Intelligent scheduling and resource management system of microbiological detection laboratory

Abstract

The invention discloses an intelligent scheduling and resource management system for a microbiological detection laboratory, and relates to the technical field of intelligent management, and the system comprises a resource distribution module which carries out the space-time resource distribution of biological safety constraints through a quantum computing adaptive modeling method, generates a task scheduling optimization problem, carries out the optimal solution search through a quantum annealing algorithm, and carries out the optimization of the optimal solution; obtaining a reference resource scheduling scheme; the scheduling module is used for analyzing through a preset public health risk monitoring threshold value, and adjusting resource application and allocation by utilizing a dynamic resource recombination algorithm to obtain a resource scheduling instruction; the collaboration module is used for predicting a laboratory reagent consumption trend by adopting a metabonomics data analysis method, constructing a cross-laboratory emergency resource sharing path through a GAN network and outputting an emergency resource sharing scheme; according to the method, the biosecurity isolation rule is converted into the quantum bit coupling relation through quantum computing adaptive modeling, and the resource scheduling efficiency and the instrument utilization rate of high-grade experiments are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent management, and particularly to an intelligent scheduling and resource management system for a microbiology testing laboratory. Background Art

[0002] In recent years, the resource scheduling and safety management technology of microbiology testing laboratories has gradually developed towards the intelligent direction. Traditional methods mainly perform manual scheduling based on static rules (such as CLSI standards), while existing technologies have introduced graph network models for spatio-temporal conflict detection and used heuristic algorithms to optimize resource allocation. The application of quantum computing in combinatorial optimization problems provides new ideas for high-dimensional resource scheduling. For example, the D-Wave system has attempted to solve simple laboratory instrument scheduling problems.

[0003] The current microbiology testing laboratory management system is difficult to achieve the dynamic coupling of biosafety levels and resource scheduling. Traditional methods usually adopt a fixed permission allocation mechanism, which cannot dynamically adjust the usage permissions of instruments and personnel according to the risks of real-time detection tasks (such as operations on BSL-3 pathogens), resulting in conflicts between the isolation measures for high-level experiments and the resource scheduling requirements. Summary of the Invention

[0004] In view of the above existing problems, the present invention is proposed.

[0005] Therefore, the present invention provides an intelligent scheduling and resource management system for a microbiology testing laboratory to solve the problem of difficult dynamic coupling of biosafety levels and resource scheduling.

[0006] To solve the above technical problems, the present invention provides the following technical solutions: The present invention provides an intelligent scheduling and resource management system for a microbiology testing laboratory, which includes The data acquisition module collects multi-dimensional resource data of the microbiological testing laboratory and performs preprocessing; the isolation module divides resource usage permissions through a dynamic bio-safety level empowerment method to obtain a resource permission constraint table, and uses a spatio-temporal graph network model for safety isolation analysis to generate a safety isolation task queue; the resource allocation module performs spatio-temporal resource allocation with bio-safety constraints through a quantum computing adaptation modeling method, generates a task scheduling optimization problem, and uses a quantum annealing algorithm to search for the optimal solution to obtain a benchmark resource scheduling plan; the scheduling module analyzes through a preset public health risk monitoring threshold, and uses a dynamic resource recombination algorithm to adjust resource usage and allocation to obtain a resource scheduling instruction; the collaboration module uses a metabolomics data analysis method to predict the consumption trend of laboratory reagents, and constructs a cross-laboratory emergency resource sharing path through a GAN network to output an emergency resource sharing plan; the optimization module combines the emergency resource sharing plan and the benchmark resource scheduling plan, and performs iterative optimization through a multi-objective reinforcement learning method to generate the best operation strategy for the microbiological testing laboratory.

[0007] As a preferred embodiment of the intelligent scheduling and resource management system for the microbiological testing laboratory of the present invention, wherein: The multi-dimensional resource data of the microbiological testing laboratory includes experimental instrument status data, personnel flow data, environmental parameter data, reagent and consumable inventory data, sample processing progress data, and bio-safety level data; The preprocessing includes data cleaning, format standardization, key parameter dimensionality reduction, and time series alignment.

[0008] As a preferred embodiment of the intelligent scheduling and resource management system for the microbiological testing laboratory of the present invention, wherein: the step of dividing resource usage permissions through a dynamic bio-safety level empowerment method to obtain a resource permission constraint table is as follows. Based on the preprocessed multi-dimensional resource data, multi-factor weighted fusion is performed through a fuzzy comprehensive evaluation algorithm to generate a safety level scoring matrix; Perform spatio-temporal correlation analysis on the safety level scoring matrix and the preprocessed multi-dimensional resource data to obtain a laboratory operation permission conflict matrix, and use a rule inference engine to match multi-dimensional constraint conditions to generate a preliminary permission mapping table; Combine the preliminary permission mapping table with the preprocessed multi-dimensional resource data, perform multi-dimensional conflict detection through a graph coloring algorithm to obtain a permission conflict list with priorities, and combine historical violation record data to perform iterative optimization of permission weights through a reinforcement learning algorithm to output a resource permission constraint table.

[0009] As a preferred embodiment of the intelligent scheduling and resource management system for the microbiological testing laboratory of the present invention, wherein: the step of generating a safety isolation task queue is as follows. Based on the resource permission constraint table, identify high-risk operations through a rule matching engine, and isolate them to obtain a list of isolated combination lists; Based on the historical isolated combination list, train the spatio-temporal graph network model through the spatio-temporal graph attention mechanism method, and output an overlapping resource conflict relationship graph with a time window; Through a heuristic scheduling algorithm, resolve resource conflicts in the overlapping resource conflict relationship graph, generate a security isolation policy, and use a multi-objective optimization algorithm to evaluate the task urgency to obtain a security isolation task queue with priority markings.

[0010] As a preferred solution of the intelligent scheduling and resource management system for the microbial detection laboratory described in the present invention, wherein: through the quantum computing adaptation modeling method, perform spatio-temporal resource allocation of biosafety constraints, generate a task scheduling optimization problem, and the specific steps are as follows. Based on the security isolation task queue, convert high-risk operations into a binary decision variable group through a quantum bit mapping algorithm, and output a conflict operation vector with weight encoding; Through the biosafety constraint embedding method, convert the isolated combination list into a chain coupling strength, generate a Hamiltonian containing spatio-temporal isolation constraints, and convert it into a task scheduling optimization problem through the QUBO-to-Ising conversion rule.

