Method, system and equipment for managing and allocating medical emergency materials based on large model, and medium

Abstract

The invention belongs to the technical field of material allocation intellectualization, and particularly relates to a medical emergency material management allocation method, system, equipment and medium based on a large model, and the method comprises the steps: receiving the related information of an emergency, and generating the type and number of recommended medical materials through calling a medical large model and a general large model; the recommended medical supplies are audited and adjusted, and a finally allocated medical supply list is determined; calculating warehouses meeting reserve conditions according to the medical material demand information and the reserve of the warehouses, and screening out a recommended warehouse list; the optimal path of the unmanned aerial vehicle from the warehouse to the event occurrence site and the optimal path of the emergency vehicle from the warehouse to the event occurrence site are planned, and the paths of the emergency vehicle are calibrated in combination with the real-time road congestion data; and sending the material allocation information and the path planning to a warehouse administrator, an emergency vehicle driver and an unmanned aerial vehicle driving system. And the emergency response speed and the rescue efficiency are improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of intelligent material allocation, and specifically relates to a method, system, equipment and medium for managing and allocating medical emergency materials based on a large model. Background Art

[0002] In today's society, emergencies such as natural disasters, public health incidents, and major accidents pose serious threats to people's lives and property. In these emergencies, the rapid and effective deployment and management of emergency medical supplies is crucial to ensuring that affected people receive timely treatment and minimizing disaster losses.

[0003] Traditional decision-making in medical emergency supply management relies primarily on manual experience and simple information systems. After an emergency occurs, personnel must spend considerable time collecting and analyzing information, including the type of incident, location, number of affected individuals, and their needs. They then manually develop supply allocation plans based on this information. This approach is not only inefficient but also susceptible to human error, resulting in inaccurate supply allocation that fails to meet actual needs.

[0004] Warehouse selection is also a crucial issue in the material allocation process. Traditional warehouse selection methods typically only consider the warehouse's inventory level, while ignoring other important factors such as storage ratio and throughput. This can result in the selected warehouse having sufficient inventory but being unable to meet the allocation needs in a timely manner due to insufficient storage space or limited processing capacity, thus delaying the rescue effort. Summary of the Invention

[0005] In response to the above-mentioned deficiencies in the medical emergency supplies allocation process, the present invention provides a method, system, device and medium for managing and allocating medical emergency supplies based on a large model.

[0006] In a first aspect, the technical solution of the present invention provides a method for managing and allocating medical emergency supplies based on a large model, comprising: Receive relevant information about emergencies, including event type, location, number of affected people, and group characteristics. Generate recommended types and quantities of medical supplies by invoking the medical and general big models. Parse the text output by the big models to extract structured keywords. Review and adjust the recommended medical supplies and determine the final list of medical supplies to be allocated; Based on the medical supplies demand information in the medical supplies list and the warehouse's reserve quantity, the system calculates warehouses that meet the reserve requirements. Furthermore, the system selects a list of recommended warehouses based on the shortest distance between the incident location and the warehouse. Obtain real-time GPS information of emergency vehicles and drones, and select the emergency vehicles and drones closest to the recommended warehouse; Plan the optimal path for drones from the warehouse to the incident site, as well as the optimal path for emergency vehicles from the warehouse to the incident site, and calibrate the emergency vehicle paths based on real-time road congestion data; Send material allocation information and route planning to warehouse managers, emergency vehicle drivers, and drone driving systems.

[0007] By calling the medical big model and the general big model, it is possible to quickly generate recommended types and quantities of medical supplies. Combined with the analysis, review and adjustment of the big model output text, it ensures that the allocated medical supplies are more in line with actual needs and improves the accuracy and efficiency of material allocation. Taking into account the medical supply demand information, warehouse reserves, shortest distance and real-time road congestion data, a list of recommended warehouses is screened and the optimal route is planned, so that emergency supplies can be delivered to the scene at the fastest speed and shortest route, improving emergency response speed and rescue efficiency. From receiving event information to the final sending of material allocation information, it covers the key links of the entire medical emergency material management and allocation. Each link is closely connected and collaboratively optimized to achieve intelligent management of the entire process, effectively solving many problems existing in the traditional manual allocation process.

[0008] As a further limitation of the technical solution of the present invention, the steps of calculating warehouses that meet the storage conditions based on the medical supply demand information in the medical supply list and the warehouse reserve, and screening a list of recommended warehouses based on the shortest distance between the event location and the warehouse include: Compare the medical supplies list with the real-time inventory data of each warehouse to select candidate warehouses that meet the demand in terms of both material types and quantities; The warehouses in the candidate set are screened again based on their storage ratio and throughput; Calculate the shortest path distance between the incident location and the filtered warehouse; A recommended priority list of warehouses is generated by combining the warehouse's material matching degree, storage capacity, and the shortest path distance to the incident location.

[0009] By comparing the medical supply inventory with real-time inventory data from each warehouse, a candidate set of warehouses with the right variety and quantity of supplies was identified, ensuring sufficient inventory to meet allocation needs and preventing allocation delays due to insufficient warehouse supplies. A comprehensive assessment of multiple factors, including warehouse storage ratios, throughput, material matching, storage capacity, and the shortest path distance to the incident site, generated a priority list of recommended warehouses. This prioritized warehouses, ensuring they not only had adequate inventory but also offered advantages in transportation efficiency and distance, further optimizing the warehouse selection process.

