second matching probability, and a recommendation result is obtained according to the first matching probability and the second matching probability; according to the disaster relief plan recommendation method and the disaster relief plan recommendation system, different types of disaster characteristics are quickly and accurately identified by combining a plan recommendation algorithm, and the personalized disaster relief plan is recommended and generated according to the disaster characteristics, so that the disaster response efficiency and effect can be remarkably improved, and the disaster relief plan recommendation method and the disaster relief plan recommendation system are high in practicability. And economic loss and casualties caused by disasters are reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of emergency rescue plan auxiliary recommendation, and in particular to a method and device for generating a disaster rescue plan based on disaster characteristics. Background Art

[0002] At present, the frequency and intensity of natural disasters are increasing, causing huge losses to human society. In order to effectively deal with disasters, a lot of resources have been invested in disaster management, including disaster warning, emergency response, resource allocation, etc. However, the existing disaster management plan system has some limitations, such as untimely plan updates, low matching degree of different disasters, and slow response speed when disasters occur.

[0003] Although the application of technologies such as data analysis, machine learning, and artificial intelligence in the field of disaster management has made significant progress in recent years, how to effectively combine these technologies to achieve rapid and accurate identification of disaster characteristics and generate or recommend appropriate plans based on them remains a technical challenge. In recent years, with the rapid development of big data and big model technology, researchers have proposed a variety of plan generation methods based on different algorithms and technologies. For example, some studies use historical disaster data and statistical methods to predict the probability and impact range of disasters. However, these methods often lack the ability to deeply analyze disaster characteristics and generate personalized plans, resulting in the generated plans being unable to meet actual conditions.

[0004] In view of this, overcoming the defects of the prior art is an urgent problem to be solved in the field of this technology. Summary of the invention

[0005] The technical problem to be solved by the present invention is how to generate a disaster emergency plan that meets the actual situation according to the disaster characteristics of the current disaster.

[0006] The present invention adopts the following technical solution:

[0007] In a first aspect, a method for generating a disaster relief plan based on disaster characteristics is provided, the method comprising:

[0008] Build a plan recommendation model;

[0009] According to the plan recommendation model, the current disaster information is matched with the custom plan and the historical disaster plan respectively to obtain a first matching probability and a second matching probability respectively;

[0010] Obtaining a recommendation result according to the first matching probability and the second matching probability;

[0011] A disaster relief plan corresponding to the current disaster is generated according to the recommendation result.

[0012] Preferably, matching the current disaster information with the custom plan and the historical disaster plan respectively according to the plan recommendation model to obtain a first matching probability and a second matching probability respectively, specifically includes:

[0013] Obtain disaster factor matrix based on current disaster information;

[0014] Calculating a first cosine similarity and a first divergence correction factor between the disaster factor matrix and the custom plan, and correcting the first cosine similarity according to the first divergence correction factor to obtain the first matching probability;

[0015] The second cosine similarity and the second divergence correction factor between the disaster element matrix and the historical disaster emergency plan are calculated, and the second cosine similarity is corrected according to the second divergence correction factor to obtain the second matching probability.

[0016] Preferably, the first cosine similarity and the second cosine similarity are calculated in the same manner, specifically as follows:

[0017] A penalty factor is added to the original cosine similarity to obtain an improved cosine similarity, so as to obtain the first cosine similarity and the second cosine similarity respectively, wherein the penalty factor is calculated as follows:

[0018]

[0019] Among them, pe i is the penalty factor, tol represents the total number of disasters, c i is the number of disaster type i; the calculation formula of the improved cosine similarity is:

[0020]

[0021] Among them, Sim(u,v) represents the similarity between decision makers u and v, I uv is the set of historical plans scored by decision makers u and v, R u,i is the score of decision maker u on historical plan i, R v,i is the score given by decision maker v to historical plan i, is the average score of all historical plans given by decision maker u, is the average score of all historical plans given by decision maker v, I u represents the set of historical plans rated by decision maker u, I v Represents the set of historical plans rated by decision maker v.

[0022] Preferably, the method for obtaining the historical disaster emergency plan includes:

[0023] Reading a historical disaster dataset from a distributed file system, and converting the historical disaster dataset into a first elastic distributed dataset;

[0024] Filtering out historical disaster numbers, type numbers, and emergency plan numbers from the first elastic distributed data set, and obtaining a second elastic distributed data set according to the historical disaster numbers, the type numbers, and the emergency plan numbers;

[0025] The second elastic distributed data set is converted into a structured data set to obtain the historical disaster emergency plan.

[0026] Preferably, obtaining a recommendation result according to the first matching probability and the second matching probability includes:

[0027] Acquire a candidate disaster relief plan corresponding to each first matching probability according to the custom plan, and establish a first disaster similarity data set according to the first matching probability, the corresponding candidate disaster relief plan and the current disaster information;

[0028] Acquire a candidate disaster relief plan corresponding to each second matching probability according to the historical disaster plan, and establish a second disaster similarity data set according to the second matching probability, the corresponding candidate disaster relief plan and the current disaster information;

[0029] Score each candidate disaster relief plan in the first disaster similarity data set and the second disaster similarity data set respectively to obtain a scoring result for each candidate disaster relief plan, sort each candidate disaster relief plan according to the scoring result, and select a preset number of candidate disaster relief plans as the final recommendation result.

