An emergency on-duty scheduling optimization method and system based on artificial intelligence

Abstract

The application relates to the technical field of emergency management, and provides an emergency on-duty scheduling optimization method and system based on artificial intelligence, which comprises the following steps: acquiring emergency knowledge of an emergency command center; constructing an emergency knowledge model based on the emergency knowledge; acquiring input environment information and operation information; predicting a chain of cascading events and a possibility of occurrence in a future specific time window by using the emergency knowledge model, and generating an early warning event chain; analyzing comprehensive capability demand based on the early warning event chain; acquiring availability, individual capability characteristics and state information of personnel, and matching the comprehensive capability demand to generate a team personnel configuration scheme meeting the comprehensive capability demand; continuously acquiring early warning information, personnel state feedback and event progress in an emergency command process; and dynamically feeding back and adjusting the team personnel configuration scheme according to the early warning information, the personnel state feedback and the event progress. The application has the effect of improving the response efficiency and dynamic adaptability of an emergency on-duty team.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of emergency management, and in particular to an emergency on-duty scheduling optimization method and system based on artificial intelligence. BACKGROUND

[0002] In critical infrastructures that bear important social functions, such as emergency command centers, uninterrupted operation is required throughout the year and around the clock to respond to various emergencies. In order to ensure the effectiveness of emergency response, personnel with specific skills need to be arranged for rotation on duty. The traditional scheduling method mainly relies on the experience of management personnel and preset rules, which can ensure basic personnel allocation, but it is difficult to fully consider the diversity of emergencies and their demand for specific skill combinations of on-duty teams.

[0003] At the same time, the real-time physiological and psychological state of the on-duty personnel has an important influence on the emergency disposal ability, and the existing system usually cannot obtain or effectively use such real-time state information, and also lacks the ability to dynamically adjust the scheduling scheme according to these information. The existing technology also fails to fully consider the familiarity of team members with emergency plans and the past execution effect, which is crucial to the speed and effectiveness of emergency response. Therefore, the existing technology has many deficiencies in emergency on-duty scheduling, and it is difficult to meet the requirements of overall team capability and dynamic adaptability in complex and variable emergency scenarios.

[0004] In view of the above problems, the existing technology needs to be improved. SUMMARY

[0005] The purpose of the present application is to provide an emergency on-duty scheduling optimization method and system based on artificial intelligence, which can predict potential event chains in advance according to dynamically changing emergency risks and personnel states, generate team personnel allocation schemes that meet specific comprehensive capability requirements, and dynamically adjust them, thereby improving the overall response efficiency and dynamic adaptability of emergency on-duty teams.

[0006] The present application provides an emergency on-duty scheduling optimization method based on artificial intelligence, comprising:

[0007] Obtaining emergency knowledge from an emergency command center; the emergency knowledge includes risk sources, event evolution paths and mutual influence relationships;

[0008] Building an emergency knowledge model based on the emergency knowledge;

[0009] Obtaining input environmental information and operational information, using the emergency knowledge model to predict the chain of events and the possibility of occurrence within a specific time window in the future, and generating a warning event chain;

[0010] Based on the early warning event chain, the corresponding comprehensive ability demand is analyzed; the comprehensive ability demand includes the specific skill combination required to respond to the early warning event chain, the familiarity with the emergency plan, and the number of personnel required;

[0011] Obtain the availability, individual ability characteristics and state information of the personnel, and match according to the comprehensive ability demand to generate a team personnel configuration scheme that meets the comprehensive ability demand;

[0012] Continuously obtain early warning information, personnel state feedback and event progress during the emergency command process;

[0013] According to the early warning information, personnel state feedback and event progress, the team personnel configuration scheme is dynamically adjusted.

[0014] Through the above scheme, the potential event chain can be predicted in advance according to the dynamically changing emergency risk and personnel state, and the team personnel configuration scheme that meets the specific comprehensive ability demand is generated and dynamically adjusted, thereby improving the overall response efficiency and dynamic adaptability of the emergency on-duty team.

[0015] Further, the present application also proposes the steps of obtaining input environment information and operation information, predicting the chain of events and the possibility of occurrence within a specific time window in the future using an emergency knowledge model, and generating an early warning event chain, comprising:

[0016] Obtain the environment information and operation information in the emergency command site;

[0017] Detect whether there is conflict data between the environment information and the operation information;

[0018] If there is conflict data, process the conflict data according to the preset conflict processing strategy;

[0019] If there is no conflict data, process the environment information and operation information to generate current situation information;

[0020] Identify the uncertainty description in the current situation information and convert it into a quantitative parameter;

[0021] Based on the current situation information and the quantitative parameter, the chain of events and the possibility of occurrence within a specific time window in the future are predicted using an emergency knowledge model to generate an early warning event chain.

[0022] Through the above scheme, the generation process of the early warning event chain is refined, and the accuracy and reliability of the early warning are improved by processing the environment information and operation information, including conflict detection and uncertainty quantification.

[0023] Further, the present application also proposes, before the step of detecting whether there is conflict data between the environment information and the operation information, further comprising:

[0024] a reliability index and a timestamp of the environment information and the operation information are identified;

[0025] based on the reliability index and the timestamp, validity evaluation is performed to obtain environment information and operation information whose validity has been evaluated;

[0026] the environment information whose validity has been evaluated is updated to the environment information, and the operation information whose validity has been evaluated is updated to the operation information.

[0027] Through the above scheme, validity evaluation of the environment information and the operation information is added before the generation of the early warning event chain, further improving the quality of the input data and the accuracy of the early warning.

[0028] Further, the application further proposes a step of analyzing corresponding comprehensive capability requirements based on the early warning event chain, comprising:

[0029] receiving the early warning event chain;

[0030] based on the number of professional fields related to knowledge and the number of disposal measures in the early warning event chain, determining whether the early warning event chain is a composite event chain;

[0031] if it is a composite event chain, the early warning event chain is decomposed into a plurality of task units with corresponding professional fields;

[0032] determining the basic capability requirements corresponding to the task units according to a preset knowledge and capability mapping rule;

[0033] based on the logical association between the task units, identifying key nodes that need to be cooperated between the task units to obtain collaborative dependency relationship information;

[0034] generating specific collaborative capability requirements according to the collaborative dependency relationship information;

[0035] integrating the basic capability requirements and the specific collaborative capability requirements of the task units to generate corresponding comprehensive capability requirements.

[0036] Through the above scheme, the analysis process of the comprehensive capability requirements is refined, especially for the composite event chain, which can decompose the task units and consider the collaborative dependency relationship, so that the capability requirement analysis is more comprehensive and more suitable for actual emergency scenarios.

[0037] Further, the application further proposes a step of identifying key nodes that need to be cooperated between the task units based on the logical association between the task units to obtain collaborative dependency relationship information, comprising:

[0038] based on the logical association between the task units and the event evolution relationship defined in the emergency knowledge model, identifying key nodes that need to be cooperated to obtain first collaborative dependency relationship;

[0039] obtain resource requirement information, environment condition information, state output information and state input information corresponding to the task unit;

[0040] According to the resource requirement information, the environment condition information, the state output information and the state input information, whether there is a collaborative demand or an indirect causal chain between the task units is identified, and if there is, it is determined that there is a second collaborative dependency relationship between the corresponding two task units;

[0041] Integrate the first collaborative dependency relationship and the second collaborative dependency relationship to generate collaborative dependency relationship information.

[0042] Through the above scheme, the identification process of the collaborative dependency relationship is refined, not only considering logical association and event evolution, but also combining resource, environment, state and other information, so that the identification of collaborative demand is more accurate.

[0043] Further, the present application also proposes that, according to the resource requirement information, the environment condition information, the state output information and the state input information, whether there is a collaborative demand or an indirect causal chain between the task units is identified, and if there is, it is determined that there is a second collaborative dependency relationship between the corresponding two task units, the step comprising:

[0044] Based on the resource requirement information, shared resource identification is performed, and if there is a shared resource between the task units, it indicates that there is a collaborative demand, and it is determined that there is a second collaborative dependency relationship between the task units;

[0045] Based on the environment condition information and the current environment state information, common environment factor identification is performed, and if there is a common environment factor between the task units, it indicates that there is a collaborative demand, and it is determined that there is a second collaborative dependency relationship between the task units;

[0046] Based on the state output information and the state input information, state output and state input association identification is performed, and if the state output of one task unit constitutes the state input of another task unit, it indicates that there is an indirect causal chain, and it is determined that there is a second collaborative dependency relationship between the task units.

