Emergency command and dispatch method based on event driving

Abstract

The invention discloses an event-driven emergency command and dispatch method, and relates to the field of emergency command and dispatch, and the method comprises the steps: carrying out the feature extraction of historical emergency event data through feature engineering; wherein the feature data set comprises an event type feature vector, an event frequency feature vector, a distribution density feature vector and a text feature vector; training an emergency event classification model based on a CART decision tree by using the feature data set; acquiring real-time emergency event data, and performing feature extraction on the real-time emergency event data to obtain a real-time event feature vector; using the obtained real-time event feature vector as an input, and using an emergency event classification model to identify an emergency event; and according to the identified emergency event, an emergency command and dispatch scheme is retrieved and generated from a preset emergency plan knowledge base by adopting a rule-based case-based reasoning method. In order to solve the problem of low individuation degree of the emergency command and dispatch scheme in the prior art, the individualized emergency command and dispatch scheme is retrieved and generated.

Description

Technical Field

[0001] The present application relates to the field of emergency command and dispatch, and in particular to an event-driven emergency command and dispatch method. Background Art

[0002] With the rapid development of social economy and the increasing frequency of human activities, various emergencies and natural disasters have shown complex characteristics such as multiple, frequent and coupled. The effective response and disposal of emergency events are facing increasing challenges. Emergency command and dispatch is the core link of emergency management, and the scientificity and effectiveness of its plan directly affect the effect of emergency disposal. However, due to the complexity and variability of emergency events, traditional emergency command and dispatch methods are difficult to adapt to the growing personalized needs. Therefore, it is of great significance to study new intelligent and precise emergency command and dispatch methods to improve the personalization level and real-time response capability of emergency plans in order to improve the level of emergency management of emergencies.

[0003] In recent years, the development of emerging information technologies such as artificial intelligence and big data has provided new ways for the intelligent transformation of emergency command. Introducing artificial intelligence technologies such as machine learning and knowledge engineering into the field of emergency command and building intelligent decision-making models driven by data and knowledge fusion are research hotspots and development trends in the field of emergency management. However, relevant research currently focuses on emergency resource scheduling optimization, rescue path planning, etc., and research on real-time generation and personalized construction of emergency plans is relatively weak.

[0004] Existing emergency command and dispatch methods mainly include matching methods based on emergency plans and manual formulation methods based on expert experience. The plan matching method establishes an emergency plan library, retrieves the plan library according to the type and characteristics of the emergency event, and provides a reference dispatch plan. However, most of the plans in the plan library are general plans designed for typical scenarios, which are difficult to accurately match complex and changeable actual emergency events, and the plans are not targeted and personalized enough. The manual formulation method based on expert experience relies on the experience and ability of the commanders, and has problems such as strong subjectivity, low formulation efficiency, and difficulty in responding to new situations. The timeliness and reliability of the plan are difficult to guarantee. Summary of the invention

[0005] In view of the problem of low personalization of emergency command and dispatch schemes in the prior art, the present application provides an event-driven emergency command and dispatch method, which retrieves and generates personalized emergency command and dispatch schemes through ontology-based and case-based reasoning.

[0006] The purpose of this application is achieved through the following technical solutions.

[0007] The present application provides an event-driven emergency command and dispatch method, comprising: obtaining historical emergency event data, the historical emergency event data including event type, event time, event location and disposal plan; using feature engineering to extract features from the historical emergency event data to obtain a feature data set; wherein the feature data set includes an event type feature vector, an event frequency feature vector, a distribution density feature vector and a text feature vector; using the feature data set to train an emergency event classification model based on a CART decision tree; obtaining real-time emergency event data, and extracting features from the real-time emergency event data to obtain a real-time event feature vector; using the obtained real-time event feature vector as input, and using the emergency event classification model to identify emergency events; and according to the identified emergency events, using a rule-based case reasoning method to retrieve and generate an emergency command and dispatch plan from a preset emergency plan knowledge base.

[0008] Among them, CART is the abbreviation of "Classification and Regression Tree", which refers to classification and regression tree. In this application, a feature data set is used to train a CART decision tree model to obtain an emergency event classification model. This classifier can be used to determine which type a new emergency event belongs to. Case-based Reasoning (CBR) is an analogy reasoning method. By retrieving existing cases similar to new problems, solutions to historical cases are obtained, and then they are adjusted or modified to adapt to new problems. In this application, according to the event classification results, a rule-based CBR method is used to find relevant emergency plans in the case library (emergency plan knowledge base), and a specific emergency dispatch plan is generated based on this. In this application, historical emergency event data: used for model training, that is, to generate an emergency event classifier. Historical data includes event type, time of occurrence, location, disposal plan taken, etc. Real-time emergency event data: refers to information on sudden emergency events, which is the input of the model. Feature extraction is performed on real-time data, and then it is identified with a trained classifier. The recognition results will be used for subsequent plan retrieval.

[0009] Furthermore, feature engineering is used to extract features from historical emergency event data, including: using One-Hot coding with the same number of binary bits as the number of event types to encode event types and obtain event type feature vectors; using the sliding time window method to perform time window statistics on event times and obtain event frequency feature vectors at different time scales; dividing the event locations into geographic grids, and using the kernel density estimation method to estimate the number of events in each grid and obtain the distribution density feature vectors of events in different regions; using the TF-IDF algorithm to extract features from the text description of the disposal plan and obtain the text feature vector of the disposal plan; and concatenating the event type feature vector, event frequency feature vector, distribution density feature vector and text feature vector through vectors to obtain a feature data set.

[0010] Among them, event type refers to the type of emergency or sudden incident. Different emergency events have different natures, scales, impacts, and handling methods, so they need to be classified. The number of event types refers to the total number of emergency event types. This requires a comprehensive sorting and summarization of various types of events based on the actual needs of emergency management. Events can usually be classified from the perspectives of time (such as forest fires often occur in summer), space (such as floods often occur in river basins), and disaster types (such as geological disasters, accidents and disasters). The handling plan refers to a series of measures and actions taken to deal with emergency events in order to control the situation, eliminate hidden dangers, and reduce losses. In the emergency plan, each emergency event will have a corresponding handling plan. The handling plan is usually described in text form, including specific arrangements for pre-event prevention, incident response, in-event handling, and post-event recovery. The content involves organization and command, team assembly, information collection, on-site investigation, force dispatch, personnel evacuation, material support, post-disaster reconstruction, summary and evaluation, etc.