[0030] Preferably, scoring each candidate disaster relief plan in the first disaster similarity dataset and the second disaster similarity dataset comprises:

[0031] Scoring the candidate disaster relief plans in the first disaster similarity dataset and the second disaster similarity dataset respectively by using a scoring prediction formula;

[0032] The rating prediction formula is:

[0033]

[0034] in, represents the predicted score of decision maker u for candidate disaster relief plan i, N(u) represents the set of decision makers similar to decision maker u, these decision makers have a common scoring history with decision maker u, and their scoring behavior is similar to that of decision maker u, sin(u,v) represents the similarity between decision makers u and decision makers v; r vi It represents the actual score given by decision maker v to candidate disaster relief plan i.

[0035] Preferably, generating a disaster relief plan corresponding to the current disaster according to the recommendation result includes:

[0036] According to the recommendation results, organizations in the surrounding areas that can provide corresponding disaster relief forces are obtained, and the geographical locations of the corresponding organizations are displayed on the map;

[0037] According to the geographical location of each organization and the road condition information of the current disaster, the shortest action route in time and distance are planned for each organization and displayed on the map to generate a disaster relief plan corresponding to the current disaster.

[0038] Preferably, the current disaster information includes the latitude and longitude information of the current disaster, casualty information, economic loss information, and data information specific to different disasters.

[0039] In a second aspect, a disaster relief plan generation device based on disaster characteristics is provided, the disaster relief plan generation device based on disaster characteristics comprising: a processor and a memory for storing processor executable instructions;

[0040] Wherein, the processor is configured to execute the method for generating a disaster relief plan based on disaster characteristics as described in the first aspect.

[0041] In a third aspect, a non-volatile computer storage medium is provided, wherein the computer storage medium stores computer executable instructions, which are executed by one or more processors to complete the method for generating disaster relief plans based on disaster characteristics as described in the first aspect.

[0042] In a fourth aspect, a chip is provided, comprising: a processor and an interface, for calling and running a computer program stored in a memory from a memory, and executing the method for generating a disaster relief plan based on disaster characteristics as described in the first aspect.

[0043] In a fifth aspect, a computer program product comprising instructions is provided. When the instructions are executed on a computer or a processor, the computer or the processor executes the method for generating a disaster relief plan based on disaster characteristics as described in the first to fourth aspects and any one of the aspects.

[0044] Compared with the prior art, the present invention has the following beneficial effects:

[0045] The present invention combines a plan recommendation algorithm to quickly and accurately identify the characteristics of different types of disasters, and recommends and generates personalized disaster relief plans based on the disaster characteristics, which can significantly improve the efficiency and effectiveness of disaster response and reduce economic losses and casualties caused by disasters. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0047] Figure 1 It is a flowchart of a method for generating a disaster relief plan based on disaster characteristics provided by an embodiment of the present invention;

[0048] Figure 2 It is a flowchart of a method for generating a disaster relief plan based on disaster characteristics provided by an embodiment of the present invention;

[0049] Figure 3 is a schematic diagram of a flow chart of obtaining a first matching probability and a second matching probability provided by an embodiment of the present invention;

[0050] Figure 4 It is a flow chart of building a plan recommendation model provided by an embodiment of the present invention;

[0051] Figure 5 is a schematic diagram of a process for obtaining recommendation results provided by an embodiment of the present invention;

[0052] Figure 6 It is a schematic diagram of a process for generating a disaster relief plan provided by an embodiment of the present invention;

[0053] Figure 7 It is a structural schematic diagram of a disaster relief plan generation device based on disaster characteristics provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0054] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0055] Unless the context requires otherwise, throughout the specification and claims, the term "including" is to be interpreted as open inclusion, that is, "including, but not limited to". In the description of the specification, the terms "one embodiment", "some embodiments", "exemplary embodiments", "examples", "specific examples" or "some examples" and the like are intended to indicate that specific features, structures, materials or characteristics associated with the embodiment or example are included in at least one embodiment or example of the present disclosure. The schematic representation of the above terms does not necessarily refer to the same embodiment or example. In addition, the specific features, structures, materials or characteristics may be included in any one or more embodiments or examples in any appropriate manner, that is, although they may be carried in the embodiments or examples of the above terms due to reasons such as the order and position of appearance, it is not limited to that they can be carried in combination by one embodiment or example.

[0056] In the description of the present invention, the terms "first" and "second" are used only for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of indicated technical features. Thus, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the embodiments of the present disclosure, unless otherwise specified, the meaning of "multiple" is two or more. In addition, for example, the same type of nouns may be described as two independent individuals by adding "A" and "B" at the end. In this case, the corresponding features defined as "A" and "B" are only used to distinguish the same type of individuals for description purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of indicated technical features.

[0057] When describing some embodiments, the expressions "coupling", "coupling" and "connection" and their derivatives may be used. For example, when describing some embodiments, the term "connection" may be used to indicate that two or more components are in direct physical or electrical contact with each other. For another example, when describing some embodiments, the term "coupling" may be used to indicate that two or more components are in direct physical or electrical contact. However, the terms "connection" or "coupling" may also refer to two or more components that are not in direct contact with each other, but still cooperate or interact with each other, such as "optical path coupling", "wireless connection", etc. The embodiments disclosed herein are not necessarily limited to the contents of the present invention.

