Intelligent airport emergency rescue resource allocation method and platform, storage medium and program product

Abstract

The invention provides an airport emergency rescue resource intelligent allocation method and platform, a storage medium and a program product, and relates to the technical field of emergency rescue resource allocation, and the method comprises the steps: carrying out the communication subnet division of all rescue departments according to communication network data under the condition of obtaining the communication network data between all rescue departments, dividing the target rescue departments meeting the preset division condition into the same communication sub-network; configuring a resource allocation center for each communication sub-network, and establishing a resource sharing cache node between the two communication sub-networks meeting a preset cache establishment condition; when the communication interruption is detected, performing communication link detection among the communication subnetworks to obtain a link interruption detection result; and triggering a network interruption emergency mode of the resource allocation center according to a link interruption detection result. The technical problem that cross-department resource coordination and allocation are difficult to carry out under the condition that a communication network is interrupted in the prior art is solved.

Description

Technical Field

[0001] This application relates to the field of emergency rescue resource allocation technology, and in particular to an intelligent allocation method, platform, storage medium and program product for airport emergency rescue resources. Background Technology

[0002] With the rapid development of the civil aviation industry and the continuous expansion of airport operations, airport emergency rescue systems have become a crucial link in ensuring aviation safety. In the face of emergencies such as flight accidents, natural disasters, and public health incidents, multiple rescue departments, including fire services, medical services, security, and ground support, need to work together. Therefore, it is imperative to establish an efficient and reliable resource allocation mechanism to optimize the allocation of emergency rescue resources and ensure rapid response.

[0003] In related technologies, a centralized resource allocation approach based on a central dispatch system is typically adopted. In practice, a central dispatch server is first deployed in the airport emergency command center. This server establishes a real-time connection with the terminal equipment of each rescue department via wired or wireless communication networks. Then, each rescue department periodically reports its resource status information, including personnel numbers, equipment types, and material reserves, to the central dispatch server through its terminal equipment. When an emergency occurs, the central dispatch server receives the event information, calculates a resource allocation plan based on a preset dispatch algorithm, and issues dispatch instructions to each rescue department through the communication network. Upon receiving the instructions, each rescue department executes the rescue mission according to the unified arrangements of the central dispatch server and provides real-time feedback on the mission execution status. Throughout the allocation process, the central dispatch server continuously monitors the resource usage of each department and dynamically adjusts resource allocation based on the on-site situation. This centralized allocation method enables unified management and coordination of resources, but the operation of the entire system is highly dependent on the communication network connection between the central dispatch server and each rescue department.

[0004] However, with the above-mentioned centralized resource allocation method, once the communication network is interrupted, the central dispatch server cannot obtain the real-time resource status of each rescue department, and the information exchange channels between the rescue departments are also lost, causing the entire allocation system to be paralyzed. Each department can only rely on its own resources to act independently, which makes it difficult to coordinate and allocate cross-departmental resources in the event of a communication network interruption. Summary of the Invention

[0005] This application provides a method, platform, storage medium, and program product for intelligent allocation of airport emergency rescue resources, which can be used to realize cross-departmental resource coordination and allocation in the event of communication network interruption.

[0006] Firstly, this application provides an intelligent allocation method for airport emergency rescue resources, applied to the aforementioned intelligent allocation platform for airport emergency rescue resources. The method includes: upon obtaining communication network data between various rescue departments, dividing each rescue department into communication subnets based on the communication network data, so as to group target rescue departments that meet preset division conditions into the same communication subnet; configuring a resource allocation center for each communication subnet, and establishing a resource-sharing cache node between two communication subnets that meet preset cache establishment conditions. The resource-sharing cache node is used to save cross-subnet resource status data within a preset time period before the communication interruption when communication between two communication subnets is interrupted; upon detecting a communication interruption, performing communication link detection between each communication subnet to obtain a link interruption detection result; and triggering a network outage emergency mode for the resource allocation center based on the link interruption detection result. The network outage emergency mode schedules each resource allocation center based on local resource status data and cross-subnet resource status data.

[0007] By adopting the above technical solution, communication subnets are divided for each rescue department based on communication network data, forming a distributed network architecture. Each communication subnet is equipped with a resource allocation center, enabling independent management capabilities. Resource-sharing cache nodes established between two communication subnets create cross-subnet data bridges within this architecture. The cross-subnet resource status data stored by these cache nodes, along with the local resource status data of each communication subnet, form a complementary two-layer data system. When a communication interruption is detected and a link interruption detection result is obtained, the resource allocation center can combine local resource status data with cached cross-subnet resource status data for scheduling decisions in emergency network outage mode, precisely because the previously established resource-sharing cache nodes have stored the necessary cross-subnet resource status data. This collaborative mechanism of pre-caching and emergency scheduling enables a smooth transition from centralized scheduling to distributed autonomy, ensuring orderly resource allocation even during communication interruptions. This solves the technical problem of difficulty in cross-departmental resource coordination and allocation in the event of a communication network interruption, achieving the technical effect of enabling cross-departmental resource coordination and allocation even under communication network interruption conditions.

[0008] Optionally, after obtaining communication network data between various rescue departments, the communication subnets of each rescue department are divided according to the communication network data to group target rescue departments that meet preset division conditions into the same communication subnet. Specifically, this includes: obtaining communication bandwidth data and communication latency data of the communication links between each rescue department; sorting the communication bandwidth data in descending order to obtain a communication bandwidth sorting list, and sorting the communication latency data in ascending order to obtain a communication latency sorting list; and determining the group of rescue departments corresponding to a group of communication links that are in the top N positions in the communication bandwidth sorting list and in the top M positions in the communication latency sorting list. The system acquires physical location information and area information of a group of rescue departments within the airport. The area information includes at least one of the terminal area, runway area, apron area, and cargo area. Based on the physical location information and area information, the system determines the physical distance between each rescue department in the group and classifies one or more rescue departments whose physical distance is less than a preset distance threshold into a neighboring rescue department group. The system acquires real-time resource status data of the neighboring rescue department group. Based on the real-time resource status data, the system identifies target rescue departments with complementary resource relationships from the neighboring rescue department group and classifies the target rescue departments into the same communication subnet.

[0009] By adopting the above technical solution, the set of rescue departments with the best communication quality is selected by sorting communication bandwidth in descending order and communication latency in ascending order. This provides candidates for high-quality communication support for subsequent geographic location analysis. Based on this, physical location information and responsible area information are obtained for distance calculation, and departments with excellent communication and geographical proximity are aggregated into neighboring rescue department groups. This dual screening ensures that members of the group have both a good communication foundation and are easy to coordinate on the ground. Then, real-time resource status data of neighboring rescue department groups are obtained to identify target rescue departments with complementary resource relationships. Finally, these departments that meet the requirements in three dimensions of communication quality, geographic location and resource complementarity are divided into the same communication subnet. This multi-dimensional progressive screening mechanism makes the final communication subnet not only stable and responsive, but also has internal resource self-balancing capabilities, which greatly improves the overall collaborative efficiency of the communication subnet.

[0010] Optionally, the target rescue departments with complementary resource relationships are determined based on real-time resource status data. This includes: extracting features from the real-time resource status data to obtain resource type vectors and resource quantity vectors for each rescue department; determining the resource complementarity coefficient between any two rescue departments based on the resource type vector, whereby the resource complementarity coefficient characterizes the degree of matching between the surplus resources of one rescue department and the missing resources of the other; determining the resource load rate of each rescue department based on the resource quantity vector; comparing the resource load rate with a preset load threshold to identify high-load rescue departments with a resource load rate greater than the preset load threshold and low-load rescue departments with a resource load rate less than the preset idle threshold; and iterating through and combining rescue departments within neighboring rescue department groups to identify combinations that satisfy a resource complementarity coefficient greater than the preset complementarity threshold and include at least one high-load rescue department and at least one low-load rescue department as target rescue departments with complementary resource relationships.

[0011] By adopting the above technical solution, resource type vectors and resource quantity vectors are obtained through feature extraction. These two vectors characterize resource features from both qualitative and quantitative dimensions. The resource type vector is used to calculate the resource complementarity coefficient, accurately quantifying the degree of resource matching between rescue departments, while the resource quantity vector is used to determine the resource load rate and identify high-load and low-load rescue departments. This dual evaluation mechanism of type matching and load balancing works together. When traversing combinations, it is required that not only the resource complementarity coefficient is greater than the preset complementarity threshold to ensure the complementarity of resource types, but also that the combination must include high-load and low-load departments to ensure the actual needs of resource allocation. This qualitative and quantitative judgment method ensures that the final target rescue departments can not only complement each other in terms of resource types, but more importantly, there is actual allocation space and urgent need in terms of resource quantity, thereby truly achieving precise matching of supply and demand.

[0012] Optionally, based on the link interruption detection results, a network outage emergency mode is triggered for the resource allocation center. The network outage emergency mode schedules each resource allocation center based on local resource status data and cross-subnet resource status data. Specifically, this includes: identifying the communication interruption type based on the link interruption detection results; when the communication interruption type is identified as inter-subnet communication interruption, determining at least two communication subnets where external communication interruption occurred; sending a network outage emergency trigger signal to at least two resource allocation centers corresponding to at least two communication subnets to obtain cross-subnet resource status data read by at least two resource allocation centers from at least one resource shared cache node; constructing a local resource pool based on the local resource status data; and, based on the cross-subnet resource status data... The system constructs a virtual external resource pool. The local resource pool stores information on the available and configurable rescue resources within at least two communication subnets, as well as their availability. The virtual external resource pool stores information on the resource allocation status and occupied resources of other communication subnets that have established caching relationships with at least two communication subnets before communication interruption, obtained from resource-sharing cache nodes. Emergency rescue task information is read from at least two local task queues corresponding to at least two resource allocation centers, and the priority of emergency tasks is determined based on the task type, urgency, and impact scope. Based on the priority of emergency tasks, the system prioritizes scheduling within the local resource pool using available and configurable rescue resource information and their availability.

[0013] By adopting the above technical solution, the link interruption detection results identify the specific fault type of inter-subnet communication interruption, clarify the scope of communication subnets that need to activate the emergency mechanism, and the network outage emergency trigger signal activates the process of the relevant resource allocation center reading cross-subnet resource status data from the resource sharing cache node. This read data and the local resource status data respectively construct a virtual external resource pool and a local resource pool, forming a hierarchical resource view. The resource allocation status and occupied resource information provided by the virtual external resource pool help avoid resource conflicts. Under this dual-pool architecture, the emergency task priority determined according to the task type, urgency, and impact scope guides the resource allocation center to prioritize scheduling within the local resource pool. This hierarchical scheduling mechanism supported by cached data enables each subnet to make independent decisions while considering the global resource occupancy situation during a network outage, achieving coordination and consistency in a distributed environment.