[0011] As a preferred solution of the intelligent scheduling and resource management system for the microbial detection laboratory described in the present invention, wherein: use the quantum annealing algorithm to search for the optimal solution to obtain a benchmark resource scheduling plan, and the specific steps are as follows. Through the D-Wave quantum processor interface, convert the optimization problem elements in the task scheduling optimization problem into quantum bit coupling parameters and local magnetic field parameters to obtain a binary input stream; Perform a ground state search on the D-Wave quantum processor through the quantum annealing algorithm to obtain an optimal set of original bit states; Decode and verify the optimal set of original bit states through a classical post-processing method, eliminate invalid original bit state solutions that do not meet the constraint conditions, and generate a feasible scheduling solution; Through a greedy matching algorithm, map the feasible scheduling solution to the three elements of the laboratory spatio-temporal resource allocation, and output a benchmark resource scheduling plan.

[0012] As a preferred solution of the intelligent scheduling and resource management system for the microbial detection laboratory described in the present invention, wherein: analyze through a preset public health risk monitoring threshold, and use a dynamic resource reorganization algorithm to adjust resource usage and allocation to obtain a resource scheduling instruction, and the specific steps are as follows. Based on the benchmark resource scheduling scheme, analyze through the preset public health risk monitoring threshold, and output a set of risk operation marks exceeding the preset public health risk monitoring threshold; Through the adaptive genetic algorithm, dynamically adjust the resource usage and allocation of the three elements of laboratory spatio-temporal resources, and generate resource scheduling instructions.

[0013] As a preferred solution of the intelligent scheduling and resource management system for the microbial detection laboratory described in the present invention, wherein: the method of metabolomics data analysis is adopted to predict the consumption trend of laboratory reagents, and the specific steps are as follows. Based on the resource scheduling instructions, extract the historical laboratory reagent consumption data through the experimental record parsing engine, perform time alignment and outlier removal processing, and output the time-stamped laboratory reagent usage sequence; Based on the laboratory reagent usage sequence, extract the characteristics of the laboratory reagent metabolite concentration and time curve through the LC-MS metabolomics peak area integration method, and perform non-linear regression fitting to output the metabolic kinetic parameters of laboratory reagent consumption; Through the reverse parsing method of experimental procedures combined with the real-time laboratory reagent inventory, perform mapping analysis and trend extrapolation of the metabolic kinetic parameters for the laboratory reagent consumption, and output the predicted consumption trend of laboratory reagents.

[0014] As a preferred solution of the intelligent scheduling and resource management system for the microbial detection laboratory described in the present invention, wherein: construct a cross-laboratory emergency resource sharing path through the GAN network, and output an emergency resource sharing solution, and the specific steps are as follows. Based on the predicted consumption trend of laboratory reagents, perform vector quantization encoding of resource demand characteristics through the generator of the GAN network, and output a standardized resource demand tensor; Through the discriminator of the GAN network, perform multi-laboratory resource matching degree analysis, and combine with the shortest path algorithm of the graph network to generate the optimal allocation path, and obtain the emergency resource sharing solution.

[0015] As a preferred solution of the intelligent scheduling and resource management system for the microbial detection laboratory described in the present invention, wherein: combine the emergency resource sharing solution and the benchmark resource scheduling solution, and perform iterative optimization through the multi-objective reinforcement learning method to generate the best operation strategy for the microbial detection laboratory, and the specific steps are as follows. Based on the emergency resource sharing solution and the benchmark resource scheduling solution, perform spatio-temporal resource conflict resolution and fusion through the multi-constraint fusion algorithm, and output the optimization objective function; Perform Monte Carlo simulation iteration of the resource scheduling strategy through multi-objective reinforcement learning, and output the best operation strategy for the microbial detection laboratory.

[0016] The beneficial effects of the present invention are as follows: By means of quantum computing adaptation modeling, the biosafety isolation rules are transformed into qubit coupling relationships, which improves the resource scheduling efficiency and instrument utilization rate of high-level experiments; Through the collaborative steps of metabolomics and GAN network, based on reagent metabolism feature analysis and cross-laboratory resource matching modeling, the prediction accuracy of reagent requirements and the sharing efficiency of emergency resources are improved. A closed-loop system is formed through dynamic parameter coupling. When insufficient resources are detected, the cross-laboratory collaboration mechanism is automatically triggered, which not only improves the overall resource utilization rate of the laboratory, but also realizes the improvement of biosafety management level, achieving the collaborative optimization goal of safety and efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0018] Figure 1 FIG. is a schematic diagram of an intelligent scheduling and resource management system for a microbiological detection laboratory; Figure 2 FIG. is a schematic diagram of the generation of a resource permission constraint table; Figure 3 FIG. is a schematic diagram of the generation of a benchmark resource scheduling plan; Figure 4 FIG. is a schematic diagram of the generation of an optimal operation strategy. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0019] In order to make the above objects, features and advantages of the present invention more obvious and understandable, the following will describe the specific embodiments of the present invention in detail with reference to the accompanying drawings of the specification.

[0020] Many specific details are set forth in the following description in order to provide a thorough understanding of the present invention, but the present invention may be practiced in other ways different from those described herein. Those skilled in the art can make similar extensions without departing from the connotation of the present invention. Therefore, the present invention is not limited by the specific embodiments disclosed below.

[0021] Secondly, the so-called "one embodiment" or "embodiment" herein refers to a specific feature, structure or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment that excludes other embodiments.

[0022] Refer to Figures 1 to 4, which is an embodiment of the present invention. This embodiment provides an intelligent scheduling and resource management system for a microbiological testing laboratory, including the following steps: A data collection module that collects multi-dimensional resource data of the microbiological testing laboratory and performs preprocessing.

[0023] The multi-dimensional resource data of the microbiological testing laboratory includes experimental instrument status data, personnel flow data, environmental parameter data, reagent and consumable inventory data, sample processing progress data, and biosafety level data.

[0024] Specifically, the multi-dimensional resource data of the microbiological testing laboratory collects the real-time operation parameters of the experimental instrument status data through a sensor network, Internet of Things terminals, and log recording devices, locates the trajectory coordinates of the personnel flow data through access control tracking devices and positioning tags, collects the temperature, humidity, and differential pressure values of the environmental parameter data through environmental monitoring probes, collects the remaining quantity and expiration information of the reagent and consumable inventory data through barcode scanning instruments and inventory management terminals, collects the stage identification and time nodes of the sample processing progress data through sample tracking devices and process management terminals, and collects the biosafety level data through security audit records.

[0025] Preprocessing includes data cleaning, format standardization, key parameter dimensionality reduction, and time series alignment.