[0058] In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0059] Embodiment 1:

[0060] The main purpose of this embodiment is to match the custom plan and the historical disaster plan according to the disaster characteristics of the current disaster, so as to assist in generating the disaster relief plan corresponding to the current disaster according to the selected custom plan and / or historical disaster plan. In view of the problems that disaster relief operations cannot be carried out in a targeted, timely and accurate manner under the constraints of multiple factors such as different geographical environments, population densities and infrastructures for different disasters, it is necessary to provide reasonable resource allocation according to the type and characteristics of the disaster, provide more accurate disaster risk assessment, and help decision makers formulate more effective disaster prevention and mitigation measures.

[0061] Next, the steps of the method for generating a disaster relief plan based on disaster characteristics will be described in detail. Figure 1 and Figure 2 As shown, the method includes:

[0062] Step 101: Construct a plan recommendation model, and match the current disaster information with the custom plan and the historical disaster plan according to the plan recommendation model to obtain a first matching probability and a second matching probability, respectively, and obtain a recommendation result according to the first matching probability and the second matching probability.

[0063] The custom plan includes custom disaster information and custom disaster relief plan. The historical disaster plan includes historical disaster information and historical disaster relief plan. The process of matching the current disaster information with the custom plan and the historical disaster plan is essentially the process of matching the current disaster information with the custom disaster information and the historical disaster information.

[0064] Among them, the emergency plan recommendation model learns the data, extracts key information such as the correlation pattern between disaster characteristics and emergency plans, and recommends custom emergency plans and historical disasters based on the extracted information. For example, for earthquake disasters, the emergency plan recommendation model learns the relationship between the earthquake magnitude, focal depth, geographical characteristics of the disaster-stricken area, population density and other characteristics and various emergency plan contents such as rescue measures, material deployment, and personnel resettlement for earthquake response.

[0065] Before the disaster arrives, the decision-making unit will formulate a corresponding custom plan to deal with possible disasters, and compare and analyze the various characteristics of the current disaster information, such as the type of disaster, intensity, geographical environment of the location of occurrence, socioeconomic conditions, etc., with the applicable conditions and response strategies in the custom plan. Calculate the similarity between the two, and then obtain the first matching probability. For example, if the custom plan is formulated for coastal typhoon disasters, and the current disaster is an inland flood, then the matching probability may be low. Make a comprehensive comparison of the characteristics of the current disaster with disasters that have occurred in history, including the attribute characteristics and derivative characteristics of the disaster. For example, comprehensively consider the scale, duration, and degree of damage caused by the disaster, find historical disaster cases similar to the current disaster, and then calculate the degree of matching with the current disaster based on the plans adopted by the historical disasters and their effects, and obtain the second matching probability.

[0066] Step 102: Generate a disaster relief plan corresponding to the current disaster according to the recommendation result.

[0067] According to the size of the first matching probability, one or more custom plans are selected as the first recommendation result; according to the size of the second matching probability, one or more disaster relief plans corresponding to historical disasters are selected as the second recommendation result. According to the first recommendation result and the second recommendation result, relevant plan contents can be screened and integrated from the custom plan library, or existing plans can be adjusted and optimized. For example, if the recommendation result shows that the plan for a certain historical earthquake disaster has a high degree of match with the current earthquake disaster, the experience of personnel rescue, material distribution, temporary resettlement, etc. in the historical plan is referred to, combined with the actual situation of the current disaster, such as the number of people affected and the degree of building damage, etc., to generate a disaster relief plan suitable for the current disaster.

[0068] In this embodiment, by combining the plan recommendation algorithm, the characteristics of different types of disasters can be quickly and accurately identified, and personalized disaster relief plans can be recommended and generated based on the disaster characteristics, which can significantly improve the efficiency and effectiveness of disaster response and reduce economic losses and casualties caused by disasters.

[0069] In one embodiment, Figure 3 and Figure 4 As shown, the current disaster information is matched with the custom plan and the historical disaster plan according to the plan recommendation model to obtain a first matching probability and a second matching probability, respectively, which specifically includes:

[0070] Step 1011: Obtain a disaster factor matrix based on current disaster information.

[0071] The current disaster information includes the latitude and longitude information of the current disaster, casualty information, economic loss information, and data information specific to different disasters. First, analyze the various characteristic information of the current disaster, which may include the type of disaster (such as earthquake, flood, fire, etc.), disaster intensity (earthquake magnitude, flood water volume, fire scale, etc.), geographical characteristics of the occurrence site (topography, altitude, longitude and latitude, etc.), time characteristics (occurrence time, duration, etc.), and socio-economic influencing factors (population density, distribution of important infrastructure, etc.).

[0072] The above different disaster characteristics are quantified and sorted, and arranged in a certain order to form a multi-dimensional disaster factor matrix. For example, if three disaster characteristics are considered, represented by numerical values ​​as A, B, and C, then the disaster factor matrix can be expressed as [[A1, B1, C1], [A2, B2, C2], ...], where each row represents a specific combination of disaster characteristics, corresponding to a possible disaster situation.