[0014] Optionally, the above method further includes: when the local resource pool does not meet the resource requirements of the current emergency task, generating a resource conflict avoidance strategy based on the resource availability and resource allocation status; decomposing the current emergency task according to the resource conflict avoidance strategy to obtain a first sub-task and a second sub-task; executing the first sub-task in the local resource pool and generating a resource requirement flag and expected completion time for the second sub-task; associating the resource requirement flag, expected completion time, and second sub-task with the second sub-task and storing them in any task cache queue of at least two resource allocation centers; when communication between at least two communication subnets is detected to be restored, retrieving the second sub-task from any task cache queue and sending a cross-subnet resource request including the resource requirement flag and expected completion time to the second resource allocation center corresponding to the other communication subnet; receiving resource allocation confirmation information returned by the second resource allocation center after evaluating the received expected completion time and its own resource scheduling status information, and scheduling the second sub-task according to the resource allocation confirmation information.

[0015] By adopting the above technical solution, when the local resource pool cannot meet the demand, the resource conflict avoidance strategy provides a scientific basis for task decomposition based on the analysis results of resource availability and allocation status. This allows the first subtask to be executed immediately with full utilization of existing local resources, while the second subtask is accurately described and cached through the generation of resource requirement tags and expected completion times. This combination of decomposed execution and delayed processing ensures timely response to urgent parts. When communication is restored, the cached resource requirement tags and expected completion times are sent along with cross-subnet resource requests, enabling the second resource allocation center to assess the urgency of the task based on the expected completion time and make reasonable resource allocation decisions in conjunction with its own resource scheduling status. This mechanism of caching during network outages and coordinating after communication is restored minimizes the impact of communication interruptions and achieves continuity of task execution and optimization of resource utilization.

[0016] Optionally, the above method further includes: when the communication interruption type is identified as an internal communication interruption within a subnet, determining the target communication subnet where the internal communication interruption occurred and at least two missing rescue departments within the target communication subnet; obtaining historical resource allocation records of at least two missing rescue departments cached by the target resource allocation center corresponding to the target communication subnet before the communication interruption; generating resource occupancy prediction information for at least two missing rescue departments based on the resource usage patterns and task execution cycles in the historical resource allocation records; marking the resources of at least two missing rescue departments as pre-occupancy status information based on the resource occupancy prediction information, and obtaining supplementary resource information from other normal communication rescue departments within the target communication subnet besides the at least two missing rescue departments; reconstructing the available resource set of the target communication subnet based on the supplementary resource information, and allocating resources for newly added emergency tasks within the target communication subnet based on the available resource set to obtain a temporary resource allocation plan; when communication between at least two missing rescue departments is detected to be restored, obtaining the actual resource occupancy status information of at least two missing rescue departments; determining the resource status occupancy deviation between the actual resource occupancy status information and the pre-occupancy status information; and correcting the temporary resource allocation plan based on the resource status occupancy deviation.

[0017] By adopting the above technical solutions, the identification of communication interruptions within the subnet triggers a special handling process for the lost rescue departments. The resource usage patterns and task execution cycles provided by historical resource allocation records make resource occupancy prediction possible. This prediction based on historical data provides a basis for subsequent decision-making. The predicted information is converted into pre-occupancy status markers to prevent the duplication of resources. At the same time, supplementary resource information obtained from departments with normal communication expands the available resource pool. The combined reconstructed available resource set considers both the potential occupancy of lost departments and the supplementation of alternative resources. The temporary resource allocation scheme generated based on the available resource set is dynamically corrected after communication is restored by analyzing the deviation between the actual resource occupancy status and the predicted status. This closed-loop mechanism of prediction, execution, and correction enables reasonable resource allocation to be maintained even in the degraded state of some nodes being out of contact, and to quickly correct deviations and restore the optimal state after communication is restored.

[0018] Optionally, a resource-sharing cache node is established between two communication subnets that meet the preset cache establishment conditions. Specifically, this includes: deploying a cache server at the network topology boundary of the two communication subnets and configuring a bidirectional communication interface for the cache server. Each bidirectional communication interface establishes a data transmission channel with at least two resource allocation centers corresponding to the two communication subnets; setting a cache data update cycle within the cache server, and obtaining resource change events from at least two resource allocation centers through the bidirectional communication interface within each cache data update cycle; determining cache priority based on the event type and scope of impact of the resource change events, where event types include resource addition, resource occupation, resource release, and resource failure; and hierarchically storing cross-subnet resource status data according to cache priority to ensure cross-subnet... The first resource status data with high cache priority is stored in the fast access area, and the second resource status data with low cache priority across subnets is stored in the normal storage area. When the communication delay between two communication subnets is detected to be greater than a preset delay threshold or the communication bandwidth is less than a preset bandwidth threshold, the current update frequency of the cache data update cycle is adjusted to the target update frequency, and the current time window of the preset time period is adjusted to the target time window. The target update frequency is greater than the current update frequency, and the target time window is greater than the current time window. Historical communication interruption records of the two communication subnets are obtained, and data is pre-filled in the fast access area and the normal storage area according to the resource access mode in the historical communication interruption records to generate resource-shared cache nodes.

[0019] By adopting the above technical solutions, the boundary deployment location and bidirectional communication interface of the cache server ensure timely acquisition of resource change events in the two communication subnets. The setting of the cache data update cycle makes data collection regular, while the cache priority determined according to the event type and the scope of change impact enables differentiated data processing. The hierarchical storage mechanism places high-priority data in the fast access area and low-priority data in the ordinary storage area. This storage strategy, in conjunction with the subsequent adaptive adjustment mechanism, increases the update frequency when a decline in communication quality is detected to ensure the timeliness of critical data. Expanding the time window allows the resource-sharing cache node to store cross-subnet resource status data for a longer period of time. The resource access pattern based on historical communication interruption records pre-fills the two storage areas, making the cached content highly matched with the access needs during actual network outages. This multi-layered optimization mechanism of dynamic monitoring, adaptive adjustment, and intelligent pre-filling jointly constructs an efficient and reliable resource-sharing cache node.

[0020] Secondly, embodiments of this application provide an intelligent allocation platform for airport emergency rescue resources. The intelligent allocation platform for airport emergency rescue resources includes: one or more processors and a memory; the memory is coupled to one or more processors, and the memory is used to store computer program code, which includes computer instructions. The one or more processors invoke the computer instructions to cause the intelligent allocation platform for airport emergency rescue resources to perform the method described in the first aspect and any possible implementation thereof.

[0021] Thirdly, embodiments of this application provide a computer program product containing instructions that, when the computer program product is run on an intelligent allocation platform for airport emergency rescue resources, cause the intelligent allocation platform for airport emergency rescue resources to execute the method described in the first aspect and any possible implementation thereof.

[0022] Fourthly, embodiments of this application provide a computer-readable storage medium including instructions that, when executed on an airport emergency rescue resource intelligent allocation platform, cause the airport emergency rescue resource intelligent allocation platform to perform the method described in the first aspect and any possible implementation thereof.

[0023] One or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages: 1. The intelligent allocation method for airport emergency rescue resources provided in this application divides communication subnets for each rescue department based on communication network data, forming a distributed network architecture. The resource allocation center configured in each communication subnet enables each communication subnet to have independent management capabilities. The resource sharing cache nodes established between two communication subnets create cross-subnet data bridges on this architecture. The cross-subnet resource status data stored by these resource cache nodes and the local resource status data of each communication subnet form a complementary two-layer data system. When a communication interruption is detected and the link interruption detection result is obtained, it is precisely because the resource sharing cache nodes established in the early stage have stored the necessary cross-subnet resource status data that the resource allocation center can combine the local resource status data with the cached cross-subnet resource status data to make scheduling decisions in the network outage emergency mode. This collaborative mechanism of pre-caching and emergency scheduling can realize a smooth switch from centralized scheduling to distributed autonomy, ensuring that orderly resource allocation can still be maintained when communication is interrupted.

[0024] 2. The intelligent allocation method for airport emergency rescue resources provided in this application selects the set of rescue departments with the best communication quality by sorting communication bandwidth in descending order and communication latency in ascending order. This provides candidates for high-quality communication support for subsequent geographical location analysis. Based on this, physical location information and responsible area information are obtained for distance calculation, and departments with excellent communication and geographical proximity are aggregated into neighboring rescue department groups. This dual screening ensures that members within the group have both a good communication foundation and are easy to coordinate on the ground. Then, real-time resource status data of neighboring rescue department groups are obtained to identify target rescue departments with complementary resource relationships. Finally, these departments that meet the requirements in three dimensions of communication quality, geographical location and resource complementarity are divided into the same communication subnet. This multi-dimensional progressive screening mechanism makes the final communication subnet not only stable in communication and rapid in response, but also has the ability to self-balance internal resources, which greatly improves the overall collaborative efficiency of the communication subnet.

[0025] 3. The intelligent allocation method for airport emergency rescue resources provided in this application obtains resource type vectors and resource quantity vectors through feature extraction. These two vectors characterize resource features from both qualitative and quantitative dimensions. The resource type vector is used to calculate the resource complementarity coefficient, accurately quantifying the degree of resource matching between rescue departments, while the resource quantity vector is used to determine the resource load rate and identify high-load and low-load rescue departments. This dual evaluation mechanism of type matching and load balancing works together. When traversing combinations, it not only requires that the resource complementarity coefficient be greater than the preset complementarity threshold to ensure the complementarity of resource types, but also requires that the combination must include high-load and low-load departments to ensure the actual needs of resource allocation. This qualitative and quantitative judgment method makes the finally determined target rescue departments not only complementary in terms of resource types, but more importantly, they have actual allocation space and urgent needs in terms of resource quantity, thereby truly achieving precise matching of supply and demand. Attached Figure Description

[0026] Figure 1 This is a flowchart illustrating an intelligent allocation method for airport emergency rescue resources in an embodiment of this application. Figure 2 This is a schematic diagram of the physical device structure of an intelligent allocation platform for airport emergency rescue resources in this application embodiment. Detailed Implementation

[0027] The terminology used in the following embodiments of this application is for the purpose of describing particular embodiments only and is not intended to be limiting of this application. As used in the specification and appended claims of this application, the singular expressions “a,” “an,” “the,” “the,” “the,” and “this” are intended to include the plural expressions as well, unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used in this application refers to any or all possible combinations including one or more of the listed items.

[0028] Hereinafter, the terms "first" and "second" are used for descriptive purposes only and should not be construed as implying or suggesting relative importance or implicitly indicating the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature, and in the description of the embodiments of this application, unless otherwise stated, "multiple" means two or more.