[0047] Through the above scheme, the method of identifying collaborative demand or indirect causal chain based on resource, environment and state information is further refined, and the accuracy of collaborative dependency relationship identification is improved.

[0048] Further, the present application also proposes that, based on the state output information and the state input information, state output and state input association identification is performed, the step comprising:

[0049] Based on the two task units, the state output information of one task unit and the state input information of another task unit are obtained;

[0050] extracting a first uncertainty parameter in the state output information, and extracting a second uncertainty parameter in the state input information;

[0051] determining whether the uncertainty of the state output information is within an uncertainty acceptance range corresponding to the second uncertainty parameter based on the first uncertainty parameter and the second uncertainty parameter, and generating a determination result;

[0052] determining whether a case that a state output of a task unit constitutes a state input of another task unit occurs between the task units according to the determination result.

[0053] Through the above scheme, the method of identifying an indirect cause-effect chain based on the association between the state output and the state input is refined, and the accuracy of the association determination is improved by considering the uncertainty parameters.

[0054] Further, the present application also proposes the steps of obtaining the availability, individual ability characteristics and state information of the personnel, and matching according to the comprehensive ability demand to generate a team personnel configuration scheme meeting the comprehensive ability demand, including:

[0055] decomposing a predetermined overall response effectiveness in the early warning event chain into a plurality of emergency response sub-targets;

[0056] taking the availability, individual ability characteristics, state information of the personnel and a specific time window as constraint conditions for optimization calculation;

[0057] iterating a team personnel configuration scheme meeting all the emergency response sub-targets through iterative processing based on the emergency response sub-targets and the constraint conditions.

[0058] Through the above scheme, the personnel matching and scheme generation process are refined, and a more optimal team configuration scheme meeting multiple demands can be generated by decomposing the overall effectiveness into sub-targets and iteratively optimizing in combination with the constraint conditions.

[0059] Further, the present application also proposes the steps of iterating a team personnel configuration scheme meeting all the emergency response sub-targets through iterative processing based on the emergency response sub-targets and the constraint conditions, including:

[0060] defining a weight relationship between the plurality of emergency response sub-targets;

[0061] iterating a team personnel configuration scheme meeting all the emergency response sub-targets through iterative processing based on the emergency response sub-targets and the constraint conditions, and the iterative processing includes:

[0062] generating a team personnel configuration scheme to be evaluated;

[0063] evaluating the satisfaction degree of the team personnel configuration scheme to be evaluated to all the emergency response sub-targets;

[0064] According to the weight relationship and the satisfaction degree, a comprehensive evaluation value of the team staffing scheme to be evaluated is calculated;

[0065] According to the comprehensive evaluation value, an iterative processing procedure is performed under the constraint condition;

[0066] Until the team staffing scheme satisfying all the emergency response sub-targets is iterated out.

[0067] Through the above scheme, the iterative optimization procedure is further refined, and by defining the sub-target weight and calculating the comprehensive evaluation value, the iterative procedure can be more effectively guided to find a scheme satisfying all the sub-targets.

[0068] Further, the present application also proposes an emergency on-duty scheduling optimization system based on artificial intelligence, used for executing the above-mentioned emergency on-duty scheduling optimization procedure based on artificial intelligence, comprising:

[0069] An emergency knowledge acquisition module is used for acquiring emergency knowledge from an emergency command center;

[0070] An emergency knowledge model module is used for constructing an emergency knowledge model based on the emergency knowledge;

[0071] An event chain generation module is used for acquiring input environmental information and running information, predicting a chain of events and a possibility of occurrence in a specific time window in the future by using the emergency knowledge model, and generating a pre-warning event chain;

[0072] A capability demand generation module is used for analyzing corresponding comprehensive capability demands based on the pre-warning event chain;

[0073] A personnel scheme generation module is used for acquiring availability, individual capability characteristics and state information of personnel, and matching according to the comprehensive capability demands to generate a team staffing scheme satisfying the comprehensive capability demands;

[0074] An emergency information acquisition module is used for continuously acquiring pre-warning information, personnel state feedback and event progress in the emergency command procedure;

[0075] A personnel scheme adjustment module is used for dynamically feeding back and adjusting the team staffing scheme according to the pre-warning information, personnel state feedback and event progress.

[0076] From the above, the application provides an emergency value guard scheduling optimization method and system based on artificial intelligence, which solves the problems that the traditional and existing AI scheduling cannot cope with complex and variable emergency scenes, cannot dynamically match the overall ability of the team, and cannot adjust using dynamic information, by acquiring emergency knowledge, constructing a knowledge model, predicting and warning event chains, analyzing comprehensive ability requirements, matching personnel and generating a scheme, and dynamically adjusting, and has the advantages of being able to predict potential event chains in advance according to dynamically changing emergency risks and personnel states, generate a team personnel configuration scheme that meets specific comprehensive ability requirements, and dynamically adjust, thereby improving the overall response efficiency and dynamic adaptability of the emergency value guard team. BRIEF DESCRIPTION OF DRAWINGS

[0077] Figure 1 A method flowchart of an emergency value guard scheduling optimization method based on artificial intelligence in one embodiment of the application;

[0078] Figure 2 A method flowchart of an emergency value guard scheduling optimization method based on artificial intelligence in another embodiment of the application;

[0079] Figure 3 A method flowchart of an emergency value guard scheduling optimization method based on artificial intelligence in another embodiment of the application;

[0080] Figure 4 A method flowchart of an emergency value guard scheduling optimization method based on artificial intelligence in another embodiment of the application;

[0081] Figure 5 A method flowchart of an emergency value guard scheduling optimization method based on artificial intelligence in another embodiment of the application;

[0082] Figure 6 A method flowchart of an emergency value guard scheduling optimization method based on artificial intelligence in another embodiment of the application;

[0083] Figure 7 A method flowchart of an emergency value guard scheduling optimization method based on artificial intelligence in another embodiment of the application;

[0084] Figure 8 A method flowchart of an emergency value guard scheduling optimization method based on artificial intelligence in another embodiment of the application;

[0085] Figure 9 A method flowchart of an emergency value guard scheduling optimization method based on artificial intelligence in another embodiment of the application;

[0086] Figure 10A system block diagram of an artificial intelligence-based emergency on-duty scheduling optimization system in another embodiment of the present application. DETAILED DESCRIPTION

[0087] The technical solutions in the present application will be described clearly and completely below in combination with the drawings in the present application. Obviously, the described embodiments are only some of the embodiments of the present application, rather than all the embodiments. The components of the present application described and shown in the drawings herein can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the claimed present application, but only represents selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative work fall within the scope of the present application.

[0088] It should be noted that similar reference numbers and letters represent similar items in the following drawings, so once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings. Meanwhile, in the description of the present application, the terms "first", "second", etc. are only used to distinguish the description, and cannot be understood as indicating or implying relative importance.

[0089] The conventional existing emergency on-duty scheduling method mainly relies on the experience of management personnel and preset rules when dealing with emergencies, and it is difficult to effectively consider complex factors such as the diversity of event types, the dynamic changes of risk occurrence, the real-time state of on-duty personnel, and the difference in familiarity with emergency plans, resulting in that the scheduling scheme cannot be matched with potential risks in a targeted manner, affecting the efficiency and effectiveness of emergency response.

[0090] For example, assuming that in an urban emergency command center, on-duty shifts need to be arranged for the next week. The existing system may only arrange shifts according to the basic skills, available time and post requirements of personnel. However, if it is predicted that the risk of urban waterlogging will increase due to extreme weather in a specific period in the future, or the risk of public safety events will increase due to large-scale activities, the existing scheduling method is difficult to identify these dynamically changing risks, and also cannot evaluate the actual ability and familiarity of different team member combinations in dealing with waterlogging or public safety event plans. At the same time, if a on-duty personnel is tired due to continuous handling of emergency events in the current shift, his state information cannot be taken into account in the arrangement of subsequent shifts, which may result in his inability to effectively perform tasks in the next shift.