[0073] Step 1012: Calculate the first cosine similarity and the first divergence correction factor between the disaster factor matrix and the user-defined plan, and correct the first cosine similarity according to the first divergence correction factor to obtain the first matching probability.

[0074] Step 1013: Calculate the second cosine similarity and the second divergence correction factor between the disaster element matrix and the historical disaster plan, and correct the second cosine similarity according to the second divergence correction factor to obtain the second matching probability.

[0075] For a vector in the disaster element matrix corresponding to the given current disaster information (representing a specific disaster situation), it is compared with the vectors corresponding to the custom plan and the historical disaster. Cosine similarity measures the similarity between two vectors by calculating the cosine value of the angle between them. The closer the value of the similarity is to 1, the more similar the two disaster situations are. This means that the current disaster is most similar to the disaster corresponding to a custom plan or a historical disaster.

[0076] The divergence correction factor is used to measure the difference between two probability distributions. In disaster analysis, the divergence correction factor can be determined by calculating the divergence between the probability distribution corresponding to the current disaster information and the probability distribution corresponding to historical disasters or custom plans. For example, methods such as Kullback-Leibler divergence or Jensen-Shannon divergence can be used. If the probability distribution of the current disaster and a historical disaster on certain features is very different, then the corresponding divergence value will be larger, resulting in a smaller divergence correction factor, which is used to adjust the cosine similarity to more accurately reflect the actual similarity between the two.

[0077] The calculation formulas of the first divergence correction factor and the second divergence correction factor are both:

[0078]

[0079] Among them, JS(M,Q) represents the Jensen-Shannon divergence between probability distribution M and probability distribution Q, D kl (M,H) represents the Kullback-Leibler divergence between probability distribution M and probability distribution H, D kl (Q, H) represents the Kullback-Leibler divergence between probability distribution Q and probability distribution H. In one embodiment, M represents the probability distribution corresponding to the current disaster information, and Q represents the probability corresponding to the historical disaster or the custom plan.

[0080] In one embodiment, considering that some common disasters occur frequently but have a limited impact range, and rescue is greatly affected by environmental factors, it is difficult for traditional filtering recommendation algorithms to give accurate recommendation predictions, so a penalty factor is added to the traditional modified cosine similarity.

[0081] Sometimes relying solely on cosine similarity may not accurately reflect the differences between disasters, especially when certain disaster characteristics are of special importance or there are some negative factors. For example, if the current disaster occurs in a densely populated area, and the historical disaster occurs in a sparsely populated area, even if other characteristics are similar, this difference needs to be penalized to reduce the similarity judgment. The penalty factor can be determined based on the specific disaster characteristics and actual needs. It can be a fixed value or a dynamic value calculated according to specific rules. For example, the size of the penalty factor can be determined based on factors such as differences in population density and differences in disaster intensity. The penalty factor is combined with the original cosine similarity to obtain the improved cosine similarity.

[0082] In one embodiment, the first cosine similarity and the second cosine similarity are calculated in the same manner, specifically, by adding a penalty factor to the original cosine similarity to obtain an improved cosine similarity, so as to obtain the first cosine similarity and the second cosine similarity respectively, wherein the penalty factor is calculated as follows:

[0083]

[0084] Among them, pe i is the penalty factor, tol represents the total number of disasters, c i is the number of disaster type i occurring.

[0085] The calculation formula of the improved cosine similarity is:

[0086]

[0087] Among them, Sim(u,v) represents the similarity between decision makers u and v, I uv is the set of historical plans scored by decision makers u and v, R u,i is the score of decision maker u on historical plan i, R v,i is the score given by decision maker v to historical plan i, is the average score of all historical plans given by decision maker u, is the average score of all historical plans given by decision maker v, U u represents the set of historical plans rated by decision maker u, I v Represents the set of historical plans rated by decision maker v.

[0088] The divergence correction factor is combined with the cosine similarity. For example, the cosine similarity can be multiplied by the divergence correction factor to obtain the first matching probability and the second matching probability. In this way, the influence of the cosine similarity and the divergence can be comprehensively considered, making the similarity measurement more accurate.

[0089] The plan recommendation model can input the disaster factor matrix corresponding to the current disaster information, and output the similarity measurement value (i.e., the first matching probability and the second matching probability) with the custom plan or the historical disaster plan. By comparing the similarity measurement values ​​between different plans, the plan recommendation result that best suits the current disaster can be determined. For example, when a new earthquake disaster information is input, the plan recommendation model can calculate the similarity with each custom plan (i.e., the first matching probability) and the similarity with the historical earthquake plan (i.e., the second matching probability), thereby recommending the most similar custom plan and historical disaster plan for reference by decision makers based on the corresponding first matching probability and second matching probability.

[0090] In one embodiment, the method for obtaining the historical disaster plan includes: reading a historical disaster data set from a distributed file system, and converting the historical disaster data set into a first elastic distributed data set; filtering out historical disaster numbers, type numbers and plan numbers from the first elastic distributed data set, and obtaining a second elastic distributed data set based on the historical disaster numbers, the type numbers and the plan numbers; converting the second elastic distributed data set into a structured data set to obtain the historical disaster plan.