[0011] Furthermore, the feature data set is used to train an emergency event classification model based on a CART decision tree, including: using the feature data set as a training set, using the Gini index to select the optimal partitioning feature, and generating a CART decision tree in a recursive manner; wherein the feature data set is used as the input of the CART decision tree, and the event type is used as the output of the CART decision tree; using a K-fold cross-validation method, the feature data set is divided into K mutually exclusive subsets, one subset is selected each time as a validation set, and the remaining K-1 subsets are used as training sets; using the training set to train the CART decision tree, and evaluating the classification performance of the CART decision tree on the validation set; repeating the cross-validation process K times to obtain an emergency event classification model.

[0012] Preferably, the SMOTE (Synthetic Minority Over-sampling Technique) algorithm is used to oversample event types with a small number of samples to balance the quantity distribution of different types of events; the balanced feature data set is randomly divided into a training set, a validation set, and a test set; the root node is initialized with all the samples in the training set and regarded as a node to be divided; for each node to be divided, the feature vectors of the samples contained therein are extracted to obtain each feature dimension and its possible values; for each possible value of each feature dimension, it is used as a partitioning threshold, and the samples of the current node are tried to be divided into two subsets above and below the threshold, and the Gini index after partition is calculated; the feature dimension with the smallest Gini index and its corresponding threshold are selected as the optimal partitioning basis, and the current node sample set is divided into two child nodes based on this; the newly generated child node is added to the list of nodes to be divided; the execution is repeated until all nodes meet the preset stopping conditions; and finally the first-layer CART decision tree is obtained; the first-layer CART decision tree is used to predict each sample in the training set, and the residual between the predicted value and the true event type label is calculated; the feature vector of the training set sample is used as input, and the residual is used as the new target value, construct a training sample set for residual learning; use the residual learning training sample set to repeat the steps to obtain the second layer of CART decision tree; repeat each iteration to generate a layer of CART decision tree until the preset maximum number of layers is reached to obtain a multi-layer CART decision tree; for each sample in the validation set, use the CART decision trees of all layers to predict and obtain the prediction probability vector of each layer of decision trees; concatenate the prediction probability vectors of each sample on each layer of decision trees to generate a new Stacking feature vector; use the Stacking feature vector as input and the event type as the target value to train a Logistic regression model as the Stacking metamodel; use the trained multi-layer CART decision tree and Stacking metamodel to classify and predict new emergency event data: extract the feature vector of the new event data; input the feature vector into each layer of CART decision tree to obtain the prediction probability vector of each layer; concatenate the prediction probability vectors of all layers into a Stacking feature vector, input it into the Stacking metamodel, and obtain the final event type prediction label.

[0013] Furthermore, an emergency command and dispatch plan is generated, including: using an ontology-based knowledge representation method to structurely represent and store emergency plans in the form of an ontology, and construct an emergency plan knowledge base; according to the identified emergency event, the emergency plan knowledge base is retrieved through the ontology reasoning mechanism to obtain multiple emergency plans corresponding to the emergency event; the semantic similarity between the identified emergency event and the obtained multiple emergency plans is calculated, and one or more emergency plans with a semantic similarity greater than a preset threshold are selected as matching results; the difference features between the identified emergency event and the matching results are extracted, and a heuristic search algorithm is used to generate a personalized emergency sub-plan for the identified emergency event; the matched one or more emergency plans are integrated with the generated personalized emergency sub-plan to obtain an emergency command and dispatch plan.

[0014] Among them, the emergency plan knowledge base is a knowledge base system for storing emergency plans. It is different from traditional relational databases. Instead, it adopts knowledge representation and knowledge organization methods, based on semantic web technology, to structure and represent and store concepts, relationships and rules in emergency plans in the form of ontology. Ontology reasoning refers to the process of deriving implicit knowledge, answering questions or solving problems by using the concept hierarchy, relationship network, constraint rules, etc. in the ontology knowledge base. By describing the semantic association of domain concepts, the ontology gives knowledge clear semantics, so that computers can perform logical reasoning based on semantics. In this application, when a new emergency event is identified, it is necessary to retrieve relevant plans from the emergency plan knowledge base. This retrieval is not a simple keyword matching, but a semantic retrieval. Through ontology reasoning, plans that are conceptually semantically related to the input event can be found. For example, the failure of offset wing aircraft often occurs on the wing. Common ontology reasoning mechanisms include: rule-based reasoning (using rules in the form of if-then), reasoning based on description logic (using formal definitions of concepts and relationships), etc. The reasoning process may involve traversal of the ontology hierarchy, search of the relationship network, matching of the rule base, etc. The inference results can be in the form of recognized concept instances, triple relations, logical predicates, etc.

[0015] In this application, the difference feature refers to the difference feature between the identified emergency event and the retrieved standard emergency plan. The emergency plan knowledge base stores some standardized and templated plans, but the actual emergency events often have their own particularities. For example, the time, location, and magnitude of an earthquake, the climate, terrain, and wind direction of a wildfire, etc., will all be different from the standard plan.

[0016] Furthermore, multiple emergency plans corresponding to emergency events are obtained, including: using an ontology construction method to represent the identified emergency events ontologies to obtain an event ontology; extracting key concepts in the event ontology to form an event query ontology; wherein the key concepts include event type, event time and event location; according to the event query ontology, using an ontology reasoning mechanism based on description logic, retrieving an emergency plan ontology containing all key concepts in the event query ontology from a constructed emergency plan knowledge base as a candidate emergency plan ontology set; wherein the emergency plan ontology includes plan types and disposal plans; for each emergency plan ontology in the candidate emergency plan ontology set, extracting the corresponding plan type and disposal plan to form a key concept set of the emergency plan ontology; calculating the semantic similarity between the key concepts in the event query ontology and the key concept set of the emergency plan ontology to obtain the similarity of each candidate emergency plan ontology; taking the candidate emergency plan ontology with a similarity greater than a preset threshold as a matched emergency plan ontology; obtaining the emergency plan corresponding to the matched emergency plan ontology as the emergency plan corresponding to the identified emergency event.

[0017] Furthermore, the similarity of each candidate emergency plan ontology is obtained, including: obtaining the key concepts in the event query ontology ; Get each key concept in the key concept set of the emergency plan ontology ; Get key concepts from the WordNet knowledge base Synonyms of ; Get key concepts from the WordNet knowledge base Synonyms of ;calculate and The semantic similarity of : ,in, express and The synset of the lowest common ancestor concept in WordNet; Represents the shortest path length from the synonym set synset to the WordNet root node; calculates key concepts All key concepts in the key concept set of the emergency plan ontology The maximum semantic similarity of , as a key concept The similarity component is calculated as: , where O represents the key concept set of the emergency plan ontology; all key concepts in the event query ontology are calculated The arithmetic mean of the similarity components is taken as the similarity between the event query ontology and the current candidate emergency plan ontology: ,in, Represents the event query ontology; represents the candidate emergency plan ontology; n is the number of key concepts in the event query ontology.

