Medical consumable management method and device based on ant colony algorithm

Abstract

The invention provides a medical consumable management method and device based on an ant colony algorithm, and relates to the technical field of ant colony algorithms, and the method comprises the steps: constructing a medical consumable turnover frequency weight model, and generating a consumable turnover priority function; generating a goods allocation weight mapping matrix through an ant colony algorithm in combination with a priority function, wherein the mapping matrix is established according to the medical consumables and the access areas; simulating ant colony individuals to perform multi-path search in a hospital channel topological structure, and dynamically updating path pheromones to form a shortest path set; a storage space is constructed into a heterogeneous graph structure with storage locations as graph nodes and channels as graph edges, and an embedding result is fed back to an ant colony optimization algorithm as a goods allocation initial constraint through embedding attribute vectors of each node by a graph neural network; goods allocation initial constraint and a trafficability score and a blocking factor generated by camera identification data are introduced, and a consumable transportation path under an emergent task is dynamically adjusted and realized. According to the invention, full-life-cycle management of the medical consumables can be realized.

Description

Technical Field

[0001] The present application relates to the technical field of ant colony algorithms, and in particular to a medical consumables management method and device based on ant colony algorithms. Background Art

[0002] High-value consumables management in hospitals refers to a comprehensive, refined control system implemented for medical consumables that are expensive, clinically dependent, have uncertain usage frequencies, and are subject to strict regulatory requirements. This system covers the entire lifecycle, from procurement and warehousing, storage location, clinical requisition, preoperative preparation, intraoperative use, postoperative settlement, and recycling review. Its core goal is to achieve optimal allocation of consumable resources and minimize risk. Through standardized usage pathways, digital traceability mechanisms, and dynamic inventory control measures, high-value consumables are ensured to meet medical safety and treatment efficiency requirements while also being economical, controllable, and compliant. Therefore, a method is needed to achieve full lifecycle management of medical consumables. Summary of the Invention

[0003] The present application provides a medical consumables management method and device based on ant colony algorithm, which can realize the full life cycle management of medical consumables.

[0004] In a first aspect of the present application, a medical consumables management method based on an ant colony algorithm is provided, the method comprising: Obtain medical consumables data and build a medical consumables turnover frequency weight model and generate a consumables turnover priority function; Generate a cargo space allocation weight mapping matrix by combining the consumables turnover priority function with an ant colony optimization algorithm, wherein the mapping matrix is ​​established based on medical consumables and storage and access areas; During the transportation of medical supplies, ant colony individuals are simulated to perform multi-path search in the hospital channel topology, and path pheromones are dynamically updated to form a shortest path set; In the cargo location graph modeling phase, the warehouse space is constructed as a heterogeneous graph structure with storage locations as graph nodes and channels as graph edges. The attribute vector of each node is embedded through a graph neural network, and the embedding result is fed back to the ant colony optimization algorithm as the initial constraint for cargo location allocation. The initial constraints of cargo space allocation and the accessibility score and blocking factor generated by camera recognition data are introduced into the dynamic update of the path to dynamically adjust the consumables transportation path under sudden tasks.

[0005] On the basis of the above technical solution, preferably, the method of generating a cargo location allocation weight mapping matrix by combining the consumables turnover priority function with an ant colony optimization algorithm specifically includes: Constructing the consumables turnover priority function based on the consumables turnover frequency, task urgency level, and spatial deployment expectation; Using the consumables turnover priority function as a basis for setting the pheromone concentration of the ant colony optimization algorithm; defining a matching function between the medical consumables and the access area as a heuristic function of the ant colony optimization algorithm; By simulating the path selection behavior of individual ant colonies on the storage location topology map, the selection probability of allocating the medical consumables to different storage locations is calculated according to the pheromone concentration and the matching function; The cargo location allocation weight mapping matrix is ​​generated based on the selection probability.

[0006] Based on the above technical solution, preferably, during the transportation of medical consumables, simulating ant colony individuals to perform multi-path search in the hospital channel topology structure and dynamically updating path pheromones to form a shortest path set specifically includes: The hospital channel topology is modeled as a path graph topology with medical supplies storage location nodes, operating area nodes, exit nodes, and transfer nodes as nodes, and access channels as connecting edges; Setting a travel cost value for the connection edge, wherein the travel cost value is composed of physical distance, channel passenger flow, travel priority and historical accessibility; Initializing multiple ant colony individuals in the path graph topology structure, starting from the medical consumables storage location node and ending at the target surgical area node, and calculating the path selection probability based on the heuristic function constructed based on the path pheromone concentration and the path accessibility score; After each round of path search is completed, the path contribution value is calculated based on the path travel time and the task time window, and is used to update the path pheromone concentration. At the same time, the path blocking factor is generated based on the crowd density recognized by the camera and the RFID node traffic, and the pheromone concentration is dynamically suppressed, thereby forming the shortest path set for the medical consumables transportation task in multiple rounds of iterations.

[0007] Based on the above technical solutions, preferably, in the cargo location graph modeling stage, the storage space is constructed as a heterogeneous graph structure with storage locations as graph nodes and channels as graph edges. The attribute vector of each node is embedded through a graph neural network, and the embedding result is fed back to the ant colony optimization algorithm as the initial constraint for cargo location allocation, specifically including: Establishing the heterogeneous graph structure to represent the medical consumables storage space, wherein the graph nodes correspond to medical consumables storage locations, the graph edges correspond to access channels between the locations, and each graph edge is associated with a travel path distance, a travel efficiency score, and an environmental interference intensity; Generate an attribute vector for each graph node including RFID reading frequency, historical access delay, current congestion status, remaining capacity of the storage location, and environmental safety index; The attribute vector is input into the graph neural network model for multiple rounds of graph embedding calculations. A node embedding vector is generated by aggregating the attribute vectors of each node and its neighboring nodes. The node embedding vector is fed back to the ant colony optimization algorithm as a cargo location state representation vector. During the ant colony path search process, the similarity calculation result between the node embedding vector and the medical consumables category attribute vector is combined as a component of the ant colony heuristic function to construct a cargo location allocation weight mapping matrix and guide the precise cargo location allocation of medical consumables.

[0008] Based on the above technical solution, preferably, the initial constraints of the cargo space allocation and the passability score and blocking factor generated by the camera recognition data are introduced into the dynamic update of the path to dynamically adjust the consumables transportation path under the emergency task, specifically including: Extracting the embedded attribute vector and injecting it into the path search starting point of the ant colony optimization algorithm as the initial constraint for cargo location allocation; Cameras deployed in hospital corridors capture image data of human and logistics flows. This data is then fed into a convolutional neural network model to identify traffic status features. The accessibility score for each edge in the path graph topology is calculated based on personnel density, handling equipment flow, obstacle interference, and channel accessibility. Calculating a path blocking factor by combining the passability score with the RFID node reading frequency. The path blocking factor includes the current channel density prediction value, historical blocking frequency, abnormal event probability, and path alternative score. The passability score and the path blocking factor are embedded together in the path selection function of the ant colony optimization algorithm to dynamically correct the path pheromone concentration and the path selection probability.

[0009] Based on the above technical solution, preferably, the initial constraints on cargo space allocation and the passability score and blocking factor generated by camera recognition data are introduced into the dynamic path update to dynamically adjust the consumables transportation path under the emergency task, specifically including: Determining the emergency consumables corresponding to the emergency task according to the triggered emergency task identification signal; Performing exponential weighted updating on the weight function of the emergency consumables corresponding to the category of consumables in the consumables turnover priority function; Based on the graph neural network, the heterogeneous graph structure is locally updated, and the state embedding of the storage location node corresponding to the emergency task and its neighboring nodes is quickly recalculated, thereby embedding the emergency storage location state vector into the storage location node corresponding to the emergency task and its neighboring nodes; The result of re-embedding the emergency storage location state vector is fed back to the ant colony algorithm as a weight compensation factor to improve the path weight of the outbound path of the emergency consumables in the ant colony path search.