[0010] As a further limitation of the technical solution of the present invention, the step of re-screening the warehouses in the candidate set based on the storage ratio and throughput of the warehouses includes: Obtain the real-time remaining capacity and total capacity of the warehouse to calculate the storage ratio; remove warehouses with storage ratios less than the set ratio threshold from the candidate warehouse set; Obtain the warehouse's material handling capacity per unit time, combine it with the total amount of allocated materials, calculate the estimated processing time, and then eliminate warehouse candidates whose centralized processing time exceeds the set time threshold; The storage capacity score of the screened warehouses is obtained by weighted summing up the storage ratio and throughput.

[0011] By calculating a warehouse's storage ratio and throughput, and screening it based on set ratio and time thresholds, we eliminate warehouses with insufficient storage capacity and low processing capabilities, ensuring that the selected warehouses have sufficient space and efficient material handling capabilities to quickly complete material allocation tasks. The weighted sum of the selected warehouses' storage ratio and throughput is used to generate a reserve capacity score, providing a scientific basis for subsequent warehouse prioritization, making warehouse selection more objective and reasonable, and improving the efficiency and reliability of the entire allocation system.

[0012] As a further limitation of the technical solution of the present invention, the step of generating a warehouse recommendation priority list based on the warehouse's material matching degree, storage capacity, and shortest path distance to the event location includes: For each candidate warehouse, calculate its material matching degree; Obtain the shortest transportation distance between the incident location and the warehouse, and calculate the distance score by introducing real-time traffic data; The material matching degree, storage capacity score, and distance score are weighted and summed to obtain the comprehensive score of each candidate warehouse; Arrange the warehouses in descending order of comprehensive scores to generate a priority list of warehouse recommendations.

[0013] The system comprehensively considers a warehouse's material matching degree, storage capacity score, and distance score, and calculates a weighted sum to derive a comprehensive score for each candidate warehouse. This comprehensive assessment of a warehouse's overall performance in terms of material storage, transportation efficiency, and distance makes the generated warehouse recommendation priority list more scientific and practical. Incorporating real-time traffic data into the distance score calculation allows for dynamic adjustment of warehouse priorities based on current traffic conditions, ensuring the optimal warehouse selection even under adverse conditions such as traffic congestion, further improving the timeliness and reliability of emergency material allocation.

[0014] As a further limitation of the technical solution of the present invention, the calculation formula for the material matching degree of the warehouse is as follows:

[0015] Where, Indicates the first i The amount of reserves of such materials, Indicates the i The demand for the material type, n is the total number of material types; Distance score

[0016] D is the shortest transportation distance, is the maximum distance among all candidate warehouses; Congestion coefficient; obtain the congestion coefficient based on the current real-time traffic, such as smooth, slow, congested, and severely congested, with different congestion coefficients;

[0017] Where, For the comprehensive rating, Score reserve capacity, is the weight coefficient.

[0018] Through a clear calculation formula, a warehouse's material matching degree, distance score, and overall score are quantitatively evaluated, making warehouse performance evaluation more objective and accurate. This facilitates the rapid selection of the most suitable warehouse from multiple candidate warehouses, improving the scientific nature and efficiency of decision-making. The introduction of the congestion coefficient allows the distance score to be adjusted based on real-time traffic conditions, further enhancing the system's adaptability to traffic congestion, ensuring that the optimal route is planned for emergency vehicles under various traffic conditions, and improving emergency response speed.

[0019] As a further limitation of the technical solution of the present invention, the step of planning the optimal path for the drone from the warehouse to the location of the incident includes: Define a path node set V, including the warehouse location node S and the incident location node T; initialize the distance array to store the currently known shortest distance from the starting point S to the node v; create a priority queue to store the nodes to be processed, and the initial state only contains the starting point S; establish an array of parent node records; Remove the node u with the smallest current distance value from the priority queue; traverse all unvisited adjacent nodes v of u and perform the following processing: calculate the candidate distance from S through u to v; if the candidate distance is less than the current shortest distance, update the distance array with the candidate distance, set the parent node record array to node u, and add node v to the priority queue; When the incident location node T is taken out and processed, the algorithm is terminated and the process starts from the incident location node T and traces back along the parent node record array to the warehouse location node S to generate a complete flight path; if the priority queue is empty and T has not been processed yet, it is determined that the path does not exist.

[0020] Using the Dijkstra algorithm to plan the optimal path for the drone from the warehouse to the incident site quickly and accurately finds the shortest path, ensuring the drone reaches the incident site in the shortest possible time, improving emergency rescue efficiency. From defining the path node set, initializing the distance array and priority queue, to traversing the nodes, updating the distance, and generating the complete flight path, the entire process is logically clear and step-by-step, effectively avoiding errors and omissions in path planning and ensuring the feasibility and reliability of the path. Furthermore, timely detection and feedback are provided if a path does not exist, facilitating the implementation of alternative emergency measures.

[0021] As a further limitation of the technical solution of the present invention, the step of calibrating the emergency vehicle's path in combination with real-time road congestion data includes: Obtain real-time road congestion data from an online map API, including the road congestion status, average vehicle speed, and estimated travel time; Dynamically adjust edge weights in path planning based on the real-time road congestion data, where the weights of congested road sections are increased and the weights of unobstructed road sections remain unchanged or decrease; Use the weighted road network information to recalculate the optimal path from the warehouse to the incident site; If the recalculated path is different from the original planned path and the estimated arrival time is shorter, the path planning result is updated.

[0022] Dynamically calibrating emergency vehicle routes based on real-time traffic congestion data allows for timely adjustments to route planning based on current traffic conditions, avoiding delays caused by congestion and ensuring that emergency vehicles can reach the scene as quickly as possible, improving the timeliness and effectiveness of emergency response. By dynamically adjusting edge weights in route planning and recalculating the optimal route based on adjusted road network information, route planning can better adapt to real-time traffic changes, improving its adaptability and reliability and further optimizing the transportation efficiency of emergency vehicles.