[0091] In view of this, the present application proposes an artificial intelligence-based emergency on-duty scheduling optimization method, which combines Figure 1 as shown, comprising:

[0092] S1, obtaining emergency knowledge from an emergency command center;

[0093] S2, constructing an emergency knowledge model based on emergency knowledge;

[0094] S3, acquiring input environmental information and operational information, using the emergency knowledge model to predict the chain of cascading events and the likelihood of occurrence within a specific time window in the future, and generating a pre-warning event chain;

[0095] S4, based on the pre-warning event chain, analyzing the corresponding comprehensive capability demand; the comprehensive capability demand includes the specific skill combination required to respond to the pre-warning event chain, the familiarity with the emergency plan, and the number of personnel required;

[0096] S5, acquiring the availability, individual capability characteristics and state information of personnel, and matching according to the comprehensive capability demand to generate a team personnel configuration scheme that meets the comprehensive capability demand;

[0097] S6, continuously acquiring pre-warning information, personnel state feedback and event progress during the emergency command process;

[0098] S7, dynamically adjusting the team personnel configuration scheme according to the pre-warning information, personnel state feedback and event progress.

[0099] The emergency knowledge refers to information about potential emergency situations, including risk sources, event evolution paths, and mutual influence relations, and is mainly used to provide a basic understanding of potential threats and their dynamics. An emergency knowledge model is constructed based on the obtained emergency knowledge. The emergency knowledge model refers to a structured representation of emergency knowledge, which can be implemented using a knowledge graph, an expert system, a Bayesian network, or a rule system, and is mainly used to support subsequent event prediction and capability demand analysis, facilitating computer processing and reasoning. The input environment information and operation information are obtained, and the emergency knowledge model is used to predict the chain of cascading events and the likelihood of occurrence within a specific time window in the future, generating a warning event chain. The warning event chain refers to a potential future event sequence and its likelihood predicted within a specific time range, and is mainly used to provide foresight of potential future scenarios, allowing proactive planning. Based on the warning event chain, the corresponding comprehensive capability demand is analyzed. The comprehensive capability demand refers to the necessary capabilities required by the team to effectively respond to the predicted warning event chain, including the specific skill combination required to respond to the warning event chain, the familiarity with emergency plans, and the number of personnel required, and is mainly used to define the target image of the optimal team configuration. The availability, individual capability characteristics, and state information of the personnel are obtained, and are matched according to the comprehensive capability demand to generate a team configuration scheme that meets the comprehensive capability demand. The team configuration scheme refers to the specific allocation of available personnel to form a team that meets the comprehensive capability demand, and is mainly used to provide a specific scheduling plan. The warning information, personnel state feedback, and event progress in the emergency command process are continuously obtained. The team configuration scheme is dynamically adjusted according to the obtained warning information, personnel state feedback, and event progress. The dynamic feedback adjustment refers to modifying the team configuration scheme according to real-time changes, and is mainly used to ensure that the team maintains the optimal configuration as the situation evolves.

[0100] The approach works by first establishing a foundational understanding of potential emergencies by acquiring and modeling emergency knowledge. This structured knowledge base is then used in conjunction with real-time environmental and operational data to predict potential future sequences of events and their likelihoods within a defined time frame, resulting in a chain of alert events. This prediction step enables the system to foresee potential challenges. Based on the characteristics of the predicted event chain, the approach analyzes and determines the specific integrated capabilities required for the response team, including necessary skills, familiarity with relevant plans, and the number of personnel. At the same time, the approach collects detailed information about available personnel, their individual skills, and their current state. By matching the required integrated capabilities with the available personnel attributes, an initial crew staffing plan is generated. This plan represents an optimized allocation aimed at meeting the predicted needs. The process does not stop there; the approach continuously monitors evolving situations, acquiring new alert information, personnel state feedback, and real-time updates on event progression. This continuous feedback loop enables the approach to dynamically adjust the generated crew staffing plan as needed, ensuring that the assigned team remains appropriate and effective even as situations change. This integrated process, from knowledge modeling and prediction to dynamic adjustment, results in a proactive, risk-aware, and adaptive emergency on-call scheduling method.

[0101] In some specific embodiments, an emergency command center is considered. Emergency knowledge can be stored in a database structured as a knowledge graph, where nodes represent risk sources and events, and edges represent relationships. An emergency knowledge model module can be implemented using a graph database and a related reasoning engine. Environmental information can include sensor readings and external reports, while operational information can include ongoing maintenance plans. An event chain generation module can use probabilistic reasoning algorithms on the knowledge graph, triggered by incoming data, to predict possible sequences of events. Integrated capability demand analysis can involve traversing the predicted event chain in the knowledge graph to identify required response actions and corresponding skills or plans defined in the knowledge base. Personnel information can be stored in a personnel database, including skill certifications, training records, availability calendars, and real-time state updates from wearable sensors or manual input. A personnel plan generation module can employ optimization algorithms, such as constraint programming solvers or genetic algorithms, to find the best match between required capabilities and available personnel under constraints such as shift length and rest time. Dynamic adjustments can be triggered by new incoming data, such as new weather alerts or personnel fatigue reports, which initiate a reevaluation of the predicted event chain or personnel state, leading to a recalculation or modification of the current or upcoming shift assignments.

[0102] Optionally, in combination with Figure 2 As shown, step S3: Acquire input environmental information and operational information, use the emergency knowledge model to predict a chain of interlocking events and the likelihood of occurrence within a specific time window in the future, and generate a chain of alert events, including:

[0103] S301, acquire environment information and running information in the emergency command site;

[0104] S302, detect whether there is conflict data between the environment information and the running information;

[0105] S303, if there is conflict data, process the conflict data according to a preset conflict processing strategy;

[0106] S304, if there is no conflict data, process the environment information and the running information to generate current situation information;

[0107] S305, identify the uncertainty description in the current situation information and convert it into a quantitative parameter;

[0108] S306, based on the current situation information and the quantitative parameter, predict the chain of interlocking events and its occurrence possibility within a specific time window in the future by using an emergency knowledge model, and generate a pre-warning event chain.

[0109] The environment information refers to data describing the physical environment state of the emergency command site and its periphery, and can specifically include temperature, humidity, air pressure, wind speed, wind direction, air quality, light intensity, noise level, and other sensor collected data, and aims to reflect the environmental factors that can affect the occurrence or disposal of emergency events. The operation information refers to data describing the activity state of various systems, devices, and personnel in the emergency command site, and can specifically include device operation parameters, system load, network state, personnel position, on-duty personnel scheduling, communication link state, and aims to reflect the internal operation condition and resource allocation of the site. The conflict data refers to data points or data groups that are contradictory, inconsistent, or obviously abnormal in the obtained environment information and operation information, and can specifically manifest as a value difference of the same parameter in different data sources exceeding the allowed range, data being seriously inconsistent with known common sense or historical data, or logical contradictions between different types of data, and aims to identify problems in the original data that can cause errors in subsequent processing and prediction. The conflict processing strategy refers to a series of rules, algorithms or methods for solving conflict data, and can specifically include data cleaning, data fusion (such as weighted average, majority voting), outlier removal, data source priority sorting, manual verification, and aims to eliminate or reduce the impact of conflict data on subsequent processing and improve data quality. The current situation information refers to the comprehensive information that can reflect the current overall state of the emergency command site, which is integrated and refined from the pre-processed (including conflict processing and conventional processing) environment information and operation information, and can specifically be a summary, analysis and structured representation of key environmental parameters, device operation state, personnel distribution, ongoing tasks, and aims to provide an accurate and comprehensive current state description for subsequent prediction. The uncertainty description refers to information in the current situation information that cannot be completely determined or has ambiguity, and can specifically manifest as low confidence of some data, uncertain information source, fuzzy judgment of future trends, or non-precise description of some states, and aims to identify the uncertain components in the information for consideration in the prediction process. The quantitative parameter refers to the conversion of the uncertainty description into a numerical or parameter form that can be mathematically calculated or model processed, and can specifically use probability value, confidence interval, fuzzy set membership degree, uncertainty index, and aims to enable the emergency knowledge model to process and utilize these uncertain information for prediction. In the present application, step S306 is used to predict the chain of interlocking events and its occurrence possibility in a specific time window in the future based on the current situation information and the quantitative parameter, using the emergency knowledge model, to generate a pre-warning event chain. This step is the core link of prediction and forward planning in the entire emergency on-duty scheduling optimization method, and its implementation principle is described as follows:

[0110] First, the current situation information obtained in step S305 and the quantification parameters are taken as input data. The current situation information is a comprehensive state description formed after the structured integration and processing of environmental information and operation information, and usually includes but is not limited to: temperature, humidity, air pressure, wind speed, wind direction, light intensity and other environmental element parameters, as well as device operation status, system load, network connectivity, personnel position and state and other operation information. At the same time, the quantification parameters are a set of mathematical quantification indexes converted from uncertainty description, used to express the uncertainty degree of information, such as probability value, confidence interval or fuzzy membership degree, etc.