[0026] Specifically, for the preprocessing operation of the multi-dimensional resource data of the microbiological testing laboratory, the data cleaning step uses the Z-score anomaly detection algorithm to calculate the mean and standard deviation of numerical fields, deletes examples in the experimental instrument status data where the voltage value deviates from the mean by more than 3σ, and eliminates records in the environmental parameter data where the continuous sampling of temperature and humidity exceeds the range of ±3σ; The format standardization step applies the ISO 8601 time parser to uniformly convert the timestamps of the sample processing progress data into a standardized timestamp format, uses the forced type conversion method to convert the coordinate unit of the personnel flow data from feet to meters (for example, 1 foot = 0.3048 meters), converts the Fahrenheit temperature in the environmental parameter data to Celsius through the unit mapping table method (for example, the formula for ℉→℃ conversion), and implements the date formatting function to reconstruct the expiration date of the reagent and consumable inventory data into the YYYY-MM-DD format; The key parameter dimensionality reduction step uses the principal component analysis method to extract the principal components of the environmental parameter data and retains dimensions with a cumulative contribution rate > 85%, uses the variance threshold feature selection method to screen the experimental instrument status data, retains the trajectory inflection point coordinates in the personnel flow data where the stay duration exceeds 5 minutes through the trajectory key point extraction algorithm, and applies the rule-driven feature screening to focus on the remaining quantity / expiration date / storage condition warning dimensions of the reagent and consumable inventory data; The time series alignment step uses the linear interpolation compensation method to fill in the missing timestamps of the environmental parameter data and the sample processing progress data (for example, the missing period is compensated by the forward filling method).

[0027] Isolation module, which divides resource usage permissions through a biosecurity level dynamic empowerment method to obtain a resource permission constraint table.

[0028] Based on the preprocessed multi-dimensional resource data, multi-factor weighted fusion is performed through a fuzzy comprehensive evaluation algorithm to generate a safety level scoring matrix.

[0029] Specifically, a membership function for each dimension of data is constructed through a fuzzy comprehensive evaluation algorithm, and the experimental instrument status data, personnel flow data, environmental parameter data, reagent and consumable inventory data, sample processing progress data, and biosecurity level data are respectively fuzzified to generate a membership matrix for each dimension; perform fuzzy matrix multiplication operations on the factor weight vector generated by the analytic hierarchy process and the membership matrix of the preprocessed multi-dimensional resource data. For example, the membership degree of the differential pressure of 4.9 Pa in the environmental parameter data is 0.98, and the membership degree of the abnormal voltage in the experimental instrument status data is 0.12. Perform row-by-row weighted fusion, and the safety dimension score of each resource is calculated by multiplying the weight by the membership degree value and then accumulating. Generate a safety level scoring matrix. Each element in the safety level scoring matrix represents the comprehensive score of the resource in a specific safety dimension. For example, the instrument risk score of the biological safety cabinet = instrument risk weight 0.3 × membership degree 0.92 + personnel qualification weight 0.2 × membership degree 0.85 +... + sample risk weight 0.1 × membership degree 0.78 = 0.92, and finally output and generate a safety level scoring matrix.

[0030] Perform spatio-temporal correlation analysis on the safety level scoring matrix and the preprocessed multi-dimensional resource data to obtain a laboratory operation permission conflict matrix, and use a rule inference engine to match multi-dimensional constraint conditions to generate a preliminary permission mapping table.

[0031] Specifically, perform spatio-temporal correlation analysis on the safety level scoring matrix and the preprocessed multi-dimensional resource data. Perform spatial overlay analysis on the resource safety score in the safety level scoring matrix and the real-time position coordinates in the personnel flow data to identify spatial conflicts between safety level requirements and the actual distribution of personnel; compare the time dimension score in the safety level scoring matrix with the operation time node in the sample processing progress data through a time window to detect conflicts between safety time requirements and the actual operation time sequence; Load the laboratory biosafety management specification terms through the rule inference engine, and match the identified spatial conflicts and temporal conflicts with the specification terms one by one; the condition matching in the rule inference engine calls the corresponding constraint conditions according to the conflict type. The spatial conflict matches the spatial isolation specification terms, and the temporal conflict matches the operation timing specification terms. For example, the sterilization time is reserved before and after the BSL-3 level sample processing; the successfully matched constraint conditions generate permission control instructions. The spatial conflict instruction restricts personnel from entering high-risk areas, and the temporal conflict instruction adjusts the instrument usage period; the output of the rule inference engine organizes the matching results into a permission mapping relationship. The spatial permission mapping records the access control relationship of resource-person-region, and the temporal permission mapping records the usage control relationship of instrument-period, generating a preliminary permission mapping table containing fields such as resource ID, personnel ID, regional permission, and period permission.

[0032] Combine the preliminary permission mapping table with the preprocessed multi-dimensional resource data, perform multi-dimensional conflict detection through the graph coloring algorithm, obtain a permission conflict list with priorities, and combine the historical violation record data to perform iterative optimization of permission weights through the reinforcement learning algorithm, and output a resource permission constraint table.

[0033] Specifically, perform permission-resource association matching between the preliminary permission mapping table and the experimental instrument status data, personnel flow data, environmental parameter data, reagent and consumable inventory data, sample processing progress data, and biosafety level data fields in the preprocessed multi-dimensional resource data; transform the permission conflict detection problem into a graph structure of resource-personnel permissions through the graph coloring algorithm: the nodes represent the resource-personnel permission pairs in the preliminary permission mapping table, and the edges represent the mutual exclusion relationship between different permission pairs. Use the greedy algorithm to assign colors to the graph nodes of resource-personnel permissions. Each color corresponds to a permission conflict type, generating a permission conflict list containing fields such as conflict type code, conflict resource ID, conflict personnel ID, and conflict priority; the conflict priority is set according to the occurrence frequency of the same type of conflict in the historical violation record data; construct a permission weight optimization environment through the reinforcement learning algorithm: the state space is the conflict type and priority in the permission conflict list, the action space is the permission weight adjustment operation in the preliminary permission mapping table, and the reward function is calculated based on the decrease ratio of the violation rate in the historical violation record data; execute Q-learning to iteratively update the permission weight parameters. When the number of high-priority conflicts in the permission conflict list drops to, obtain the conflict number threshold through the average number of occurrences of the same type of conflict event in the laboratory safety audit report in the past 30 days, with an example historical mean of 20 times, and the conflict number threshold is set to 10 times, terminate the optimization when ; output the resource permission constraint table.

[0034] Use the spatio-temporal graph network model to perform safety isolation analysis and generate a safety isolation task queue.

[0035] Based on the resource permission constraint table, high-risk operations are identified through a rule matching engine and isolated to obtain a list of isolated combinations.

[0036] Specifically, the resource ID field in the resource permission constraint table is encoded as a resource identifier matching condition. For example, resource ID = EQ001, the permission type field is encoded as an operation type matching condition. For example, permission type = centrifuge operation, the optimized weight value field is encoded as a weight threshold matching condition. For example, optimized weight value = 0.4, and the conflict priority field is encoded as a priority determination condition. For example, conflict priority = 0.9. Each rule in the rule matching engine's rule library is defined using a "when-then" structure. For example, the rule is defined as "when the permission type is centrifuge operation and the optimized weight value is less than the example value of 0.5, then mark it as a high-risk operation". When the rule matching engine executes, it first loads all the records in the resource permission constraint table and matches them one by one with the conditions in the rule matching engine's rule library. The Rete algorithm is used to optimize the search efficiency during the matching process, and the pattern matching network is used to quickly locate the data records that meet the conditions. For the high-risk operation records that match successfully, they are classified according to the value of the conflict priority field: those with a priority greater than the example value of 0.8 are immediately terminated and an isolation instruction is generated (example isolation period = 08:00 - 10:00), and those with a priority less than or equal to the example value of 0.8 are allowed to complete the current operation step before generating an isolation instruction (example isolation area range = Experimental Area No. 3). The isolation instruction contains three necessary parameters: the operation resource ID example EQ001, the isolation period example 14:00 - 16:00, and the isolation area range example Biosafety Cabinet No. 2. Finally, all the generated isolation instructions are summarized into a list of isolated combinations.