[0023] In a second aspect, the technical solution of the present invention further provides a system for managing and allocating medical emergency supplies based on a large model, comprising: The information receiving module receives relevant information about the emergency, including the type of event, location, number of affected people and group characteristics; The large model calling module is used to generate recommended types and quantities of medical supplies by calling the medical large model and the general large model; The text parsing module is used to parse the text output by the large model and extract structured keywords; The material review module is used to review and adjust the recommended medical materials and determine the final list of medical materials to be allocated; The warehouse screening module is used to calculate the warehouses that meet the storage conditions based on the medical supply demand information in the medical supply list and the warehouse's reserve quantity, and to filter out a list of recommended warehouses based on the shortest distance between the incident location and the warehouse; The path planning module is used to obtain real-time GPS information of emergency vehicles and drones, select the emergency vehicles and drones closest to the recommended warehouse, plan the optimal path for the drone from the warehouse to the incident location, and the optimal path for the emergency vehicle from the warehouse to the incident location, and calibrate the emergency vehicle's path based on real-time road congestion data; The message center module is used to send material allocation information and route planning to warehouse managers, emergency vehicle drivers, and drone driving systems.

[0024] A complete large-scale model-based medical emergency supply management and allocation system has been constructed, encompassing multiple modules including information reception, large-scale model invocation, text parsing, supply auditing, warehouse screening, route planning, and a message center. This system enables systematic management and collaborative operations across all links, improving the efficiency and reliability of the entire allocation process. Through the collaborative work of these modules, comprehensive intelligent decision-making support is provided for emergency supply allocation. From recommending supply types and quantities, to selecting warehouses and planning routes, to sending allocation information, the system is able to quickly and accurately complete various tasks, providing strong support for emergency rescue efforts.

[0025] As a further limitation of the technical solution of the present invention, the warehouse screening module includes: The initial screening unit is used to compare the medical supplies list with the real-time reserve data of each warehouse to select a candidate set of warehouses whose material types and quantities meet the requirements; A secondary screening unit, used to re-screen the warehouses in the candidate set based on their storage ratio and throughput; A distance calculation unit calculates the shortest path distance between the event location and the screened warehouse; The priority list generation unit is used to generate a warehouse recommendation priority list based on the warehouse's material matching degree, storage capacity and the shortest path distance to the event location.

[0026] As a further limitation of the technical solution of the present invention, the secondary screening unit includes: The storage ratio screening submodule is used to obtain the real-time remaining capacity and total capacity of the warehouse to calculate the storage ratio; and eliminate warehouses with storage ratios less than the set ratio threshold from the warehouse candidate set; The processing time screening submodule is used to obtain the material processing capacity value of the warehouse per unit time, combine it with the total amount of allocated materials, calculate the expected processing time, and then eliminate warehouse candidates whose centralized processing time is greater than the set time threshold.

[0027] As a further limitation of the technical solution of the present invention, the priority list generating unit includes: The reserve capacity calculation submodule is used to calculate the weighted sum of the selected warehouses according to the storage ratio and throughput to obtain the reserve capacity score; The material matching degree calculation submodule is used to calculate the material matching degree of each candidate warehouse; The distance score calculation submodule is used to obtain the shortest transportation distance between the event location and the warehouse, and calculate the distance score after introducing real-time traffic data; The comprehensive score calculation submodule calculates the weighted sum of the material matching degree, storage capacity score, and distance score to obtain the comprehensive score of each candidate warehouse; The list generation submodule is used to sort warehouses in descending order according to their comprehensive scores and generate a warehouse recommendation priority list.

[0028] The material matching degree calculation formula of the warehouse is as follows:

[0029] Where, Indicates the first i The amount of reserves of such materials, Indicates the i The demand for the material type, n is the total number of material types; Distance score

[0030] D is the shortest transportation distance, is the maximum distance among all candidate warehouses; Congestion coefficient; obtain the congestion coefficient based on the current real-time traffic, such as smooth, slow, congested, and severely congested, with different congestion coefficients;

[0031] Where, For the comprehensive rating, Score reserve capacity, is the weight coefficient.

[0032] As a further limitation of the technical solution of the present invention, the path planning module includes a UAV path planning unit, which is used to plan the optimal path of the UAV from the warehouse to the location of the incident; specifically, it is used to define a path node set V, including a warehouse location node S and an incident location node T; initialize the distance array to store the currently known shortest distance from the starting point S to the node v; create a priority queue to store the nodes to be processed, and the initial state only contains the starting point S; establish a parent node record array; take out the node u with the smallest current distance value from the priority queue; traverse all unvisited adjacent nodes v of u and perform the following processing: calculate the candidate distance from S through u to v; if the candidate distance is less than the currently known shortest distance, update the distance array with the candidate distance, set the parent node record array to node u, and add node v to the priority queue; when the incident location node T is taken out for processing, terminate the algorithm and start from the incident location node T and trace back along the parent node record array to the warehouse location node S to generate a complete flight path; if the priority queue is empty and T has not been processed yet, it is determined that the path does not exist.

[0033] As a further limitation of the technical solution of the present invention, the path planning module also includes a path calibration unit for calibrating the path of the emergency vehicle in combination with real-time road congestion data; specifically, it is used to obtain real-time road congestion data from an online map API, and the real-time road congestion data includes the congestion status of the road, the average vehicle speed and the estimated travel time; according to the real-time road congestion data, the edge weights in the path planning are dynamically adjusted, wherein the weight of the congested section is increased, and the weight of the unobstructed section remains unchanged or decreases; using the road network information with adjusted weights, the optimal path from the warehouse to the scene of the incident is recalculated; if the recalculated path is different from the originally planned path and the estimated arrival time is shorter, the path planning result is updated.