[0111] Secondly, the above input data is introduced into the emergency knowledge model. The emergency knowledge model is a modeling system for structured expression and reasoning of emergency knowledge, which can be implemented by knowledge graph, Bayesian network, expert system or rule-based system. In a typical implementation, the knowledge graph is constructed by a graph database, in which the nodes represent events, risk sources or system states, and the edges represent causal relationships or influence paths; each node is attached with attribute information such as historical occurrence probability, influence range, response time limit, etc. The Bayesian network constructs a directed acyclic graph, and the probability dependence relationship between events is defined by conditional probability table, which can infer the occurrence probability of other nodes when receiving part of the state information.

[0112] Thirdly, in the model reasoning stage, the related nodes in the knowledge model are activated according to the current situation information, and the event chain is deduced. Taking the knowledge graph as an example, the system traverses the nodes matched with the current input data, simulates the possible future event path according to the event propagation path defined in the graph, and combines the historical data or probability parameters. Taking the Bayesian network as an example, the system updates according to the Bayesian inference method, and deduces the occurrence probability of each event within a set time window according to the prior probability and conditional probability.

[0113] In the reasoning process, the system considers the uncertainty quantification parameters in the input, such as indicators with low data confidence, which will participate in reasoning with lower weight, or express the possible state space in the form of probability distribution. By introducing these uncertainty parameters, the emergency knowledge model can use Monte Carlo simulation, fuzzy logic reasoning and other methods to improve the adaptability and prediction robustness to complex scenarios.

[0114] After reasoning, the system will generate a set of possible event sequences that may occur within a given future time window (e.g., 1 hour, 3 hours, or 24 hours in the future), each event sequence constituting a pre-warning event chain, and attaching a probability of occurrence. Each pre-warning event chain represents a chain of events that is likely to occur under current conditions, and its form can be structured as: "Event A → Event B → Event C, with a probability of P", where the logical relationship between events and the causal chain are determined by the model, and the probability P is calculated by the model.

[0115] Finally, the pre-warning event chains output by this step not only serve as input for subsequent comprehensive capability demand analysis, but can also be directly used for early warning and intervention decision-making at the emergency command center. The output results can include event sequences, probabilities of occurrence, prediction time periods, potential impact areas, and recommended response levels. The system supports outputting pre-warning event chains in structured data formats (such as JSON or XML) for other systems to call, and can also display event evolution paths and prediction results through a visual interface to assist on-duty personnel in making more effective response strategies and personnel configuration adjustments.

[0116] In summary, step S306 achieves the prediction of potential event chains and their probabilities of occurrence within a specific future time window by inputting current situation information and quantified uncertainty into a structured emergency knowledge model, combining probability reasoning and event evolution modeling, thereby providing key support for proactive response and dynamic scheduling optimization.

[0117] In some preferred embodiments, specifically, acquiring the environmental information and the operation information in the emergency command site can be receiving data streams in real time or at a fixed time from various channels such as temperature sensors, smoke sensors, monitoring systems (such as video monitoring, access control systems), equipment management systems (such as SCADA systems), personnel positioning systems, and manual input interfaces, etc. by the data acquisition module. Detecting whether there is conflict data between the environmental information and the operation information can be achieved by the data verification rule library and the conflict detection algorithm, for example, defining the difference threshold of different temperature sensor readings in the same area, the logical relationship between the equipment operation state and the energy consumption data, the consistency of personnel location information and access control records, etc. When the received data violates these rules, it is determined that there is conflict data. If there is conflict data, processing the conflict data according to the preset conflict processing strategy can be executed by calling the corresponding data processing module, for example, for numerical conflict, the weighted average value of multiple data sources can be used as the final value, and the weight can be preset according to the reliability of the data source; for logical conflict, the exception can be marked and the manual review process can be triggered, or the data source with higher credibility can be selected according to the preset priority. If there is no conflict data, processing the environmental information and the operation information to generate the current situation information can be achieved by the data integration and analysis module, for example, associating and aggregating different types of data to generate comprehensive indicators reflecting the regional safety state, equipment health status, personnel distribution heat map, etc. to form a structured situation information report. Identifying the uncertainty description in the current situation information and converting it into a quantitative parameter can be achieved by natural language processing technology to identify fuzzy words (such as "may", "probably", "partly") in the text description or by statistical analysis to calculate the confidence of the data, and then these uncertainties are converted into quantitative forms such as probability values or confidence intervals. Based on the current situation information and the quantitative parameter, the future chain of events and its occurrence probability within a specific time window are predicted using the emergency knowledge model to generate a warning event chain, which can be achieved by inputting the processed current situation information and the quantitative parameter into the reasoning engine based on the emergency knowledge, the engine according to the event evolution path and the mutual influence relationship defined in the knowledge graph, combined with the input uncertainty parameters, running simulation or reasoning algorithms (such as Bayesian inference, Monte Carlo simulation), predicting the event sequence that may occur in the future within a specific time period (such as 1 hour in the future, 3 hours in the future) and its corresponding occurrence probability, and finally outputting a warning list containing multiple predicted event chains and their probabilities.

[0118] Optionally, in combination with Figure 3 As shown in FIG. 3, before step S302 of detecting whether there is conflict data between the environmental information and the operation information, the method further includes:

[0119] S3011, identifying the reliability indicators and time stamps of the environmental information and the operation information;

[0120] S3012, performing validity evaluation based on the reliability index and the time stamp, obtaining the environment information and the operation information whose validity has been evaluated;

[0121] S3013, updating the environment information whose validity has been evaluated to the environment information; updating the operation information whose validity has been evaluated to the operation information.

[0122] Wherein, the reliability index refers to a parameter for measuring the credibility or measurement accuracy of the environment information and the operation information, which can be a numerical value, a level or a confidence interval, and its purpose is to quantify the quality of information; the time stamp refers to a mark recording the time when the environment information and the operation information are generated or collected, which can be a time value accurate to seconds or milliseconds, and its purpose is to reflect the timeliness of information; the validity evaluation refers to the process of screening and judging the environment information and the operation information according to the reliability index and the time stamp, which can be marking the information below the reliability threshold or earlier than the time threshold as invalid by setting the reliability threshold and the time threshold, and its purpose is to identify and eliminate unreliable or outdated information; the environment information and the operation information whose validity has been evaluated refer to the information set judged as valid after the validity evaluation process, and its purpose is to provide high-quality data for subsequent processing; updating to the environment information and the operation information refers to replacing the original environment information and the operation information with the environment information and the operation information whose validity has been evaluated, and its purpose is to ensure that the data used in the subsequent steps is verified and screened.