[0018] Furthermore, one or more emergency plans whose semantic similarity is greater than a preset threshold are selected as matching results, including: obtaining the event type of the identified emergency event, and converting the event type into an event type vector through One-hot encoding; performing word segmentation and part-of-speech tagging on the obtained multiple emergency plans, extracting nouns and verbs as keywords, and extracting the title of the emergency plan; concatenating the extracted keywords and titles as an emergency plan vector; mapping each element in the event type vector to a word embedding vector corresponding to the event type, and obtaining the event type embedding vector by weighted averaging the word embedding vectors corresponding to the non-zero elements; wherein the weight is the frequency of occurrence of each event type in historical data; mapping each word in the emergency plan vector to a corresponding word embedding vector, and obtaining the emergency plan embedding vector by weighted averaging all word embedding vectors; wherein the weight is the TF-IDF value of each word; using cosine similarity to calculate the similarity between the event type embedding vector and the emergency plan embedding vector to obtain a similarity list; according to the preset threshold, obtaining one or more emergency plans from the similarity list as matching results.

[0019] Furthermore, a heuristic search algorithm is used to generate personalized emergency sub-plans for the identified emergency events, including: performing attribute decomposition on the identified emergency events, extracting the event type, event time and location of the events, and forming an event attribute vector; performing attribute decomposition on one or more matched emergency plans, extracting the event type, event time and location of the plans, and forming a plan attribute vector; calculating the weighted Euclidean distance between the event attribute vector and each plan attribute vector as the degree of difference, and generating a difference attribute set; based on the difference attribute set, retrieving plan fragments whose semantic similarity with the difference attribute is greater than a threshold from the emergency plan knowledge base to form a plan fragment library; wherein the plan fragment represents a step or measure in the emergency plan; and using the event attribute vector in the difference attribute set as the initial state of the search. , take the corresponding plan attribute vector in the difference attribute set as the target state of the search ; In the plan fragment library, search from the initial state To the target state The optimal path is used as a personalized emergency sub-plan. Among them, a plan fragment refers to an independent operation step or disposal measure in an emergency plan, which is the basic unit of a complete plan. An emergency plan is usually composed of multiple plan fragments, and there is a certain logical relationship and execution order between the fragments.

[0020] Furthermore, the search starts from the initial state To the target state The optimal path is used as a personalized emergency sub-plan, including: for each plan fragment in the plan fragment library , calculate from the initial state arrive The cost : ,in, From the initial state arrive the actual cost of For To the target state The estimated cost of Value Scenario Snippet , mark it as visited, and Updated to The corresponding plan segment attribute vector , denoted as ;by is the new initial state, repeat the above marking steps until the target state is reached Or preset termination conditions; if the target state is reached , then the sequence of plan fragments accessed during the search process As the optimal path, extract the plan fragment content corresponding to the corresponding path, and combine them in sequence to form a personalized emergency sub-plan; if the preset termination condition is reached, the personalized emergency sub-plan is empty.

[0021] Furthermore, the actual cost , calculated by the following formula: ,in, and Represents the initial state event attribute vector And the plan fragment The attribute vector The kth component of ; is the weight of the kth component, reflecting the importance of the component; is the value range of the kth component, used for normalization; estimated cost , calculated by the following formula: ,in, and Represent the attribute vectors of the plan fragments and the target state plan attribute vector The kth component of ; is the weight of the kth component.

[0022] Compared with the prior art, the advantages of this application are:

[0023] The ontology knowledge representation method is used to structure the emergency plan, and a semantically rich and structured emergency plan knowledge base is constructed. By calculating the semantic similarity between real-time emergency events and the plans in the plan knowledge base, the most relevant emergency plans are automatically retrieved and matched, and a personalized reference plan for event handling is given.

[0024] By analyzing the difference characteristics of real-time events and matching plans, combined with heuristic search algorithms, targeted personalized plan fragments are searched from the knowledge base to automatically generate event-driven personalized emergency sub-plans. This method makes full use of the plan fragments in the knowledge base, and generates plans through the assembly of fragments driven by difference characteristics, so that the generated scheduling plan can accurately match the personalized needs of real-time events.

[0025] When generating personalized emergency sub-plans, by defining the cost function in the heuristic search algorithm, mechanisms such as event attribute correlation weights and plan fragment attribute value normalization are introduced, so that the search process can comprehensively weigh the matching degree between event characteristics and plan characteristics, search for the optimal personalized plan combination, and improve the applicability and effectiveness of the emergency dispatch plan.

[0026] The emergency plans retrieved from the knowledge base are integrated with the personalized sub-plans generated by reasoning to form a complete event-driven emergency command and dispatch plan. The integrated plan takes into account both the general plans in the plan knowledge base and the personalized plans generated specifically, and can meet both the general and specific needs of emergency event handling. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] The present application will be further described in the form of exemplary embodiments, which will be described in detail by the accompanying drawings. These embodiments are not restrictive, and in these embodiments, the same number represents the same structure, wherein:

[0028] Figure 1 is an exemplary flow chart of an event-driven emergency command and dispatch method according to some embodiments of the present application;

[0029] Figure 2 is an exemplary flow chart of generating a feature data set according to some embodiments of the present application;

[0030] Figure 3 It is an exemplary flowchart for generating an emergency command and dispatch plan according to some embodiments of the present application. DETAILED DESCRIPTION

[0031] The method and system provided in the embodiments of the present application are described in detail below with reference to the accompanying drawings.

[0032] like Figure 1 As shown, historical emergency event data is obtained, and the historical emergency event data includes event type, event time, event location and disposal plan; feature engineering is used to extract features from the historical emergency event data to obtain a feature data set; wherein the feature data set includes an event type feature vector, an event frequency feature vector, a distribution density feature vector and a text feature vector; the feature data set is used to train an emergency event classification model based on a CART decision tree; real-time emergency event data is obtained, and features are extracted from the real-time emergency event data to obtain a real-time event feature vector; the obtained real-time event feature vector is used as input, and the emergency event classification model is used to identify the emergency event; according to the identified emergency event, a rule-based case reasoning method is used to retrieve and generate an emergency command and dispatch plan from a preset emergency plan knowledge base.