[0029] This application provides a method for intelligent allocation of airport emergency rescue resources, see reference. Figure 1 , Figure 1 This is a flowchart illustrating an intelligent allocation method for airport emergency rescue resources in an embodiment of this application, including the following steps: Step S101: After obtaining the communication network data between various rescue departments, divide the communication subnets of each rescue department according to the communication network data, so as to divide the target rescue departments that meet the preset division conditions into the same communication subnet. Step S102: Configure a resource allocation center for each communication subnet and establish a resource sharing cache node between two communication subnets that meet the preset cache establishment conditions. The resource sharing cache node is used to save cross-subnet resource status data within a preset time period before the communication interruption when communication between two communication subnets is interrupted. Step S103: When a communication interruption is detected, communication link detection is performed between each communication subnet to obtain the link interruption detection result; Step S104: Trigger the network outage emergency mode of the resource allocation center based on the link interruption detection result. The network outage emergency mode performs scheduling for each resource allocation center based on local resource status data and cross-subnet resource status data.

[0030] The communication network data refers to the performance parameters and topology information of the network connections between various rescue departments, specifically including network bandwidth, transmission delay, packet loss rate, network topology, and routing information. Rescue departments refer to the various functional units within the airport responsible for emergency response, specifically including the airport fire brigade, medical emergency center, security department, ground services department, airline emergency response team, and airport operations command center. Communication subnets represent local communication networks divided according to specific conditions, specifically including fire and rescue subnets, medical rescue subnets, and security emergency subnets. Preset division conditions refer to the set of standards used to determine whether rescue departments should belong to the same subnet, specifically including communication quality thresholds, geographical distance restrictions, and business relevance requirements. The resource allocation center represents the control node responsible for the unified management and scheduling of resources within the subnet, specifically including the subnet scheduling server, resource management system, and task... Allocation modules, etc.; Preset cache establishment conditions refer to the set of conditions for determining whether a cache node needs to be established between two subnets, specifically including inter-subnet communication frequency thresholds, historical interruption counts, and business dependency levels; Resource sharing cache nodes refer to intermediate storage devices used to store cross-subnet resource information, specifically including edge cache servers, distributed storage nodes, and data synchronization centers; Cross-subnet resource status data refers to resource information shared between different subnets, specifically including personnel configuration information, equipment usage status, material inventory data, and task execution progress; Link interruption detection results represent diagnostic information obtained after detecting network connection status, specifically including interruption location, interruption type, impact range, and estimated recovery time; Network outage emergency mode refers to a special operating mechanism activated when the network is interrupted, specifically including local autonomy mode, cache data scheduling mode, and degraded service mode.

[0031] This step is primarily executed during the initialization phase of the airport emergency rescue system (hereinafter referred to as the system) or when the network topology changes. It is used to establish a distributed resource allocation architecture and prepare for potential network outages. Specifically, the system first collects real-time communication data between various rescue departments through network monitoring equipment, including ping test results between nodes, bandwidth usage, and data transmission success rates. Then, it uses clustering or graph partitioning algorithms to group the rescue departments, assigning departments with good communication quality and close business connections to the same communication subnet. A resource allocation center is deployed within each subnet, responsible for collecting and managing resource information from all rescue departments within its subnet and possessing independent scheduling decision-making capabilities. The system assesses the communication needs and historical fault conditions between subnets, establishing resource-sharing cache nodes between subnet pairs that meet the criteria. The system continuously monitors and caches cross-subnet resource status data, with the retention period determined by a preset time period, ranging from 30 minutes to 2 hours, without limitation. When the network monitoring system detects any communication anomalies, it immediately initiates a link detection program to determine the specific location and type of interruption by sending probe packets, checking routing tables, and analyzing traffic patterns. Based on the detection results, the relevant resource allocation centers automatically switch to network outage emergency mode. In this mode, resource allocation centers no longer rely on real-time cross-subnet communication but instead make resource scheduling decisions based on locally stored resource data and historical cross-subnet data obtained from resource-shared cache nodes.

[0032] In some embodiments, communication subnetting and emergency dispatch mechanisms can be implemented in multiple ways: Optionally, a community detection algorithm based on graph theory is adopted. In practice, a weighted undirected graph G=(V,E,W) of the rescue department's communication network is first constructed, where V represents the set of rescue department nodes, E represents the set of communication link edges, and W represents the edge weight matrix. Bandwidth utilization, average latency, and packet loss rate data for each link are collected periodically using network monitoring tools. A normalized formula is used to calculate the comprehensive weight, considering a weighted combination of bandwidth, latency, and packet loss rate. The Louvain algorithm is run for community detection, setting appropriate modularity thresholds and resolution parameters. After multiple iterations, the department is divided into several communication subnets. A containerized resource allocation center based on a container orchestration system is deployed at the center of each subnet, configured with multiple master nodes and worker nodes. A distributed key-value storage system is used to store the cluster state, and a reasonable data backup cycle is set. An in-memory database cluster is deployed at the boundaries of adjacent subnets, configured with a master-slave node architecture, set appropriate memory capacity, and uses an LRU (Least Recently Used) eviction policy, enabling a sentinel mechanism for automatic failover. Optionally, a deep learning-based intelligent partitioning approach is adopted, collecting historical communication log data as a training set. A multi-layer network structure is constructed using a Graph Convolutional Network (GCN), with appropriate hidden layer dimensions configured. ReLU activation function and dropout are used to prevent overfitting. During training, the Adam optimizer is used, with appropriate learning rate and batch size set. After sufficient training, the model reaches stable performance. The resource allocation center adopts a microservice architecture, including components such as a service registry, API gateway, configuration center, and circuit breaker. Each microservice is configured with multiple instances to achieve load balancing. Cache nodes are built using a distributed stream processing platform, creating multiple topics corresponding to each subnet. Each topic is configured with appropriate partitions and replicas, and a reasonable data retention time is set. Network outage detection uses periodic TCP heartbeat packets combined with SNMP protocol trap messages. When multiple consecutive heartbeat failures or a link disconnection trap is received, it is determined that the link is interrupted. It is understood that other methods can also be used to achieve dynamic partitioning of communication subnets and emergency response to network outages; this is not limited here.

[0033] Through the above steps, communication subnets are divided for each rescue department based on communication network data, forming a distributed network architecture. Each communication subnet is equipped with a resource allocation center, enabling independent management. Resource-sharing cache nodes established between two communication subnets create cross-subnet data bridges within this architecture. The cross-subnet resource status data stored by these cache nodes, along with the local resource status data of each communication subnet, form a complementary two-layer data system. When a communication interruption is detected and a link interruption detection result is obtained, the resource allocation center, thanks to the necessary cross-subnet resource status data stored by the previously established resource-sharing cache nodes, can combine local resource status data with the cached cross-subnet resource status data to make scheduling decisions in emergency network outage mode. This collaborative mechanism of pre-caching and emergency scheduling enables a smooth transition from centralized scheduling to distributed autonomy, ensuring orderly resource allocation even during communication interruptions. This solves the technical problem of difficulty in cross-departmental resource coordination and allocation in the event of a communication network interruption, achieving the technical effect of enabling cross-departmental resource coordination and allocation even under such circumstances.

[0034] The entity executing the above steps can be a platform, such as an airport emergency rescue resource intelligent allocation platform, or a control system or control equipment, or a controller or processor in the equipment or system, or a standalone controller or processor, or other processing equipment or processing units with similar processing functions, but is not limited to these.

[0035] In an optional embodiment, after obtaining communication network data between various rescue departments, the communication subnets of each rescue department are divided according to the communication network data to divide target rescue departments that meet preset division conditions into the same communication subnet. Specifically, this includes: obtaining communication bandwidth data and communication latency data of the communication links between each rescue department; sorting the communication bandwidth data in descending order to obtain a communication bandwidth sorting list; and sorting the communication latency data in ascending order to obtain a communication latency sorting list; and determining a set of communication links corresponding to a communication link that is in the top N positions in the communication bandwidth sorting list and in the top M positions in the communication latency sorting list. The system establishes a rescue department group; it acquires the physical location information and responsible area information of a group of rescue departments within the airport, including at least one of the terminal area, runway area, apron area, and cargo area; it determines the physical distance between each rescue department in the group based on the physical location information and responsible area information, and classifies one or more rescue departments whose physical distance is less than a preset distance threshold into a neighboring rescue department group; it acquires real-time resource status data of the neighboring rescue department group; based on the real-time resource status data, it identifies target rescue departments with complementary resource relationships from the neighboring rescue department group and classifies the target rescue departments into the same communication subnet.

[0036] Among them, communication bandwidth data represents the data transmission capacity of a network connection, specifically including uplink bandwidth, downlink bandwidth, average available bandwidth, peak bandwidth, etc., usually measured in Mbps (Megabits per second) or Gbps (Gigabits per second); communication latency data refers to the time required for data packets to travel in the network, specifically including propagation latency, processing latency, queuing latency, and round-trip time (RTT), usually measured in milliseconds; the communication bandwidth sorting list represents a list of links arranged in descending order of bandwidth, specifically including link identifier, origin and destination nodes, bandwidth value, sorting sequence number, etc.; the communication latency sorting list represents a list of links arranged in ascending order of latency, specifically including link number, average latency, latency jitter, sorting position, etc.; physical location information refers to the actual geographical coordinates of the rescue department at the airport, specifically including latitude and longitude coordinates, building location, floor information, GPS... (Global Positioning System) positioning data, etc.; Responsible area information indicates the scope of responsibility and jurisdiction of the rescue department, specifically including terminal areas (T1, T2, T3 terminals, etc.), runway areas (east runway, west runway, taxiway, etc.), apron areas (remote stands, near stands, cargo aprons, etc.), and cargo areas (cargo terminals, bonded warehouses, dangerous goods warehouses, etc.); Physical distance refers to the actual spatial distance between rescue departments, specifically including straight-line distance, road distance, emergency access distance, etc.; Neighboring rescue department groups indicate groups of departments with similar geographical locations, specifically including department groups within the same terminal, service groups on adjacent aprons, and emergency groups at both ends of the same runway, etc.; Real-time resource status data refers to the current resource allocation of the rescue department, specifically including the number of on-duty personnel, available vehicle list, medical supply inventory, and fire equipment status, etc.; Complementary resource relationships indicate the relationships between departments where resources can complement each other, specifically including personnel mutual assistance relationships, equipment sharing relationships, and material allocation relationships, etc.

[0037] This step is performed when initial partitioning or re-optimization of the communication subnet is required, ensuring the rationality of the partitioning results through multi-dimensional evaluation. Specifically, firstly, network performance monitoring tools (such as iperf, netperf) are used to measure the actual communication bandwidth between various rescue departments, recording bandwidth changes at different times and calculating the average value; simultaneously, the ping command or professional network latency testing tools are used to obtain round-trip latency data for each link, with a sampling frequency typically once per minute, continuously collected for 24 hours to obtain stable statistical results; the collected bandwidth data is sorted from largest to smallest, and latency data is sorted from smallest to largest, generating sorted lists respectively; based on preset screening parameters N and M (N can be 30% of the total number of links, etc., and M can be 40%, etc.), high-quality links that simultaneously meet the requirements of high bandwidth and low latency are identified, and the rescue departments connected to these links are determined as the candidate set; then, the airport geographic information system (GIS) is called. The system uses a Geographic Information System (GIS) to obtain the precise location coordinates and jurisdictional data of candidate departments, calculates the Euclidean or Manhattan distance between departments, and categorizes departments with a distance less than a preset threshold (such as 500 meters) into neighboring groups. Then, it uses the API (Application Programming Interface) of the resource management system to query the resource configuration data of each neighboring department in real time, including the number and skill level of various rescue personnel, the model and status of various rescue vehicles, and the types and quantities of medical emergency supplies. Finally, it analyzes the resource data to identify departmental combinations with complementary resource types (such as one department having sufficient fire-fighting equipment but insufficient medical resources, and another department having the opposite) and actual cooperation needs, and ultimately determines these departments as members of the same communication subnet.