[0123] In some preferred embodiments, specifically, the environment information and the operation information can be obtained from different sensors, systems or manual inputs, and the reliability index and the time stamp associated with these information are obtained at the same time. For example, a temperature sensor reading can be accompanied by a reliability index indicating its calibration status or measurement accuracy, and a time stamp recording the time when the reading occurs. Then, a reliability threshold can be set, for example, a reliability index below a certain value indicates that the data is not reliable; at the same time, a time threshold can be set, for example, a time stamp earlier than the current time by a certain interval (such as 5 minutes) indicates that the data has expired. Based on these thresholds, each piece of environment information and operation information obtained is checked, and if the reliability index of an information is below the threshold or the time stamp is earlier than the threshold, the information is considered invalid. All the information judged as valid is collected to form the environment information and the operation information whose validity has been evaluated. Finally, this set of environment information and operation information whose validity has been evaluated is used to replace the original environment information and operation information, ensuring that the data used in the subsequent conflict detection and situation generation steps is quality controlled.

[0124] Optionally, in combination with Figure 4 As shown in FIG. 4, step S4: based on the early warning event chain, the corresponding comprehensive capability demand is analyzed, including:

[0125] S401, receiving an early warning event chain;

[0126] S402, judging whether the early warning event chain is a compound event chain based on the number of professional fields involved in the early warning event chain and the number of disposal measures;

[0127] S403, if it is a compound event chain, the early warning event chain is decomposed into several task units with corresponding professional fields;

[0128] S404, determining the basic ability requirements corresponding to the task units according to the preset knowledge and ability mapping rules;

[0129] S405, based on the logical association between each task unit, identifying the key nodes that need to be mutually coordinated between the task units to obtain collaborative dependency relationship information;

[0130] S406, generating specific collaborative ability requirements according to the collaborative dependency relationship information;

[0131] S407, integrating the basic ability requirements and specific collaborative ability requirements of the task units to generate corresponding comprehensive ability requirements.

[0132] The compound event chain refers to an early warning event chain composed of multiple interrelated events involving different professional fields or requiring multiple disposal measures, which can be determined by analyzing the event node types, association relationships and required disposal action types contained in the event chain. The task unit refers to the smallest processing unit obtained by decomposing a complex early warning event chain, which has a clear target and required professional field, which can be defined and decomposed based on the event chain structure, professional field division or disposal process node. The knowledge and ability mapping rules refer to a set of corresponding relationships pre-set for converting specific knowledge or task requirements into specific ability requirements, which can be stored and applied using a database, rule set or ontology model. The collaborative dependency relationship information refers to relationship data describing the need for mutual cooperation, information sharing or resource coordination between different task units during execution, which can be represented using a graph structure, matrix or list. The specific collaborative ability requirement refers to the non-basic ability requirements such as communication, coordination and team collaboration required to complete the task units with collaborative dependency relationship, which can be determined using an ability model, behavior index or historical data analysis.

[0133] In some preferred embodiments, for example, when the system receives a pre-warning event chain describing a situation that a certain area has power facility failure and local fire due to extreme weather, the system first analyzes the event chain and finds that it involves multiple professional fields such as power engineering and fire rescue, and requires multiple disposal measures such as power repair and fire disposal, and thus judges it as a composite event chain. The system then decomposes the event chain into a "power facility repair" task unit and a "fire disposal" task unit, where the "power facility repair" task unit corresponds to the power engineering professional field and the "fire disposal" task unit corresponds to the fire rescue professional field. According to the preset knowledge and ability mapping rules, the basic ability requirements of the "power facility repair" task unit may include high-voltage operation skills, fault diagnosis skills, etc., and the basic ability requirements of the "fire disposal" task unit may include fire extinguishing skills, emergency medical skills, etc. At the same time, the system analyzes and finds that the execution of the "power facility repair" task unit needs to ensure the safety of the scene, and the progress of the "fire disposal" task unit will affect the safety of the scene, and there is a logical association between them, and the key nodes that need to be cooperated between them are identified, and the coordination dependency relationship information is obtained. According to the coordination dependency relationship information, specific coordination ability requirements are generated, for example, efficient communication and coordination ability and cross-department collaboration ability are required between the two task units. Finally, the system integrates the basic ability requirements of the "power facility repair" and "fire disposal" task units and the generated specific coordination ability requirements to generate comprehensive ability requirements required to deal with the composite pre-warning event chain, which will guide the subsequent team personnel configuration.

[0134] Optionally, in combination with Figure 5 As shown in FIG. 4, step S405: based on the logical association between each task unit, the key nodes that need to be cooperated between the task units are identified, and the coordination dependency relationship information is obtained, including:

[0135] S4051, based on the logical association between the task units and the event evolution relationship defined in the emergency knowledge model, the key nodes that need to be cooperated are identified, and the first coordination dependency relationship is obtained;

[0136] S4052, obtaining resource requirement information, environmental condition information, state output information and state input information corresponding to the task unit;

[0137] S4053, according to the resource requirement information, the environmental condition information, the state output information and the state input information, whether there is a coordination requirement or an indirect causal chain between the task units is identified, if there is, it is determined that there is a second coordination dependency relationship between the corresponding two task units;

[0138] S4054, integrating the first coordination dependency relationship and the second coordination dependency relationship to generate the coordination dependency relationship information.

[0139] wherein the first coordination dependency refers to direct cooperation requirements determined based on logical order or causal relationships of task units in the emergency event chain, which can be achieved by analyzing the event evolution path and explicit associations between task units defined in the emergency knowledge model, and aims to identify the sequence or parallel relationship of task execution; the resource requirement information refers to the need of material, equipment, tool or personnel resources in the execution process of the task unit, which can be obtained by consulting the standard operating procedures or plans corresponding to the task unit, and aims to reflect the dependence of task execution on external resources; the environmental condition information refers to the physical environment, climate condition or external situation and other factors when the task unit is executed, which can be obtained by real-time monitoring or querying environmental sensor data, and aims to reflect the influence of environmental factors on task execution; the state output information refers to the key results, intermediate states or feedback signals generated after the task unit is executed, which can be obtained by log records or manual reports of the task execution system, and aims to reflect the progress and results of task execution; the state input information refers to specific conditions or preconditions that need to be met for the task unit to start or continue execution, which can be obtained by the start condition definition or flow specification of the task unit, and aims to reflect the prerequisite conditions of task execution; the second coordination dependency refers to indirect cooperation requirements or causal relationships formed based on the demand of task units for shared resources, the influence of common environmental factors or through state transmission, which can be achieved by analyzing the resource allocation, environmental impact assessment or state information flow path, and aims to supplement the implicit or indirect coordination requirements not covered by the first coordination dependency.

[0140] In some preferred embodiments, for example, it is assumed that the early warning event chain contains task unit A and task unit B. First, based on the emergency knowledge model, it is found that task unit A must be completed before task unit B, which is a logical sequence, based on which it is identified that there is a first collaborative dependency relationship between task unit A and task unit B. At the same time, it is obtained that task unit A needs to use a specific device X (resource requirement information), and task unit B also needs to use device X. It is obtained that task unit A needs to be executed in a high-temperature environment (environment condition information), and task unit B also needs to be executed in a high-temperature environment. It is obtained that task unit A outputs a "device ready" signal after execution (state output information), and task unit B needs to receive the "device ready" signal to start (state input information). According to this information, it is identified that there is a collaborative requirement between task unit A and task unit B based on the shared resource device X, there is a collaborative requirement based on the common environment of high temperature, and there is an indirect causal chain based on the state output and the state input, based on which it is determined that there is a second collaborative dependency relationship between task unit A and task unit B. Finally, the first collaborative dependency relationship and the second collaborative dependency relationship are integrated to obtain comprehensive collaborative dependency relationship information between task unit A and task unit B, which not only contains the logical sequence, but also contains the mutual dependence of resources, environment and state.

[0141] Optionally, in combination with Figure 6 As shown in FIG. 4, step S4053: According to the resource requirement information, the environment condition information, the state output information and the state input information, it is identified whether there is a collaborative requirement or an indirect causal chain between the task units, if there is, it is determined that there is a second collaborative dependency relationship between the corresponding two task units, including:

[0142] S40531, based on the resource requirement information, shared resource identification is performed, if there is a shared resource between the task units, it means that there is a collaborative requirement, and it is determined that there is a second collaborative dependency relationship between the task units;

[0143] S40532, based on the environment condition information and the current environment state information, common environment factor identification is performed, if there is a common environment factor between the task units, it means that there is a collaborative requirement, and it is determined that there is a second collaborative dependency relationship between the task units;

[0144] S40533, based on the state output information and the state input information, state output and state input association identification is performed, if the state output of one task unit constitutes the state input of another task unit, it means that there is an indirect causal chain, and it is determined that there is a second collaborative dependency relationship between the task units.