[0010] On the basis of the above technical solution, preferably, the method of generating a cargo location allocation weight mapping matrix by combining the consumables turnover priority function with an ant colony optimization algorithm specifically includes: According to the triggered emergency task identification signal, the current cargo space allocation process is interrupted, and the preemptive scheduling mechanism is started. The consumables turnover priority function is dynamically weighted and modified based on the emergency task call status. The modified consumables turnover priority function includes the historical call frequency, task urgency level, surgical area proximity factor and emergency call status factor; Adjusting the pheromone concentration in the ant colony optimization algorithm according to the modified consumables turnover priority function, and introducing the emergency channel priority function in the heuristic function to enhance the path selection probability of the emergency target path; The path weight value in the emergency state is calculated by the pheromone concentration and the heuristic function, and the cargo location allocation weight mapping matrix in the emergency state is updated and generated to achieve preferential allocation of medical consumables to storage locations with short paths, stable environments and close to the emergency operating area.

[0011] In a second aspect of the present application, a medical consumables management device based on an ant colony algorithm is provided, wherein the device is used to execute any one of the above-mentioned medical consumables management methods based on an ant colony algorithm. The device includes an acquisition module, a processing module, and an output module, wherein: The acquisition module is used to acquire medical consumables data, build a medical consumables turnover frequency weight model, and generate a consumables turnover priority function; The processing module is used to generate a cargo space allocation weight mapping matrix by combining the consumables turnover priority function with an ant colony optimization algorithm, wherein the mapping matrix is ​​established based on medical consumables and storage and access areas; The processing module is used to simulate ant colony individuals to perform multi-path search in the hospital channel topology during the transportation of medical consumables, and dynamically update path pheromones to form a shortest path set; The processing module is used to construct the storage space into a heterogeneous graph structure with storage locations as graph nodes and channels as graph edges during the cargo location graph modeling phase, embed the attribute vector of each node through a graph neural network, and feed the embedding result into the ant colony optimization algorithm as the initial constraint for cargo location allocation; The output module is used to introduce the initial constraints of the cargo space allocation and the passability score and blocking factor generated by the camera recognition data into the dynamic update of the path, and dynamically adjust the consumables transportation path under the sudden task.

[0012] On the basis of the above technical solution, preferably, the acquisition module is used to construct the consumable turnover priority function based on the consumable turnover frequency, task urgency level and spatial deployment expectation; The processing module is configured to use the consumables turnover priority function as a basis for setting the pheromone concentration of the ant colony optimization algorithm; The processing module is configured to define a matching function between the medical consumables and the access area as a heuristic function of the ant colony optimization algorithm; The processing module is configured to calculate the selection probability of allocating the medical consumables to different storage locations based on the pheromone concentration and the matching function by simulating the path selection behavior of individual ant colonies on the storage location topology map; The processing module is used to generate the cargo location allocation weight mapping matrix based on the selection probability.

[0013] On the basis of the above technical solution, preferably, the processing module is used to model the hospital channel topology structure as a path graph topology structure with medical consumables storage location nodes, surgical area nodes, exit nodes and transfer nodes as nodes, and passages as connecting edges; The processing module is configured to set a travel cost value for the connection edge, wherein the travel cost value is composed of physical distance, channel passenger flow, travel priority, and historical accessibility; The processing module is configured to initialize multiple ant colony individuals in the path graph topology structure, perform path search starting from the medical consumables storage location node and ending at the target surgical area node, and calculate the path selection probability based on a heuristic function constructed based on the path pheromone concentration and the path accessibility score; The processing module is used to calculate the path contribution value based on the path travel time and the task time window after each round of path search is completed, and to update the path pheromone concentration. At the same time, it generates a path blocking factor based on the crowd density recognized by the camera and the RFID node traffic, and dynamically suppresses the pheromone concentration, thereby forming the shortest path set for the medical consumables transportation task in multiple rounds of iterations.

[0014] Based on the above technical solution, preferably, the processing module is used to establish the heterogeneous graph structure to represent the medical consumables storage space, wherein the graph nodes correspond to medical consumables storage locations, the graph edges correspond to access channels between the storage locations, and each graph edge is associated with the passage path distance, the passage efficiency score and the environmental interference intensity; The processing module is used to generate an attribute vector for each graph node, including RFID reading frequency, historical access delay, current congestion status, remaining capacity of the storage location, and environmental safety index; The processing module is used to input the attribute vector into the graph neural network model for multiple rounds of graph embedding calculations, generate a node embedding vector by aggregating the attribute vectors of each node and its neighboring nodes, and feed the node embedding vector back to the ant colony optimization algorithm as a cargo location state representation vector. During the ant colony path search process, the similarity calculation result between the node embedding vector and the medical consumables category attribute vector is combined as a component of the ant colony heuristic function to construct a cargo location allocation weight mapping matrix and guide the precise cargo location allocation of medical consumables.

[0015] On the basis of the above technical solution, preferably, the acquisition module is used to extract the embedded attribute vector as the initial constraint for cargo location allocation and inject it into the path search starting point of the ant colony optimization algorithm; The processing module is used to obtain human and logistics image data through cameras deployed in the hospital corridor area, input the image data into the convolutional neural network model to identify traffic status characteristics, and calculate the trafficability score of each edge in the path graph topology structure based on personnel density, handling equipment flow, obstacle interference, and channel accessibility; The processing module is configured to calculate a path blocking factor by combining the passability score and the RFID node reading frequency, wherein the path blocking factor includes a current channel density prediction value, a historical blocking frequency, an abnormal event probability, and a path alternative score; The processing module is used to embed the passability score and the path blocking factor into the path selection function of the ant colony optimization algorithm to dynamically correct the path pheromone concentration and the path selection probability.

[0016] On the basis of the above technical solution, preferably, the processing module is used to determine the emergency consumables corresponding to the emergency task according to the triggered emergency task identification signal; The processing module is configured to perform exponential weighted update on the weight function of the emergency consumables corresponding to the category of consumables in the consumables turnover priority function; The processing module is used to perform a local graph update on the heterogeneous graph structure based on the graph neural network, and quickly recalculate the state embedding of the storage location node corresponding to the emergency task and its neighboring nodes, thereby embedding the emergency storage location state vector into the storage location node corresponding to the emergency task and its neighboring nodes; The processing module is used to feed back the result of re-embedding the emergency storage location state vector to the ant colony algorithm as a weight compensation factor, thereby increasing the path weight of the outbound path of the emergency consumables in the ant colony path search.

[0017] On the basis of the above technical solution, preferably, the processing module is used to interrupt the current cargo space allocation process according to the triggered emergency task identification signal, start the preemptive scheduling mechanism, and dynamically weightedly correct the consumables turnover priority function based on the emergency task call status. The corrected consumables turnover priority function includes the historical call frequency, task urgency level, surgical area proximity factor and emergency call status factor; Adjusting the pheromone concentration in the ant colony optimization algorithm according to the modified consumables turnover priority function, and introducing the emergency channel priority function in the heuristic function to enhance the path selection probability of the emergency target path; The processing module is used to calculate the path weight value under the emergency state through the pheromone concentration and the heuristic function, and update and generate the cargo location allocation weight mapping matrix under the emergency state, so as to realize the priority allocation of medical consumables to storage locations with short paths, stable environments and close to the emergency operating area.

[0018] In the third aspect of the present application, an electronic device is provided, including a processor, a memory, a user interface and a network interface, the memory is used to store instructions, the user interface and the network interface are both used to communicate with other devices, and the processor is used to execute the instructions stored in the memory so that the electronic device performs any of the methods described above.

[0019] In a fourth aspect of the present application, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores instructions. When the instructions are executed, any one of the methods described above is executed.

[0020] In summary, one or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages: 1. This application constructs a medical consumables turnover frequency weight model and generates a consumables turnover priority function, combines the ant colony optimization algorithm to realize intelligent allocation of cargo locations, and then uses a graph neural network to model the warehouse heterogeneous graph structure to obtain high-dimensional warehouse status embedding, and integrates path pheromones, camera recognition accessibility scores and blocking factors in the transportation path to dynamically update the transportation path. At the same time, a preemptive scheduling mechanism is introduced to deal with emergency tasks, and the entire process from warehousing, storage, allocation, outbound to emergency response is realized. A dynamic closed-loop system integrating perception, decision-making and execution is constructed, thereby realizing refined management and intelligent scheduling of the entire life cycle of medical consumables.