[0034] In a third aspect, the technical solution of the present invention also provides an electronic device, comprising: at least one processor; and a memory communicatively connected to the at least one processor; the memory stores computer program instructions that can be executed by the at least one processor, and the computer program instructions are executed by the at least one processor so that the at least one processor can execute the method for management and allocation of medical emergency materials based on a large model as described in the first aspect.

[0035] In a fourth aspect, the technical solution of the present invention also provides a non-transitory computer-readable storage medium, which stores computer instructions, and the computer instructions enable the computer to execute the method for medical emergency material management and allocation based on a large model as described in the first aspect.

[0036] The present invention has the beneficial effects of utilizing a large medical model and a general large model, combined with information related to an emergency, to generate recommendations for types and quantities of medical supplies, improving the accuracy of these recommendations and ensuring that the supplies provided are more aligned with actual needs. The text output by the large model is parsed and structured keywords are extracted, facilitating the efficient processing and analysis of the recommended supplies, saving manual processing time and improving overall efficiency. Recommended medical supplies are reviewed and adjusted to determine the final allocation list, avoiding blind allocation of supplies, making allocation more scientific and rational, and improving resource utilization. Recommended warehouses are filtered based on medical supply demand information and warehouse inventory, taking into account distance, comprehensively considering supply and transportation costs. This facilitates the rapid and efficient acquisition of required supplies, reducing transportation time and costs. Optimal routes are planned for drones and emergency vehicles, and emergency vehicle routes are calibrated, taking into account real-time traffic congestion data, ensuring that supplies can be delivered to the incident site as quickly as possible, improving the timeliness of emergency rescue efforts. The allocation information and route plans are distributed to relevant personnel and systems, ensuring timely information sharing and communication, enabling collaborative work and improving emergency response capabilities. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0038] Figure 1 A schematic flow chart of a method according to an embodiment of the present invention.

[0039] Figure 2 A schematic block diagram of a system according to an embodiment of the present invention. DETAILED DESCRIPTION

[0040] In order to make the purpose, features, and advantages of the present invention more obvious and easy to understand, the technical solutions of the present invention will be clearly and completely described below in conjunction with the drawings in the specific embodiments. Obviously, the embodiments described below are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0041] like Figure 1 As shown, an embodiment of the present invention provides a method for managing and allocating medical emergency supplies based on a large model, including: S1. Receive relevant information about the emergency, including the type of incident, location, number of affected people, and group characteristics. Generate recommended types and quantities of medical supplies by invoking the medical and general models. Parse the text output by the model and extract structured keywords. It's important to note that emergency information typically comes from on-site rescue workers or witnesses entering event information into the system via mobile devices (e.g., mobile phones or tablets). This information can be entered through a dedicated emergency app, which provides a standardized form for users to fill in the event type (e.g., earthquake, flood), location (via GPS or manual address entry), number of affected people, and group characteristics (e.g., whether there are pregnant women, infants, the elderly, or other special groups).

[0042] Automatic monitoring equipment (such as sensors and cameras) is deployed in high-risk areas (such as earthquake-prone and flood-prone areas). These devices can monitor environmental changes in real time and automatically trigger an event reporting mechanism upon detecting an anomaly, sending the event information to the system.

[0043] Design a standardized information entry form, ensuring all required fields are filled out. The form may include the following fields: Event Type: Select from the drop-down menu (e.g. natural disaster, accident, etc.).

[0044] Location: GPS positioning or manual address entry (map selection function supported).

[0045] Number of people affected: Enter the specific number in the input box.

[0046] Group characteristics: Check box selection (such as pregnant women, infants, elderly people, etc.).

[0047] Data Verification: During the information entry process, user-entered data is verified to ensure accuracy and completeness. For example, the system checks whether the number of affected people is a positive integer or whether the address meets format requirements. If data does not meet the requirements, the user is prompted to re-enter the data.

[0048] Choose specialized large-scale models in the medical field, such as MedGPT and Lingyi. These models are trained on extensive medical data and can generate precise medical supply recommendations based on the type of incident and the characteristics of the affected population. Choose large-scale models in general fields, such as DeepSeek and Baichuan 3. These models have extensive domain knowledge and can combine event context (such as geographic location and disaster scale) to generate more comprehensive supply recommendations.

[0049] The big model is called through an API. Collected emergency information (type of incident, location, number of affected people, and group characteristics) is sent as input parameters to the big model's API. The big model typically returns output in text format, including recommended types and quantities of medical supplies.

[0050] To parse the text output by large models and extract structured keywords, a bidirectional long short-term memory (BiLSTM) model combined with a conditional random field (CRF) model is typically used. BiLSTM can capture contextual information in the text, while CRF can effectively identify keywords and structured information in the text.

[0051] Model training and deployment require a large amount of labeled data for training the BiLSTM-CRF model. This labeled data includes key information such as the name and quantity of medical supplies. Using this labeled data, the BiLSTM-CRF model is trained to optimize model parameters and improve accuracy and recall. The trained model is deployed to the server, and parsing services are provided through the API.

[0052] The text returned by the large model is fed into the BiLSTM-CRF model. The model outputs structured keyword information and extracts the names and quantities of medical supplies.