[0145] The shared resource identification refers to judging whether multiple task units need to use the same resource at the same time or in a similar time period by analyzing specific demand lists of the resource of different task units, which can be realized by comparing the resource identifiers or resource types in the resource demand lists of the task units, and the purpose is to find the potential collaborative demand caused by resource competition or sharing. The common environmental factor identification refers to judging whether multiple task units are simultaneously affected by the same environmental factor by analyzing the dependence or sensitivity of the task units on environmental conditions and combining the current actual environmental state information, which can be realized by comparing the required environmental conditions of the task units with the current environmental state information and identifying the common influencing factors, and the purpose is to find the collaborative demand that needs to be coped with due to external environmental changes. The state output and state input association identification refers to judging whether the state information generated in the completion or execution process of a task unit is the necessary input condition for another task unit to start execution or proceed smoothly by analyzing the state information, which can be realized by establishing a state flow chart or a dependency relationship diagram of the task units, and the purpose is to find the implicit causal chain or pre-post relationship between the task units.

[0146] In some preferred embodiments, the specific implementation is as follows: the system first obtains the resource demand information of all task units to be analyzed, for example, task A needs "device X" and "consumable Y", and task B needs "device X" and "personnel Z". By comparison, it is found that task A and task B both need "device X", so "device X" is identified as a shared resource, and it is determined that there is a second collaborative dependency relationship between task A and task B. Then, the system obtains the environmental condition information of the task units and the current environmental state information, for example, task C needs to be executed "outdoors", task D also needs to be executed "outdoors", and the current environmental state information is "heavy rain". By comparison, it is found that task C and task D are both affected by the common environmental factor of "heavy rain", so it is determined that there is a second collaborative dependency relationship between task C and task D. Subsequently, the system obtains the state output information and the state input information of the task units, for example, the state output information of task E is "system initialization is completed", and the state input information of task F requires "system initialization is completed". By comparison, it is found that the state output of task E constitutes the state input of task F, so it is determined that there is a second collaborative dependency relationship between task E and task F. Through the identification of the above three aspects, a list of second collaborative dependency relationships between the task units can be obtained.

[0147] Optionally, in combination with Figure 7 As shown in FIG. 4, the step of performing the state output and state input association identification based on the state output information and the state input information in step S40533 includes:

[0148] S405331, based on two task units, obtaining state output information of one task unit and state input information of another task unit;

[0149] S405332, extracting a first uncertainty parameter in the state output information and a second uncertainty parameter in the state input information;

[0150] S405333, based on the first uncertainty parameter and the second uncertainty parameter, determining whether the uncertainty of the state output information is within an uncertainty acceptance range corresponding to the second uncertainty parameter, and generating a determination result;

[0151] S405334, according to the determination result, determining whether the state output of one task unit constitutes the state input of another task unit.

[0152] Wherein, the state output information refers to the data generated after one task unit completes its function, reflecting its internal state or external influence, which can include device operating parameters, processing results, environmental perception data, etc., and its purpose is to provide a basis for the current state for other task units; the state input information refers to the data that one task unit needs to receive to execute its function, reflecting the state of other task units or external environment, which can include device state requirements, processing conditions, environmental parameter thresholds, etc., and its purpose is to enable the task unit to adjust its behavior according to the external state; the first uncertainty parameter refers to the numerical value or index associated with the state output information, quantifying the reliability or accuracy of the information, which can take the form of confidence interval, probability distribution, fuzzy measure or error range, etc., and its purpose is to represent the inherent uncertainty of the state output information itself; the second uncertainty parameter refers to the numerical value or index associated with the state input information, quantifying the tolerance of the task unit receiving the state input to uncertainty, which can take the form of maximum allowable error, minimum confidence threshold or acceptable fuzzy range, etc., and its purpose is to define the reliability requirements that the state input information needs to meet; the uncertainty acceptance range refers to the interval or threshold of the uncertainty degree that the state input information can accept, which is defined by the second uncertainty parameter, and its purpose is to determine whether the uncertainty of the state output information meets the reliability requirements of the state input; the determination result refers to the judgment conclusion about whether the uncertainty of the state output information is within the uncertainty acceptance range based on the comparison of the first uncertainty parameter and the second uncertainty parameter, which can be a Boolean value or a numerical value indicating the degree of compliance, and its purpose is to provide a basis for subsequent determination of state transmission relationship.

[0153] Optionally, in combination with Figure 8As shown, step S5: obtaining the availability, individual ability characteristics and state information of personnel, and matching according to the comprehensive ability demand to generate a team personnel configuration scheme that meets the comprehensive ability demand, including:

[0154] S501, decompose the predetermined overall response performance in the early warning event chain into several emergency response sub-goals;

[0155] S502, taking the availability, individual ability characteristics, state information of personnel and specific time window as the constraint conditions of optimization calculation;

[0156] S503, based on the emergency response sub-goal and the constraint condition, through iterative processing, iteratively output a team personnel configuration scheme that meets several emergency response sub-goals.

[0157] Wherein, the predetermined overall response performance in the early warning event chain refers to the overall effect or performance level expected to be achieved when responding to a specific early warning event chain, which can be measured by a series of key performance indicators (KPI) or comprehensive evaluation scores; the emergency response sub-goal refers to a more specific and measurable small goal obtained by decomposing the overall response performance, which can be determined by analyzing each link or task unit of the early warning event chain, for example, it can include information collection completion time, field control range, personnel evacuation efficiency, etc.; the constraint condition of optimization calculation refers to the limitation condition that must be met when calculating the personnel configuration scheme, which can be obtained and set by reading the personnel database, sensor data, scheduling rules and other information; iterative processing refers to an algorithm that gradually approaches the optimal solution or a solution that meets the conditions through repeated calculation process, which can be implemented by using optimization algorithms such as genetic algorithm, simulated annealing algorithm, tabu search algorithm or rule-based heuristic search; the team personnel configuration scheme refers to the specific team members and their division arrangement determined for responding to the early warning event chain, which can be a structured data set containing personnel ID, task allocation, time arrangement and other information.

[0158] In some preferred embodiments, the following is illustrated by a specific example. Assume that there is a pre-warning event chain predicting that an event of "device overheating causing local fire" is likely to occur within a specific time window in the future. The predetermined overall response effectiveness of this pre-warning event chain can be decomposed into several emergency response sub-goals, such as "locating the fire source within 5 minutes", "starting initial fire extinguishing measures within 15 minutes", and "completing the isolation of the affected area within 30 minutes". At the same time, the system obtains the availability information of multiple personnel (e.g., Zhang San is available, Li Si is available, and Wang Wu is unavailable), individual ability characteristics (e.g., Zhang San is proficient in device structure and initial fire extinguishing, Li Si is good at on-site isolation and personnel evacuation, and Wang Wu is proficient in device maintenance), state information (e.g., Zhang San is in good condition, and Li Si is slightly fatigued), and the specific time window for responding to this event (e.g., it is required to control the spread of the fire within 1 hour after the event occurs). The system inputs these sub-goals and constraints into the optimization calculation module and processes them iteratively. During the iteration process, the system can generate a team personnel configuration scheme to be evaluated, such as a team consisting of Zhang San and Li Si, with Zhang San responsible for locating the fire source and initial fire extinguishing, and Li Si responsible for area isolation. The system evaluates the degree of satisfaction of this scheme for the three sub-goals, such as assessing the likelihood of Zhang San locating the fire source within 5 minutes and starting initial fire extinguishing within 15 minutes based on his ability and state, and assessing the likelihood of Li Si completing the area isolation within 30 minutes based on his ability and state. If the evaluation result shows that this scheme can satisfy all sub-goals and comply with the constraints of personnel availability, state, and overall time window, it is output as a team personnel configuration scheme that meets the conditions. If it fails to meet the conditions, the system adjusts the scheme based on the evaluation result, such as considering adding other available personnel or adjusting task allocation, and then proceeds to the next iteration until a scheme that satisfies all sub-goals is found.