[0033] Specifically, historical emergency event data is obtained from public databases. The historical emergency event data includes event type, event time, event location and disposal plan. For the selected data source, data collection tools such as crawlers can be used to automatically extract the required structured event information from unstructured or semi-structured data such as web pages or files. For example, for the accident notification news page published on the official website of the emergency management department, targeted web page parsing rules are written to extract key fields such as the time, location, accident type, process description, and disposal of the accident. For the existing structured accident database, historical event records can be directly obtained in batches through database query statements. The obtained data is preprocessed, and the preprocessed historical emergency event data is imported into relational databases such as MySQL, PostgreSQL, etc. for centralized storage and management to establish an emergency event history database. The database table structure design must cover but is not limited to the following fields: Event ID: a serial number that uniquely identifies an emergency event. Event type: such as fire, traffic accident, geological disaster, etc., which can be represented by numerical codes. Event time: the time of the incident, with a unified format of year-month-day hour: minute: second. Incident location: Detailed description of the incident address, such as No. XX, XX Road, XX District, XX City, XX Province. Disposal plan: A text field that records the specific disposal measures during the incident disposal process.

[0034] like Figure 2As shown in the figure, feature engineering is used to extract features from historical emergency event data to obtain a feature data set; wherein the feature data set includes an event type feature vector, an event frequency feature vector, a distribution density feature vector and a text feature vector. Specifically, event type feature extraction uses the One-Hot encoding method to extract features for event type attributes in historical emergency event data. First, enumerate the event types to obtain an event type set, such as {fire, traffic accident, geological disaster}. Then, the number of binary bits of the One-Hot encoding is determined according to the size of the event type set, and the number of bits is the same as the number of types. For each emergency event, the binary position corresponding to its event type is set to 1, and the remaining positions are set to 0 to obtain a feature vector representation of the event type. For example, assuming that the event type set is {fire, traffic accident, geological disaster}, the feature vector of the "fire" event is [1, 0, 0], and the feature vector of the "traffic accident" event is [0, 1, 0]. Event frequency feature extraction uses the sliding time window method to extract frequency features for event time attributes in historical emergency event data. First, set different time window scales according to application requirements, such as hours, days, weeks, months, etc. Then, we use the sliding window method to count the number of events in each time window and get the event frequency at different time scales. For example, we set a 24-hour sliding window, count the number of events within every 24 hours, and get the event frequency vector at the hourly scale. Then, we set a 7-day sliding window, count the number of events within every 7 days, and get the event frequency vector at the daily scale. By concatenating the frequency vectors at multiple time scales, we get the event frequency feature vector.

[0035] Specifically, event distribution density feature extraction, input: historical emergency event dataset D, each event data contains attributes such as event ID, event location (longitude and latitude coordinates). According to the given target area geographic range R and grid granularity parameter g, the area R is divided into m×n grid cells of equal size. Each grid cell is represented as G(i, j), where i is the row number, j is the column number, i∈[1, m], j∈[1, n]. Let the grid cell set be G. For each historical event e in the dataset D, calculate the grid cell number (i, j) where the event is located according to its location coordinates (longitude lng, latitude lat). Add the event ID to the event list L(i, j) of the corresponding grid. After traversing all events, get the event list corresponding to each grid cell. For each grid cell G(i, j) in the grid set G, with its center coordinates As the core, the kernel density estimation method is used to estimate the event density of the grid. Set the kernel function to K. Common kernel functions include Gaussian kernel, Epanechnikov kernel, etc. Here, Gaussian kernel is taken as an example. The formula is: ; Set the bandwidth parameter of the kernel function to h. According to the rule of thumb, we can set , where σ is the standard deviation of the event coordinates and n is the total number of events. Then, for the grid For each event e' in the grid, calculate its distance to the center of the grid : ; Calculate the event based on the distance value d and bandwidth h To Grid Contribution coefficient of density estimation : ; Finally, the grid Sum the contribution coefficients of all events in the grid to get the estimated event density of the grid. : Traverse all grid cells in the grid set G, obtain the event density estimate of each grid in turn, and construct the event distribution density feature vector : ,in, The length of is m×n, indicating the total number of grid cells. It depicts the spatial distribution pattern of emergency events in the entire target area, which helps to explore the regional characteristics and hot spots of the events.

[0036] Specifically, extract the text feature vector of the treatment plan, extract the text description of the treatment plan of all events in the data set D, and form the original corpus C. Perform text preprocessing on corpus C; segment each text description according to semantic content words to obtain a word segmentation sequence. You can use Chinese word segmentation tools such as Jieba and THULAC. After preprocessing, you will get a standardized corpus According to the corpus , count the frequency of occurrence of each word w in the corpus, and get the word frequency TF(w): TF(w) = (number of occurrences of w in the corpus) / (total number of words in the corpus); count the number of texts containing word w, and get the inverse document frequency IDF(w): IDF(w) = log((total number of texts in the corpus + α) / (number of texts containing w + α)), where α is a smoothing parameter, usually 1. Calculate the TF-IDF weight of each word w: ; Select the N words with the highest TF-IDF weight to form the feature dictionary W of the treatment plan text. For each event e in the data set D, extract its treatment plan text description t(e), perform preprocessing such as word segmentation and stop word removal to obtain a normalized word sequence. Then, according to the feature dictionary W, count each feature word in the word sequence TF-IDF weights to construct text feature vectors : ; in, is the i-th feature word in the dictionary W, n is the dictionary size N, for TF-IDF weight in text t(e). Traverse all events in dataset D and obtain the text feature vector of each event handling solution. The vector of text features of historical emergency event handling solutions It characterizes the semantic information and key measures of event handling solutions, which helps to explore the similarities and differences in handling measures for different events.