[0053] S2. Review and adjust the recommended medical supplies and determine the final list of medical supplies to be allocated; In this step, medical experts assess whether the recommended medical supplies meet medical standards based on the type of event (e.g., earthquake, flood, etc.) and the characteristics of the affected population (e.g., whether there are any special populations involved). They also check whether the recommended medical supplies are comprehensive and whether there are any omissions. For example, in earthquake relief, in addition to standard first aid medicines and equipment, specialized supplies such as fracture fixation devices and tourniquets are also necessary. Material management experts assess whether the recommended quantities of medical supplies meet actual needs based on the number of people affected and their demographics. They communicate with on-site rescue personnel in real time through video conferencing, phone calls, or instant messaging to understand actual needs and special circumstances on the ground. For example, specific models of medical equipment or specialized medications may be required on-site. Based on on-site feedback, the recommended medical supply list is adjusted promptly. For example, if demand for a particular supply is significantly higher than the recommended quantity, the quantity of that supply should be increased immediately.

[0054] If some recommended medical supplies are not suitable for the current type of incident or on-site conditions, select appropriate alternative supplies. For example, in remote mountainous areas, some large medical equipment may not be available due to transportation difficulties and can be replaced with portable medical equipment. Based on the audit results, supplement the medical supplies that are missing from the recommended list. Based on the audit results, adjust the quantity of recommended medical supplies. If the quantity of certain supplies is insufficient, increase the quantity; if the quantity of certain supplies is excessive, reduce it appropriately to optimize resource allocation. When adjusting the quantity of supplies, appropriate redundancy should be considered to cope with possible emergencies. Summarize the opinions and suggestions of the audit team to form the final medical supply allocation list. The final list will be confirmed and signed for approval by the head of the audit team. The confirmed list will serve as the basis for emergency supply allocation and will be sent to relevant personnel such as warehouse managers, emergency vehicle drivers, and drone driving systems. The audit process and final list will be recorded and archived for subsequent traceability and evaluation.

[0055] S3. Based on the medical supply demand information in the medical supply list and the warehouse's reserve quantity, calculate the warehouses that meet the reserve requirements. Combined with the shortest distance between the incident location and the warehouse, select a list of recommended warehouses. This includes: S31. Compare the medical supplies list with the real-time reserve data of each warehouse to select a candidate set of warehouses whose material types and quantities meet the requirements; S32. Rescreen the warehouses in the candidate set based on their storage ratio and throughput; S33, calculating the shortest path distance between the location where the event occurred and the screened warehouse; To calculate the shortest path distance, you typically use an online mapping service that provides an API. Follow the documentation for the online mapping service to obtain the API for calculating the shortest path distance between two points. Send a request to the API with the event location and warehouse location as parameters, and extract the shortest path distance from the API response.

[0056] S34. Generate a warehouse recommendation priority list based on the warehouse's material matching degree, storage capacity, and shortest path distance to the incident location.

[0057] S4. Obtain real-time GPS information of emergency vehicles and drones, and select the emergency vehicles and drones closest to the recommended warehouse; S5. Plan the optimal path for the drone from the warehouse to the incident location, as well as the optimal path for the emergency vehicle from the warehouse to the incident location, and calibrate the emergency vehicle's path based on real-time road congestion data; S6. Send material allocation information and path planning to warehouse managers, emergency vehicle drivers, and drone driving systems.

[0058] In some embodiments, the step of re-screening the warehouses in the candidate set based on the storage ratio and throughput of the warehouses includes: S321. Obtain the real-time remaining capacity and total capacity of the warehouse to calculate the storage ratio; remove warehouses whose storage ratio is less than a set ratio threshold from the candidate warehouse set; S322. Obtain the warehouse's material handling capacity per unit time, combine it with the total amount of allocated materials, calculate the estimated handling time, and again eliminate candidate warehouses whose centralized handling time exceeds the set time threshold. S323. Perform a weighted summation of the selected warehouses according to the storage ratio and throughput to obtain a reserve capacity score.

[0059] In some embodiments, the step of generating a warehouse recommendation priority list based on the warehouse's material matching degree, storage capacity, and shortest path distance to the incident location includes: S341. For each candidate warehouse, calculate its material matching degree; The material matching degree calculation formula of the warehouse is as follows:

[0060] Where, Indicates the first i The amount of reserves of such materials, Indicates the i The demand for the material type, n is the total number of material types; S342. Obtain the shortest transportation distance between the incident location and the warehouse, and calculate the distance score after introducing real-time traffic data; Distance score

[0061] D is the shortest transportation distance, is the maximum distance among all candidate warehouses; Congestion coefficient; obtain the congestion coefficient based on the current real-time traffic, such as smooth, slow, congested, and severely congested, with different congestion coefficients; S343. Perform a weighted summation of the material matching degree, storage capacity score, and distance score to obtain a comprehensive score for each candidate warehouse;

[0062] Where, For the comprehensive rating, Score reserve capacity, is the weight coefficient.

[0063] S344. Arrange the warehouses in descending order according to the comprehensive scores and generate a warehouse recommendation priority list.

[0064] In some embodiments, the steps of planning an optimal path for a drone from a warehouse to the location of an incident include: Define a set of path nodes V, including the warehouse location node S and the incident location node T; initialize a distance array to store the currently known shortest distance from the starting point S to node v; create a priority queue to store the nodes to be processed, and the initial state only includes the starting point S; establish a parent node record array; S51. Take out the node u with the smallest current distance value from the priority queue; traverse all unvisited adjacent nodes v of u and perform the following processing: calculate the candidate distance from S through u to v; if the candidate distance is less than the currently known shortest distance, update the distance array with the candidate distance, set the parent node record array to node u, and add node v to the priority queue; S52. When the incident location node T is taken out for processing, terminate the algorithm and start from the incident location node T to trace back along the parent node record array to the warehouse location node S to generate a complete flight path; if the priority queue is empty and T has not been processed yet, it is determined that the path does not exist.