[0037] Based on the historical list of isolated combinations, the spatio-temporal graph network model is trained through the spatio-temporal graph attention mechanism method, and an overlapping resource conflict relationship graph with a time window is output.

[0038] Specifically, the isolated resource ID, associated operator ID, isolation start time, and estimated release time fields in the historical list of isolated combinations are input into the spatio-temporal graph attention mechanism method. The node features are encoded as a combined vector of resource type and personnel qualification, and the edge features are encoded as the time window of the isolation period, such as 14:00 - 15:00, and the conflict type code; the spatio-temporal graph attention mechanism method calculates the conflict association weights for different time windows through the multi-head attention layer. For example, the conflict weight for the 14:00 - 14:30 period is 0.92, and the graph convolutional layer aggregates the spatio-temporal features of adjacent nodes to generate node embedding vectors; During the training phase, the mean squared error loss function is adopted, and the optimization objective is to match the predicted conflict relationships with the true records in the historical isolation combination list. The Adam optimizer is used to update the network parameters, and the training stops when the number of training rounds reaches a preset value, such as 100 rounds. After training, the spatio-temporal graph network model performs forward inference on the fields of the isolated resource ID, associated operator ID, isolation start time, and expected release time in the input isolation combination list, and outputs an overlapping resource conflict relationship graph with a time window.

[0039] The heuristic scheduling algorithm is used to resolve resource conflicts in the overlapping resource conflict relationship graph, generate a safety isolation strategy, and use the multi-objective optimization algorithm to evaluate the task urgency, resulting in a safety isolation task queue with priority markings.

[0040] Specifically, the conflict time window, resource ID, personnel ID, and emergency index fields in the overlapping resource conflict relationship graph with a time window are input into the heuristic scheduling algorithm. Based on the overlapping degree of the conflict time window and the emergency index value, the earliest schedulable time window priority strategy is applied, for example, tasks with an emergency index > 0.8 are scheduled first, and the time window of the conflict resources is shifted, for example, the original time window 14:00 - 14:30 is adjusted to 14:30 - 15:00. The generated safety isolation strategy includes the resource ID, adjusted time window, and isolated operator ID fields. The safety isolation strategy and the isolation reason code field in the historical isolation combination list are input into the multi-objective optimization algorithm. The optimization objectives are set as minimizing the total isolation duration, maximizing the resource utilization rate, and equalizing the emergency index distribution. The NSGA-II algorithm is used to calculate the Pareto optimal solution set, for example, generating 3 groups of candidate solutions. The laboratory task priority determination criteria, for example, a solution that simultaneously satisfies an isolation duration < 2 hours and an emergency index > 0.7 is marked as high priority. The candidate solutions are marked with priorities, and a safety isolation task queue with priority markings is output.

[0041] The resource allocation module generates a task scheduling optimization problem by performing spatio-temporal resource allocation with biosecurity constraints through a quantum computing adaptation modeling method.

[0042] Based on the safety isolation task queue, the high-risk operations are transformed into a binary decision variable group through the quantum bit mapping algorithm, and a conflict operation vector with weight encoding is output.

[0043] Specifically, based on the resource ID, adjusted time window, executor ID, and priority code field in the security isolation task queue, the qubit mapping algorithm encodes the high-risk operation status of the security isolation task queue into binary decision variables: the priority codes P1 - P3 are mapped to weighted coding examples, where P1 corresponds to 0.9, P2 corresponds to 0.7, and P3 corresponds to 0.5. The combination of the resource ID and the adjusted time window generates a unique operation identifier example: centrifuge_2023 - 10 - 05T14:30 - 15:00; the high-risk operation status is assigned according to the execution result of the security isolation policy example: the isolated operation is 1, and the non-isolated operation is 0. The weighted coding and the binary status combination form a conflict operation vector example: resource ID = centrifuge, binary status = 1, weighted coding = 0.9; the conflict operation vector with weighted coding is output.

[0044] Through the biosecurity constraint embedding method, the isolation combination list is transformed into a chain coupling strength, generating a Hamiltonian containing spatio-temporal isolation constraints, and through the QUBO-to-Ising conversion rule, it is transformed into a task scheduling optimization problem.

[0045] It should be noted that the expression for generating the Hamiltonian containing spatio-temporal isolation constraints is: ; Among them, is the Hamiltonian, is the isolation combination list, is the set of all isolation operations, is the operation pair number, such as = centrifugation operation, = sterilization operation, is the centrifugation operation and the sterilization operation between the coupling strength, is the centrifugation operation qubit (taking values -1 or +1), is the sterilization operation qubit (taking values -1 or +1), is the centrifugation operation local magnetic field strength; Specifically, based on the resource ID, time window, weighted coding, and binary status field in the security isolation task queue and the conflict operation vector with weighted coding, the biosecurity constraint embedding method transforms the isolation operation pairs in the isolation combination list, such as the centrifuge operation and the sterilization operation, into a chain coupling strength: according to the overlap degree of the time windows of the resource ID, for example, the time window of the centrifuge operation 14:30 - 15:00 and the time window of the sterilization operation 14:45 - 15:15 have a 15-minute overlap, and the coupling strength value is set, for example =0.8. According to the weight coding difference, the weight of the example centrifuge is 0.9 and the weight of sterilization is 0.7, and the local magnetic field strength example is adjusted. =0.9 - 0.7 = 0.2. When generating the Hamiltonian with space-time isolation constraints, through the QUBO-to-Ising conversion rule, the binary states (0 or 1) in the conflict operation vector of the binary decision variable group are converted into qubit spin states (-1 or +1), and the mapping relationship is that binary state 0 corresponds to spin -1 and binary state 1 corresponds to spin +1; the task scheduling optimization problem is transformed into solving the ground state configuration of the Hamiltonian H, and a task scheduling scheme with qubit spin states matching the coupling strength is output.

[0046] The quantum annealing algorithm is used to search for the optimal solution to obtain the benchmark resource scheduling scheme.

[0047] Through the D-Wave quantum processor interface, the optimization problem elements in the task scheduling optimization problem are converted into qubit coupling parameters and local magnetic field parameters to obtain a binary input stream.