[0091] Among them, the historical disaster dataset contains metadata of historical disasters (such as type, impact range, level, etc.) and basic information of custom plans (plan type, material equipment information, personnel information, etc.). Distributed file systems are usually used to store large-scale data. The historical disaster dataset contains detailed information on various historical disaster events, such as the time, location, type, intensity, damage caused by the disaster, and the corresponding disaster relief plan number.

[0092] Reading the above data from the distributed file system can ensure efficient processing of a large amount of historical data. The historical disaster data set is converted into a first elastic distributed data set (such as in the Spark framework, the first elastic distributed data set is a data set that can be operated in parallel on a cluster). The advantage of this is that the advantages of distributed computing can be used to quickly process large-scale data. For example, if there are a large number of historical disaster records, the data can be distributed to multiple computing nodes for parallel processing through the first elastic distributed data set, thereby improving the efficiency and speed of data processing.

[0093] In one embodiment, Figure 5 As shown, obtaining a recommendation result according to the first matching probability and the second matching probability includes:

[0094] Step 201: Obtain a candidate disaster relief plan corresponding to each first matching probability according to the custom plan, and establish a first disaster similarity data set according to the first matching probability, the corresponding candidate disaster relief plan and the current disaster information.

[0095] Step 202: Obtain a candidate disaster relief plan corresponding to each second matching probability according to the historical disaster plan, and establish a second disaster similarity data set according to the second matching probability, the corresponding candidate disaster relief plan and the current disaster information.

[0096] In step 201 and step 202, the improved cosine similarity calculation method is used to calculate the similarity between the current disaster information and the historical disaster plan and the custom plan. Using the historical disaster number and type number as indexes, the feature vector of the current disaster is compared with the feature vector of the historical disaster to obtain a similarity value. For example, if the current disaster is an earthquake, its characteristics (such as magnitude, focal depth, affected area, etc.) are compared with the characteristics of historical earthquakes, and the similarity is calculated using the improved cosine similarity formula. The calculated disaster similarity value is combined with the corresponding historical disaster number, type number, and plan number to form a disaster similarity data set.

[0097] The first disaster similarity dataset records the similarity between the current disaster information and each custom disaster (i.e., the first matching probability), as well as the corresponding plan information (i.e., the corresponding candidate disaster relief plan), providing a basis for the subsequent custom plan recommendation. The second disaster similarity dataset records the similarity between the current disaster information and each historical disaster (i.e., the second matching probability), as well as the corresponding plan information (i.e., the corresponding candidate disaster relief plan), providing a basis for the subsequent historical disaster plan recommendation.

[0098] Step 203: Score each candidate disaster relief plan in the first disaster similarity data set and the second disaster similarity data set respectively to obtain a scoring result for each candidate disaster relief plan, sort each candidate disaster relief plan according to the scoring result, and select a preset number of candidate disaster relief plans as the final recommendation result.

[0099] A score prediction is performed for each candidate disaster relief plan in the disaster similarity data set (including the first disaster similarity data set and the second disaster similarity data set). The calculation can be based on the similarity value and other factors. For example, different weights can be given according to the size of the similarity value, or factors such as the execution effect of the historical plan can be considered. The predicted score of each plan is calculated by the corresponding score prediction formula (such as weighted average, machine learning model, etc.). The score predictions of all plans are combined to form a score prediction set. The score prediction set contains the predicted score value of each plan, as well as the corresponding historical disaster number, type number and other information. According to the score prediction set, a plan with a higher score is selected as the recommended result. A score threshold can be set to select a plan with a score greater than the score threshold as the recommended result or to select the first few plans with the highest score as the recommended result. In one embodiment, if the score of a plan in the score prediction set is significantly higher than that of other plans, it can be used as the recommended plan that best suits the current disaster, providing a reference for disaster relief decision-making.

[0100] In one embodiment, after obtaining the first disaster similarity data set and the second disaster similarity data set, the optimized overall process is as follows, wherein the first matching probability and the second matching probability are collectively referred to as matching probability, and the first disaster similarity data set and the second disaster similarity data set are collectively referred to as disaster similarity data sets.

[0101] Specifically, the method includes: sorting all candidate disaster plans in the disaster similarity data set according to the matching probability, and obtaining a neighbor plan data set based on the sorted candidate disaster plans; performing two join operator connection operations on the neighbor plan data set and the historical disaster data set to obtain a first set including the number of the target plan, the number of the neighbor plan, the disaster similarity, and the numbers and scoring values ​​of all scored plans in the disaster plan, and a second set including the numbers and scoring values ​​of all scored plans in the neighbor plan data set.

[0102] Perform conversion and mapping operations on the first set and the second set to obtain corresponding key-value pair type data sets; use a grouping operator to group the key values ​​in the key-value pair type data sets to obtain a read-only distributed data set; divide the read-only distributed data set into nodes of the cluster for execution, with each task processing a data shard, traverse the read-only distributed data set, obtain the plan numbers of all historical plans in the target plan that are similar to the recommended disaster plan, and obtain a score prediction set according to a score prediction formula.