[0065] In addition, using the A* algorithm to plan the optimal path for an emergency vehicle from a warehouse to the incident location specifically includes: (1) Initialization stage: Define a set of path nodes V, including the warehouse location node S and the incident location node T; Initialize the cost evaluation function: g(v): The actual cost from the starting point S to node v; h(v): The heuristic estimated cost from node v to the end point T; Set g(S)=0 at the starting point, and g(v)=+∞ for other nodes; Create a priority queue OpenList, sorted by f(v)=g(v)+h(v), and the initial state includes the starting point S; Establish a parent node record array prev[v] for path backtracking; (2) Node processing stage: Take out the node u with the smallest f value from OpenList; If u is the target node T, terminate the algorithm and perform path backtracking; Traverse all adjacent nodes v of u: a) Calculate the actual cost g_temp = g(u) + d(u, v) from S through u to v, where d(u, v) is the actual driving cost from u to v; b) If g_temp < g(v), then: Update g(v) = g_temp; Calculate f(v) = g(v) + h(v); Set prev[v] = u; If v is not in OpenList, add it to OpenList; (3) Termination judgment stage: When the target node T is taken out and processed, the algorithm is terminated and the path backtracking is performed; If OpenList is empty and T has not been processed yet, it is determined that the path does not exist; (4) Path generation stage: Starting from T and tracing back to S along the prev array, a complete driving path is generated; The path is dynamically optimized based on real-time traffic data, and the final navigation route is output.

[0066] It should be noted that the steps for calibrating the emergency vehicle's path using real-time road congestion data include: Real-time road congestion data is obtained from an online map API, including the road congestion status, average vehicle speed, and estimated travel time. Based on the real-time road congestion data, edge weights in path planning are dynamically adjusted, with the weights of congested sections increased and the weights of unobstructed sections remaining unchanged or decreased. The optimal path from the warehouse to the incident site is recalculated using the weighted road network information. If the recalculated path differs from the originally planned path and has a shorter estimated arrival time, the path planning result is updated.

[0067] like Figure 2 As shown, an embodiment of the present invention further provides a system for managing and allocating medical emergency supplies based on a large model, including: The information receiving module receives relevant information about the emergency, including the type of event, location, number of affected people and group characteristics; The large model calling module is used to generate recommended types and quantities of medical supplies by calling the medical large model and the general large model; The text parsing module is used to parse the text output by the large model and extract structured keywords; The material review module is used to review and adjust the recommended medical materials and determine the final list of medical materials to be allocated; The warehouse screening module is used to calculate the warehouses that meet the storage conditions based on the medical supply demand information in the medical supply list and the warehouse's reserve quantity, and to filter out a list of recommended warehouses based on the shortest distance between the incident location and the warehouse; The path planning module is used to obtain real-time GPS information of emergency vehicles and drones, select the emergency vehicles and drones closest to the recommended warehouse, plan the optimal path for the drone from the warehouse to the incident location, and the optimal path for the emergency vehicle from the warehouse to the incident location, and calibrate the emergency vehicle's path based on real-time road congestion data; The message center module is used to send material allocation information and route planning to warehouse managers, emergency vehicle drivers, and drone driving systems.

[0068] In some embodiments, the warehouse screening module includes: The initial screening unit is used to compare the medical supplies list with the real-time reserve data of each warehouse to select a candidate set of warehouses whose material types and quantities meet the requirements; A secondary screening unit, used to re-screen the warehouses in the candidate set based on their storage ratio and throughput; A distance calculation unit calculates the shortest path distance between the event location and the screened warehouse; The priority list generation unit is used to generate a warehouse recommendation priority list based on the warehouse's material matching degree, storage capacity and the shortest path distance to the event location.

[0069] In some embodiments, the secondary screening unit comprises: The storage ratio screening submodule is used to obtain the real-time remaining capacity and total capacity of the warehouse to calculate the storage ratio; and eliminate warehouses with storage ratios less than the set ratio threshold from the warehouse candidate set; The processing time screening submodule is used to obtain the material processing capacity value of the warehouse per unit time, combine it with the total amount of allocated materials, calculate the expected processing time, and then eliminate warehouse candidates whose centralized processing time is greater than the set time threshold.

[0070] In some embodiments, the priority list generating unit includes: The reserve capacity calculation submodule is used to calculate the weighted sum of the selected warehouses according to the storage ratio and throughput to obtain the reserve capacity score; The material matching degree calculation submodule is used to calculate the material matching degree of each candidate warehouse; The distance score calculation submodule is used to obtain the shortest transportation distance between the event location and the warehouse, and calculate the distance score after introducing real-time traffic data; The comprehensive score calculation submodule calculates the weighted sum of the material matching degree, storage capacity score, and distance score to obtain the comprehensive score of each candidate warehouse; The list generation submodule is used to sort warehouses in descending order according to their comprehensive scores and generate a warehouse recommendation priority list.

[0071] The material matching degree calculation formula of the warehouse is as follows:

[0072] Where, Indicates the first i The amount of reserves of such materials, Indicates the i The demand for the material type, n is the total number of material types; Distance score

[0073] D is the shortest transportation distance, is the maximum distance among all candidate warehouses; Congestion coefficient; obtain the congestion coefficient based on the current real-time traffic, such as smooth, slow, congested, and severely congested, with different congestion coefficients;

[0074] Where, For the comprehensive rating, Score reserve capacity, is the weight coefficient.