[0037] The feature data set is used to train the emergency event classification model based on the CART decision tree; specifically, the feature data set D is divided into a feature matrix X and a label vector y. Each row of X represents a feature vector of an event, including event type features, frequency features, density features, and text features; each element of y represents the type label of the corresponding event, such as "fire", "traffic accident", etc. The CART (Classification and Regression Tree) algorithm is used to train and generate a decision tree based on the feature matrix X and the label vector y. Calculate the Gini index Gini (D) of the current node, and select the optimal partition feature a and the optimal partition point v based on the feature matrix X, so that the sum of the Gini indexes of the divided sub-nodes is minimized. The calculation formula of the Gini index is: , ; Where D is the sample set of the current node, is the sample subset of category k, K is the total number of categories. According to the optimal partition feature a and partition point v, the sample set D of the current node is divided into the sample sets of the left and right child nodes and The sub-nodes are divided recursively until one of the following stopping conditions is met: the sample set of the current node belongs to the same category; the number of samples of the current node is less than the preset threshold ; The depth of the current node reaches the preset maximum depth ; No better partitioning features can be found. For each leaf node, the node is marked as the corresponding category according to the majority category label in its sample set. The K-fold cross-validation method is used to evaluate the generalization performance of the decision tree model. Divide the feature data set D into K mutually exclusive subsets Perform K cross validations, selecting a subset each time As the validation set, the remaining K-1 subsets constitute the training set Using the training set Training and generating decision tree models , in the validation set Evaluate the model The classification performance of the model is evaluated. Common evaluation indicators include accuracy, precision, recall, and F1 value. Repeat K times to obtain K decision tree models. And the corresponding performance evaluation results, take the average of K performance evaluation results, and get the comprehensive performance evaluation result of the decision tree model. According to the evaluation results of cross-validation, select the decision tree model with the best performance As the final emergency event classification model.

[0038] Get the latest emergency event report data. The data content includes the time, location, and preliminary description of the event. Parse the acquired real-time event data and extract key field information. According to the data format of different data sources, use regular expressions, XML parsing, JSON parsing and other methods to extract structured event attributes such as event type, time, and location from unstructured or semi-structured data. According to the keywords in the event description, use the pre-trained event type classification model to identify the possible types of events and generate event type features. According to the event time, update the event frequency statistical features within the time window. According to the coordinates of the event location, locate the geographic grid where the event is located, update the event density estimate of the grid, and generate distribution density features. Perform pre-processing such as word segmentation and stop word removal on the text description of the event, and use the pre-trained TF-IDF model to extract text feature vectors. Concatenate the feature vectors of event type, frequency, density, text, etc. to obtain the comprehensive feature vector of the real-time event as the input for subsequent event classification. Use stream processing frameworks such as Apache SparkStreaming and Flink to organize the steps of real-time event data access, parsing, and feature extraction into a real-time processing pipeline. S5, taking the obtained real-time event feature vector as input, and using the emergency event classification model to identify emergency events.

[0039] like Figure 3As shown in the figure, according to the identified emergency events, a rule-based case reasoning method is used to retrieve and generate emergency command and dispatch plans from the preset emergency plan knowledge base; the emergency plan ontology is constructed, and the emergency plan text is collected from emergency management departments, public databases and other channels, including the type of plan, applicable conditions, organizational structure, disposal process, etc. The emergency plan text is processed by natural language, and named entity recognition, keyword extraction and other technologies are used to identify key concepts in the plan, such as "earthquake", "rescue", "evacuation", etc. The semantic relationship between concepts is analyzed, and the attributes and relationships of the ontology are defined, such as "is-a", "belongs to" (part-of), etc. The concept classification hierarchy and relationship network are constructed. The ontology description language such as OWL (Web Ontology Language) is used to represent concepts, attributes, relationships, etc. in the form of formalized triples (subject-predicate-object). For example: (earthquake is-a natural disaster), (rescue part-of emergency disposal). The constructed emergency plan ontology is stored in a graph database such as Neo4j in the format of RDF (Resource Description Framework) to form an emergency plan knowledge base.

[0040] Ontology-based emergency plan retrieval uses an ontology construction method to formally represent the identified emergency events as event ontology E. Extract key attributes such as event type, occurrence time, and occurrence location from the event description. According to ontology languages ​​such as OWL, event attributes are represented as ontology concepts (Class) and relationships (Property). For example, "A 5.8-magnitude earthquake occurred somewhere on May 12, 2023" is represented as: (Erdf: type Earthquake), (E has Occur Time "2023-05-12"), (E has Occur Location somewhere), (E has Magnitude "5.8"). Generate an event query ontology, extract key concepts such as event type (event Type), event time (event Time), and event location (event Loc) from the event ontology E, and form a query ontology Q. The query ontology Q is expressed as: (Q rdf: type [event Type]), (Q hasOccur Time [event Time]), (Q has Occur Location [event Loc]), where [eventType], [event Time], and [event Loc] are specific event attribute values.

[0041] Retrieve candidate plan ontologies and use the ontology reasoning mechanism based on description logic (DL) to retrieve plan ontologies that meet the query ontology Q constraints from the emergency plan knowledge base. DL ontology reasoning defines the semantic relationship between ontologies through preset axioms. Common axioms include: Sub Class Of (C1, C2): Concept C1 is a subclass of concept C2. Equivalent Classes (C1, C2): Concepts C1 and C2 are equivalent. Dis joint Classes (C1, C2): Concepts C1 and C2 do not intersect. Define axioms: Sub Class Of (Earthquake Plan, Emergency Plan), that is, "Earthquake Plan" is a subclass of "Emergency Plan". Query the plan ontology P that meets Q: (P rdf: type [event Type]), (P has Refer Loc [eventLoc]). Find all Ps that meet the conditions through semantic relationships such as subclasses and equivalent classes as the candidate plan ontology set {P1, P2, ...}. Extract the key concepts of candidate plans. For each plan ontology Pi in the candidate plan set, extract the key concepts such as plan type (plan Type), disposal plan (plan Scheme), etc.

[0042] An example of the concept of plan type: (Pi rdf: type Building Collapse), indicating that plan Pi is applicable to building collapse events. An example of the concept of disposal plan: (Pi has Step Evacuation), (Pi has Step RescueTrapped), indicating that plan Pi contains the steps of "evacuation of personnel" and "rescue of trapped personnel". The extracted key concepts constitute the concept set {Ci1, Ci2, ...}. Calculate the event-plan similarity, and calculate the semantic similarity between each key concept in the query ontology Q and the concept set of the candidate plan Pi. For each concept q in Q: obtain the synonym set synset(q) of q; for each concept c in Pi: obtain the synonym set synset(c) of c; calculate the similarity between synset(q) and synset(c): ,in, is the most recent common ancestor concept of q and c on WordNet; is the distance from q to the root node of WordNet; calculate q and The maximum similarity of all concepts ; Calculate the average similarity of all concepts in Q as the similarity between Q and Similarity: . Match emergency plans and set similarity thresholds , select from the candidate plans that meet Plan ,extract The corresponding plan content (i.e., non-ontology form) is used as the final search result. Events and plans are uniformly represented as ontology, and DL reasoning is used to quickly narrow the search scope. The semantic similarity between the query concept and the plan concept is calculated on the candidate plan set, breaking through the limitations of keyword matching and achieving similarity judgment at the semantic level.