[0103] Among them, the target plan is recommended by the plan recommendation model, the neighbor plan is the historical plan that is similar to the target plan, and the disaster plan is all the plans built into the system (custom plans and plans for historical disasters).

[0104] In one embodiment, the disaster similarity data set can be arranged in descending order according to the values ​​of the third column of the disaster similarity data set, and the first L nearest neighbor plans of the disaster plan are selected from the disaster similarity data set to obtain the nearest neighbor plan data set.

[0105] The neighbor plan dataset and the historical disaster dataset are connected by the join operator, which can integrate the information in the neighbor plan with the detailed information of the historical disaster. After the connection, the first set of information including the target plan number (the plan number currently being considered for recommendation), the neighbor plan number (the historical plan number similar to the target plan), the disaster similarity, and the plan numbers and scores of all the rated plans in the disaster plan is obtained. At the same time, the second set of information including the plan numbers and scores of all the rated plans in the neighbor plan dataset is also obtained. Specifically, it can be achieved by performing two join operator operations.

[0106] The first set and the second set are converted to adjust the format of the data to make it more suitable for subsequent processing. For example, the data is converted from one data structure to another structure that is more convenient for calculation. Then a mapping operation is performed to map the data into a key-value pair type data set. In the key-value pair, the key can be a specific plan number or other features, and the value can be the corresponding score, similarity and other information. This key-value pair format facilitates grouping and aggregation operations.

[0107] Use the grouping operator to group the key values ​​in the key-value pair type data set according to the key value. For example, if the key is the plan number, the key-value pairs with the same plan number will be grouped into the same group. After grouping, a read-only distributed data set is obtained, which can be efficiently processed in parallel in a distributed computing environment.

[0108] Divide the read-only distributed data set into the nodes of the cluster for execution. In a distributed computing environment, data can be divided into different computing nodes, and each task processes a data shard. In this way, the computing resources of the cluster can be fully utilized and the processing speed can be improved. Traverse the read-only distributed data set to obtain the plan numbers of the scores of all historical plans in the target plan that are similar to the recommended disaster plan. By analyzing and comprehensively considering the scores of historical plans, a basis can be provided for the score prediction of the target plan. According to the score prediction formula, combined with the scores, similarities and other information of the historical plans, the predicted score of the target plan is calculated. The predicted scores of all target plans are combined to obtain a score prediction set, which is used to provide a quantitative reference for the final plan recommendation. Among them, the candidate disaster relief plans in the first disaster similarity data set and the second disaster similarity data set are scored respectively by the score prediction formula; the score prediction formula can be:

[0109]

[0110] in, represents the predicted score of decision maker u for candidate disaster relief plan i, N(u) represents the set of decision makers similar to decision maker u, these decision makers have a common scoring history with decision maker u, and their scoring behavior is similar to that of decision maker u, sin(u,v) represents the similarity between decision makers u and decision makers v; r vi It represents the actual score given by decision maker v to candidate disaster relief plan i.

[0111] After obtaining the score prediction set, the score prediction set is divided into multiple task slices for processing, and a mapping operator is used to perform a mapping operation on each row of data in each task slice; the score prediction sets corresponding to all disaster plans are arranged in descending order, and a preset number of disaster plans are selected as the final recommendation results. Among them, the score prediction set is divided into multiple task slices, and the purpose of doing so is to be able to process data in parallel in a distributed computing environment and improve computing efficiency. For example, if there are a large number of disaster plans that need to be scored and predicted, they can be divided into multiple task slices and allocated to different computing nodes for simultaneous processing. The size of the task slice can be adjusted according to the computing resources and the amount of data to achieve the best performance balance. Then a mapping operation is performed on each row of data in each task slice. The mapping operator is a commonly used operation in distributed computing, which can perform specific transformations on each element in the data set. In the above scenario, the mapping operation may be to further process or transform the score of each disaster plan in the score prediction set, such as standardizing the score, extracting specific features, etc.

[0112] Finally, the disaster response plans are ranked according to their scores. Plans with higher scores usually mean higher similarity to the current disaster, or better historical execution results. Therefore, they are more likely to be recommended as the final disaster relief plan, which can provide an important reference for disaster relief decisions corresponding to the current disaster.

[0113] In one embodiment, Figure 6 As shown, step 102 specifically includes:

[0114] Step 1021: Obtain organizations in the surrounding area that can provide corresponding disaster relief forces based on the recommendation results, and display the geographical locations of the corresponding organizations on the map.

[0115] The recommendation results include custom plans applicable to current disasters and plans corresponding to historical disasters.

[0116] When the recommended results are obtained, the most suitable disaster relief plan for the current disaster is determined. Based on this plan, the type of disaster relief force required can be further analyzed, such as medical rescue, fire rescue, material transportation, etc. Then, according to the demand for disaster relief forces, search for organizations that can provide corresponding disaster relief forces in the surrounding areas. Organizations may include government departments (such as fire brigades, earthquake bureaus, health bureaus, etc.), non-governmental organizations (such as the Red Cross, charities, etc.), and enterprises (such as logistics companies, construction companies, etc.).

[0117] Through database query, network search, information sharing with relevant departments, etc., the information of the organization is obtained, including the name of the organization, contact information, type and scale of disaster relief forces that can be provided, etc.