[0021] 2. By introducing the consumables turnover frequency, task urgency level and spatial deployment expectations, a consumables turnover priority function is constructed, and it is used as the dual control basis of pheromone concentration and heuristic function. Combined with the ant colony optimization algorithm, a multi-factor comprehensive cargo space allocation based on frequency, timeliness and spatial adaptation is realized, thereby improving the priority response capability and accuracy of high-frequency and high-urgency medical consumables in the warehouse layout.

[0022] 3. By constructing the topological structure of the hospital channel path map, setting the access cost value and dynamically updating the pheromone concentration during the path search, and combining the real-time path accessibility score and the blocking factor to form a reinforcement-inhibition joint control mechanism, the medical consumables transportation path can be fully perceived and adaptively optimized in terms of efficiency, congestion and risk during the task execution, ensuring the convergence stability and access efficiency of the transportation path during peak hours.

[0023] 4. Use graph neural networks to model the graph structure of the warehouse space and generate high-dimensional storage location embedding vectors. Combine the multidimensional characteristics of the storage location status with the similarity of consumable attributes as the path guidance factor of the ant colony optimization algorithm, and achieve dynamic optimization of the cargo location allocation accuracy from the perspectives of spatial structure, capacity status and environmental adaptation, thereby enhancing the algorithm's robustness and adaptability to environmental changes and task dynamics.

[0024] 5. Through the fusion of camera image recognition and RFID traffic, the accessibility score and blocking factor are generated and embedded into the path selection function of the ant colony optimization algorithm, forming a path feasibility modeling and pheromone concentration dynamic correction mechanism based on the real-time traffic status, thereby significantly improving the immediacy, accessibility and task congestion avoidance capabilities of path selection, and meeting the high-frequency task scheduling needs in complex campus scenarios.

[0025] 6. After an emergency task is triggered, emergency consumables are dynamically identified and their turnover priority function is exponentially weighted. At the same time, local graph updates and emergency status embedding feedback are performed on emergency-related nodes through graph neural networks to achieve rapid focusing on the emergency task resource path and pheromone weight adjustment, while ensuring the response efficiency of the graph model and strengthening the path search's ability to instantly dispatch sudden demands.

[0026] 7. Based on the preemptive scheduling process triggered by the emergency task identification signal, through priority function correction, pheromone concentration enhancement and emergency path heuristic function construction, high-intensity path weight integration of emergency task target storage locations is achieved in the cargo location allocation stage, forming a real-time reallocation mechanism with task triggering and response, layout and adjustment, thereby improving the intelligent scheduling level and system response speed of warehousing distribution under emergency tasks. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 This is a flow chart of a medical consumables management method based on ant colony algorithm disclosed in an embodiment of the present application; Figure 2 This is a module diagram of a medical consumables management device based on ant colony algorithm disclosed in an embodiment of the present application; Figure 3 This is a structural diagram of an electronic device disclosed in an embodiment of the present application.

[0028] Explanation of the reference numerals: 201, acquisition module; 202, processing module; 203, output module; 301, processor; 302, communication bus; 303, user interface; 304, network interface; 305, memory. DETAILED DESCRIPTION

[0029] In order to enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below in conjunction with the drawings in the embodiments of this specification. Obviously, the described embodiments are only part of the embodiments of this application, not all of the embodiments.

[0030] In the description of the embodiments of this application, words such as "for example" or "for instance" are used to indicate examples, illustrations, or explanations. Any embodiment or design described as "for example" or "for instance" in the embodiments of this application should not be construed as being preferred or advantageous over other embodiments or designs. Rather, the use of words such as "for example" or "for instance" is intended to present the relevant concepts in a concrete manner.

[0031] In the description of the embodiments of the present application, the term "multiple" means two or more. For example, multiple systems refer to two or more systems, and multiple screen terminals refer to two or more screen terminals. In addition, the terms "first" and "second" are used for descriptive purposes only and are not to be understood as indicating or implying relative importance or implicitly indicating the indicated technical features. Thus, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. The terms "including", "comprising", "having" and their variations all mean "including but not limited to", unless otherwise specifically emphasized.

[0032] A hospital's logistics robot is an intelligent device deployed inside medical facilities to automatically complete tasks such as medical supplies distribution, sample transfer, drug supply, and high-value consumables handling. It has autonomous navigation, multi-point scheduling, path obstacle avoidance, and task coordination capabilities. It usually integrates technical modules such as SLAM positioning, visual recognition, semantic understanding, and IoT communication. It can operate efficiently in the complex spatial layout of the hospital, significantly improve logistics efficiency, reduce the burden on medical staff, and demonstrate high stability and intelligence in scenarios such as hospital infection prevention and control, emergency task response, and nighttime operation and maintenance.

[0033] This embodiment discloses a medical consumables management method based on ant colony algorithm. Figure 1 , including the following steps S110-S150: S110, obtaining medical consumables data, building a medical consumables turnover frequency weight model, and generating a consumables turnover priority function.

[0034] The embodiments of the present application disclose an ant colony algorithm-based medical consumables management method applied to a server. The server includes, but is not limited to, electronic devices such as mobile phones, tablet computers, wearable devices, and personal computers (PCs). It can also be a backend server running the ant colony algorithm-based medical consumables management method. The server can be implemented as a standalone server or a server cluster consisting of multiple servers.

[0035] First, recent historical surgical data, department call records, and consumables in-and-out logs are extracted from hospital information. The call frequency of each type of medical consumable within a unit time window is counted as the basic turnover frequency value. The task urgency level parameter is introduced, and the task driving factor is weighted according to the probability of use of the consumable in emergency surgery, intensive care, and interventional procedures. At the same time, combined with the storage space structure and surgical area distribution information, the path length and spatial proximity between each type of medical consumable and its most commonly used department are calculated to generate a spatial deployment factor. Subsequently, the above three factors are set as frequency parameters, urgency parameters, and spatial factors, respectively. According to the empirical weight parameter setting, they are linearly combined to generate the turnover priority function of the medical consumable, forming the following structure: ,in Indicates the The turnover priority value of medical consumables. is the frequency parameter, is the urgency parameter, is the spatial factor, 、 、 To adjust the weight; finally, the turnover priority function is used as the initial decision basis in the subsequent cargo allocation and path planning modules to achieve priority response and optimal space configuration for high-frequency and high-urgency medical consumables.

[0036] S120 , generating a storage location allocation weight mapping matrix by combining an ant colony optimization algorithm with a consumables turnover priority function.

[0037] In one possible implementation, an ant colony optimization algorithm is combined with a consumables turnover priority function to generate a cargo location allocation weight mapping matrix, specifically including: constructing a consumables turnover priority function based on the consumables turnover frequency, task urgency level, and spatial deployment expectations; using the consumables turnover priority function as the basis for setting the pheromone concentration of the ant colony optimization algorithm; defining a matching function between medical consumables and access areas as an inspiration function of the ant colony optimization algorithm; calculating the selection probability of medical consumables being allocated to different storage locations based on the pheromone concentration and the matching function by simulating the path selection behavior of ant colony individuals on the storage location topology map; and generating a cargo location allocation weight mapping matrix based on the selection probability.

[0038] Specifically, we first extract the frequency statistics of each type of medical consumables in and out of the warehouse within a specific time window from the medical information, and record it as the turnover frequency parameter , reflecting the average number of calls per unit time; secondly, by analyzing historical clinical task data, the usage ratio of the medical consumables in emergency surgery, critical intervention or high-risk operations is mapped to the task urgency level parameter , quantify its time sensitivity and emergency dependence; then, based on the physical layout of the hospital, extract the path distance between the commonly used departments of medical consumables and each storage area, and combine the environmental stability and temperature and humidity adaptability of the storage area to construct the expected parameters of space deployment. The smaller the parameter value, the closer the storage location is to the operation area and the higher the physical adaptability. The final turnover priority function is defined as:

[0039] in, For the The turnover priority function value of medical consumables, 、 、 In order to adjust the weight coefficients according to the task scenario, all parameters are normalized to maintain the function convergence.