[0043] Calculate the semantic similarity between the identified emergency event and the multiple emergency plans obtained, and select one or more emergency plans with semantic similarity greater than a preset threshold as the matching result; including: vectorizing the event type, defining a set of emergency event types, such as {fire, earthquake, rainstorm, explosion, ...}, a total of N types; using One-hot encoding to convert the event type into an N-dimensional binary vector. For example, the identified event type is "earthquake", and the corresponding vector is [0, 1, 0, 0, ...]. Perform word segmentation and part-of-speech tagging on each candidate plan text. Extract noun and verb part-of-speech words as keywords. Extract the title of the plan (if any). Concatenate the keywords and title into a string to represent the plan text.

[0044] Generate event type embedding vectors and load pre-trained word embedding models, such as Word2Vec, GloVe, etc. Count the frequency of occurrence of each event type in historical emergency event data For each non-zero element of the event type vector: obtain the event type through the word embedding model The word vector .calculate The weighted embedding vector of : ; Average the weighted embedding vectors of all non-zero elements to obtain the event type embedding vector: Generate emergency plan embedding vectors, and a keyword list for each plan text : Get each keyword through the word embedding model The word vector . Calculate each keyword TF-IDF weight , as the weighting coefficient. The weighted average of all word vectors is used to obtain the embedding vector of the plan text: .

[0045] Get the final matching results and embed the vector for each plan : Calculate its embedding vector with the event type The cosine similarity of: ; Save the similarity results to the similarity list. Output similarity list . Set similarity threshold . Select from the similarity list All plans are taken as the final matching results. If the similarity of no plan exceeds the threshold, the threshold can be appropriately lowered, or the Top-K plans with the highest similarity can be selected. By embedding both the event type and the plan text into the same semantic space, their similarity in the space can be calculated, overcoming the limitations of keyword matching. When embedding events, the prior distribution of different event types is considered, and higher weights are given to high-frequency event types to adapt to the long-tail distribution of the data. When embedding plans, TF-IDF weighting is used to highlight the key information in the plan text.

[0046] Extract the difference features between the identified emergency events and the matching results, and use the heuristic search algorithm to generate personalized emergency sub-plans for the identified emergency events. First, extract the event attributes, decompose the input emergency events, and extract the following key attributes: event type: such as earthquake, fire, etc.; event time: the time when the event occurred; event location: the geographical location where the event occurred; map the extracted attribute values ​​into numerical vectors to form an event attribute vector Extract the attributes of the plan, decompose the attributes of each matching plan text input, and extract the following attributes: the type of event to which the plan applies: such as earthquake plan, fire plan, etc.; the time when the plan applies: such as weekdays, nighttime, etc.; the location where the plan applies: such as school, community, etc.; map the extracted attribute values ​​to numerical vectors to form a plan attribute vector .

[0047] Calculate the attribute difference, and for each matching plan, calculate its plan attribute vector With event attribute vector The weighted Euclidean distance of : ,in, is the weight of the kth attribute, is the value range of the kth attribute. The plan attributes with significant differences (such as diff greater than a certain threshold) and their difference values ​​are taken as difference attributes and saved in the difference attribute set.

[0048] Retrieve the plan fragments, which are calculated based on the difference between the event attributes and the plan attributes, indicating the mismatch between the current event and the existing plan, such as {"Event Type": "Flash Flood", "Event Time": "Early Morning", "Event Location": "Village"}; Emergency Plan Knowledge Base: Contains a large number of plan texts, which are divided into several plan fragments according to the granularity of chapters, paragraphs, etc. Therefore, the difference attributes are expanded into richer semantic representations to improve the recall rate of retrieval. Use pre-trained word embedding models, such as Word2Vec, GloVe, etc.; for each attribute value in the difference attribute set, find semantically similar words in the word embedding space; select the K words with the highest similarity as the extended words of the attribute value; combine the original attribute value and the extended word to form the extended difference attribute. For example, the original difference attribute: "Event Type": "Flash Flood"; the extended difference attribute: "Event Type": ["Flash Flood", "Flood", "Mudslide", "Flood"]. Use the bag-of-words model or document vectorization method, such as TF-IDF, Doc2Vec, etc., to perform preprocessing such as word segmentation and stop word removal on each plan fragment in the plan knowledge base; based on the bag-of-words model, count the word frequency in each fragment and generate a word frequency vector; or use the document vectorization method to directly map the fragment into a dense vector; normalize the fragment vector, such as L2 normalization. For example, the plan fragment: "When a flash flood occurs, the early warning system should be activated immediately to notify residents to evacuate to a safe area." Fragment vector: [0.2, 0.0, 0.1, ..., 0.0, 0.3].

[0049] The expanded difference attribute is regarded as a "virtual document". The same vectorization method as the plan fragment is used. For each difference attribute, its attribute name and attribute value (including the expanded word) are spliced ​​into a text string. The text string is preprocessed by word segmentation and stop word removal. Based on the bag-of-words model or document vectorization method, the text string is mapped into a vector, and the attribute vector is normalized. For example, difference attribute: "event type": ["mountain torrent", "flood", "mudslide", "flood", ...]; attribute text string: "event type mountain torrent flood mudslide flood"; attribute vector: [0.1, 0.0, 0.2, ..., 0.0, 0.4].

[0050] Calculate the similarity between each plan segment and the difference attribute, select the segments with high similarity, and use similarity measurement methods such as cosine similarity to calculate the similarity between each plan segment vector. and the difference attribute vector , calculate their cosine similarity: ; Traverse all the plan segments, select the segments with similarity greater than the preset threshold with any difference attribute, and add them to the candidate set; sort the candidate segments from high to low according to the similarity score. For example, the plan segment vector : [0.2, 0.0, 0.1, ..., 0.0, 0.3]; difference attribute vector : [0.1, 0.0, 0.2, ..., 0.0, 0.4]; cosine similarity: ; If the threshold is set to 0.6, then Selected into the alternative set, forming the library of alternative plan fragments.

[0051] Generate personalized emergency sub-plans based on the search algorithm, and transform the event attribute vector Marked as the initial state of the search, the plan attribute vector Mark as target state; define the open list (Open List) and closed list (Closed List) of the search. The open list stores the plan fragments to be visited and is initially empty. The closed list stores the plan fragments that have been visited and initially contains only For each scenario segment , record its parent fragment , which is the precursor fragment in the search process, : From the initial state arrive The actual cost is initially :from To the target state The estimated cost is initially , , that is, from via arrive The estimated total cost.