[0048] Specifically, through the D-Wave quantum processor interface, the chain coupling strength and local magnetic field parameters in the task scheduling optimization problem are mapped to qubit physical parameters: the chain coupling strength example, the coupling strength between the centrifuge operation and the sterilization operation, matches the connectable qubit pair example qubit number combination in the physical topology of the D-Wave quantum processor, and the chain coupling strength value is directly assigned to the coupling parameter field of the corresponding qubit pair; for the logical qubit that requires a chain structure example, a long chain spanning multiple physical qubits, the D-Wave chain coupling configuration method is used to map the same logical qubit to multiple physical qubits, and the qubit coupling parameter is given a fixed high strength value to ensure logical consistency. The local magnetic field example, the magnetic field strength of the centrifuge operation, is mapped to the magnetic field parameter field of the corresponding qubit number example qubit identifier; when the logical qubit is mapped to multiple physical qubits, the local magnetic field value is evenly distributed to each physical qubit example distribution rule. Call the standard data encapsulation function of the D-Wave quantum processor interface to encode the qubit number, qubit coupling parameter, and local magnetic field parameter into a binary input stream in the QMASM protocol format.

[0049] The ground state search is performed on the D-Wave quantum processor through the quantum annealing algorithm to obtain the optimal set of original bit states.

[0050] Specifically, the quantum annealing algorithm is executed by a D-Wave quantum processor. The qubit numbers, qubit numbers, and local magnetic field parameters in the binary input stream are loaded into the physical qubits of the D-Wave quantum processor to initialize the qubit spin states. The quantum annealing algorithm constructs an energy landscape based on the coupling parameters and magnetic field parameters, and evolves from a high temperature to a low temperature by adjusting the strength of the quantum tunneling effect, so that the qubit spin states converge to the ground state configuration with the lowest energy. After the annealing process ends, the readout circuit of the D-Wave quantum processor measures the qubit spin states (-1 or +1). After 1000 examples of multiple annealing cycles, the spin state combination with the highest occurrence probability is statistically selected as the optimal set of original bit states. The qubit spin states are converted into binary states (-1 corresponds to 0, +1 corresponds to 1), and it is verified whether the ground state energy of the optimal set of original bit states conforms to the Hamiltonian definition. The optimal set of original bit states is output.

[0051] The optimal set of original bit states is decoded and verified by a classical post-processing method, and the invalid original bit state solutions that do not meet the constraint conditions are eliminated to generate a feasible scheduling solution.

[0052] Specifically, the resource ID, time window, and binary state field in the optimal set of original bit states are decoded by a classical post-processing method: the binary state field is mapped to the actual operation instruction, for example, binary state 1 corresponds to "allowed operation", and 0 corresponds to "forbidden operation"; it is verified whether the decoded operation instruction meets the spatio-temporal isolation constraint conditions in the isolation combination list; the invalid solutions that violate the biosecurity operation specification clauses are eliminated, for example, the number of high-risk instrument users exceeds the limit in the same time period; the resource utilization rate of the remaining valid original bit state solutions is evaluated, for example, the total isolation duration and the instrument idle rate are obtained, and the original bit state solutions that simultaneously meet the minimum isolation duration and the maximum resource utilization rate are retained; a feasible scheduling solution is generated.

[0053] The feasible scheduling solution is mapped to the three elements of the laboratory spatio-temporal resource allocation by a greedy matching algorithm, and a benchmark resource scheduling plan is output.

[0054] Specifically, the resource ID, adjusted time window, and operation status field in the feasible scheduling solution are processed through a greedy matching algorithm, and the feasible tasks in the feasible scheduling solution are sorted from high to low. For example, the centrifuge operation with an emergency index of 0.9 takes precedence over the sterilization operation with an emergency index of 0.7; the feasible task list after sorting is traversed, and the resource ID of the feasible task is matched with the available resource status in the three elements of laboratory space-time resource allocation, which are the equipment registration form, personnel scheduling form, and experimental task reservation form in the laboratory resource management database (the three elements of laboratory space-time resource allocation include resource ID, allocated time window, and executor ID). For example, the biosafety cabinet is available during the period from 09:00 to 10:00; when allocating resources exemplarily, continuous time windows are preferentially occupied. For example, the centrifuge operation is allocated to the complete time period from 14:30 to 15:00; the time window boundary alignment strategy is adopted for conflict detection. For example, the time window of the sterilization operation is adjusted to 15:00 - 15:30 to avoid overlapping with the centrifuge operation; the allocation result is recorded in the benchmark resource scheduling plan, and the fields include resource ID, allocated time window, and executor ID; before output, it is verified whether the allocation result meets the priority marks in the safety isolation task queue. For example, the P1 task must be allocated to the designated safety area.

[0055] The scheduling module analyzes through a preset public health risk monitoring threshold and adjusts the resource usage and allocation using a dynamic resource reorganization algorithm to obtain a resource scheduling instruction.

[0056] Based on the benchmark resource scheduling plan, it analyzes through a preset public health risk monitoring threshold and outputs a set of risk operation marks that exceed the preset public health risk monitoring threshold.

[0057] Specifically, for the setting of the public health risk monitoring threshold, the single-concurrency operation threshold of high-risk instruments is extracted. For example, the upper limit of the number of centrifuge operations in a single time window is 2 times; the restriction clause on the personnel carrying capacity in the same physical area of the laboratory in the same regulation is parsed, and the personnel density threshold is extracted. For example, the upper limit of the number of people in the experimental area at the same time is 5 people; the violation scenarios described in the clause are directly mapped to the preset risk type codes. For example, "instrument overloading operation" in the clause corresponds to the instrument overload risk code, and "overstaffed operation" corresponds to the personnel density risk code. The public health risk monitoring threshold is strictly associated with the operation times and the upper limit of the number of personnel defined in the original regulation of the clause, and the risk code is bound to the clause violation description one by one. Finally, the setting of the public health risk monitoring threshold is completed by referring to the public content of the specification clause; Based on the resource ID, allocation time window, and executor ID fields in the benchmark resource scheduling plan, parse the number of concurrent operations of high-risk instruments within the same time window in the benchmark resource scheduling plan. For example, the centrifuge is allocated 3 operations during the period from 14:30 to 15:00, exceeding the operation threshold by 2 times. Count the number of operators in the same physical area. For example, the number of people in the BSL-3 experimental area during the same period exceeds the personnel density threshold of 5 people; compare the statistical values item by item with the public health risk monitoring threshold to identify the operation records that exceed the public health risk monitoring threshold. For example, the number of centrifuge operations = 3 > the operation threshold of 2, and the number of people = 6 > the personnel density threshold of 5; add risk type labels to the operation records that exceed the public health risk monitoring threshold. For example, the instrument overload risk code R001 and the personnel density risk code R002; generate a risk operation mark set.