[0040] The consumables turnover priority function is used as the basis for setting the pheromone concentration of the ant colony optimization algorithm. In the initialization phase of the ant colony optimization algorithm, according to the turnover priority function Initialize the pheromone concentration matrix with the normalized value of ,in Indicates the Medical consumables and The initial pheromone concentration between storage locations. High-priority medical consumables will be given a higher initial pheromone concentration, making ant individuals more inclined to search around such medical consumables in subsequent path selection, thus strengthening their clustering effect in storage location allocation.

[0041] A matching function between the medical consumables and the access area is defined as a heuristic function of the ant colony optimization algorithm. Matching function To measure the Medical consumables and The degree of adaptation between the storage locations is calculated as follows:

[0042] in Indicates the shortest travel distance from the storage location to the corresponding operation area, represents the environmental risk factor of the storage location (including a weighted combination of parameters such as temperature fluctuation and electromagnetic interference), represents the channel congestion of the storage location during peak hours; all variables are normalized, with smaller values ​​indicating higher matching. The matching function is used as a heuristic input into the path transition probability model of the ant colony optimization algorithm, guiding ants to prioritize exploring highly suitable areas.

[0043] By simulating the path selection behavior of individual ants on the storage location topology graph, the selection probability of the medical consumables being allocated to different storage locations is calculated based on the pheromone concentration and the matching function. In each round of search iteration, the ant individuals are simulated as intelligent agents for the medical consumables allocation behavior, starting from the initial node and moving within the graph structure based on the following path selection probabilities:

[0044] in For the Medical consumables in the In the iteration, select The probability of a storage location, 、 is the parameter for adjusting the pheromone concentration and the weight of the matching function, is the set of currently selectable storage locations; the path selection behavior simulates the independent movement process of multiple ant colony individuals to achieve global search and high-adaptation distribution of the entire graph space state.

[0045] The fifth step is to generate the cargo space allocation weight mapping matrix based on the selection probability. Count the selection probabilities of all ant colony individuals for each pair of medical consumables and storage space combinations during the search process. , and perform weighted accumulation in a certain number of search iterations to finally generate a cargo location allocation weight mapping matrix , where each element Indicates the Medical consumables are allocated first to This matrix serves as a reference for the final allocation results and is input into the scheduling engine to guide the execution of cargo location deployment, enabling the rapid deployment of high-priority medical consumables to highly adaptable and accessible areas, thereby improving overall warehouse allocation efficiency and clinical responsiveness.

[0046] In one possible implementation, an ant colony optimization algorithm is combined with a consumables turnover priority function to generate a cargo space allocation weight mapping matrix, specifically including: according to the triggered emergency task identification signal, interrupting the current cargo space allocation process, starting a preemptive scheduling mechanism, and dynamically weighting and correcting the consumables turnover priority function based on the emergency task call status, wherein the corrected consumables turnover priority function includes historical call frequency, task urgency level, surgical area proximity factor, and emergency call status factor; adjusting the pheromone concentration in the ant colony optimization algorithm according to the corrected consumables turnover priority function, and introducing an emergency channel priority function into the heuristic function to enhance the path selection probability of the emergency target path; calculating the path weight value under the emergency state through the pheromone concentration and the heuristic function, and updating and generating the cargo space allocation weight mapping matrix under the emergency state, so as to realize the priority allocation of medical consumables to storage locations with short paths, stable environments, and close to the emergency surgical area.

[0047] Specifically, according to the triggered emergency task identification signal, the current cargo space allocation process is interrupted, the preemptive scheduling mechanism is started, and the consumables turnover priority function is dynamically weighted and modified based on the emergency task call status. After receiving the emergency task identification signal from the surgical scheduling, medical order or emergency resource scheduling platform, the preemptive task management module in the central scheduling engine is immediately activated to interrupt the currently executing cargo space allocation process and switch the task context to the emergency scheduling mode. For the types of medical consumables involved in this emergency, the historical call frequency parameters are extracted in real time. , Mission urgency level parameters , surgical area proximity factor and emergency call status factor ,in is a Boolean variable that takes the value 1 when the medical consumable is used in an emergency order, otherwise it takes the value 0. The consumable turnover priority function under the emergency state is then generated according to the following weighted correction function:

[0048] Where α, β, γ, and δ are preset weight coefficients that adjust the relative contribution of each parameter to the final priority. This correction function ensures that the emergency task status has absolute dominance in the weight expression, enabling instant reconstruction of scheduling priorities.

[0049] According to the modified consumables turnover priority function, the pheromone concentration in the ant colony optimization algorithm is adjusted, and the emergency channel priority function is introduced into the heuristic function to enhance the path selection probability of the emergency target path. value, reset the pheromone concentration matrix , the pheromone concentration of emergency-related medical consumables is increased, and the channel path set pointing from the current storage area to the emergency surgery area is re-marked in the path graph topology structure; in order to guide the ant colony individuals to give priority to this channel path, the emergency channel priority function is defined , which is used to enhance the selection probability of the path segment in the heuristic function. The specific heuristic function is updated as follows:

[0050] in represents the path distance, Indicates the environmental risk factor of the target location of the path, Indicates its current congestion level, is the emergency path indicator function, only when the edge When it belongs to the emergency channel, the value is 1. is the priority enhancement coefficient. Through this mechanism, the priority ratio of the emergency path in the search process can be further improved on the basis of pheromone enhancement.

[0051] The path weight value in the emergency state is calculated by the pheromone concentration and the heuristic function, and the cargo location allocation weight mapping matrix in the emergency state is updated to achieve the priority allocation of medical consumables to the storage location with the shortest path, stable environment and close to the emergency operation area. and emergency heuristic function Reconstruct the path selection probability function:

[0052] In multiple rounds of simulated search, ant colony individuals complete the dynamic planning of the path from medical supplies to the target emergency area based on the above path selection probability, and finally integrate all search results to generate a weight mapping matrix for the cargo space allocation under the emergency state. , each of which Indicates the Class I emergency medical supplies are allocated to The scheduling engine uses the priority weight of each storage location to guide the handling execution to complete the layout of consumables, ensuring that they are preferentially deployed in highly adaptable storage locations with stable environments, unobstructed channels, and close to emergency operating areas within the time window, achieving the minimum delay configuration for responding to emergency tasks.

[0053] S130, during the transportation of medical supplies, simulate ant colony individuals to perform multi-path search in the hospital channel topology structure, and dynamically update path pheromones to form a shortest path set.

[0054] In one possible implementation, during the transportation of medical consumables, ant colony individuals are simulated to perform multi-path search in a hospital channel topology structure, and path pheromones are dynamically updated to form a shortest path set. Specifically, the method includes: modeling the hospital channel topology structure as a path graph topology structure with medical consumables storage location nodes, surgical area nodes, exit nodes, and transfer nodes as nodes, and access channels as connecting edges; setting access cost values ​​for the connecting edges, where the access cost values ​​are composed of physical distance, channel traffic, access priority, and historical accessibility; initializing multiple ant colony individuals in the path graph topology structure, starting from the medical consumables storage location node and ending at the target surgical area node to perform path search, and calculating the path selection probability based on a heuristic function constructed from the path pheromone concentration and the path accessibility score; after each round of path search is completed, the path contribution value is calculated based on the path travel time and the task time window, and is used to update the path pheromone concentration. At the same time, a path blocking factor is generated based on the crowd density recognized by the camera and the RFID node traffic, and the pheromone concentration is dynamically suppressed, thereby forming the shortest path set for the medical consumables transportation task in multiple rounds of iterations.

[0055] Specifically, in the hospital channel topology modeling stage, based on the actual building layout, functional area configuration and logistics channel direction of the hospital, the medical consumables storage location nodes, surgical area nodes, exit nodes and transfer nodes are uniformly abstracted as nodes in the graph structure. All reachable channels between nodes are defined in the form of connecting edges, and a path graph topology structure covering the entire path range of medical consumables allocation is constructed to ensure that the graph structure meets the accessibility and connectivity of medical consumables from any storage location to any surgical area node.