[0038] In some embodiments, precise division of communication subnets based on multi-dimensional indicators can be achieved in various ways: Optionally, in the implementation of the weighted scoring mechanism, firstly, network performance testing tools are used to conduct bandwidth tests between various rescue departments, setting reasonable test durations and parallel flow numbers, and conducting regular tests to obtain stable data; for potential links formed between departments, bandwidth values ​​within different ranges are measured; network diagnostic tools are used to measure latency, sending standard-sized ICMP packets at a certain frequency and calculating the average RTT value; the precise coordinates of each department are obtained through the airport CAD (Computer-Aided Design) system, and the Euclidean distance matrix is ​​calculated; a scoring function is constructed, comprehensively considering factors such as bandwidth, latency, distance, and resource complementarity, with different weights assigned to each factor; the K-means++ algorithm is used to initialize cluster centers, setting appropriate maximum iteration counts and convergence thresholds, and the algorithm converges after iteration; a genetic algorithm is introduced for optimization, setting reasonable population size, crossover probability, and mutation probability, using tournament selection as the selection strategy, and the fitness function comprehensively considering inter-subnet communication volume and intra-subnet cooperation efficiency, resulting in an optimized partitioning scheme after multiple generations of evolution; Optionally, in implementing a hierarchical partitioning strategy, firstly, based on the airport's functional area division, departments in different terminals, cargo areas, and runway areas are grouped into initial groups. Within each group, the Prim algorithm is used to construct a minimum spanning tree, selecting the department with better communication quality as the root node. The degree centrality, betweenness centrality, and proximity centrality of each department are calculated, defining the top-ranked nodes based on multiple centrality indicators as core nodes, and the rest as edge nodes. Cross-group resource complementarity relationships are analyzed to identify groups with complementary resource allocations, and whether to merge related departments is determined based on the degree of complementarity. The rationality of the partitioning is verified using a decision tree algorithm, with historical collaboration success rate as the label and communication quality, geographical distance, and resource matching degree as features, and the partitioning effect is evaluated through cross-validation. It is understood that other methods can also be used to achieve intelligent grouping of rescue departments and dynamic construction of communication subnets, which are not limited here.

[0039] In an optional embodiment, the process of identifying target rescue departments with complementary resource relationships based on real-time resource status data includes: extracting features from the real-time resource status data to obtain resource type vectors and resource quantity vectors for each rescue department; determining a resource complementarity coefficient between any two rescue departments based on the resource type vector, whereby the resource complementarity coefficient characterizes the degree of matching between the surplus resources of one rescue department and the missing resources of the other; determining the resource load rate of each rescue department based on the resource quantity vector; comparing the resource load rate with a preset load threshold to identify high-load rescue departments with a resource load rate greater than the preset load threshold and low-load rescue departments with a resource load rate less than a preset idle threshold; and iterating through and combining rescue departments within adjacent rescue department groups to identify combinations that satisfy a resource complementarity coefficient greater than the preset complementarity threshold and include at least one high-load rescue department and at least one low-load rescue department as target rescue departments with complementary resource relationships.

[0040] Feature extraction refers to the process of extracting key features from raw data, specifically including Principal Component Analysis (PCA), feature encoding, vectorization, and dimensionality reduction. The resource type vector represents the distribution of resource types possessed by the rescue department, specifically represented as a multi-dimensional vector, such as [number of fire trucks, number of ambulances, number of medical personnel, number of firefighters, types of medical equipment, types of fire-fighting equipment]. The resource quantity vector represents the specific quantity of each type of resource, specifically including available quantity, occupied quantity, maintenance quantity, and total quantity, forming a numerical vector such as [10, 5, 2, 17]. The resource complementarity coefficient is a numerical indicator that quantifies the degree of resource matching between two departments, specifically including cosine similarity, Pearson correlation coefficient, and Jaccard coefficient, with values ​​typically ranging from 0 to 1. Surplus resources represent idle resources exceeding normal demand, specifically including backup rescue vehicles, standby personnel, and reserve supplies. Missing resources refer to resource gaps that cannot meet current needs. The system includes the number of personnel shortages, the types of equipment shortages, and a list of material gaps; the resource load rate represents the intensity of resource utilization, specifically calculated as the ratio of occupied resources to total resources, such as 0.8 indicating that 80% of resources are in use; the preset load threshold refers to the critical value for judging a high load state, specifically including personnel load threshold (e.g., 0.85), equipment load threshold (e.g., 0.90), and comprehensive load threshold (e.g., 0.87); the preset idle threshold represents the critical value for judging a low load state, specifically including personnel idle threshold (e.g., 0.30), equipment idle threshold (e.g., 0.25), and comprehensive idle threshold (e.g., 0.28); high-load rescue departments refer to departments whose resource utilization is close to saturation, specifically characterized by task backlog, overtime work, and equipment operating at full capacity; low-load rescue departments refer to departments with relatively idle resources, specifically characterized by personnel on standby, idle equipment, and fewer tasks; the preset complementarity threshold refers to the minimum coefficient value for judging the establishment of a resource complementarity relationship, specifically such as 0.6, 0.65, 0.7, etc., set according to actual collaboration needs.

[0041] This step is performed when it is necessary to accurately identify resource complementarity relationships from neighboring rescue department groups, ensuring the scientific nature of resource pairing through quantitative analysis. Specifically, the acquired real-time resource status data is first preprocessed, including data cleaning, missing value imputation, and outlier handling. Then, feature engineering techniques are used to extract resource features, converting the textual descriptions of resource information into numerical vectors. For example, "5 fire trucks, 3 ambulances" is converted into a vector form of [5,3,0,0,...]. For resource type vectors, one-hot encoding or embedding representation is used to map each resource type to a point in a high-dimensional space. The dot product or cosine similarity of the resource type vectors of any two departments is calculated to obtain a resource complementarity coefficient matrix. The closer the coefficient is to 1, the more complementary the resource structures of the two departments are. Simultaneously, the resource load rate of each department is calculated based on the resource quantity vector. The formula is to divide the current amount of resources occupied by the total amount of resources in the department. For multiple resource types, a weighted average method can be used to calculate the overall load rate. The calculated load rate is compared with the preset load threshold and idle threshold. Departments with a load rate higher than 0.85 are marked as high load, and those lower than 0.30 are marked as low load. Finally, all departments in the adjacent rescue department group are traversed in pairs, and each pair is checked to see if it meets three conditions at the same time: the resource complementarity coefficient is greater than 0.65 (it can also be 0.7, 0.8, 0.85, etc.), it contains at least one high load department, and it contains at least one low load department. The combination that meets all the conditions is identified as the target rescue department with complementary resource relationship.

[0042] In some embodiments, the accurate identification of complementary resource relationships between rescue departments can be achieved through a variety of methods: Optionally, a graph neural network-based relationship prediction method is adopted, specifically constructing a resource relationship graph containing 35 nodes. The feature vector of each node has a dimension of 48, including the number of 12 resource types (fire trucks, ambulances, command vehicles, medical personnel, firefighters, security personnel, etc.) and 36 derived features (such as resource occupancy rate, average response time, etc.). A 3-layer GAT network is used, with the number of attention heads in each layer being [8, 16, 8] and the hidden dimensions being [64, 128, 64]. Batch normalization and Dropout of 0.3 are added after each layer. The training data includes 86 data points from the past 90 days. Forty resource allocation records were divided into training, validation, and test sets in a 7:2:1 ratio. Negative samples were generated using negative sampling techniques, with a positive-to-negative sample ratio of 1:3, and a binary cross-entropy loss function was employed. The weight matrix learned by the attention mechanism showed that the average attention weight for the medical-fire department pair was 0.72, while that for similar departments was only 0.31. An LSTM layer was introduced to process temporal features. Given a resource load sequence from the past 24 hours and a hidden layer dimension of 128, the system predicted resource demand for the next 4 hours, achieving an RMSE (Root Mean Square Error) of 8.3%. Optionally, in the implementation of the multi-objective optimization algorithm, three optimization objectives are defined: f1 = maximize(Σ resource complementarity coefficient), f2 = minimize(max(load rate) - min(load rate)), f3 = minimize(average response time); the NSGA-II algorithm is used, the population size is 200, the maximum number of generations is 100, the crossover operator is SBX (Simulated Binary Crossover) with a distribution exponent of 20, and the mutation operator is polynomial mutation with a distribution exponent of 20. After 100 generations of evolution, 48 Pareto optimal solutions were obtained, forming a Pareto front. Simulated annealing was used for local search on each solution, with an initial temperature of 100°C and a cooling rate of 0.95. The iterations were performed 50 times at each temperature, and a 2-opt exchange strategy was used for neighborhood search. A Q-learning optimization strategy was introduced. The state space includes eight features such as current resource distribution and historical matching success rate, the action space represents possible department pairings, and the reward function is the percentage increase in actual collaboration efficiency. After training for 5000 episodes, the Q-table converged, improving resource utilization under the optimal strategy. It is understood that other methods can also be used to achieve intelligent matching and complementary relationship recognition based on resource features; this is not limited here.

[0043] In an optional embodiment, a network outage emergency mode is triggered for the resource allocation center based on the link interruption detection result. The network outage emergency mode schedules each resource allocation center based on local resource status data and cross-subnet resource status data. Specifically, it includes: identifying the communication interruption type based on the link interruption detection result; and when the communication interruption type is identified as inter-subnet communication interruption, determining at least two communication subnets where external communication interruption occurred; sending a network outage emergency trigger signal to at least two resource allocation centers corresponding to the at least two communication subnets to obtain cross-subnet resource status data read by the at least two resource allocation centers from at least one resource shared cache node; constructing a local resource pool based on the local resource status data; and, based on the cross-subnet resource... Status data is used to construct a virtual external resource pool. The local resource pool is used to store the information on available rescue resources and their availability within at least two communication subnets. The virtual external resource pool is used to store the resource allocation status and occupied resource information of other communication subnets that have established caching relationships with at least two communication subnets before communication interruption, obtained from resource sharing cache nodes. Emergency rescue task information to be processed is read from at least two local task queues corresponding to at least two resource allocation centers, and the priority of emergency tasks is determined based on the task type, urgency, and impact scope in the emergency rescue task information. Based on the emergency task priority, the available rescue resources and their availability are used for scheduling within the local resource pool.