[0159] Optionally, in combination with Figure 9 As shown in FIG. 5, step S503: based on the emergency response sub-goals and constraints, a team personnel configuration scheme that satisfies all emergency response sub-goals is iteratively generated through iterative processing, including:

[0160] S5031, defining the weight relationship between the several emergency response sub-goals;

[0161] Based on the emergency response sub-goals and constraints, the iterative processing includes:

[0162] S5032, generating a team personnel configuration scheme to be evaluated;

[0163] S5033, evaluating the degree of satisfaction of the team personnel configuration scheme to be evaluated for all emergency response sub-goals;

[0164] S5034, according to the weight relationship and the satisfaction degree, calculating the comprehensive evaluation value of the team staffing scheme to be evaluated;

[0165] S5035, according to the comprehensive evaluation value, performing the iteration process according to the constraint condition;

[0166] S5036, until the team staffing scheme satisfying all the emergency response sub-targets is iterated out.

[0167] The weight relationship refers to a numerical value or rule set for measuring the relative importance of different emergency response sub-targets in the overall scheme evaluation, which can be realized by assigning a numerical weight to each sub-target. The purpose is to distinguish the priority of different sub-targets when evaluating the team staffing scheme, so that the more important sub-targets have a greater impact on the final evaluation result. The satisfaction degree refers to a quantitative representation of how much a team staffing scheme meets the requirements or expected level of a certain or a group of emergency response sub-targets. It can be realized by scoring or grading the performance of the scheme on a specific sub-target based on pre-set evaluation indicators and calculation methods. The purpose is to provide basic data for subsequent comprehensive evaluation. The comprehensive evaluation value refers to a single numerical value or indicator that measures the overall performance of a team staffing scheme by combining the satisfaction degree of each emergency response sub-target and its corresponding weight. It can be calculated by weighted summation, weighted average or other multi-objective evaluation algorithms. The purpose is to provide a unified standard for comparing the overall performance of different schemes and guiding the optimization direction of the iteration process.

[0168] In some preferred embodiments, the implementation is as follows. First, the several emergency response sub-goals obtained by decomposing the early warning event chain, such as "fast response time", "key skill coverage", "personnel fatigue control", etc., can be defined according to the emergency plan, historical event data or expert experience, and the weight relationship between them can be defined. For example, in a certain high-risk period, the weight of "fast response time" can be set to a higher value, and the weight of "personnel fatigue control" can be set to a relatively lower value. Then, based on these sub-goals and personnel constraints, an iterative process is started. In each iteration, a team staffing plan to be evaluated can be generated first, for example, randomly selected from available personnel, or fine-tuned based on the better-performing plan in the previous round. Next, the degree of satisfaction of the plan for each sub-goal is evaluated, for example, the estimated time for the team to arrive at the scene is calculated as the degree of satisfaction of "fast response time", it is checked whether the team members cover all the necessary key skills as the degree of satisfaction of "key skill coverage", and the continuous working hours of the team members are calculated as the degree of satisfaction of "personnel fatigue control". Then, according to the preset weight relationship, the weighted sum of these satisfaction degrees is calculated to obtain the comprehensive evaluation value of the plan. Finally, according to the comprehensive evaluation value, if the comprehensive evaluation value does not reach the preset threshold, the current plan is adjusted according to the constraints, and a new plan to be evaluated is generated by using an optimization algorithm (such as selection, crossover, and mutation operations of genetic algorithm), and the next iteration is entered. This process continues until a team staffing plan that meets all the emergency response sub-goals is found, for example, the satisfaction degree scores of all sub-goals reach the preset pass line.

[0169] An artificial intelligence-based emergency on-duty scheduling optimization system for performing an artificial intelligence-based emergency on-duty scheduling optimization process, in combination with Figure 10 as shown, comprising:

[0170] An emergency knowledge acquisition module for acquiring emergency knowledge from an emergency command center;

[0171] An emergency knowledge model module for constructing an emergency knowledge model based on the emergency knowledge;

[0172] An event chain generation module for acquiring input environmental information and operational information, predicting a chain of interlocking events and the likelihood of occurrence within a specific time window in the future using the emergency knowledge model, and generating a pre-warning event chain;

[0173] A capability demand generation module for analyzing corresponding comprehensive capability demands based on the pre-warning event chain;

[0174] The personnel scheme generation module is configured to acquire availability, individual ability characteristics and state information of personnel, and match according to comprehensive ability requirements to generate a team personnel configuration scheme meeting the comprehensive ability requirements.

[0175] The emergency information acquisition module is configured to continuously acquire early warning information, personnel state feedback and event progress in the emergency command process.

[0176] The personnel scheme adjustment module is configured to dynamically feedback and adjust the team personnel configuration scheme according to the early warning information, personnel state feedback and event progress.

[0177] The emergency knowledge acquisition module is a unit for receiving and processing emergency-related information from external sources, which can be a data interface module or a data acquisition and processing unit. Its purpose is to provide basic emergency knowledge data for the system. The emergency knowledge model module is a unit for converting emergency knowledge into a computable form, which can be a rule-based reasoning engine or a knowledge graph database. Its purpose is to provide structured knowledge for event prediction and ability analysis. The event chain generation module is a unit for predicting potential event occurrence and its evolution path, which can be a prediction service based on a probability model or a simulation and deduction engine. Its purpose is to identify future possible risks. The ability requirement generation module is a unit for converting the predicted event risk into specific personnel ability requirements, which can be an analysis service based on rule matching or a machine learning model. Its purpose is to clarify the personnel ability composition required to cope with risks. The personnel scheme generation module is a unit for generating specific shift arrangements according to ability requirements and personnel information, which can be a solver based on optimization algorithms or an intelligent matching engine. Its purpose is to generate team configurations meeting the requirements. The emergency information acquisition module is a unit for continuously receiving and processing real-time information in the emergency process, which can be a message bus or a data stream processing unit. Its purpose is to provide real-time data required for dynamic adjustment for the system. The personnel scheme adjustment module is a unit for correcting the existing shift arrangement according to real-time information, which can be a feedback control unit or a rapid re-optimization service. Its purpose is to adapt the shift arrangement to changing conditions.

[0178] The scheme of the present application converts each step in the artificial intelligence-based emergency on-duty scheduling optimization method into a specific system functional unit by setting an emergency knowledge acquisition module, an emergency knowledge model module, an event chain generation module, a capability requirement generation module, a personnel scheme generation module, an emergency information acquisition module, and a personnel scheme adjustment module. The emergency knowledge acquisition module and the emergency knowledge model module work together to provide the system with basic knowledge and reasoning capability. The event chain generation module uses these knowledge and real-time information to predict potential risks. The capability requirement generation module converts the risks into specific capability requirements. The personnel scheme generation module generates a preliminary scheme according to the capability requirements and personnel information. The emergency information acquisition module continuously monitors the on-site situation, and the personnel scheme adjustment module makes real-time corrections to the scheme according to the feedback information. This modular system architecture makes it possible to decompose and implement the complex method process, and the data flow and functional cooperation between the modules ensure that the generation and dynamic adjustment of the scheduling scheme can be automated, thereby solving the problem that the method is difficult to be practically applied only by relying on the method description, and improving the intelligent level and response efficiency of the emergency scheduling.