[0056] In the pass cost setting phase, a pass cost value is assigned to each connecting edge in the path graph topology. The pass cost value is calculated using the following formula:

[0057] in is the physical path distance between nodes, It is the real-time density of people and logistics based on camera recognition and RFID node traffic statistics. Indicates the priority level of the path. Score the historical accessibility stability of the path, to is the weight adjustment coefficient; the lower the cost value, the more priority the path has in transportation execution.

[0058] In the ant colony individual path search execution phase, multiple ant colony individuals are generated for each medical consumables transportation task instance, starting from the designated medical consumables storage location node, and the target node is the designated surgical area node, and multi-path search is performed in the path graph topology in a bionic way. In each path selection step, the ant colony individual selects the path based on the current path pheromone concentration. and path heuristic function Calculate the path selection probability. The selection probability formula is:

[0059] in and Represent the adjustment sensitivity to pheromone concentration and heuristic function, respectively.

[0060] During the path pheromone update phase, the travel time of each path that is fully searched and meets the time window constraint is counted. , calculate the path contribution value based on the path efficiency and use it to update the path pheromone concentration:

[0061] in is the pheromone volatility coefficient; at the same time, the blocking factor is generated according to the current camera crowd recognition density of the path node and the RFID node traffic , suppress the path attractiveness of high-risk paths and form a dynamic adjustment mechanism:

[0062] Finally, a stable set of shortest paths is converged through multiple rounds of path search iterations, achieving optimal dynamic planning and highly robust execution of the medical supplies transportation path under high-frequency task disturbances.

[0063] S150, in the cargo location graph modeling stage, the storage space is constructed as a heterogeneous graph structure with storage locations as graph nodes and channels as graph edges. The attribute vector of each node is embedded through the graph neural network, and the embedding result is fed back to the ant colony optimization algorithm as the initial constraint for cargo location allocation.

[0064] In one possible implementation, during the cargo location graph modeling stage, the storage space is constructed as a heterogeneous graph structure with storage locations as graph nodes and channels as graph edges. The attribute vectors of each node are embedded through a graph neural network, and the embedding results are fed back to the ant colony optimization algorithm as the initial constraints for cargo location allocation. Specifically, the method includes: establishing a heterogeneous graph structure to represent the medical consumables storage space, wherein the graph nodes correspond to the medical consumables storage locations, the graph edges correspond to the access channels between the storage locations, and each graph edge is associated with the travel path distance, the travel efficiency score and the environmental interference intensity; generating an attribute vector for each graph node including the RFID reading frequency, the historical access delay, the current congestion status, the remaining capacity of the storage location and the environmental safety index; inputting the attribute vector into the graph neural network model for multiple rounds of graph embedding calculation, generating a node embedding vector by aggregating the attribute vectors of each node and its neighboring nodes, and feeding the node embedding vector as the cargo location status representation vector back to the ant colony optimization algorithm; in the ant colony path search process, combining the similarity calculation results between the node embedding vector and the medical consumables category attribute vector as a component of the ant colony heuristic function, thereby constructing a cargo location allocation weight mapping matrix and guiding the precise cargo location allocation of medical consumables.

[0065] Specifically, in the medical consumables storage space modeling stage, all medical consumables storage locations are abstracted as graph nodes in the graph structure according to the physical structure of the warehouse, and all the actual access channels between the graph nodes are abstracted as graph edges in the graph structure to establish a complete heterogeneous graph structure. , where the graph node set Represents all medical consumables storage locations, graph edge collection Represents the accessible paths connecting storage locations. Each graph edge is assigned three attributes: path distance, which describes the absolute length of the physical space; a path efficiency score, which quantifies the historical accessibility and operation time of handling tasks along the path; and environmental interference intensity, which measures the temperature and humidity fluctuations, electromagnetic interference risk, or radiation safety level of the area where the path segment is located.

[0066] During the graph node attribute initialization phase, a high-dimensional attribute vector is generated for each graph node. The attribute vector includes five types of standardized characteristic indicators: RFID reading frequency represents the stability of tag recognition and the continuity of tag existence; historical access delay is the average access time of the storage location in the recent stage, which is used to measure its operational efficiency; the current congestion status is obtained by the joint recognition result of RFID and camera, which is used to reflect the pressure density of neighborhood tasks; the remaining capacity of the storage location represents the current proportion of space available for placing medical consumables, which is processed by volume standardization; the environmental safety index represents the temperature and humidity stability and the degree of distance from interference sources in the storage area, which is converted into numerical expression through hierarchical coding.

[0067] In the graph embedding calculation phase, the above attribute vector is input into the graph neural network model (GraphSAGE, GAT, or GIN can be used), and the attribute information of the graph node neighbor nodes is aggregated in each round of propagation, and the following embedding operation is performed:

[0068] in For the Layer graph node embedding vector, For nodes The set of adjacent nodes of represents a class of aggregation functions (such as average, maximum, attention weighted), is the learnable parameter matrix, is the activation function. After the layer graph embedding propagation, a stable node embedding representation vector is obtained , which is the state representation vector of the medical consumables storage location.

[0069] In the joint embedding feedback and path search phase, each graph node output by the graph neural network is embedded into a vector As the storage location status expression, it is input into the ant colony optimization algorithm. During the ant colony path search process, the embedding vector is combined with the medical consumables category attribute vector. Similarity calculation is usually performed in the form of cosine similarity:

[0070] The higher the similarity value, the higher the attribute matching degree between the storage location and the required medical consumables. The ant colony optimization algorithm uses the similarity value as part of the heuristic function to participate in the path selection probability calculation, and then dynamically constructs the storage location allocation weight mapping matrix. , guiding the priority layout of medical consumables to storage locations with stable status, efficient access, strong environmental adaptability and high task relevance, to achieve high-precision cargo location allocation under task-driven.

[0071] S150 introduces the initial constraints of cargo space allocation and the accessibility score and blocking factor generated by camera recognition data into the dynamic update of the path, and dynamically adjusts the consumables transportation path under sudden tasks.

[0072] In one possible implementation, the initial constraints of cargo space allocation and the accessibility score and blocking factor generated by camera recognition data are introduced into the dynamic update of the path to dynamically adjust the consumables transportation path under emergency tasks, specifically including: extracting the embedded attribute vector and injecting it into the path search starting point of the ant colony optimization algorithm as the initial constraints of cargo space allocation; obtaining human and logistics image data through cameras deployed in the hospital passage area, inputting the image data into the convolutional neural network model to identify the traffic status characteristics, and calculating the accessibility score of each edge in the path graph topology structure based on personnel density, handling equipment flow, obstacle interference and channel accessibility; combining the accessibility score with the RFID node reading frequency to calculate the path blocking factor, which includes the current channel density prediction value, historical blocking frequency, abnormal event probability and path alternative score; and embedding the accessibility score and path blocking factor into the path selection function of the ant colony optimization algorithm to dynamically correct the path pheromone concentration and path selection probability.

[0073] Specifically, in the dynamic path update process, the graph neural network embedded attribute vector of the medical consumables storage location node is first extracted. This embedded attribute vector reflects the high-dimensional state of the node in the current storage environment, including five dimensions: RFID reading frequency, historical access delay, current congestion status, remaining capacity of the location, and environmental safety index. This embedded attribute vector is directly injected into the ant colony optimization algorithm as the initial constraint for cargo location allocation to determine the starting node state of the path search, and serves as the benchmark data for dynamically evaluating the rationality of path passage in subsequent path selection, thereby enhancing the response sensitivity of path search to high-priority nodes.

[0074] Real-time human and logistics image data is captured by cameras deployed in the hospital's main logistics channels. The image data includes information such as static background, personnel movement, equipment operation, and the presence of obstacles. All image frames are input into the convolutional neural network model for real-time processing. The model uses feature extraction and target recognition to identify the current channel's personnel density, handling equipment flow, obstacle type, and degree of occlusion. Based on the channel's accessible width and the recognition results, the model calculates the path edge's accessibility score. The accessibility score is defined as:

[0075] in is the path physical distance, Indicates the real-time crowd density, Indicates the device traffic density, Indicates the obstacle intensity factor. A larger value indicates that the path is congested or blocked.