[0044] Among them, the communication interruption type indicates the specific category of network failure, including inter-subnet communication interruption, intra-subnet communication interruption, single-point link interruption, and regional network paralysis; inter-subnet communication interruption refers to the disconnection between different communication subnets, specifically manifested as the inability to transmit cross-subnet data packets, unreachable routes between subnets, and gateway device failure; external communication interruption indicates the connection failure between a subnet and the external network, specifically including the interruption of connection with other subnets, loss of connection with the central server, and disconnection from the Internet; network outage emergency trigger signal indicates the control signal to activate the emergency mode, specifically including emergency broadcast messages, priority interruption signals, and state switching instructions; local resource pool refers to the data structure that stores the internal resource information of the subnet, specifically implemented as a hash table, B+ tree, or in-memory database, containing fields such as resource ID (Identifier), resource type, current status, and availability time; virtual external resource pool represents the external resource view built based on cached data, specifically including resource snapshots of other subnets, historical allocation records, and predicted available resources; and deployable rescue resource information refers to the details of resources that can be scheduled and used. Specifically, this includes resource ID, department, skill level, response time, and allocation cost; resource availability status indicates whether a resource is currently available for allocation, including idle, occupied, under maintenance, faulty, and reserved statuses; resource allocation status indicates the allocation status of resources, including allocated tasks, allocation time, estimated release time, and allocation priority; occupied resource information refers to the resource record of tasks currently being executed, including a list of occupied resources, task association information, and occupation duration statistics; the local task queue represents an ordered set of tasks to be processed, specifically implemented as a priority queue, circular buffer, or linked list structure; emergency rescue task information refers to a detailed description of the task, including task ID, location, accident type, casualties, and required resource list; task type indicates the category of emergency event, including fire rescue, medical emergency, security threats, equipment failure, and natural disasters; task urgency indicates the urgency of the task, specifically divided into levels such as extremely urgent, level one, level two, and level three; and task impact scope refers to the scope of the event, including the number of affected flights, the number of passengers involved, and estimated economic losses.

[0045] This step is executed immediately upon detecting a communication interruption, used to quickly switch to autonomous dispatch mode and maintain rescue capabilities. Specifically, it first analyzes the fault characteristics in the link interruption detection results, including the network layer where the interruption occurred, the affected IP (Internet Protocol) address range, routing table changes, etc., and identifies the specific interruption type through a pattern matching algorithm. When an inter-subnet communication interruption is confirmed, the fault diagnosis module immediately identifies the list of affected communication subnets and sends an emergency trigger signal to the resource allocation centers of these subnets. The signal is broadcast using UDP (User Datagram Protocol) to ensure rapid delivery. Upon receiving the trigger signal, each resource allocation center immediately reads the most recently updated cross-subnet resource status data from the locally connected resource-sharing cache node. The reading process is optimized using batch queries to reduce I / O (Input / Output) latency. A local resource pool is constructed based on the real-time data reported by each rescue department within the subnet, using an in-memory database (such as R...). The system uses EDISI (Eddis) to store and establishes an index from resource ID to resource details, supporting O(1) time complexity queries. At the same time, it organizes external subnet data read from the cache into a virtual external resource pool, uses timestamps to mark the validity period of the data, and automatically downgrades data that has exceeded the validity period. It extracts tasks to be processed from the local task queue according to FIFO (First In First Out) or priority rules, parses the key fields in the task information, assigns basic priority scores according to the task type (e.g., 100 points for fire rescue, 90 points for medical emergency), adds weighting coefficients according to the degree of urgency (2.0 for special emergency, 1.5 for first-level emergency), and makes secondary adjustments according to the scope of impact (add 20 points for impacting more than 100 people). Finally, it calculates the comprehensive priority. It processes tasks in order of priority from high to low, prioritizes finding matching available resources from the local resource pool, checks the availability status and capability matching degree of the resources, generates a resource allocation plan and executes it immediately.

[0046] In some embodiments, intelligent resource scheduling in a network outage state can be achieved in a variety of ways: Optionally, a rule-based scheduling strategy is adopted, predefining 158 resource allocation rules in the rule engine. These rules are written in a dedicated rule language format, such as "ruleFireEmergencywhenTask(type=='Fire',priority>8)andResource(type=='Fire Truck',status=='Idle')thenallocate()". Rule priorities are divided into 5 levels, with priority values ​​ranging from 100 to 500. In case of conflict, the higher-priority rule is triggered first. Fuzzy logic is used to handle boundary conditions, defining a fuzzy set "urgency level" = {low: 0-3, medium: 3-...}. 7, High: 7-10}, The membership function adopts a triangular function; a Rete network is constructed during rule execution, with alpha nodes storing factual constraints and beta nodes processing rule condition combinations, resulting in an average rule matching time of 12ms; a Monte Carlo simulation evaluation scheme is introduced, with each scheme running 1000 random simulations, setting the task arrival rate to follow a Poisson distribution λ=5 / hour, and the service time to follow an exponential distribution μ=0.5 / hour, calculating the average task completion time and resource utilization rate; in actual testing, when handling a fire accident, the rule engine matched 3 relevant rules within 8ms, allocating 2 fire trucks and 8 firefighters, significantly reducing the task completion time compared to manual scheduling. Optionally, an adaptive scheduling method based on DQN is adopted, constructing a deep Q-network containing four fully connected layers with a number of neurons [256, 512, 512, 256]. The input layer receives a 168-dimensional state vector (including resource states of 35 departments, task queue states, etc.), and the output layer corresponds to 42 possible scheduling actions. An experience replay buffer stores 100,000 experiences, and 32 experiences are randomly sampled for training each time. An ε-greedy policy is used for exploration, with an initial ε=1.0, decaying by 0.995 every 1000 steps, and a minimum... The value is 0.01; the target network updates every 500 steps using a soft update method, τ=0.001; the reward function is designed as r=-0.1×wait time-0.2×resource idle rate+10×(task completion ? 1:0); training uses historical scheduling data from the past 2 years, containing approximately 15,000 scheduling cases, and after 50,000 episodes of training, the average reward improved from -25 to +45; in the network outage test scenario, the DQN model's average decision time is 3ms, and the resource allocation accuracy is significantly improved. It is understandable that other methods can be used to achieve optimized resource scheduling in network outage emergency situations, which are not limited here.

[0047] In an optional embodiment, the method further includes: when the local resource pool does not meet the resource requirements of the current emergency task, generating a resource conflict avoidance strategy based on the resource availability status and resource allocation status; decomposing the current emergency task according to the resource conflict avoidance strategy to obtain a first sub-task and a second sub-task; executing the first sub-task in the local resource pool and generating a resource requirement flag and an expected completion time for the second sub-task; associating the resource requirement flag, expected completion time, and second sub-task with the second sub-task and storing them in any task cache queue of at least two resource allocation centers; when communication between at least two communication subnets is detected to be restored, retrieving the second sub-task from any task cache queue and sending a cross-subnet resource request including the resource requirement flag and expected completion time to the second resource allocation center corresponding to the other communication subnet; receiving resource allocation confirmation information returned by the second resource allocation center after evaluating the received expected completion time and its own resource scheduling status information, and scheduling the second sub-task according to the resource allocation confirmation information.

[0048] Among them, resource conflict avoidance strategies refer to scheduling schemes that avoid resource competition, specifically including time-shifting strategies, spatial isolation strategies, priority yielding strategies, and resource reservation strategies; task decomposition refers to the process of breaking down a complex task into multiple sub-tasks, specifically including functional decomposition, temporal decomposition, spatial decomposition, and resource decomposition; the first sub-task represents the part of the task that can be executed immediately, specifically characterized by low resource requirements, high priority, strong independence, and short execution time; the second sub-task represents the part of the task that needs to be delayed, specifically characterized by high resource requirements, being able to wait, depending on external resources, and being a non-critical path; resource requirement marking refers to the identification information describing the resources required by the task, specifically including resource type marking, quantity requirement marking, skill requirement marking, and time window marking; expected completion time represents the planned completion time of the task, specifically including the earliest start time, the latest... Late completion time, standard duration, buffer time, etc.; Task cache queue refers to a data structure that temporarily stores tasks to be processed, specifically implemented as a circular queue, priority queue, dequeue, etc., supporting task insertion, deletion, and query operations; Cross-subnet resource request represents a message requesting resources from other subnets, specifically including request headers (request ID, source address, destination address), request body (resource requirements, task description, priority), etc.; Second resource allocation center refers to the resource management node of other subnets, specifically responsible for evaluating external requests, allocating local resources, and returning confirmation information, etc.; Resource scheduling status information represents the operating status of the scheduling system, specifically including current load, queue length, average response time, resource utilization, etc.; Resource allocation confirmation information refers to the response message for resource allocation, specifically including allocation result (agree / reject), allocated resource list, estimated arrival time, usage restrictions, etc.

[0049] This step is executed when local resources are insufficient to complete the entire emergency task, maximizing resource utilization efficiency through task splitting and delayed processing strategies. Specifically, when it is detected that the available resources in the local resource pool cannot meet all the requirements of the current task, the specific details of the resource gap are first analyzed, including the type of missing resources, the quantity difference, time conflicts, etc.; based on the resource availability status and the existing resource allocation status, a conflict detection algorithm is used to identify potential resource contention points, such as multiple tasks requesting the same resource simultaneously, or a mismatch between resource release time and demand time; based on the conflict analysis results, an avoidance strategy is generated, which may include staggering task execution times by 15 minutes, reserving 20% ​​of resources for high-priority tasks, and setting resource-sharing time slices; according to the avoidance strategy, the original task is intelligently decomposed. The decomposition principle is to divide the part that can be completed using local resources into the first sub-task, and the part that requires external resources or needs to wait into the second sub-task. The decomposition process considers the logical dependencies of tasks to ensure the independent executableness of sub-tasks; the first sub-task immediately allocates resources in the local resource pool and starts... The process begins, simultaneously generating detailed resource requirement tags for the second subtask, including the required rescue vehicle models, personnel skill levels, and special equipment requirements. The expected completion time is calculated based on the task's urgency and standard operating time. The second subtask and its associated information are packaged and stored in a task cache queue, using persistent storage to ensure no data loss. Communication link status is continuously monitored; once communication with other subnets is restored, the second subtask is immediately retrieved from the cache queue, a standard-format cross-subnet resource request message is constructed, and sent to the target subnet's resource allocation center via the restored communication link. Upon receiving the request, the second resource allocation center assesses its own resource availability and task queue status, comprehensively considering the request's priority and expected completion time, makes an allocation decision, and returns confirmation information. Based on the received confirmation information, the original requester, if resource allocation is successful, schedules the execution of the second subtask; if rejected, it re-evaluates and may send requests to other subnets.