[0058] Through the adaptive genetic algorithm, dynamically adjust the resource usage and allocation of the three elements of laboratory spatio-temporal resource allocation to generate resource scheduling instructions.

[0059] Specifically, use the resource ID and risk time window fields in the risk operation mark set as input parameters; initialize the population to include multiple candidate adjustment plans, each plan consisting of adjustment actions for the three elements of laboratory spatio-temporal resource allocation. The three elements of laboratory spatio-temporal resource allocation include resource ID, allocation time window, and executor ID. The adjustment action is to reallocate the usage period of the instrument corresponding to the resource ID or replace the executor ID; the evaluation method is to count the total number of risk type codes in the risk operation mark set, count the number of resource conflict events after adjustment, and combine the two values of the total number of risk type codes and the number of resource conflicts according to the weight ratio. For example, the weight of the total number of risk type codes is 0.6, and the weight of the number of resource conflicts is 0.4 to generate a fitness score. The lower the fitness score value, the better the comprehensive risk and conflict level of the candidate adjustment plan; Execute the selection operation to retain the candidate plans ranked in the top 30% in terms of fitness; perform crossover operations on the selected plans, randomly exchange the time window segments of the same resource ID in two plans. For example, exchange the allocation of the centrifuge from 14:30 to 15:00 in plan A and from 15:30 to 16:00 in plan B; perform mutation operations to randomly replace the executor ID with a 5% probability or reset the time window in the instrument available time pool; iteratively execute the fitness evaluation, selection, crossover, and mutation steps until the termination condition is reached. For example, iterate 100 times or the fitness does not change continuously for 10 times; output the candidate adjustment plan with the highest fitness as the resource scheduling instruction.

[0060] The collaboration module uses metabolomics data analysis methods to predict the consumption trend of laboratory reagents.

[0061] Based on resource scheduling instructions, the historical laboratory reagent consumption data is extracted through the experimental record parsing engine, and time alignment and outlier elimination are performed to output a timestamped laboratory reagent usage sequence.

[0062] Specifically, the experimental record parsing engine reads the resource ID, executor ID, and adjusted time window fields in the resource scheduling instruction; associates the historical laboratory reagent consumption data storage table according to the resource ID, and filters the corresponding operation records according to the combination of resource ID and executor ID; aligns the timestamp of the filtered reagent consumption data with the adjusted time window in the resource scheduling instruction, and intercepts the records whose consumption data timestamps fall within the adjusted time window (for example, the adjusted time window is 14:30-15:00), and only retains the reagent consumption data of 14:30≤t<15:00; performs a removal operation on the reagent consumption data outside the time window; uses the Laida criterion to identify outliers. If the recorded reagent consumption exceeds the historical mean ±3 times the standard deviation range (for example, the historical mean reagent consumption of the centrifuge in the same period is 20ml and the standard deviation is 5ml), then the records with consumption <5ml or >35ml are removed, marked as outliers and removed; the processed reagent consumption is combined and counted according to the adjusted time window to generate a laboratory reagent usage sequence with a timestamp.

[0063] Based on the laboratory reagent usage sequence, the LC-MS metabolomics peak area integration method was used to extract the concentration and time curve characteristics of laboratory reagent metabolites, and nonlinear regression fitting was performed to output the metabolic kinetic parameters of laboratory reagent consumption.

[0064] Specifically, the reagent consumption field in the timestamped laboratory reagent usage sequence is associated with the original metabolite mass spectrometry data detected by LC-MS, and the data are aligned and matched according to the timestamp; the mass spectrometry data collected at each time point are integrated to identify the characteristic ion peak of the target metabolite (for example, the peak of trypsin metabolite m / z=567.3), and the peak area value is generated by integration and converted into the metabolite concentration (for example, the peak area = 12000 corresponds to the concentration = 25μM); the laboratory reagent metabolite concentration values ​​of all time points are combined to generate a concentration-time curve; the nonlinear regression fitting method is used to input the laboratory reagent metabolite concentration value to generate the concentration-time curve into the kinetic equation (for example, the Michaelis-Menten equation or the first-order kinetic equation), and the equation parameters are iteratively optimized by the least squares method (for example, the maximum number of iterations is 1000 times and the convergence standard is 1e-6), and the metabolic kinetic parameters (for example, half-life t1 / 2=4.2h and maximum reaction rate Vmax=58μM / h) are fitted; it is verified that the residual sum of squares of the fitting result reaches the preset standard (for example, the residual sum of squares ≤0.05 is considered valid), and the verified metabolic kinetic parameters are output.

[0065] Through the reverse parsing method of experimental procedures combined with real-time laboratory reagent inventory, perform mapping analysis and trend extrapolation of the consumption of laboratory reagents for metabolic kinetic parameters, and output the predicted consumption trend of laboratory reagents.

[0066] Specifically, associate the half-life and maximum reaction rate fields in the metabolic kinetic parameters of laboratory reagent consumption with the current stock and procurement cycle fields of real-time laboratory reagent inventory; analyze the relationship between the time stamp and consumption in the historical laboratory reagent usage sequence, and generate a reagent consumption rate relationship based on the half-life and reaction rate of the metabolic kinetic parameters. Fit the rate relationship with the historical consumption sequence; combine the current stock of real-time inventory and the procurement cycle example current stock 50ml, procurement cycle 7 days, and extrapolate the future consumption according to the time window example future 7-day predicted consumption = daily average consumption 3.5ml × 7 = 24.5ml; when the predicted consumption exceeds the preset standard of real-time inventory example predicted consumption > 50ml - safety stock 10ml = 40ml, mark the time node for replenishment; output the predicted consumption trend of laboratory reagents including time window, predicted consumption, and inventory warning status fields example time window = day 5, predicted consumption = 42ml, inventory warning status = need to replenish.

[0067] Construct a cross-laboratory emergency resource sharing path through the GAN network and output an emergency resource sharing plan.

[0068] Based on the predicted consumption trend of laboratory reagents, perform vector quantization encoding of resource demand features through the generator of the GAN network and output a standardized resource demand tensor.

[0069] Specifically, convert the time window, predicted consumption, and inventory warning status fields in the predicted consumption trend of laboratory reagents into multi-dimensional numerical vectors example the time window is encoded as the starting hour number example 14 represents 14:00, the predicted consumption retains the original value, and the inventory warning status uses binary encoding, need to replenish = 1, no warning = 0; input the multi-dimensional numerical vector into the input layer of the generator of the GAN network. The generator of the GAN network includes a fully connected layer and a convolutional layer. The fully connected layer maps the input vector to a hidden feature space example the input dimension of 3 dimensions is mapped to a 128-dimensional hidden vector, and the convolutional layer enhances the features of the hidden vector to generate a high-dimensional tensor; Perform standardization processing on the tensor output by the generator of the GAN network, and use the mean square error normalization method defined in the laboratory resource encoding standard example scale the numerical values of each channel of the tensor to the [-1, 1] interval; verify that the dimension of the generated tensor is consistent with the input format required by the laboratory resource scheduling interface example the output tensor shape is example batch size, 64, 64, 3, corresponding to a 64×64 grid resource distribution map and 3 feature channels, and output a standardized resource demand tensor.