[0076] The path blocking factor is further calculated by integrating the accessibility score with the real-time reading frequency of the RFID node. The blocking factor is used to quantify the potential inaccessibility risk of the path. It is composed of four types of indicators: the current channel density prediction value is derived from the time series trend model fitted by historical images and sensor data, which is used to predict the channel saturation in the short term; the historical blocking frequency is the statistical probability of the path failing to pass, abnormal delays or detours in previous tasks; the abnormal event probability is the alarm score corresponding to the area where the path is located being marked as having an emergency (such as outage, elevator failure, closure); the path substitutability score represents the number and feasibility score of the paths that can be replaced in the task graph. Finally, the path blocking factor is calculated comprehensively:

[0077] Each item represents four types of blocking indicators, is the adjustment coefficient.

[0078] In the path search process, the path permeability score and path blocking factor are embedded in the path selection function of the ant colony optimization algorithm, and the pheromone concentration update and path selection probability are corrected. The path selection probability is updated as follows:

[0079] Through this mechanism, blocked paths can be avoided in real time, unobstructed and efficient paths can be searched preferentially, and a set of high-passage shortest paths that meet the needs of emergency task response can be dynamically formed. In subsequent path iterations, the shortest paths can be converged to a robust set of transport paths.

[0080] In one possible implementation, the initial constraints on cargo location allocation and the accessibility score and blocking factor generated by camera recognition data are introduced into the dynamic update of the path, and the consumables transportation path under emergency tasks is dynamically adjusted. Specifically, it also includes: determining the emergency consumables corresponding to the emergency task according to the triggered emergency task identification signal; exponentially weighting the weight function of the emergency consumables corresponding to the category of consumables in the consumables turnover priority function; performing local graph updates on the heterogeneous graph structure based on the graph neural network, and quickly recalculating the state embedding of the storage location nodes and their neighboring nodes corresponding to the emergency task, thereby embedding the storage location nodes and their neighboring nodes corresponding to the emergency task into the emergency storage location state vector; and feeding back the result of the re-embedded emergency storage location state vector to the ant colony algorithm as a weight compensation factor to improve the path weight of the outbound path of the emergency consumables in the ant colony path search.

[0081] Specifically, in the emergency task identification stage, by connecting to the hospital's surgical schedule, medical orders and clinical information, the emergency task identification signal from the emergency task module is received, and the emergency path optimization mechanism is immediately triggered to identify the list of emergency consumables involved in the emergency task, map it with the consumables information library through a unique identifier, confirm the category code of each emergency consumable, and construct the emergency consumables collection for subsequent path priority reconstruction and graph structure local state update.

[0082] During the dynamic update phase of the consumables turnover priority function, based on the established consumables turnover priority function:

[0083] Implement exponential weighted updates on the priority values ​​of the categories corresponding to emergency consumables, and introduce emergency correction coefficients , forming the turnover priority function after emergency update:

[0084] in represents the exponential weight growth factor, Indicates the difference between the current task time and the emergency task trigger time, Represents the emergency call status factor. This function ensures that the priority of emergency consumables rises rapidly during the mission's emergency period and occupies the pheromone distribution first.

[0085] In the local graph update stage of the heterogeneous graph structure, the graph neural network model is called to perform state update and embedding propagation on only the local subgraph consisting of the current storage location node of emergency consumables and its first-order adjacent nodes in the heterogeneous graph structure, avoiding the high computational delay caused by recalculation of the entire graph; the graph node attribute vector is re-extracted in the local subgraph, including RFID reading frequency, historical access delay, current congestion status, storage location remaining capacity and environmental safety index, and the node state vector is injected into it in combination with the emergency weight factor, and the graph embedding propagation operation is performed to form the emergency storage location state vector under the emergency state. .

[0086] In the path update phase of the ant colony optimization algorithm, the regenerated emergency storage location state vector is fed back to the ant colony optimization algorithm to compensate the heuristic function weight of the path corresponding to the emergency consumables and improve the attraction strength of the path in the path search. The ant colony optimization algorithm introduces this compensation factor into the path selection probability function ,in is the attribute vector of emergency consumables category, The emergency state vector is embedded in the target location, and the similarity calculation result is directly multiplied by the path heuristic function to make the path selection probability become:

[0087] This gives the emergency consumables outbound routes a higher priority during the search process, quickly focusing on a set of accessible routes with short paths, smooth passages, stable conditions, and close proximity to the operating area, thereby improving the real-time response capability and overall scheduling efficiency of emergency tasks.

[0088] This embodiment also discloses a medical consumables management device based on ant colony algorithm, referring to Figure 2 , comprising an acquisition module 201, a processing module 202 and an output module 203, the device is used to execute any of the above-mentioned medical consumables management methods based on ant colony algorithm, wherein: The acquisition module 201 is used to acquire medical consumables data, build a medical consumables turnover frequency weight model, and generate a consumables turnover priority function.

[0089] The processing module 202 is used to generate a cargo space allocation weight mapping matrix by combining an ant colony optimization algorithm with a consumables turnover priority function. The mapping matrix is ​​established based on medical consumables and storage and access areas.

[0090] The processing module 202 is used to simulate ant colony individuals to perform multi-path search in the hospital channel topology during the transportation of medical consumables, and dynamically update the path pheromone to form a shortest path set.

[0091] Processing module 202 is used to construct the storage space into a heterogeneous graph structure with storage locations as graph nodes and channels as graph edges during the cargo location map modeling stage. The attribute vector of each node is embedded through the graph neural network, and the embedding result is fed back to the ant colony optimization algorithm as the initial constraint for cargo location allocation.

[0092] The output module 203 is used to introduce the initial constraints of cargo space allocation and the passability score and blocking factor generated by camera recognition data into the dynamic update of the path, and dynamically adjust the consumables transportation path under the sudden task.

[0093] In a possible implementation, the acquisition module 201 is configured to construct a consumables turnover priority function based on consumables turnover frequency, task urgency level, and spatial deployment expectation.

[0094] The processing module 202 is configured to use the consumables turnover priority function as a basis for setting the pheromone concentration of the ant colony optimization algorithm.

[0095] The processing module 202 is used to define a matching function between medical consumables and access areas as a heuristic function of the ant colony optimization algorithm.

[0096] The processing module 202 is used to simulate the path selection behavior of ant colony individuals on the storage location topology map, and calculate the selection probability of allocating medical consumables to different storage locations based on the pheromone concentration and the matching function.

[0097] The processing module 202 is configured to generate a cargo location allocation weight mapping matrix based on the selection probability.

[0098] In one possible implementation, the processing module 202 is used to model the hospital channel topology structure as a path graph topology structure with medical consumables storage location nodes, surgical area nodes, exit nodes, and transfer nodes as nodes, and access channels as connecting edges.

[0099] The processing module 202 is used to set a travel cost value for the connection edge. The travel cost value is composed of physical distance, channel passenger flow, travel priority and historical accessibility.

[0100] Processing module 202 is used to initialize multiple ant colony individuals in the path graph topology structure, start from the medical supplies storage location node, and perform path search with the target surgical area node as the end point. The path selection probability is calculated based on the heuristic function constructed by the path pheromone concentration and the path accessibility score.

[0101] Processing module 202 is used to calculate the path contribution value based on the path travel time and the task time window after each round of path search is completed, and to update the path pheromone concentration. At the same time, it generates a path blocking factor based on the crowd density recognized by the camera and the RFID node traffic, and dynamically suppresses the pheromone concentration, thereby forming the shortest path set for the medical consumables transportation task in multiple rounds of iterations.

[0102] In one possible implementation, the processing module 202 is used to establish a heterogeneous graph structure to represent the medical consumables storage space, wherein the graph nodes correspond to the medical consumables storage locations, the graph edges correspond to the access channels between the locations, and each graph edge is associated with the travel path distance, the travel efficiency score and the environmental interference intensity.

[0103] The processing module 202 is used to generate an attribute vector for each graph node, including RFID reading frequency, historical access delay, current congestion status, remaining capacity of the storage location, and environmental safety index.