[0050] In some embodiments, intelligent task processing under resource-constrained conditions can be achieved in a variety of ways: Optionally, a task decomposition method based on integer linear programming can be adopted to model complex rescue tasks as ILP problems, defining decision variables x. ij ∈{0,1} indicates whether resource i is allocated to subtask j, y j ∈{0,1} indicates whether subtask j is executed; the objective function is minΣ(c ij ×x ij )+Σ(d j ×(1-y j ), where c ij To allocate costs, d jThe penalty is for delay; constraints include: resource capacity constraint Σx ij ≤R i (Capacity of resource i), task requirement constraint Σx ij ≥D j ×y j (Requirements of subtask j), priority constraint y1≥y2 (subtask 1 takes precedence over 2), time window constraint s j +p j ≤e j (Start time + processing time ≤ end time); An optimization solver is used, with an optimality gap of 0.01 and a time limit of 60 seconds. The branch and bound strategy adopts the best boundary priority approach. In a real-world case, a large-scale fire task involving 15 rescue points is decomposed into 5 sub-tasks. The first sub-task (emergency evacuation) is assigned to 3 fire trucks for immediate execution, while the second sub-task (fire extinguishing) requires a temporary queue of 8 fire trucks. Sub-tasks are managed through message queues, with priority queues set, message lifetimes of 3600 seconds, and manual confirmation as the acknowledgment mechanism. Optionally, a knowledge graph-based intelligent decomposition strategy is used to construct an emergency task knowledge graph containing 1200 entity nodes and 3500 relation edges. Entity types include task type, resource type, execution steps, and constraints. The system uses a graph database for storage and a graph query language for graph traversal and pattern matching. During task decomposition, the most similar historical task template is first found through similarity calculation. Semantic similarity of the task descriptions is then calculated using word vector technology, with a cosine similarity threshold of 0.75. A graph segmentation algorithm is applied, setting the number of segments k to 3-5, a balance constraint of 1.03, and edge weights considering dependency strength and resource sharing. An ontology reasoning engine is used for semantic reasoning, automatically identifying the resource requirement types of subtasks, with reasoning rules such as "if task type = medical rescue then required resources include (ambulance, medical personnel)". The decomposition process is recorded on a distributed ledger platform, with each subtask generating a unique hash value. Smart contracts verify the legality of resource allocation, with a block generation time of 3 seconds and a throughput of 1000 TPS. It is understood that other methods can also be used to achieve flexible decomposition of complex tasks and cross-subnet resource coordination; these are not limited here.

[0051] In an optional embodiment, the method further includes: when the communication interruption type is identified as an internal communication interruption within a subnet, determining the target communication subnet where the internal communication interruption occurred and at least two missing rescue departments within the target communication subnet; obtaining historical resource allocation records of at least two missing rescue departments cached by the target resource allocation center corresponding to the target communication subnet before the communication interruption; generating resource occupancy prediction information for at least two missing rescue departments based on the resource usage patterns and task execution cycles in the historical resource allocation records; marking the resources of at least two missing rescue departments as pre-occupancy status information based on the resource occupancy prediction information, and obtaining supplementary resource information from other normal communication rescue departments within the target communication subnet besides the at least two missing rescue departments; reconstructing the available resource set of the target communication subnet based on the supplementary resource information, and allocating resources for newly added emergency tasks within the target communication subnet based on the available resource set to obtain a temporary resource allocation plan; when communication between at least two missing rescue departments is detected to be restored, obtaining the actual resource occupancy status information of at least two missing rescue departments; determining the resource status occupancy deviation between the actual resource occupancy status information and the pre-occupancy status information; and correcting the temporary resource allocation plan based on the resource status occupancy deviation.

[0052] Among them, internal subnet communication interruption indicates that some nodes within the same subnet have lost connection, specifically manifested as unreachable routing, switch failure, or node disconnection; target communication subnet refers to the specific subnet where internal communication failure occurred, identified by subnet ID, IP address range, VLAN (Virtual Local Area Network) identifier, etc.; lost contact rescue department refers to department nodes that have lost communication connection, with specific characteristics including heartbeat timeout, ping failure, interrupted data reporting, and unknown status; historical resource allocation records are used to represent past resource usage, specifically including allocation time, resource type, usage duration, task type, completion status, etc.; resource usage pattern refers to the regularity of resource usage, specifically including periodic patterns (e.g., a cycle every 4 hours), trend patterns (e.g., gradual increase), and burst patterns (e.g., random peaks), etc.; task execution cycle represents the temporal regularity of tasks, specifically including task start time distribution, average execution duration, task interval time, and number of concurrent tasks, etc.; resource occupancy prediction information refers to the estimation of future resource usage, specifically including predicted occupancy and confidence interval. The information includes: prediction timeliness, probability of anomalies, etc.; pre-occupancy status information represents resource status markers based on predictions, specifically including pre-occupancy flags, expected occupancy duration, occupancy probability, substitutability, etc.; normal communication rescue departments refer to departments that maintain uninterrupted communication, with specific criteria including normal heartbeat, stable data transmission, and qualified response time, etc.; supplementary resource information represents additional resources that can be used for supplementation, specifically including backup personnel, backup equipment, emergency supplies, and external support resources, etc.; available resource set refers to the reorganized resource pool, specifically including confirmed available resources, predicted available resources, conditionally available resources, and restricted resources, etc.; temporary resource allocation plan represents the resource allocation plan in emergency situations, specifically including resource allocation tables, execution schedules, alternative plans, and fallback strategies, etc.; actual resource occupancy status information refers to the actual resource usage, specifically including actual occupancy amount, actual usage duration, and actual completed tasks, etc.; resource status occupancy deviation represents the difference between prediction and reality, specifically including quantity deviation, time deviation, type deviation, and distribution deviation, etc.; resource allocation correction refers to adjusting the allocation plan based on deviations, specifically including resource reallocation, task rescheduling, priority adjustment, and compensation measures, etc.

[0053] This step is executed when a local communication failure is detected within the subnet, and maintains the subnet's operational capability through prediction and dynamic correction mechanisms. Specifically, once link detection confirms the communication interruption type as an internal subnet interruption, the first step is to determine the specific list of missing rescue departments by traversing the connectivity of all nodes within the subnet, recording the time of loss of contact and the last known status. The target resource allocation center immediately queries the local database to extract historical resource allocation records of the missing departments over the past 7 or 30 days, including detailed data such as resource usage for each task, task type distribution, and resource occupancy duration. Time series analysis methods (such as ARIMA (Autoregressive Integrated Moving Average) models or the Prophet algorithm) are used to analyze resource usage patterns in historical data, identifying periodic patterns (such as routine inspections at 10 AM every day) and trend changes (such as reduced resource usage at night). Based on the identified patterns and the current time, a prediction algorithm is used to estimate the resource occupancy of the missing departments in the next 2 or 4 hours, generating resource occupancy prediction information containing predicted values ​​and confidence levels. The prediction results are converted into pre-occupancy status markers, and the corresponding resource markers are added to the resource management system. This is recorded as a pre-occupancy state to prevent these resources from being incorrectly allocated to other tasks. Simultaneously, resource requisition requests are sent to departments within the subnet that maintain normal communication, collecting supplementary resource information from these departments, including temporarily available backup personnel, equipment, and materials. Pre-occupied resources, identified available resources, and supplementary resources are integrated to form a new set of available resources. This set considers uncertainties and sets an availability probability for each resource. Based on the reconstructed resource set, temporary allocation plans are developed for newly added emergency tasks. These plans clearly indicate which resources are definitively available and which are based on predictions, and alternative plans are prepared. Network status is continuously monitored. When communication is restored to a lost department, its actual resource occupancy status is immediately queried, including the tasks actually executed, resources consumed, and currently available resources. The actual status is compared with the previous predicted status, calculating deviation values ​​for various dimensions, such as resource occupancy deviation rate and time prediction accuracy. Based on the deviation analysis results, the still-executing temporary allocation plans are corrected. Possible corrective actions include releasing excess reserved resources, adjusting task allocation, and supplementing resource gaps to ensure the system quickly returns to its optimal operating state.

[0054] In some embodiments, intelligent response to subnet communication interruptions can be achieved in a variety of ways: Optionally, a distributed prediction approach based on federated learning is adopted. This assumes lightweight prediction models are deployed locally across 35 rescue departments, using a mobile-optimized architecture to reduce computational overhead, with model size kept under 50MB. Each department trains an ARIMA(2,1,2) model using local resource usage data from the past 30 days, with a data sampling frequency of 15 minutes, including 8 features such as resource utilization and task type distribution. Federated aggregation is performed every 24 hours using a federated averaging algorithm, with a client-side learning rate of 0.01 and server-side aggregation weights weighted by data volume. The communication rounds are set to 10 rounds, with 70% of departments randomly selected in each round. Differential privacy technology is used to add noise ε=0.5 to protect data privacy. When a department loses contact, an aggregation model is used to predict its resource occupancy for the next 4 hours. The prediction results include point estimates and 95% confidence intervals. Bayesian inference quantifies uncertainty. The prior distribution uses Beta(2,2), the likelihood function is based on historical prediction accuracy, and the posterior distribution is obtained through MCMC (Markov Chain Monte Carlo) sampling. Predictions with a confidence level >0.8 are marked as "strong pre-occupancy," locking resources that cannot be allocated. Predictions with a confidence level between 0.5 and 0.8 are marked as "weak pre-occupancy," which can be requisitioned in emergencies. In actual testing, a fire station lost contact for 2 hours. The model predicted that it would occupy 3 fire trucks. After communication was restored, it was verified that only 2 fire trucks were occupied, significantly improving the prediction accuracy. Optionally, a digital twin-based virtual simulation approach is adopted, using a 3D rendering engine to construct a digital twin of the airport rescue system, recreating the spatial layout and resource configuration of 35 departments at a 1:1 scale. Each department's digital twin includes a state machine model, defining five states (idle, standby, dispatch, return, and maintenance) and a state transition probability matrix. Real-time data, including GPS location, resource status, and task execution progress, is synchronized every second via the MQTT protocol, with latency controlled within 100ms. The behavioral model is based on an HMM (Hidden Markov Model) trained on historical data. The model uses a Hidden Markov Model (HMM) with 8 states and 12 observation symbols, trained using the Baum-Welch algorithm. When the physical entity loses contact, the digital twin continues to run according to a pre-defined behavioral model, with each simulation step lasting 1 minute. The simulation is run 100 times using the Monte Carlo method, and the average value is taken. Resource usage data generated during the simulation is stored in a time-series database with a 7-day retention policy and an aggregation granularity of 5 minutes. After communication is restored, Kalman filtering is used to fuse the actual and simulated data, with a process noise covariance Q = 0.1I and a measurement noise covariance R = 0.05I, to quickly correct the digital twin model parameters. It is understood that other methods can also be used to achieve fault tolerance in the event of local faults within the subnet; this is not limited here.