[0052] Calculate the actual cost , for the plan fragment and the attribute vector of its parent fragment and , calculate the difference between them, , , is the weight of the kth attribute component, reflecting the importance of the component. is the value range of the kth component, used for normalization. Calculate the estimated cost , for the plan fragment The attribute vector and target state , calculate the similarity between them, , is the weight of the kth attribute component.

[0053] During the search process, select the plan fragment p with the smallest f(p) from the open list, mark it as the current fragment, remove the current fragment p from the open list, and add it to the closed list. If the attribute vector of the current fragment p With target status If the difference is less than the preset threshold, the search is successful; otherwise, continue searching. Generate the subsequent segments of the current segment p, that is, select several segments with semantic relevance to p in the plan segment library The subsequent segments can be selected by keyword matching, semantic similarity, etc. :if Already in the closed list, it means it has been visited before, skip it; if If it is not in the open list, add it to the open list and set its parent fragment to p. Calculate , and ;if Already in the open list, compare its original g value with the newly calculated g value, and update if the new g value is smaller The parent fragment is p, and recalculate and ; If the open list is empty, the search fails.

[0054] Generate personalized emergency sub-plans, if the search is successful: from the target fragment Start by backtracking along the parent pointer until you reach the initial state ; Extract the plan fragment contents on the backtracking path in reverse order and combine them into a complete personalized emergency sub-plan. Output the emergency sub-plan. If the search fails: the personalized emergency sub-plan is empty, and a search failure prompt is output. This application defines the actual cost and estimated cost between plan fragments, uses the f(p) value to guide the search direction, and thus finds the optimal path from event attributes to target plan attributes in the plan fragment library, and extracts the fragment contents on the path to form a personalized plan.

[0055] The matched one or more emergency plans are integrated with the generated personalized emergency sub-plans to obtain an emergency command and dispatch plan; specifically, for each matched emergency plan, its key information is extracted and converted into a structured plan object, including: metadata such as plan name, version, and applicable conditions; the organizational structure of the plan, including hierarchical information such as chapters and paragraphs; the key attributes of each chapter or paragraph, such as risk sources, disposal measures, resource requirements, etc.; key entities in the plan and their relationships, such as events, departments, personnel, places, etc.; for personalized emergency sub-plans, their key information is also extracted and represented as one or more plan fragments; each plan fragment includes its description content, key attributes, contextual relationships and other information.

[0056] Traverse the chapters and paragraphs of each emergency plan and compare them with the fragments of personalized sub-plans. The basis for comparison includes the similarity of key attributes, co-occurrence of entities, etc. You can set a similarity threshold and establish connections between plan paragraphs and plan fragments that exceed the threshold. Through comparison, find the correspondence between personalized sub-plans and plan contents, identify the personalized disposal measures already included in the plan, and identify the supplements or corrections to the standard plan in the personalized sub-plan.

[0057] Based on a matching plan, an initial command and dispatch plan is generated, and each fragment in the personalized emergency sub-plan is traversed: if the fragment is highly matched with the attributes of a paragraph in the plan, the original paragraph is replaced with the personalized fragment; if the fragment is a supplement to the plan, it is inserted into the appropriate position of the plan according to its context; if the fragment conflicts with the plan (such as incompatible disposal measures), a warning is thrown, and emergency experts decide how to deal with it, and an emergency command and dispatch plan containing all newly added and modified fragments is generated. This incremental fusion method can retain the complete structure of the standard plan to the greatest extent, while supplementing and optimizing personalized disposal measures in a targeted manner, generating a comprehensive, dynamic, and systematic command and dispatch plan.

[0058] Preferably, the generated command and dispatch plan is subjected to a global consistency and feasibility check, mainly including: resource demand conflicts: for example, the demand for the same emergency resource by different disposal measures exceeds the existing supply capacity; timing logic conflicts: for example, the order of actions required by different fragments is inconsistent or there are circular dependencies; spatial location conflicts: for example, there are location conflicts between emergency sites and evacuation routes involved in different fragments; automatic or semi-automatic identification of potential conflicts, prompting emergency commanders to make decisions, and further modifying and optimizing the plan through human-computer interaction until all serious conflicts are eliminated; and providing a variety of decision-making alternatives for inevitable resource contention and other situations, and giving risk assessments.

[0059] For the final emergency command and dispatch plan, its key information is extracted and converted into a friendly visual interface. Graphical methods such as Gantt charts and flow charts are used to display the plan's time schedule and task dependencies. Spatial information such as emergency resources, risks, and disposal actions are superimposed on the map. A text view of the plan content is provided, allowing the complete dispatch plan document to be viewed and exported. Through visualization methods, emergency command personnel are helped to quickly understand and implement the plan, and the progress of the plan's implementation can be monitored in real time.

Claims

1. An event-driven emergency command and dispatch method, characterized in that: include: Obtain historical emergency event data, which includes event type, event time, event location and disposal plan; Feature engineering is used to extract features from historical emergency event data to obtain a feature data set; wherein the feature data set includes event type feature vectors, event frequency feature vectors, distribution density feature vectors, and text feature vectors; Use the feature data set to train the emergency event classification model based on the CART decision tree; Acquire real-time emergency event data, and perform feature extraction on the real-time emergency event data to obtain a real-time event feature vector; The obtained real-time event feature vector is used as input to identify emergency events using the emergency event classification model; According to the identified emergency events, a rule-based case reasoning method is used to retrieve and generate emergency command and dispatch plans from the preset emergency plan knowledge base.

2. The event-driven emergency command and dispatch method according to claim 1 is characterized in that: Feature engineering is used to extract features from historical emergency event data, including: One-hot encoding with the same number of binary bits as the event type is used to encode the event type and obtain the event type feature vector; The sliding time window method is used to perform time window statistics on event time to obtain event frequency feature vectors at different time scales; The event locations are divided into geographic grids, and the number of events in each grid is estimated using the kernel density estimation method to obtain the distribution density feature vectors of events in different regions. The TF-IDF algorithm is used to extract features from the text description of the treatment plan to obtain the text feature vector of the treatment plan; The event type feature vector, event frequency feature vector, distribution density feature vector and text feature vector are concatenated through vectors to obtain a feature data set.

3. The event-driven emergency command and dispatch method according to claim 2 is characterized in that: The feature data set is used to train the emergency event classification model based on the CART decision tree, including: The feature data set is used as the training set, the Gini index is used to select the optimal partitioning features, and a CART decision tree is generated recursively; the feature data set is used as the input of the CART decision tree, and the event type is used as the output of the CART decision tree; The K-fold cross-validation method is used to divide the feature data set into K mutually exclusive subsets. One subset is selected as the validation set each time, and the remaining K-1 subsets are used as training sets. The CART decision tree is trained using the training set, and the classification performance of the CART decision tree is evaluated on the validation set. The cross-validation process is repeated K times to obtain the emergency event classification model.