[0104] The processing module 202 is used to input the attribute vector into the graph neural network model for multiple rounds of graph embedding calculations, generate a node embedding vector by aggregating the attribute vectors of each node and its neighboring nodes, and feed the node embedding vector as a storage location status representation vector to the ant colony optimization algorithm. During the ant colony path search process, the similarity calculation result between the node embedding vector and the medical consumables category attribute vector is combined as a component of the ant colony heuristic function to construct a storage location allocation weight mapping matrix and guide the precise storage location allocation of medical consumables.

[0105] In a possible implementation, the acquisition module 201 is configured to extract an embedded attribute vector as an initial constraint for cargo location allocation and inject it into a path search starting point of an ant colony optimization algorithm.

[0106] Processing module 202 is used to obtain human and logistics image data through cameras deployed in the hospital passage area, input the image data into the convolutional neural network model to identify traffic status characteristics, and calculate the trafficability score of each edge in the path graph topology structure based on personnel density, handling equipment flow, obstacle interference and channel accessibility.

[0107] The processing module 202 is configured to calculate a path blocking factor by combining the passability score and the RFID node reading frequency. The path blocking factor includes a current channel density prediction value, a historical blocking frequency, an abnormal event probability, and a path alternative score.

[0108] The processing module 202 is used to embed the passability score and the path blocking factor into the path selection function of the ant colony optimization algorithm to dynamically modify the path pheromone concentration and the path selection probability.

[0109] In a possible implementation, the processing module 202 is configured to determine the emergency consumables corresponding to the emergency task according to the triggered emergency task identification signal.

[0110] The processing module 202 is configured to perform exponential weighting update on the weight function of the emergency consumables corresponding to the category of consumables in the consumables turnover priority function.

[0111] Processing module 202 is used to perform local graph updates on heterogeneous graph structures based on graph neural networks, and quickly recalculate the state embedding of the storage location nodes and their neighboring nodes corresponding to the emergency tasks, thereby embedding the emergency storage location state vectors for the storage location nodes and their neighboring nodes corresponding to the emergency tasks.

[0112] The processing module 202 is used to feed back the result of re-embedding the emergency storage location state vector to the ant colony algorithm as a weight compensation factor, thereby increasing the path weight of the outbound path of emergency consumables in the ant colony path search.

[0113] In one possible implementation, the processing module 202 is used to interrupt the current cargo space allocation process according to the triggered emergency task identification signal, start the preemptive scheduling mechanism, and dynamically weightedly correct the consumables turnover priority function based on the emergency task call status. The corrected consumables turnover priority function includes the historical call frequency, task urgency level, surgical area proximity factor and emergency call status factor.

[0114] The pheromone concentration in the ant colony optimization algorithm is adjusted according to the revised consumables turnover priority function, and the emergency channel priority function is introduced into the heuristic function to enhance the path selection probability of the emergency target path.

[0115] Processing module 202 is used to calculate the path weight value in the emergency state through pheromone concentration and heuristic function, and update and generate the cargo location allocation weight mapping matrix in the emergency state to achieve preferential allocation of medical consumables to storage locations with short paths, stable environments and close to the emergency operating area.

[0116] It should be noted that the above embodiments provide devices that implement their functions using only the division of the above functional modules as examples. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. In addition, the device and method embodiments provided in the above embodiments are based on the same concept. The specific implementation process is detailed in the method embodiment and will not be repeated here.

[0117] This embodiment also discloses an electronic device, referring to Figure 3 The electronic device may include: at least one processor 301 , at least one communication bus 302 , a user interface 303 , a network interface 304 , and at least one memory 305 .

[0118] The communication bus 302 is used to implement the connection and communication between these components.

[0119] The user interface 303 may include a display screen (Display) and a camera (Camera). Optionally, the user interface 303 may also include a standard wired interface and a wireless interface.

[0120] The network interface 304 may optionally include a standard wired interface or a wireless interface (such as a WI-FI interface).

[0121] The processor 301 may include one or more processing cores. Using various interfaces and circuits, the processor 301 connects to various components within the server. It executes instructions, programs, code sets, or instruction sets stored in the memory 305, as well as accesses data stored in the memory 305, to perform various server functions and process data. Optionally, the processor 301 may be implemented using at least one of the following hardware forms: a digital signal processing (DSP), a field-programmable gate array (FPGA), or a programmable logic array (PLA). The processor 301 may integrate one or a combination of a central processing unit (CPU), a graphics processing unit (GPU), and a modem. The CPU primarily processes the operating system, user interface, and application programs; the GPU is responsible for rendering and drawing content displayed on the display screen; and the modem handles wireless communications. It is understood that the modem may not be integrated into the processor 301 but implemented as a separate chip.

[0122] Memory 305 may include random access memory (RAM) or read-only memory (ROM). Optionally, the memory may include non-transitory computer-readable storage medium. Memory 305 may be used to store instructions, programs, code, code sets, or instruction sets. Memory 305 may include a program storage area and a data storage area. The program storage area may store instructions for implementing an operating system, instructions for at least one function (such as touch control, sound playback, image playback, etc.), and instructions for implementing the aforementioned method embodiments. The data storage area may store data related to the aforementioned method embodiments. Memory 305 may also optionally be at least one storage device located remotely from the processor 301. Memory 305, as a computer storage medium, may include an operating system, a network communication module, a user interface 303 module, and an application for a medical consumables management method based on an ant colony algorithm.

[0123] exist Figure 3In the electronic device shown, the user interface 303 is mainly used to provide an input interface for the user and obtain data input by the user; and the processor 301 can be used to call an application stored in the memory 305 that stores a medical consumables management method based on an ant colony algorithm. When executed by one or more processors 301, the electronic device executes one or more methods in the above embodiments.

[0124] It should be noted that for the aforementioned method embodiments, for simplicity of description, they are all expressed as a series of action combinations, but those skilled in the art should be aware that this application is not limited by the order of the actions described, because according to this application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required for this application.

[0125] In the above embodiments, the description of each embodiment has its own focus. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0126] In the several embodiments provided in this application, it should be understood that the disclosed devices can be implemented in other ways. For example, the device embodiments described above are merely schematic, such as the division of units, which is only a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some service interface, and the indirect coupling or communication connection of devices or units can be electrical or other forms.

[0127] Units described as separate components may or may not be physically separate, and components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0128] In addition, the functional units in the various embodiments of the present application may be integrated into a single processing unit, or each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0129] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable memory. Based on this understanding, the technical solution of this application, or the portion that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a memory 305 and includes several instructions for causing a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the various embodiments of the method of this application. The aforementioned memory 305 includes various media that can store program code, such as a USB flash drive, a mobile hard drive, a magnetic disk, or an optical disk.

[0130] The present application also discloses a computer-readable storage medium storing instructions, which, when executed by one or more processors 301 , enable an electronic device to execute one or more methods in the above embodiments.

[0131] The above are merely exemplary embodiments of the present disclosure and are not intended to limit the scope of the present disclosure. That is, any equivalent changes and modifications made in accordance with the teachings of the present disclosure are still within the scope of the present disclosure. After considering the disclosure of the specification and the truth of practice, those skilled in the art will easily think of other embodiments of the present disclosure. This application is intended to cover any variations, uses or adaptations of the present disclosure, which follow the general principles of the present disclosure and include common knowledge or customary technical means in the art that are not recorded in the present disclosure. The description and examples are to be regarded as exemplary only, and the scope and spirit of the present disclosure are defined by the claims.

Claims

1. A medical consumables management method based on ant colony algorithm, characterized in that: The method comprises: Obtain medical consumables data and build a medical consumables turnover frequency weight model and generate a consumables turnover priority function; Generate a cargo space allocation weight mapping matrix by combining the consumables turnover priority function with an ant colony optimization algorithm, wherein the mapping matrix is ​​established based on medical consumables and storage and access areas; During the transportation of medical supplies, ant colony individuals are simulated to perform multi-path search in the hospital channel topology, and path pheromones are dynamically updated to form a shortest path set; In the cargo location graph modeling phase, the warehouse space is constructed as a heterogeneous graph structure with storage locations as graph nodes and channels as graph edges. The attribute vector of each node is embedded through a graph neural network, and the embedding result is fed back to the ant colony optimization algorithm as the initial constraint for cargo location allocation. The initial constraints of cargo space allocation and the accessibility score and blocking factor generated by camera recognition data are introduced into the dynamic update of the path to dynamically adjust the consumables transportation path under sudden tasks.