[0055] In an optional embodiment, a resource-sharing cache node is established between two communication subnets that meet preset cache establishment conditions. Specifically, this includes: deploying a cache server at the network topology boundary of the two communication subnets and configuring a bidirectional communication interface for the cache server. The bidirectional communication interface establishes data transmission channels with at least two resource allocation centers corresponding to each of the two communication subnets; setting a cache data update cycle within the cache server, and obtaining resource change events from at least two resource allocation centers through the bidirectional communication interface within each cache data update cycle; determining cache priority based on the event type and impact scope of the resource change events, where event types include resource addition, resource occupation, resource release, and resource failure; and hierarchically storing cross-subnet resource status data according to cache priority. The first resource status data with high cache priority from the cross-subnet resource status data is stored in the fast access area, and the second resource status data with low cache priority from the cross-subnet resource status data is stored in the ordinary storage area. When the communication delay of the two communication subnets is detected to be greater than a preset delay threshold or the communication bandwidth is less than a preset bandwidth threshold, the current update frequency of the cache data update cycle is adjusted to the target update frequency, and the current time window of the preset time period is adjusted to the target time window. The target update frequency is greater than the current update frequency, and the target time window is greater than the current time window. The historical communication interruption records of the two communication subnets are obtained, and the data in the fast access area and the ordinary storage area are pre-filled according to the resource access mode in the historical communication interruption records to generate resource shared cache nodes.

[0056] In this context, network topology boundaries represent the connection interface between two communication subnets, specifically including gateway locations, border routers, DMZ (Demilitarized Zone) areas, VLAN interfaces, etc.; cache servers refer to hardware devices specifically used for data caching, specifically including high-performance servers, storage arrays, edge computing nodes, in-memory database servers, etc.; bidirectional communication interfaces represent network interfaces that support bidirectional data transmission, specifically including gigabit Ethernet ports, 10-gigabit fiber optic ports, InfiniBand interfaces, RDMA (Remote Direct Memory Access) interfaces, etc.; and data transmission channels are used to represent data transmission. Logical connections specifically include TCP (Transmission Control Protocol) connections, UDP channels, message queue channels, and RPC (Remote Procedure Call) channels; cached data update cycle refers to the time interval for data refresh, specifically including fixed cycles (e.g., every 5 minutes), adaptive cycles (dynamically adjusted according to load), and event-driven cycles (updating immediately upon change); resource change events represent events in which the resource state changes, specifically including resource online events, resource offline events, state transition events, and attribute modification events; event type refers to the category of change event, specifically including resource addition (new device entry)... The impact of a change is categorized into several levels: network access, personnel on-site, resource usage (task execution begins), resource release (task completion), and resource failure (equipment damage, personnel injury). The scope of impact indicates the degree of influence of an event, specifically including the number of affected departments, the scale of resources involved, the number of associated tasks, and the estimated recovery time. Cache priority indicates the importance of data, specifically divided into critical (priority value 90-100), important (70-89), general (50-69), and low (0-49). The fast access area refers to the high-speed cache storage area, specifically implemented through SSD (Solid State Drive) storage, memory caching, and CPU (Central Processing Unit) caching. IT (Central Processing Unit) cache, dedicated cache chips, etc.; ordinary storage area refers to the regular cache storage area, specifically implemented including mechanical hard drives, distributed storage, object storage, cold storage, etc.; preset latency threshold refers to the latency standard for judging communication quality, such as one-way latency of 50ms, round-trip latency of 100ms, jitter of 20ms, etc.; preset bandwidth threshold refers to the bandwidth standard for judging communication quality, such as minimum bandwidth of 100Mbps, average bandwidth of 500Mbps, peak bandwidth of 1Gbps, etc.; current update frequency refers to the current data update rate, such as once per minute, 10 times per second, etc.; target update frequency refers to the adjusted update rate, which is dynamically calculated based on network conditions, usually 1 / 3 of the current frequency.5-3 times; the current time window represents the time range of currently saved data, such as the last 30 minutes, the last hour, etc.; the target time window refers to the expanded time range, such as the last 2 hours, the last 4 hours, the last 24 hours, etc.; historical communication interruption records represent past fault records, specifically including interruption time, duration, cause, recovery process, and scope of impact; resource access patterns refer to the patterns of resource queries, specifically including hot resource lists, access frequency distribution, query time characteristics, and related query patterns.

[0057] This step is executed during system initialization or when frequent interaction needs are detected between two subnets, and is used to establish an efficient and reliable cross-subnet data caching mechanism.Specifically, first, select a suitable deployment point at the physical or logical boundary of the two communication subnets, typically choosing a location with minimal network latency and maximum bandwidth. Deploy high-performance cache servers, with server configurations including multi-core CPUs (e.g., 32 cores or more), large-capacity memory (e.g., 256GB or more), and high-speed SSD storage (e.g., NVMe (Non-Volatile Memory Express) SSD arrays). Configure redundant bidirectional communication interfaces for the cache servers, with each interface connected to a core switch of a subnet. Use link aggregation technology to improve bandwidth and reliability, and configure VRRP (Virtual Router Redundancy Protocol). Interface-level failover is achieved using either the Virtual Router Redundancy Protocol (VRRP) or the Hot Standby Router Protocol (HSRP). Persistent data transmission channels are established between these interfaces and the resource allocation centers of the two subnets, using long-lived connections to reduce connection establishment overhead and enabling the TCP_NODELAY option to reduce transmission latency. Cache management software runs on the cache server, with an initial cache data update cycle set to 5 minutes, which can be adjusted according to actual needs. Within each update cycle, the cache server proactively sends data synchronization requests to the two resource allocation centers, using incremental synchronization to transmit changed data. The allocation center responds to the request by packaging and sending all resource change events that occurred within the current period. Each event includes a timestamp, event type, and change details. The cache server receives and analyzes the events, assigning a basic priority score based on the event type (e.g., 100 points for resource failure, 80 points for resource usage, 60 points for resource addition, 40 points for resource release, etc.). Further adjustments are made based on the scope of the change's impact (adding 20 points for affecting more than 3 departments). Based on the calculated cache priority, data is stored in tiers: data with a priority greater than 70 is stored in the fast access zone (using a Redis in-memory database), while data with a priority less than 70 is stored in the regular storage zone (using MongoDB or PostgreSQL Post-In). GresSQL (later Ingres Structured Query Language) continuously monitors the communication quality between the two subnets. When a round-trip latency exceeding 100ms or available bandwidth falling below 100Mbps is detected, an adaptive adjustment mechanism is triggered, increasing the update frequency from every 5 minutes to every 2 minutes and extending the data retention window from 1 hour to 4 hours to ensure data timeliness even when network quality deteriorates. Simultaneously, the historical communication interruption record database is queried to analyze all interruption events within the past 6 months, statistically analyzing the resource types and specific resource IDs frequently queried during the interruption. Based on this statistical information, the cache is pre-populated, preloading historically frequently accessed data into the fast access area to improve cache hit rate.

[0058] In some embodiments, efficient and reliable resource-sharing cache node construction can be achieved in a variety of ways: Optionally, a distributed caching architecture based on consistent hashing is adopted. Assuming a caching cluster is formed by deploying 5 high-performance servers, each configured with a 32-core, 64-thread processor, 256GB of memory, and 8 2TB SSDs in a RAID 10 configuration; data is allocated using the consistent hashing algorithm, with 150 virtual nodes per physical node, and MurmurHash3 as the hash function to ensure a data distribution standard deviation of <5%; each data item is stored in 3 replicas using a chained replication strategy, with writes synchronized in the order of master -> slave 1 -> slave 2, and reads from any replica; inter-node replication uses... The Gossip protocol synchronizes state, randomly selecting 3 nodes to exchange information every 3 seconds, and using vector clocks to resolve concurrent conflicts; fault detection employs a φ cumulative failure detector with a threshold φ=8, triggering data migration within 15 seconds when a node fails, with the migration speed limited to 100MB / s to avoid network congestion; a read-write separation architecture is implemented, with write requests routed to the master node and two-phase commit used to ensure consistency, while read requests are routed to the nearest node using consistent hashing; a Bloom filter is deployed, 64MB in size, with 4 hash functions and a false positive rate of 0.01%, directly returning results for non-existent keys, reducing invalid I / O; Optionally, a machine learning-based intelligent caching strategy is adopted. Assuming a 2-layer LSTM network with hidden layer dimensions [128, 64], the input is the access sequence of the past 48 hours, and the output is the access prediction for the next 4 hours. The training data contains 6 months of access logs, approximately 200 million records. A sliding window is used to generate time-series samples with a window size of 48 and a stride of 1. The model training uses the Adam optimizer with a learning rate of 0.001 and a batch size of 128. An early stopping strategy is used to prevent overfitting, and the model stops if the validation set loss does not decrease after 5 epochs. The predicted hotspot data is migrated from the mechanical hard drive to the solid-state drive in advance. Upgrading from SSDs to RAM, the tiered storage migration thresholds are access frequencies >10 times / hour and >50 times / hour, respectively. Reinforcement learning optimizes cache replacement, with states including 12-dimensional features such as cache occupancy and access frequency distribution. The action space includes four strategies: {LRU, LFU, FIFO, ARC}. Training uses the SARSA algorithm with a learning rate of 0.1, a discount factor of 0.9, an ε-greedy exploration rate of 0.1, and a reward function equal to the percentage increase in hit rate. Federated learning aggregates models from five cache nodes, using a proximal regularization algorithm to handle non-independent and identically distributed data, with a proximal term parameter μ=0.01. It is understood that other methods can also be used to achieve high-performance data caching and intelligent pre-filling across subnets; this is not limited here.

[0059] It should also be noted that the examples of all the specific values ​​mentioned above are merely exemplary embodiments, and the specific values ​​mentioned above are not limited to the examples.

[0060] This application's embodiments intelligently divide rescue departments into multiple communication subnets and establish resource-sharing cache nodes, achieving autonomous resource scheduling capabilities in the event of network interruption. A multi-dimensional evaluation mechanism accurately identifies resource complementarity relationships between departments, ensuring the rationality of subnet division. A hierarchical caching mechanism deployed at subnet boundaries guarantees reliable storage and rapid access to cross-subnet resource status data. Automatic triggering of network outage emergency modes and intelligent scheduling based on local and cached data maintain rescue response capabilities during communication interruptions. When inter-subnet or intra-subnet communication interruptions are detected, the system can adopt corresponding response strategies based on different interruption types, including intelligent task decomposition, resource consumption prediction, and temporary solution formulation, effectively reducing the impact of network failures on emergency rescue efficiency.

[0061] The intelligent airport emergency rescue resource allocation platform in the embodiments of this invention is described below from the perspective of hardware processing. (See attached document.) Figure 2 , Figure 2 This is a schematic diagram of the physical device structure of an intelligent allocation platform for airport emergency rescue resources in this application embodiment.

[0062] It should be noted that, Figure 2 The structure of the intelligent allocation platform for airport emergency rescue resources shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of the present invention.