[0070] The discriminator of the GAN network is used to analyze the matching degree of multi-laboratory resources, and combined with the shortest path algorithm of the graph network to generate the optimal allocation path, so as to obtain the emergency resource sharing plan.

[0071] Specifically, the standardized resource demand tensor is input into the discriminator of the GAN network. The discriminator structure includes a convolutional layer and a fully connected layer. The convolutional layer extracts features of instrument usage density, personnel distribution density, and reagent consumption density in the tensor; the fully connected layer compares the features with the available reagent stock and instrument idle time fields in the real-time resource list of multiple laboratories, and outputs the matching degree score. For example, the matching degree between the resource list of laboratory A and the demand tensor is 0.92. Traverse all laboratory resource lists, and screen candidate laboratories whose matching degree scores reach the preset effective standard. For example, the matching degree of laboratory A is 0.92, and the matching degree of laboratory B is 0.85; construct a graph network structure for the transfer path between laboratories, where the nodes are the laboratory location coordinates, and the edge weights are the transportation time, transportation cost, and path reliability parameters between laboratories. For example, the transportation time from laboratory A to laboratory B is 2 hours, and the cost is 500 yuan; run the Dijkstra shortest path algorithm in the graph network, with the goal of minimizing the weighted sum of transportation time and cost, to generate the optimal transfer path from the demanding laboratory to the candidate laboratory. For example, the path is the demanding party → laboratory C → laboratory A; bind the optimal transfer path with the resource allocation quantity and time window fields of the matching laboratory to generate the emergency resource sharing plan. For example, the plan includes the transfer path "demanding party → laboratory A", the allocation quantity = 42 ml, and the time window = 14:30 - 16:00.

[0072] The optimization module combines the emergency resource sharing plan and the benchmark resource scheduling plan, and performs iterative optimization through the multi-objective reinforcement learning method to generate the best operation strategy for the microbiological testing laboratory.

[0073] Based on the emergency resource sharing plan and the benchmark resource scheduling plan, the spatio-temporal resource conflict resolution and fusion are carried out through the multi-constraint fusion algorithm, and the optimization objective function is output.

[0074] Specifically, extract the transfer path, allocation quantity, and time window fields in the emergency resource sharing plan and the resource ID, allocation time window, and executor ID fields in the benchmark resource scheduling plan; according to the conflict types defined by the laboratory resource scheduling conflict detection rules, such as instrument usage time overlap and personnel area density exceeding the limit, compare item by item whether the instrument usage time of the resource ID corresponding to the emergency resource sharing plan and the benchmark resource scheduling plan overlaps. For example, the centrifuge resource ID = EQ001 is occupied by both plans at 14:30 - 15:00, and whether the executor ID exceeds the carrying limit in the same physical area. For example, the number of people in the BSL-3 experimental area = 7 > the upper limit of 5. Apply the conflict resolution rules to the detected conflict types. For the emergency resource sharing plan example that covers the benchmark resource scheduling plan, the centrifuge time period occupied by the allocation path takes precedence, and the conflict time period in the benchmark plan is automatically shifted backward; input the resource allocation data after conflict resolution into the multi-constraint fusion algorithm. The multi-constraint fusion algorithm performs weighted fusion on the resource utilization rate, operation delay time, and conflict resolution times indicators based on the preset operation delay time weight of 0.4 and resource utilization rate weight of 0.6. Example: resource utilization rate × 0.6 + (1 - operation delay coefficient) × 0.4; output the optimization objective function.

[0075] Perform Monte Carlo simulation iterations on the resource scheduling strategy through multi-objective reinforcement learning, and output the optimal operation strategy for the microbiology testing laboratory.

[0076] Specifically, map the comprehensive evaluation value, resource utilization rate, and operation delay coefficient fields in the optimization objective function to the state space; initialize the policy parameters, and define the action as adjusting the instrument usage time period, personnel allocation ratio, and reagent allocation path fields. Example: action = "Adjust the usage time period of centrifuge resource ID = EQ001 from 14:30 - 15:00 to 15:00 - 15:30". Execute the Monte Carlo simulation iteration until the number of simulations reaches the standard. Example: run 1000 sampling times; in each simulation, select the action according to the current policy, update the laboratory resource allocation status, and calculate the immediate reward value for the reduction of biosafety violation times, cost reduction, and penalty for increased time delay. Example: reward for reduction of biosafety violation times +5, reward for cost reduction +3, penalty for increased time delay -2; update the policy parameters through the temporal difference error, and the error calculation method uses the weighted average of the long-term rewards of the state-action pairs in the historical simulation data; screen the non-dominated solution strategies that simultaneously meet the biosafety requirements, cost limit, and time efficiency standards. The screening method is Pareto front analysis. Example: retain the policy combinations with biosafety violation times ≤ 1, cost ≤ 800 yuan, and delay ≤ 60 minutes. Output the optimal operation strategy for the microbiology testing laboratory, including the instrument usage time period adjustment table, personnel allocation plan, and reagent allocation priority list. Example: adjust the time period of centrifuge resource ID = EQ001 to 15:00 - 15:30, the upper limit of personnel allocation in the BSL-3 experimental area is 5 people, and the priority of the trypsin allocation path is demand side → laboratory A.

[0077] In summary, the present invention achieves the following: Through quantum computing adaptation modeling, the biosafety isolation rules are transformed into qubit coupling relationships, improving the resource scheduling efficiency and instrument utilization rate of high-level experiments; through the collaborative steps of metabolomics and the GAN network, based on reagent metabolism feature analysis and cross-laboratory resource matching modeling, the prediction accuracy of reagent requirements and the sharing efficiency of emergency resources are improved. A closed-loop system is formed through dynamic parameter coupling, automatically triggering the cross-laboratory collaboration mechanism when resource shortages are detected, not only improving the overall resource utilization rate of the laboratory, but also achieving an improvement in biosafety management level and attaining the collaborative optimization goal of safety and efficiency.

[0078] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention, and they should all be covered within the scope of the claims of the present invention.