[0118] Combine the acquired organizational information with the geographic information system to display the geographical location of the corresponding organization on the map. Electronic map software or a dedicated map platform can be used to display the location of the organization on the map in the form of icons, marks, etc. This can provide an intuitive understanding of the distribution of organizations that can provide disaster relief forces in the surrounding areas, providing a basis for subsequent action route planning and disaster relief coordination.

[0119] Step 1022: According to the geographical location of each organization and the road condition information of the current disaster, the shortest action route in terms of time and distance are planned for each organization respectively, and displayed on a map to generate a disaster relief plan corresponding to the current disaster.

[0120] According to the geographical location of each organization and the road condition information of the current disaster, the shortest action route and the shortest distance action route are planned for each organization. Road condition information can be obtained through the real-time monitoring system of the transportation department, satellite navigation system, social media and other channels, including road congestion, road damage, traffic control, etc. For the shortest action route, factors such as road speed, traffic flow, and signal light waiting time can be considered, and path planning algorithms (such as Dijkstra algorithm, A* algorithm, etc.) can be used to calculate the fastest path from the organization's location to the disaster site. For the shortest distance action route, the geographical distance is mainly considered, and the shortest path algorithm is used to calculate the shortest distance path from the organization's location to the disaster site.

[0121] The shortest action routes and the shortest action routes of each organization are displayed on the map. Different colors and line styles can be used to distinguish different types of routes for the convenience of decision makers and rescuers. The key nodes of the route can also be marked on the map, such as road intersections, bridges, tunnels, etc., as well as factors that may affect the action, such as gas stations, rest areas, and relief material supply points.

[0122] Combined with the organizations and their geographical locations determined in step 1021, and the action routes planned in step 1022, a disaster relief plan corresponding to the current disaster is generated. The disaster relief plan may include the following: contact information of each organization and the disaster relief forces that can be provided; detailed information on the action route with the shortest time and the action route with the shortest distance; key nodes and precautions along the way; timetable and task allocation of disaster relief operations, etc. The obtained disaster relief plan can serve as an important basis for guiding rescue operations, helping various organizations to collaborate efficiently, improve disaster relief efficiency, and minimize disaster losses.

[0123] In summary, this embodiment combines the plan recommendation algorithm to quickly and accurately identify the characteristics of different types of disasters, and recommends and generates personalized disaster relief plans based on the disaster characteristics, which can significantly improve the efficiency and effectiveness of disaster response and reduce the economic losses and casualties caused by disasters.

[0124] Embodiment 2:

[0125] In Example 1, a method for generating a disaster relief plan based on disaster characteristics is provided. In this embodiment, a device for generating a disaster relief plan based on disaster characteristics is proposed. The device for generating a disaster relief plan based on disaster characteristics includes: a processor and a memory for storing processor executable instructions; wherein the processor is configured to execute the method for generating a disaster relief plan based on disaster characteristics described in Example 1.

[0126] like Figure 7 As shown, the disaster relief plan generation device based on disaster characteristics includes a processor 21 and a memory 22, wherein the processor 21 and the memory 22 can be connected via a bus or other means.

[0127] The processor 21 may be a central processing unit (CPU). The processor 21 may also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, or a combination of the above chips.

[0128] The memory 22, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs, non-transitory computer executable programs and modules, such as program instructions / modules corresponding to the disaster relief plan generation method based on disaster characteristics in Embodiment 1 of the present invention. The processor executes various functional applications and training processes of the processor by running the non-transitory software programs, instructions and modules stored in the memory.

[0129] The memory 22 may include a program storage area and a training storage area, wherein the program storage area may store an operating system and application programs required for at least one function; the training storage area may store training created by the processor, etc. In addition, the memory may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some embodiments, the memory 22 may optionally include a memory remotely located relative to the processor, and these remote memories may be connected to the processor via a network. Examples of the above-mentioned networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof. The one or more modules are stored in the memory 22, and when executed by the processor 21, execution such as Figure 1 The method for generating a disaster relief plan based on disaster characteristics in Example 1 is shown. The specific details of the above-mentioned method for generating a disaster relief plan based on disaster characteristics can be found in Figure 1 , Figure 2 and Figure 3 The corresponding descriptions and effects in the embodiments shown in the figure can be understood and will not be repeated here.

[0130] This embodiment also provides a computer storage medium, which stores a computer program. The computer program can be executed by a processor to complete the method for generating a disaster relief plan based on disaster characteristics described in Example 1.

[0131] The computer storage medium stores computer executable instructions, and the computer executable instructions can execute the disaster relief plan generation method based on disaster characteristics in any of the above method embodiments. Among them, the storage medium can be a disk, an optical disk, a read-only memory (ROM), a random access memory (RAM), a flash memory, a hard disk (HDD), or a solid-state drive (SSD); the storage medium can also include a combination of the above types of memory.

[0132] The specific steps of the method for generating a disaster relief plan based on disaster characteristics are shown in Example 1 and will not be repeated in this embodiment.

[0133] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for generating a disaster relief plan based on disaster characteristics, characterized in that: The disaster relief plan generation method comprises: Build a plan recommendation model; According to the plan recommendation model, the current disaster information is matched with the custom plan and the historical disaster plan respectively to obtain a first matching probability and a second matching probability respectively; Obtaining a recommendation result according to the first matching probability and the second matching probability; A disaster relief plan corresponding to the current disaster is generated according to the recommendation result.