[0063] like Figure 2 As shown, the airport emergency rescue resource intelligent allocation platform includes a Central Processing Unit (CPU) 201, which can perform various appropriate actions and processes based on programs stored in Read-Only Memory (ROM) 202 or programs loaded from storage section 208 into Random Access Memory (RAM) 203, such as executing the methods described in the above embodiments. The RAM 203 also stores various programs and data required for system operation. The CPU 201, ROM 202, and RAM 203 are interconnected via bus 204. Input / output (I / O) interface 205 is also connected to bus 204.

[0064] The following components are connected to I / O interface 205: input section 206 including audio input devices, push-button switches, etc.; output section 207 including a liquid crystal display (LCD) and audio output devices, indicator lights, etc.; storage section 208 including a hard disk, etc.; and communication section 209 including a network interface card such as a LAN (Local Area Network) card, modem, etc. Communication section 209 performs communication processing via a network such as the Internet. Drive 210 is also connected to I / O interface 205 as needed. Removable media 211, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., are installed on drive 210 as needed so that computer programs read from them can be installed into storage section 208 as needed.

[0065] In particular, according to embodiments of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present invention include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing computer programs for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 209, and / or installed from removable medium 211. When the computer program is executed by central processing unit (CPU) 201, it performs the various functions defined in the present invention.

[0066] It should be noted that specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), flash memory, optical fiber, portable compact disc read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this invention, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0067] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present invention. Each block in a flowchart or block diagram may represent a module, program segment, or portion of code, which contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those shown in the drawings.

[0068] Specifically, the airport emergency rescue resource intelligent allocation platform of this embodiment includes a processor and a memory. The memory stores a computer program. When the computer program is executed by the processor, it implements the airport emergency rescue resource intelligent allocation method provided in the above embodiment.

[0069] In another aspect, the present invention also provides a computer-readable storage medium, which may be included in the intelligent airport emergency rescue resource allocation platform described in the above embodiments; or it may exist independently and not assembled into the intelligent airport emergency rescue resource allocation platform. The storage medium carries one or more computer programs, which, when executed by a processor of the intelligent airport emergency rescue resource allocation platform, enable the intelligent airport emergency rescue resource allocation platform to implement the intelligent airport emergency rescue resource allocation method provided in the above embodiments.

[0070] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit it. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

[0071] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. This program can be stored in a computer-readable storage medium, and when executed, it can include the processes described in the above method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as ROM or random access memory (RAM), magnetic disks, or optical disks.

Claims

1. A method for intelligent allocation of airport emergency rescue resources, characterized in that, include: Upon obtaining communication network data between various rescue departments, communication subnets are divided for each rescue department based on the communication network data, so that target rescue departments that meet the preset division conditions are divided into the same communication subnet. Each communication subnet is configured with a resource allocation center, and a resource sharing cache node is established between two communication subnets that meet the preset cache establishment conditions. The resource sharing cache node is used to save cross-subnet resource status data within a preset time period before the communication interruption when the communication between the two communication subnets is interrupted. When a communication interruption is detected, communication link detection is performed between each communication subnet to obtain the link interruption detection result; The network outage emergency mode of the resource allocation center is triggered based on the link interruption detection result. The network outage emergency mode is used to schedule resources for each resource allocation center based on local resource status data and cross-subnet resource status data.

2. The method according to claim 1, characterized in that, Upon obtaining communication network data between various rescue departments, the process of dividing each rescue department into communication subnets based on the communication network data, so as to group target rescue departments that meet preset division conditions into the same communication subnet, specifically includes: Obtain communication bandwidth and communication latency data of the communication links between the various rescue departments; The communication bandwidth data is sorted in descending order to obtain a communication bandwidth sorting list, and the communication delay data is sorted in ascending order to obtain a communication delay sorting list. Identify a group of rescue departments corresponding to a group of communication links that are ranked in the top N positions in the communication bandwidth ranking list and in the top M positions in the communication delay ranking list; Obtain the physical location information and area of ​​responsibility information of the group of rescue departments within the airport, wherein the area of ​​responsibility information includes at least one of the terminal area, runway area, apron area and cargo area; Based on the physical location information and the area of ​​responsibility information, the physical distance between each rescue department in the group of rescue departments is determined, and one or more rescue departments whose physical distance is less than a preset distance threshold are divided into a neighboring rescue department group. Obtain real-time resource status data of the nearby rescue department group; Based on the real-time resource status data, the target rescue department with complementary resource relationships is identified from the group of neighboring rescue departments, and the target rescue department is divided into the same communication subnet.

3. The method according to claim 2, characterized in that, The step of determining the target rescue department with complementary resource relationships based on the real-time resource status data specifically includes: Feature extraction is performed on the real-time resource status data to obtain the resource type vector and resource quantity vector of each rescue department; The resource complementarity coefficient between any two rescue departments is determined based on the resource type vector. The resource complementarity coefficient is used to characterize the degree of matching between the surplus resources of one rescue department and the missing resources of the other rescue department. The resource load rate of each rescue department is determined based on the resource quantity vector; The resource load rate is compared with a preset load threshold to determine the high-load rescue departments whose resource load rate is greater than the preset load threshold and the low-load rescue departments whose resource load rate is less than the preset idle threshold. The rescue departments within the adjacent rescue department group are traversed and combined to identify the target rescue department that has the complementary resource relationship, and that satisfies the resource complementarity coefficient being greater than a preset complementarity threshold and includes at least one high-load rescue department and at least one low-load rescue department.

4. The method according to claim 1, characterized in that, The step of triggering the network outage emergency mode of the resource allocation center based on the link interruption detection result, wherein the network outage emergency mode performs scheduling for each resource allocation center based on local resource status data and cross-subnet resource status data, specifically including: Based on the link interruption detection results, the communication interruption type is identified, and when the communication interruption type is identified as inter-subnet communication interruption, at least two communication subnets where external communication interruption has occurred are determined. Send a network outage emergency trigger signal to at least two resource allocation centers corresponding to the at least two communication subnets, so as to obtain the cross-subnet resource status data read by the at least two resource allocation centers from at least one resource sharing cache node; A local resource pool is constructed based on the local resource status data, and a virtual external resource pool is constructed based on the cross-subnet resource status data. The local resource pool is used to store the information on the available and available resources within each of the at least two communication subnets. The virtual external resource pool is used to store the resource allocation status and occupied resource information of other communication subnets that have established a caching relationship with the at least two communication subnets before the communication interruption, obtained from the resource sharing cache node. Read the emergency rescue task information to be processed from at least two local task queues corresponding to the at least two resource allocation centers, and determine the priority of the emergency task based on the task type, urgency and impact scope in the emergency rescue task information; Based on the emergency task priority, the scheduling is performed preferentially within the local resource pool using the available rescue resource information and the resource availability status.

5. The method according to claim 4, characterized in that, The method further includes: When the local resource pool does not meet the resource requirements of the current emergency task, a resource conflict avoidance strategy is generated based on the resource availability status and the resource allocation status. The current emergency task is decomposed according to the resource conflict avoidance strategy to obtain the first sub-task and the second sub-task. Execute the first subtask within the local resource pool, and generate resource requirement tags and expected completion times for the second subtask; The resource requirement tag, the expected completion time, and the second subtask are associated and stored in any of the task cache queues of the at least two resource allocation centers; Upon detecting the restoration of communication between the at least two communication subnets, the second subtask is extracted from any of the task cache queues, and a cross-subnet resource request including the resource requirement flag and the expected completion time is sent to the second resource allocation center corresponding to the other communication subnet. The system receives the resource allocation confirmation information returned by the second resource allocation center after evaluating the expected completion time and its own resource scheduling status information, and schedules the second subtask according to the resource allocation confirmation information.

6. The method according to claim 4, characterized in that, The method further includes: When the communication interruption type is identified as an internal subnet communication interruption, the target communication subnet where the internal communication interruption occurred and at least two lost rescue departments within the target communication subnet are determined. Obtain the historical resource allocation records of the at least two missing rescue departments cached by the target resource allocation center corresponding to the target communication subnet before the communication interruption; Based on the resource usage patterns and task execution cycles in the historical resource allocation records, generate resource usage prediction information for the at least two missing rescue departments; Based on the resource occupancy prediction information, the resources of the at least two missing rescue departments are marked as pre-occupancy status information, and supplementary resource information is obtained from other normal communication rescue departments in the target communication subnet, excluding the at least two missing rescue departments. Based on the supplementary resource information, the available resource set of the target communication subnet is reconstructed, and resources are allocated to the newly added emergency tasks within the target communication subnet based on the available resource set to obtain a temporary resource allocation scheme. Upon detecting the restoration of communication between the at least two missing rescue departments, obtain the actual resource occupancy status information of the at least two missing rescue departments; Determine the resource status deviation between the actual resource occupancy status information and the pre-occupancy status information; The temporary resource allocation scheme is corrected based on the resource status occupancy deviation.

7. The method according to claim 1, characterized in that, The establishment of a resource-sharing cache node between two communication subnets that meet the preset cache establishment conditions specifically includes: A cache server is deployed at the network topology boundary of the two communication subnets, and a bidirectional communication interface is configured for the cache server. The bidirectional communication interface establishes a data transmission channel with at least two resource allocation centers that correspond one-to-one with the two communication subnets. A cache data update cycle is set within the cache server, and resource change events are obtained from the at least two resource allocation centers through the bidirectional communication interface during each cache data update cycle. The cache priority is determined based on the event type and scope of impact of the resource change event. The event types include resource addition, resource occupation, resource release, and resource failure. The cross-subnet resource status data is stored hierarchically according to the cache priority, so that the first resource status data with high cache priority is stored in the fast access area, and the second resource status data with low cache priority is stored in the normal storage area. When it is detected that the communication delay of the two communication subnets is greater than a preset delay threshold or the communication bandwidth is less than a preset bandwidth threshold, the current update frequency of the cached data update cycle is adjusted to the target update frequency, and the current time window of the preset time period is adjusted to the target time window, wherein the target update frequency is greater than the current update frequency and the target time window is greater than the current time window; The historical communication interruption records of the two communication subnets are obtained, and the data in the fast access area and the ordinary storage area are pre-filled according to the resource access mode in the historical communication interruption records to generate the resource sharing cache node.

8. An intelligent allocation platform for airport emergency rescue resources, characterized in that, The intelligent allocation platform for airport emergency rescue resources includes: one or more processors and a memory; the memory is coupled to the one or more processors, and the memory is used to store computer program code, the computer program code including computer instructions, and the one or more processors call the computer instructions to cause the intelligent allocation platform for airport emergency rescue resources to perform the method as described in any one of claims 1-7.

9. A computer-readable storage medium comprising instructions, characterized in that, When the instruction is executed on the airport emergency rescue resource intelligent allocation platform, the airport emergency rescue resource intelligent allocation platform performs the method as described in any one of claims 1-7.

10. A computer program product, characterized in that, When the computer program product is run on the airport emergency rescue resource intelligent allocation platform, the airport emergency rescue resource intelligent allocation platform performs the method as described in any one of claims 1-7.

Images