Claims

1. An intelligent scheduling and resource management system for a microbiological testing laboratory, characterized in that: Including, A data acquisition module that acquires multi-dimensional resource data of a microbiology testing laboratory and performs preprocessing; An isolation module that divides resource usage permissions through a dynamic bio-safety level empowerment method to obtain a resource permission constraint table, and uses a spatio-temporal graph network model for safety isolation analysis to generate a safety isolation task queue; A resource allocation module that performs spatio-temporal resource allocation with bio-safety constraints through a quantum computing adaptation modeling method, generates a task scheduling optimization problem, and uses a quantum annealing algorithm to search for the optimal solution to obtain a benchmark resource scheduling plan; A scheduling module that analyzes through a preset public health risk monitoring threshold, and uses a dynamic resource recombination algorithm to adjust resource usage and allocation to obtain a resource scheduling instruction; A collaboration module that uses metabolomics data analysis methods to predict the consumption trend of laboratory reagents, and constructs a cross-laboratory emergency resource sharing path through a GAN network to output an emergency resource sharing plan; An optimization module that combines the emergency resource sharing plan and the benchmark resource scheduling plan, and performs iterative optimization through a multi-objective reinforcement learning method to generate the best operation strategy for the microbiology testing laboratory.

2. The intelligent scheduling and resource management system for a microbial detection laboratory according to claim 1, wherein: The multi-dimensional resource data of the microbiology testing laboratory includes experimental instrument status data, personnel flow data, environmental parameter data, reagent and consumable inventory data, sample processing progress data, and bio-safety level data; The preprocessing includes data cleaning, format standardization, key parameter dimension reduction, and time series alignment.

3. The intelligent scheduling and resource management system for a microbiological testing laboratory according to claim 2, wherein: The method of dividing resource usage permissions through the dynamic bio-safety level empowerment method to obtain a resource permission constraint table is as follows: Based on the preprocessed multi-dimensional resource data, multi-factor weighted fusion is performed through a fuzzy comprehensive evaluation algorithm to generate a safety level scoring matrix; The safety level scoring matrix and the preprocessed multi-dimensional resource data are subjected to spatio-temporal correlation analysis to obtain a laboratory operation permission conflict matrix, and a rule inference engine is used to match multi-dimensional constraint conditions to generate a preliminary permission mapping table; The preliminary permission mapping table is combined with the preprocessed multi-dimensional resource data, and multi-dimensional conflict detection is performed through a graph coloring algorithm to obtain a permission conflict list with priorities. Combining historical violation record data, permission weight iterative optimization is performed through a reinforcement learning algorithm to output a resource permission constraint table.

4. The intelligent scheduling and resource management system for a microbial detection laboratory according to claim 3, characterized in that: The steps of generating the safety isolation task queue are as follows: Based on the resource permission constraint table, a rule matching engine is used to identify high-risk operations and isolate them to obtain an isolation combination list; Based on the historical isolation combination list, the spatio-temporal graph network model is trained through the spatio-temporal graph attention mechanism method, and an overlapping resource conflict relationship graph with a time window is output; The overlapping resource conflict relationship graph is subjected to resource conflict resolution through a heuristic scheduling algorithm to generate a safety isolation strategy, and a multi-objective optimization algorithm is used to evaluate the task urgency to obtain a safety isolation task queue with priority markings.

5. The intelligent scheduling and resource management system for a microbiological testing laboratory according to claim 4, wherein: The steps of performing spatio-temporal resource allocation with bio-safety constraints through the quantum computing adaptation modeling method to generate a task scheduling optimization problem are as follows: Based on the secure isolation task queue, the high-risk operations are converted into binary decision variable groups through the quantum bit mapping algorithm, and the conflict operation vector with weight encoding is output; Through the biosafety constraint embedding method, the isolation combination list is transformed into chain coupling strength to generate a Hamiltonian containing spatiotemporal isolation constraints, which is then converted into a task scheduling optimization problem through the QUBO-to-Ising conversion rule.

6. The intelligent scheduling and resource management system for a microbiological testing laboratory according to claim 5, characterized in that: The specific steps of obtaining the benchmark resource scheduling solution are as follows: Through the D-Wave quantum processor interface, the optimization problem elements in the task scheduling optimization problem are converted into quantum bit coupling parameters and local magnetic field parameters to obtain a binary input stream; The quantum annealing algorithm is used to search for the ground state on the D-Wave quantum processor to obtain the optimal original bit state solution set; The optimal original bit state solution set is decoded and verified through classical post-processing methods, invalid original bit state solutions that do not meet the constraints are eliminated, and feasible scheduling solutions are generated; Through the greedy matching algorithm, the feasible scheduling solution is mapped to the three elements of laboratory spatiotemporal resource allocation, and the benchmark resource scheduling plan is output.

7. The intelligent scheduling and resource management system for a microbial detection laboratory according to claim 6, characterized in that: The above-mentioned steps are as follows: Based on the benchmark resource scheduling scheme, the preset public health risk monitoring threshold is used for analysis, and a risk operation marker set that exceeds the preset public health risk monitoring threshold is output; Through adaptive genetic algorithms, the resource usage and allocation of the three elements of laboratory space-time resource allocation are dynamically adjusted to generate resource scheduling instructions.

8. The intelligent scheduling and resource management system for a microbiological testing laboratory according to claim 7, characterized in that: The metabolomics data analysis method is used to predict the trend of laboratory reagent consumption. The specific steps are as follows: Based on resource scheduling instructions, the experimental record parsing engine extracts historical laboratory reagent consumption data, performs time alignment and outlier elimination, and outputs a laboratory reagent usage sequence with a timestamp; Based on the laboratory reagent usage sequence, the LC-MS metabolomics peak area integration method was used to extract the laboratory reagent metabolite concentration and time curve characteristics, and nonlinear regression fitting was performed to output the metabolic kinetic parameters of laboratory reagent consumption; Through the reverse analysis of experimental procedures combined with real-time laboratory reagent inventory, the metabolic kinetic parameters are mapped and analyzed for laboratory reagent consumption and trend extrapolation, and the predicted consumption trend of laboratory reagents is output.

9. The intelligent scheduling and resource management system for a microbiological testing laboratory according to claim 8, characterized in that: The cross-laboratory emergency resource sharing path is constructed through the GAN network, and the emergency resource sharing plan is output. The specific steps are as follows: Based on the predicted consumption trend of laboratory reagents, the resource demand feature vector is quantized and encoded through the generator of the GAN network, and the standardized resource demand tensor is output; The discriminator of the GAN network is used to analyze the matching degree of multi-laboratory resources, and the optimal allocation path is generated by combining the graph network shortest path algorithm to obtain an emergency resource sharing plan.

10. The intelligent scheduling and resource management system for a microbiological testing laboratory according to claim 9, characterized in that: The emergency resource sharing scheme and the benchmark resource scheduling scheme are combined, and iterative optimization is performed through a multi-objective reinforcement learning method to generate the optimal microbial testing laboratory operation strategy. The specific steps are as follows: Based on the emergency resource sharing plan and the benchmark resource scheduling plan, the spatio-temporal resource conflict resolution and fusion are carried out through a multi-constraint fusion algorithm, and the optimized objective function is output; Through the Monte Carlo simulation iteration of the resource scheduling strategy by multi-objective reinforcement learning, the best operation strategy of the microbial detection laboratory is output.

Images