[0075] In some embodiments, the path planning module includes a drone path planning unit, which is used to plan the optimal path of the drone from the warehouse to the location of the incident; specifically, it is used to define a path node set V, including a warehouse location node S and an incident location node T; initialize the distance array to store the currently known shortest distance from the starting point S to the node v; create a priority queue to store the nodes to be processed, and the initial state only contains the starting point S; establish a parent node record array; take out the node u with the smallest current distance value from the priority queue; traverse all unvisited adjacent nodes v of u and perform the following processing: calculate the candidate distance from S through u to v; if the candidate distance is less than the currently known shortest distance, update the distance array with the candidate distance, set the parent node record array to node u, and add node v to the priority queue; when the incident location node T is taken out for processing, terminate the algorithm and start from the incident location node T and trace back along the parent node record array to the warehouse location node S to generate a complete flight path; if the priority queue is empty and T has not been processed yet, it is determined that the path does not exist.

[0076] In some embodiments, the path planning module also includes a path calibration unit for calibrating the path of the emergency vehicle in combination with real-time road congestion data; specifically, it is used to obtain real-time road congestion data from an online map API, wherein the real-time road congestion data includes the congestion status of the road, the average vehicle speed and the estimated travel time; according to the real-time road congestion data, the edge weights in the path planning are dynamically adjusted, wherein the weight of the congested road section is increased, and the weight of the unobstructed road section remains unchanged or decreases; using the road network information with adjusted weights, the optimal path from the warehouse to the scene of the incident is recalculated; if the recalculated path is different from the originally planned path and the estimated arrival time is shorter, the path planning result is updated.

[0077] An embodiment of the present invention further provides an electronic device comprising: a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other via the communication bus. The communication bus can be used to transmit information between the electronic device and a sensor. The processor can call the logic instructions in the memory to execute the following method: S1. Receive relevant information of the emergency event, including the event type, location, number of affected people and group characteristics, and generate recommended types and quantities of medical supplies by calling the medical big model and the general big model; parse the text output by the big model and extract structured keywords; S2. Review and adjust the recommended medical supplies to determine the final list of medical supplies to be allocated; S3. Based on the medical supply demand information and warehouse reserves in the medical supply list, calculate the warehouses that meet the reserve conditions, and combine the shortest distance between the event location and the warehouse to filter out the recommended warehouse list; S4. Obtain real-time GPS information of emergency vehicles and drones, and select the emergency vehicles and drones closest to the recommended warehouses; S5. Plan the optimal path for the drone from the warehouse to the event location, and the optimal path for the emergency vehicle from the warehouse to the event location, and calibrate the path of the emergency vehicle in combination with real-time road congestion data; S6. Send the material allocation information and path planning to the warehouse manager, emergency vehicle driver and drone driving system.

[0078] Furthermore, the logical instructions in the aforementioned memory can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage media include various media capable of storing program code, such as USB flash drives, mobile hard drives, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical disks.

[0079] An embodiment of the present invention provides a non-transitory computer-readable storage medium, which stores computer instructions. The computer instructions enable a computer to execute the method provided by the above method embodiment, for example, including: S1, receiving relevant information of an emergency, including the type of event, the location of the event, the number of affected people and group characteristics, generating recommended types and quantities of medical supplies by calling a medical big model and a general big model; parsing the text output by the big model and extracting structured keywords; S2, reviewing and adjusting the recommended medical supplies, and determining the final list of medical supplies to be allocated; S3, Based on the medical supplies demand information and warehouse reserves in the list, the warehouses that meet the reserve conditions are calculated, and the recommended warehouse list is screened out in combination with the shortest distance between the incident site and the warehouse; S4, the real-time GPS information of the emergency vehicle and the drone is obtained, and the emergency vehicle and drone closest to the recommended warehouse are selected; S5, the optimal path for the drone from the warehouse to the incident site, and the optimal path for the emergency vehicle from the warehouse to the incident site are planned, and the path of the emergency vehicle is calibrated in combination with real-time road congestion data; S6, the material allocation information and path planning are sent to the warehouse manager, emergency vehicle driver and drone driving system.

[0080] Although the present invention has been described in detail with reference to the accompanying drawings and in conjunction with preferred embodiments, the present invention is not limited thereto. Without departing from the spirit and essence of the present invention, persons of ordinary skill in the art may make various equivalent modifications or substitutions to the embodiments of the present invention, and such modifications or substitutions shall be within the scope of the present invention. Any changes or substitutions that can be easily conceived by persons skilled in the art within the technical scope disclosed in the present invention shall be within the scope of protection of the present invention.

Claims

1. A method for managing and allocating medical emergency supplies based on a large model, characterized in that: include: Receive relevant information about emergencies, including the type of incident, location, number of affected people, and group characteristics, and generate recommended types and quantities of medical supplies by calling the medical and general models; Review and adjust the recommended medical supplies and determine the final list of medical supplies to be allocated; Based on the medical supplies demand information in the medical supplies list and the warehouse's reserve quantity, the system calculates warehouses that meet the reserve requirements. Furthermore, the system selects a list of recommended warehouses based on the shortest distance between the incident location and the warehouse. Obtain real-time GPS information of emergency vehicles and drones, and select the emergency vehicles and drones closest to the recommended warehouse; Plan the optimal path for drones from the warehouse to the incident site, as well as the optimal path for emergency vehicles from the warehouse to the incident site, and calibrate the emergency vehicle paths based on real-time road congestion data; Send material allocation information and route planning to warehouse managers, emergency vehicle drivers, and drone driving systems.