2. The medical consumables management method based on ant colony algorithm according to claim 1, characterized in that: The generating of the cargo location allocation weight mapping matrix by combining the consumables turnover priority function with the ant colony optimization algorithm specifically includes: Constructing the consumables turnover priority function based on the consumables turnover frequency, task urgency level, and spatial deployment expectation; Using the consumables turnover priority function as a basis for setting the pheromone concentration of the ant colony optimization algorithm; defining a matching function between the medical consumables and the access area as a heuristic function of the ant colony optimization algorithm; By simulating the path selection behavior of individual ant colonies on the storage location topology map, the selection probability of allocating the medical consumables to different storage locations is calculated according to the pheromone concentration and the matching function; The cargo location allocation weight mapping matrix is ​​generated based on the selection probability.

3. The medical consumables management method based on ant colony algorithm according to claim 1, characterized in that: During the transportation of medical supplies, ant colony individuals are simulated to perform multi-path search in the hospital channel topology structure, and path pheromones are dynamically updated to form a shortest path set, specifically including: The hospital channel topology is modeled as a path graph topology with medical supplies storage location nodes, operating area nodes, exit nodes, and transfer nodes as nodes, and access channels as connecting edges; Setting a travel cost value for the connection edge, wherein the travel cost value is composed of physical distance, channel passenger flow, travel priority and historical accessibility; Initializing multiple ant colony individuals in the path graph topology structure, starting from the medical consumables storage location node and ending at the target surgical area node, and calculating the path selection probability based on the heuristic function constructed based on the path pheromone concentration and the path accessibility score; After each round of path search is completed, the path contribution value is calculated based on the path travel time and the task time window, and is used to update the path pheromone concentration. At the same time, the path blocking factor is generated based on the crowd density recognized by the camera and the RFID node traffic, and the pheromone concentration is dynamically suppressed, thereby forming the shortest path set for the medical consumables transportation task in multiple rounds of iterations.

4. The medical consumables management method based on ant colony algorithm according to claim 1, characterized in that: In the cargo location graph modeling stage, the storage space is constructed as a heterogeneous graph structure with storage locations as graph nodes and channels as graph edges. The attribute vector of each node is embedded through the graph neural network, and the embedding result is fed back to the ant colony optimization algorithm as the initial constraint for cargo location allocation. Specifically, it includes: Establishing the heterogeneous graph structure to represent the medical consumables storage space, wherein the graph nodes correspond to medical consumables storage locations, the graph edges correspond to access channels between the locations, and each graph edge is associated with a travel path distance, a travel efficiency score, and an environmental interference intensity; Generate an attribute vector for each graph node including RFID reading frequency, historical access delay, current congestion status, remaining capacity of the storage location, and environmental safety index; The attribute vector is input into the graph neural network model for multiple rounds of graph embedding calculations. A node embedding vector is generated by aggregating the attribute vectors of each node and its neighboring nodes. The node embedding vector is fed back to the ant colony optimization algorithm as a cargo location state representation vector. During the ant colony path search process, the similarity calculation result between the node embedding vector and the medical consumables category attribute vector is combined as a component of the ant colony heuristic function to construct a cargo location allocation weight mapping matrix and guide the precise cargo location allocation of medical consumables.

5. The medical consumables management method based on ant colony algorithm according to claim 1, characterized in that: The dynamic update of the path introduces the initial constraints of the cargo space allocation and the passability score and the blocking factor generated by the camera recognition data to dynamically adjust the consumables transportation path under the emergency task, specifically including: Extracting the embedded attribute vector and injecting it into the path search starting point of the ant colony optimization algorithm as the initial constraint for cargo location allocation; Cameras deployed in hospital corridors capture image data of human and logistics flows. This data is then fed into a convolutional neural network model to identify traffic status features. The accessibility score for each edge in the path graph topology is calculated based on personnel density, handling equipment flow, obstacle interference, and channel accessibility. Calculating a path blocking factor by combining the passability score with the RFID node reading frequency. The path blocking factor includes the current channel density prediction value, historical blocking frequency, abnormal event probability, and path alternative score. The passability score and the path blocking factor are embedded together in the path selection function of the ant colony optimization algorithm to dynamically correct the path pheromone concentration and the path selection probability.

6. The medical consumables management method based on ant colony algorithm according to claim 1, characterized in that: The dynamic path update includes introducing the initial constraints of the cargo space allocation and the passability score and the blocking factor generated by the camera recognition data to dynamically adjust the consumables transportation path under the emergency task, specifically including: Determining the emergency consumables corresponding to the emergency task according to the triggered emergency task identification signal; Performing exponential weighted updating on the weight function of the emergency consumables corresponding to the category of consumables in the consumables turnover priority function; Based on the graph neural network, the heterogeneous graph structure is locally updated, and the state embedding of the storage location node corresponding to the emergency task and its neighboring nodes is quickly recalculated, thereby embedding the emergency storage location state vector into the storage location node corresponding to the emergency task and its neighboring nodes; The result of re-embedding the emergency storage location state vector is fed back to the ant colony algorithm as a weight compensation factor to improve the path weight of the outbound path of the emergency consumables in the ant colony path search.

7. The medical consumables management method based on ant colony algorithm according to claim 1, characterized in that: The generating of the cargo location allocation weight mapping matrix by combining the consumables turnover priority function with the ant colony optimization algorithm specifically includes: According to the triggered emergency task identification signal, the current cargo space allocation process is interrupted, and the preemptive scheduling mechanism is started. The consumables turnover priority function is dynamically weighted and modified based on the emergency task call status. The modified consumables turnover priority function includes the historical call frequency, task urgency level, surgical area proximity factor and emergency call status factor; Adjusting the pheromone concentration in the ant colony optimization algorithm according to the modified consumables turnover priority function, and introducing the emergency channel priority function in the heuristic function to enhance the path selection probability of the emergency target path; The path weight value in the emergency state is calculated by the pheromone concentration and the heuristic function, and the cargo location allocation weight mapping matrix in the emergency state is updated and generated to achieve preferential allocation of medical consumables to storage locations with short paths, stable environments and close to the emergency operating area.

8. A medical consumables management device based on ant colony algorithm, characterized in that: The device is used to execute a medical consumables management method based on an ant colony algorithm as claimed in any one of claims 1 to 7, and the device comprises an acquisition module (201), a processing module (202) and an output module (203), wherein: The acquisition module (201) is used to acquire medical consumables data and construct a medical consumables turnover frequency weight model and generate a consumables turnover priority function; The processing module (202) is used to generate a cargo space allocation weight mapping matrix by combining the consumables turnover priority function with an ant colony optimization algorithm, wherein the mapping matrix is ​​established based on medical consumables and storage and access areas; The processing module (202) is used to simulate ant colony individuals to perform multi-path search in the hospital channel topology during the transportation of medical consumables, and dynamically update the path pheromone to form a shortest path set; The processing module (202) is used to construct the storage space into a heterogeneous graph structure with storage locations as graph nodes and channels as graph edges during the cargo location graph modeling phase, embed the attribute vector of each node through a graph neural network, and feed the embedding result back to the ant colony optimization algorithm as an initial constraint for cargo location allocation; The output module (203) is used to introduce the initial constraints of the cargo space allocation and the passability score and blocking factor generated by the camera recognition data into the dynamic update of the path, and dynamically adjust the consumable material transportation path under the emergency task.

9. An electronic device, characterized in that: The electronic device comprises a processor (301), a communication bus (302), a user interface (303), a network interface (304) and a memory (305), wherein the memory (305) is used to store instructions, the user interface (303) and the network interface (304) are both used to communicate with other devices, the communication bus (302) is used to realize connection and communication between components in the electronic device, and the processor (301) is used to execute the instructions stored in the memory (305) so that the electronic device executes the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores instructions, and when the instructions are executed, the method according to any one of claims 1 to 7 is executed.

Images