[0179] The above merely describes the embodiments of the present application and is not used to limit the protection scope of the present application. For those skilled in the art, the present application can have various modifications and changes. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. An emergency duty scheduling optimization method based on artificial intelligence, characterized in that, include: Acquire emergency knowledge from the emergency command center; this emergency knowledge includes risk sources, event evolution paths, and interrelationships. Construct an emergency knowledge model based on the aforementioned emergency knowledge; The system acquires input environmental and operational information, utilizes the emergency knowledge model to predict the chain of events and their probability of occurrence within a specific future time window, and generates an early warning event chain. Based on the aforementioned early warning event chain, the corresponding comprehensive capability requirements are analyzed. These comprehensive capability requirements include specific skill combinations required to cope with the early warning event chain, familiarity with emergency plans, and the required number of personnel. In the process of analyzing the corresponding comprehensive capability requirements, collaborative dependency information needs to be obtained. This collaborative dependency information includes a first collaborative dependency and a second collaborative dependency. The first collaborative dependency refers to the direct cooperation requirement determined based on the logical order or causal relationship of task units in the emergency event chain. The second collaborative dependency refers to the indirect cooperation requirement or causal relationship formed based on the task units' need for shared resources, the influence of common environmental factors, or through state transmission. The availability, individual ability characteristics, and status information of personnel are obtained, and matched according to the comprehensive ability requirements to generate a team personnel configuration plan that meets the comprehensive ability requirements. Continuously acquire early warning information, personnel status feedback, and event progress during the emergency command process; The team personnel configuration plan is dynamically adjusted based on the early warning information, personnel status feedback, and event progress.

2. The emergency duty scheduling optimization method based on artificial intelligence as described in claim 1, characterized in that, The steps of acquiring input environmental and operational information, using the emergency knowledge model, predicting the chain of events and their probability of occurrence within a specific future time window, and generating an early warning event chain include: Acquire environmental and operational information from emergency command sites; Detect whether there are conflicting data between the environmental information and the operational information; If conflicting data exists, the conflicting data will be processed according to a preset conflict handling strategy. If no conflicting data exists, the environmental and operational information is processed to generate current situation information; Identify the uncertainty descriptions in the current situation information and convert them into quantitative parameters; Based on the current situation information and the quantitative parameters, the emergency knowledge model is used to predict the chain of events and their probability of occurrence within a specific future time window, and the early warning event chain is generated.

3. The emergency duty scheduling optimization method based on artificial intelligence as described in claim 2, characterized in that, Before the step of detecting whether there is conflicting data between the environmental information and the operational information, the method further includes: Reliability indicators and timestamps for identifying the environmental and operational information; Based on the reliability index and timestamp, an effectiveness assessment is performed to obtain environmental and operational information on which the effectiveness has been assessed. Update the environmental information to reflect the effectiveness of the assessment; update the operational information to reflect the effectiveness of the assessment.

4. The emergency duty scheduling optimization method based on artificial intelligence as described in claim 1, characterized in that, The step of analyzing the corresponding comprehensive capability requirements based on the early warning event chain and collaborative dependency information includes: Receive the aforementioned warning event chain; Based on the number of professional fields of knowledge involved and the number of response measures in the early warning event chain, determine whether the early warning event chain is a composite event chain; If it is a complex event chain, the early warning event chain is decomposed into several task units with corresponding professional fields; Based on the preset knowledge and ability mapping rules, the basic ability requirements corresponding to the task unit are determined; Based on the logical connections between the various task units, the key nodes that require mutual cooperation between the task units are identified, and the collaborative dependency information is obtained. Based on the collaborative dependency information, specific collaborative capability requirements are generated; The basic capability requirements and specific collaborative capability requirements of the task units are integrated to generate corresponding comprehensive capability requirements.

5. The emergency duty scheduling optimization method based on artificial intelligence as described in claim 4, characterized in that, The step of identifying key nodes requiring cooperation between task units based on the logical relationships between them, and obtaining the collaborative dependency information, includes: Based on the logical connections between the task units and the event evolution relationships defined in the emergency knowledge model, key nodes that need to cooperate with each other are identified, and the first collaborative dependency relationship is obtained. Obtain the resource requirement information, environmental condition information, status output information, and status input information corresponding to the task unit; Based on the resource demand information, environmental condition information, status output information, and status input information, identify whether there is a collaborative requirement or indirect causal chain between the task units. If so, determine that there is a second collaborative dependency relationship between the two corresponding task units. Integrate the first collaborative dependency and the second collaborative dependency to generate collaborative dependency information.

6. The emergency duty scheduling optimization method based on artificial intelligence as described in claim 5, characterized in that, The step of identifying whether there is a collaborative requirement or indirect causal chain between the task units based on the resource demand information, environmental condition information, status output information, and status input information, and determining that there is a second collaborative dependency relationship between the corresponding two task units if such a chain exists, includes: Based on the resource requirement information, shared resources are identified. If there are shared resources among the task units, it indicates that there is a collaborative requirement, and a second collaborative dependency relationship is determined to exist among the task units. Based on the environmental condition information and the current environmental state information, common environmental factors are identified. If there are common environmental factors among the task units, it indicates that there is a need for collaboration, and it is determined that there is a second collaborative dependency relationship among the task units. Based on the state output information and the state input information, the association between state output and state input is identified. If the state output of one task unit constitutes the state input of another task unit, it indicates that there is an indirect causal chain, and it is determined that there is a second cooperative dependency relationship between the task units.

7. The emergency duty scheduling optimization method based on artificial intelligence as described in claim 6, characterized in that, The step of associating and identifying the state output with the state input based on the state output information and the state input information includes: Based on the two task units, obtain the status output information of one task unit and the status input information of the other task unit; Extract the first uncertainty parameter from the state output information, and extract the second uncertainty parameter from the state input information; Based on the first uncertainty parameter and the second uncertainty parameter, determine whether the uncertainty of the state output information is within the uncertainty acceptance range corresponding to the second uncertainty parameter, and generate a determination result. Based on the determination result, it is determined whether there is a situation where the state output of one task unit constitutes the state input of another task unit.

8. The emergency duty scheduling optimization method based on artificial intelligence as described in claim 1, characterized in that, The steps of acquiring personnel availability, individual ability characteristics, and status information, and matching them according to the comprehensive ability requirements to generate a team personnel configuration plan that meets the comprehensive ability requirements, include: The predetermined overall response effectiveness in the aforementioned early warning event chain is decomposed into several emergency response sub-objectives; The availability of the personnel, their individual ability characteristics, status information, and the specific time window are used as constraints for the optimization calculation. Based on the emergency response sub-objectives and the constraints, through iterative processing, several team personnel configuration schemes that satisfy all of the emergency response sub-objectives are obtained.

9. The emergency duty scheduling optimization method based on artificial intelligence as described in claim 8, characterized in that, The step of iteratively generating a team personnel configuration scheme that satisfies several of the emergency response sub-objectives through iterative processing based on the emergency response sub-objectives and the constraints includes: Define the weight relationships among several of the aforementioned emergency response sub-objectives; Based on the aforementioned emergency response sub-objectives and the aforementioned constraints, an iterative process is employed, the iterative process including: Generate the team staffing plan to be evaluated; Assess the degree to which the personnel configuration plan of the team to be evaluated meets all the aforementioned emergency response sub-objectives; Based on the weighting relationship and the degree of satisfaction, calculate the comprehensive evaluation value of the team personnel allocation plan to be evaluated; Based on the comprehensive evaluation value, the iterative processing procedure is performed according to the constraints. The process continues until a team personnel configuration scheme that satisfies all the aforementioned emergency response sub-objectives is found.

10. An artificial intelligence-based emergency duty scheduling optimization system, used to execute the artificial intelligence-based emergency duty scheduling optimization method according to any one of claims 1-9, characterized in that, include: The emergency knowledge acquisition module is used to acquire emergency knowledge from the emergency command center. An emergency knowledge model module is used to construct an emergency knowledge model based on the emergency knowledge. The event chain generation module is used to acquire input environmental and operational information, and using the emergency knowledge model, predict the chain of events and their probability of occurrence within a specific future time window, and generate an early warning event chain. The capability requirement generation module is used to analyze the corresponding comprehensive capability requirements based on the early warning event chain. The personnel plan generation module is used to obtain personnel availability, individual ability characteristics and status information, and match them according to the comprehensive ability requirements to generate a team personnel configuration plan that meets the comprehensive ability requirements. The emergency information acquisition module is used to continuously acquire early warning information, personnel status feedback, and event progress during the emergency command process; The personnel plan adjustment module is used to dynamically adjust the team personnel configuration plan based on the early warning information, personnel status feedback, and event progress.

Images