2. The method for managing and allocating medical emergency supplies based on a large model according to claim 1 is characterized in that: Based on the medical supply demand information in the medical supply list and the warehouse's reserve quantity, the steps for calculating warehouses that meet the reserve conditions and combining the shortest distance between the incident location and the warehouse to screen out the recommended warehouse list include: Compare the medical supplies list with the real-time inventory data of each warehouse to select candidate warehouses that meet the demand in terms of both material types and quantities; The warehouses in the candidate set are screened again based on their storage ratio and throughput; Calculate the shortest path distance between the incident location and the filtered warehouse; A recommended priority list of warehouses is generated by combining the warehouse's material matching degree, storage capacity, and the shortest path distance to the incident location.

3. The method for managing and allocating medical emergency supplies based on a large model according to claim 2 is characterized in that: The steps for re-screening the candidate warehouses based on their storage ratio and throughput include: Obtain the real-time remaining capacity and total capacity of the warehouse to calculate the storage ratio; remove warehouses with storage ratios less than the set ratio threshold from the candidate warehouse set; Obtain the warehouse's material handling capacity per unit time, combine it with the total amount of allocated materials, calculate the estimated processing time, and then eliminate warehouse candidates whose centralized processing time exceeds the set time threshold; The storage capacity score of the screened warehouses is obtained by weighted summing up the storage ratio and throughput.

4. The method for managing and allocating medical emergency supplies based on a large model according to claim 3 is characterized in that: The steps to generate a recommended warehouse priority list based on the warehouse's material matching degree, storage capacity, and shortest path distance to the incident location include: For each candidate warehouse, calculate its material matching degree; Obtain the shortest transportation distance between the incident location and the warehouse, and calculate the distance score by introducing real-time traffic data; The material matching degree, storage capacity score, and distance score are weighted and summed to obtain the comprehensive score of each candidate warehouse; Arrange the warehouses in descending order of comprehensive scores to generate a priority list of warehouse recommendations.

5. The method for managing and allocating medical emergency supplies based on a large model according to claim 4 is characterized in that: The material matching degree calculation formula of the warehouse is as follows: Where, Indicates the first i The amount of reserves of such materials, Indicates the i The demand for the material type, n is the total number of material types; Distance score D is the shortest transportation distance, is the maximum distance among all candidate warehouses; is the congestion coefficient; Where, For the comprehensive rating, Score reserve capacity, is the weight coefficient.

6. The method for managing and allocating medical emergency supplies based on a large model according to claim 5 is characterized in that: The steps for planning the optimal path for the drone from the warehouse to the incident location include: Define a path node set V, including the warehouse location node S and the incident location node T; initialize the distance array to store the currently known shortest distance from the starting point S to the node v; create a priority queue to store the nodes to be processed, and the initial state only contains the starting point S; establish an array of parent node records; Remove the node u with the smallest current distance value from the priority queue; traverse all unvisited adjacent nodes v of u and perform the following processing: calculate the candidate distance from S through u to v; if the candidate distance is less than the current shortest distance, update the distance array with the candidate distance, set the parent node record array to node u, and add node v to the priority queue; When the incident location node T is taken out and processed, the algorithm is terminated and the process starts from the incident location node T and traces back along the parent node record array to the warehouse location node S to generate a complete flight path; if the priority queue is empty and T has not been processed yet, it is determined that the path does not exist.

7. The method for managing and allocating medical emergency supplies based on a large model according to claim 6 is characterized in that: The steps for calibrating emergency vehicle routes using real-time road congestion data include: Obtain real-time road congestion data from an online map API, including the road congestion status, average vehicle speed, and estimated travel time; Dynamically adjust edge weights in path planning based on the real-time road congestion data, where the weights of congested road sections are increased and the weights of unobstructed road sections remain unchanged or decrease; Use the weighted road network information to recalculate the optimal path from the warehouse to the incident site; If the recalculated path is different from the original planned path and the estimated arrival time is shorter, the path planning result is updated.

8. A system for managing and allocating medical emergency supplies based on a large model, characterized by: include: The information receiving module receives relevant information about the emergency, including the type of event, location, number of affected people and group characteristics; The large model calling module is used to generate recommended types and quantities of medical supplies by calling the medical large model and the general large model; The text parsing module is used to parse the text output by the large model and extract structured keywords; The material review module is used to review and adjust the recommended medical materials and determine the final list of medical materials to be allocated; The warehouse screening module is used to calculate the warehouses that meet the storage conditions based on the medical supply demand information in the medical supply list and the warehouse's reserve quantity, and to filter out a list of recommended warehouses based on the shortest distance between the incident location and the warehouse; The path planning module is used to obtain real-time GPS information of emergency vehicles and drones, select the emergency vehicles and drones closest to the recommended warehouse, plan the optimal path for the drone from the warehouse to the incident location, and the optimal path for the emergency vehicle from the warehouse to the incident location, and calibrate the emergency vehicle's path based on real-time road congestion data; The message center module is used to send material allocation information and route planning to warehouse managers, emergency vehicle drivers, and drone driving systems.

9. An electronic device, characterized in that: The electronic device includes: at least one processor; and a memory communicatively connected to the at least one processor; the memory stores computer program instructions that can be executed by the at least one processor, and the computer program instructions are executed by the at least one processor to enable the at least one processor to execute the method for medical emergency material management and allocation based on a large model as described in any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium, characterized in that The non-transitory computer-readable storage medium stores computer instructions, and the computer instructions enable the computer to execute the method for managing and allocating medical emergency supplies based on a large model as described in any one of claims 1 to 7.

Images