4. The event-driven emergency command and dispatch method according to claim 3 is characterized in that: Generate emergency command and dispatch plan, including: Adopting ontology-based knowledge representation method, the emergency plan is structured and stored in the form of ontology to build an emergency plan knowledge base; According to the identified emergency events, the emergency plan knowledge base is retrieved through the ontology reasoning mechanism to obtain multiple emergency plans corresponding to the emergency events; Calculate the semantic similarity between the identified emergency event and the acquired multiple emergency plans, and select one or more emergency plans with semantic similarity greater than a preset threshold as the matching result; Extract the difference features between the identified emergency events and the matching results, and use the heuristic search algorithm to generate personalized emergency sub-plans for the identified emergency events; The matched one or more emergency plans are integrated with the generated personalized emergency sub-plans to obtain an emergency command and dispatch plan.

5. The event-driven emergency command and dispatch method according to claim 4 is characterized in that: Obtain multiple emergency response plans corresponding to emergency events, including: The ontology construction method is used to represent the identified emergency events and obtain the event ontology; Extract key concepts from the event ontology to form an event query ontology; the key concepts include event type, event time and event location; According to the event query ontology, the ontology reasoning mechanism based on description logic is adopted to retrieve the emergency plan ontology containing all the key concepts in the event query ontology from the constructed emergency plan knowledge base as the candidate emergency plan ontology set; wherein, the emergency plan ontology contains the plan type and disposal plan; For each emergency plan ontology in the candidate emergency plan ontology set, the corresponding plan type and disposal plan are extracted to form a key concept set of the emergency plan ontology; Calculate the semantic similarity between the key concepts in the event query ontology and the key concept set of the emergency plan ontology to obtain the similarity of each candidate emergency plan ontology; The candidate emergency plan ontology with a similarity greater than a preset threshold is taken as the matched emergency plan ontology; The emergency plan corresponding to the matched emergency plan ontology is obtained as the emergency plan corresponding to the identified emergency event.

6. The event-driven emergency command and dispatch method according to claim 5 is characterized in that: The similarity of each candidate emergency plan ontology is obtained, including: Get the key concepts in the event query ontology ; Get each key concept in the key concept set of the emergency plan ontology ; Get key concepts from the WordNet knowledge base Synonyms of ; Get key concepts from the WordNet knowledge base Synonyms of ; calculate and The semantic similarity of : ； in, express and The synset of the lowest common ancestor concept in WordNet; Indicates the shortest path length from the synonym set synset to the WordNet root node; Computing Key Concepts All key concepts in the key concept set of the emergency plan ontology The maximum semantic similarity of , as a key concept The similarity component is calculated as: ； Among them, O represents the key concept set of the emergency plan ontology; Compute all key concepts in the event query ontology The arithmetic mean of the similarity components is taken as the similarity between the event query ontology and the current candidate emergency plan ontology: ； in, Represents the event query ontology; represents the candidate emergency plan ontology; n is the number of key concepts in the event query ontology.

7. The event-driven emergency command and dispatch method according to claim 4 is characterized in that: One or more emergency plans with semantic similarity greater than a preset threshold are selected as matching results, including: Obtain the event type of the identified emergency event and convert the event type into an event type vector through One-hot encoding; Perform word segmentation and part-of-speech tagging on the multiple emergency plans obtained, extract nouns and verbs as keywords, and extract the titles of the emergency plans; The extracted keywords and titles are concatenated as emergency plan vectors; Map each element in the event type vector to the word embedding vector of the corresponding event type, and obtain the event type embedding vector by taking the weighted average of the word embedding vectors corresponding to the non-zero elements; the weight is the frequency of each event type in the historical data; Map each word in the emergency plan vector to a corresponding word embedding vector, and obtain the emergency plan embedding vector by weighted averaging all word embedding vectors; where the weight is the TF-IDF value of each word; The cosine similarity is used to calculate the similarity between the event type embedding vector and the emergency plan embedding vector to obtain a similarity list; According to a preset threshold, one or more emergency plans are obtained from the similarity list as matching results.

8. The event-driven emergency command and dispatch method according to claim 4 is characterized in that: A heuristic search algorithm is used to generate personalized emergency sub-plans for the identified emergency events, including: Decompose the attributes of the identified emergency events, extract the event type, event time and location, and form an event attribute vector; Decompose the attributes of one or more matched emergency plans, extract the event type, event time and location of the plan, and form a plan attribute vector; Calculate the weighted Euclidean distance between the event attribute vector and each plan attribute vector as the difference degree, and generate a difference attribute set; According to the difference attribute set, the emergency plan fragments whose semantic similarity with the difference attribute is greater than a threshold are retrieved from the emergency plan knowledge base to form an emergency plan fragment library; wherein the emergency plan fragment represents a step or measure in the emergency plan; Use the event attribute vector in the difference attribute set as the initial state of the search , take the corresponding plan attribute vector in the difference attribute set as the target state of the search ; In the plan fragment library, search for the initial state To the target state The optimal path is used as a personalized emergency sub-plan.

9. The event-driven emergency command and dispatch method according to claim 8 is characterized in that: Search from the initial state To the target state The optimal path, as a personalized emergency sub-plan, includes: For each plan segment in the plan segment library , calculate from the initial state arrive The cost : ； in, From the initial state arrive the actual cost of For To the target state the estimated cost of Select the one with the minimum Value Scenario Snippet , mark it as visited, and Updated to The corresponding plan segment attribute vector , denoted as ; by is the new initial state, repeat the above marking steps until the target state is reached or preset termination conditions; If the target state is reached , then the sequence of plan fragments accessed during the search process As the optimal path, extract the plan fragment content corresponding to the corresponding path, and combine them in order to form a personalized emergency sub-plan; If the preset termination condition is reached, the personalized emergency sub-plan is empty.

10. The event-driven emergency command and dispatch method according to claim 9, characterized in that: Actual cost , calculated by the following formula: ； in, and Represents the initial state event attribute vector And the plan fragment The attribute vector The kth component of ; is the weight of the kth component, reflecting the importance of the component; is the value range of the kth component, used for normalization processing; Estimated cost , calculated by the following formula: ； in, and Represent the attribute vectors of the plan fragments and the target state plan attribute vector The kth component of ; is the weight of the kth component.

Images