2. The method for generating a disaster relief plan based on disaster characteristics according to claim 1, characterized in that: The matching of the current disaster information with the custom plan and the historical disaster plan according to the plan recommendation model to obtain a first matching probability and a second matching probability respectively includes: Obtain disaster factor matrix based on current disaster information; Calculating a first cosine similarity and a first divergence correction factor between the disaster factor matrix and the custom plan, and correcting the first cosine similarity according to the first divergence correction factor to obtain the first matching probability; The second cosine similarity and the second divergence correction factor between the disaster element matrix and the historical disaster emergency plan are calculated, and the second cosine similarity is corrected according to the second divergence correction factor to obtain the second matching probability.

3. The method for generating a disaster relief plan based on disaster characteristics according to claim 2, characterized in that: The first cosine similarity and the second cosine similarity are calculated in the same way, specifically as follows: A penalty factor is added to the original cosine similarity to obtain an improved cosine similarity, so as to obtain the first cosine similarity and the second cosine similarity respectively, wherein the penalty factor is calculated as follows: Among them, pe i is the penalty factor, tol represents the total number of disasters, c i is the number of disaster type i; the calculation formula of the improved cosine similarity is: Among them, Sim(u,v) represents the similarity between decision makers u and v, I uν is the set of historical plans scored by decision makers u and v, R u,i is the score of decision maker u on historical plan i, R v,i is the score given by decision maker v to historical plan i, is the average score of all historical plans given by decision maker u, is the average score of all historical plans given by decision maker v, I u represents the set of historical plans rated by decision maker u, I v Represents the set of historical plans rated by decision maker v.

4. The method for generating a disaster relief plan based on disaster characteristics according to claim 1, characterized in that: The methods for obtaining the historical disaster emergency plan include: Reading a historical disaster dataset from a distributed file system, and converting the historical disaster dataset into a first elastic distributed dataset; Filtering out historical disaster numbers, type numbers, and emergency plan numbers from the first elastic distributed data set, and obtaining a second elastic distributed data set according to the historical disaster numbers, the type numbers, and the emergency plan numbers; The second elastic distributed data set is converted into a structured data set to obtain the historical disaster emergency plan.

5. The method for generating a disaster relief plan based on disaster characteristics according to claim 1, characterized in that: The obtaining of a recommendation result according to the first matching probability and the second matching probability includes: Acquire a candidate disaster relief plan corresponding to each first matching probability according to the custom plan, and establish a first disaster similarity data set according to the first matching probability, the corresponding candidate disaster relief plan and the current disaster information; Acquire a candidate disaster relief plan corresponding to each second matching probability according to the historical disaster plan, and establish a second disaster similarity data set according to the second matching probability, the corresponding candidate disaster relief plan and the current disaster information; Score each candidate disaster relief plan in the first disaster similarity data set and the second disaster similarity data set respectively to obtain a scoring result for each candidate disaster relief plan, sort each candidate disaster relief plan according to the scoring result, and select a preset number of candidate disaster relief plans as the final recommendation result.

6. The method for generating a disaster relief plan based on disaster characteristics according to claim 5, characterized in that: The scoring of each candidate disaster relief plan in the first disaster similarity dataset and the second disaster similarity dataset comprises: Scoring the candidate disaster relief plans in the first disaster similarity dataset and the second disaster similarity dataset respectively by using a scoring prediction formula; The rating prediction formula is: in, represents the predicted score of decision maker u for candidate disaster relief plan i, N(u) represents the set of decision makers similar to decision maker u, these decision makers have a common scoring history with decision maker u, and their scoring behavior is similar to that of decision maker u, sin(u,v) represents the similarity between decision makers u and decision makers v; r vi It represents the actual score given by decision maker v to candidate disaster relief plan i.

7. The method for generating a disaster relief plan based on disaster characteristics according to claim 1, characterized in that: Generating a disaster relief plan corresponding to the current disaster according to the recommendation result includes: According to the recommendation results, organizations in the surrounding areas that can provide corresponding disaster relief forces are obtained, and the geographical locations of the corresponding organizations are displayed on the map; According to the geographical location of each organization and the road condition information of the current disaster, the shortest action route in time and distance are planned for each organization and displayed on the map to generate a disaster relief plan corresponding to the current disaster.

8. The method for generating a disaster relief plan based on disaster characteristics according to claim 1, characterized in that: The current disaster information includes the latitude and longitude information of the current disaster, casualty information, economic loss information, and data information specific to different disasters.

9. A disaster relief plan generation device based on disaster characteristics, characterized in that: The disaster relief plan generation device based on disaster characteristics includes: a processor and a memory for storing instructions executable by the processor; The processor is configured to execute the method for generating a disaster relief plan based on disaster characteristics as described in any one of claims 1 to 8.

10. A non-volatile computer storage medium, characterized in that: The computer storage medium stores computer executable instructions, and the computer executable instructions are executed by one or more processors to complete the method for generating a disaster relief plan based on disaster characteristics as described in any one of claims 1-8.

Images