Method, device and equipment for arranging and testing laboratory orders and storage medium

Abstract

The invention relates to the technical field of software development, and discloses a method, a device and equipment for arranging and testing laboratory orders, and a storage medium. The method comprises the following steps: according to a current order arrangement strategy generated based on a reinforcement learning model, performing chromosome coding on current laboratory order information, current experimenter information and current experimental equipment information to obtain a corresponding initial population; the fitness of individuals in the initial population is evaluated based on a current fitness function, the initial population is iteratively evolved by selecting and improving crossover and mutation operation of chromosomes, and the current fitness function is obtained based on examination parameters of laboratory orders predicted by a first AI model; and under the condition that a set genetic termination condition is met, an optimal laboratory order scheduling and testing scheme is obtained. Therefore, through the reinforcement learning model and the AI model and in combination with the genetic algorithm, comprehensive optimization of laboratory order arrangement and measurement is realized, and the order arrangement efficiency and the resource utilization rate are improved.

Description

Technical Field

[0001] This application relates to the field of software development technology, such as methods, apparatus, devices, and storage media for laboratory order scheduling. Background Technology

[0002] Laboratory order scheduling systems play a crucial role in modern laboratory management, often handling a large volume of testing orders daily with varying priorities, deadlines, and testing times. Without an efficient scheduling system, order processing can become chaotic, leading to inefficient testing.

[0003] Relevant laboratory order scheduling systems typically rely on predefined rules, such as First-Come, First-Served (FCFS), Deadline-First (EDD), and Shortest-Place-First (SPT). Resource allocation is usually static, failing to consider dynamic changes such as machine failures or changes in operator availability. Furthermore, special circumstances like retesting requirements or urgent orders often necessitate manual intervention from operators to adjust the schedule. Clearly, these systems lack flexibility, cannot effectively handle dynamic changes, and have limited optimization capabilities. Relying on a single rule makes it difficult to achieve multi-objective optimization, such as simultaneously optimizing order completion time, operator load balancing, and station utilization. They also struggle to adaptively adjust scheduling strategies in complex scenarios, such as retesting, early completion, or machine failures.

[0004] It should be noted that the information disclosed in the background section above is only used to enhance the understanding of the background of this application, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0005] To provide a basic understanding of some aspects of the disclosed embodiments, a brief summary is given below. This summary is not intended as a general commentary, nor is it intended to identify key / important components or describe the scope of protection of these embodiments, but rather as a prelude to the detailed description that follows.

[0006] This disclosure provides a method, apparatus, device, and storage medium for laboratory order scheduling to address the technical problems of inflexible and inefficient laboratory order scheduling.

[0007] In some embodiments, the method includes: Based on the current order scheduling strategy generated by the reinforcement learning model, chromosome encoding is performed on the current laboratory order information, current experimental personnel information, and current experimental equipment information to obtain the corresponding initial population; The fitness of individuals in the initial population is evaluated based on the current fitness function, and the initial population is iteratively evolved by selecting and improving chromosome crossover and mutation operations. The current fitness function is obtained based on the examination parameters of the laboratory order predicted by the first AI model. Under the condition of meeting the set genetic termination conditions, the optimal laboratory order scheduling scheme is obtained.

[0008] In this way, the scheduling strategy can be dynamically adjusted based on real-time data and a reinforcement learning model to adapt to the dynamic changes in the laboratory. By combining reinforcement learning and AI models and using the global search capability of genetic algorithms, the optimal scheduling solution can be found, thus achieving comprehensive optimization of laboratory order scheduling and improving scheduling efficiency and resource utilization.

[0009] In some embodiments, the current order placement strategy is generated based on a reinforcement learning model, including: The definition includes a state space containing current laboratory order information, current laboratory personnel information, and current laboratory equipment information, as well as an action space for assigning orders to designated laboratory personnel and designating laboratory equipment. Based on the reward function and the reinforcement learning model, the optimal order scheduling strategy that matches the state space and action space is gradually learned and determined as the current order scheduling strategy. The reward function is determined based on the evaluation factors of laboratory orders, which include one or more of the following: order completion status, experimenter load balancing, and experimental equipment utilization rate.

[0010] It is evident that by defining the state space and action space, the reinforcement learning model can learn the optimal order-queueing strategy, improving the adaptability and effectiveness of the strategy. Furthermore, the reward function comprehensively considers multiple factors, ensuring the comprehensiveness and balance of the order-queueing strategy.

[0011] In some embodiments, it also includes: Collect historical laboratory order information, historical laboratory personnel information, historical laboratory equipment information, and historical assessment parameters, and extract corresponding feature information. The feature information includes one or more of the following: order submission time, deadline, testing time, whether retesting is required, laboratory personnel workload and available time, and laboratory equipment usage information. The model is trained based on machine learning or deep learning algorithms. Based on feature information, a first AI model is trained to predict the examination parameters of laboratory orders. The examination parameters include one or more of the following: order detection time, retest probability, laboratory technician work efficiency, and laboratory equipment utilization rate.

[0012] Therefore, by collecting and analyzing historical data, AI models can predict key assessment parameters, providing more accurate input to the fitness function. Machine learning or deep learning algorithms improve the accuracy of assessment parameter prediction, thereby increasing the reliability of the scheduling scheme.

[0013] In some embodiments, evaluating the fitness of individuals in the initial population based on the current fitness function includes: According to formula (1), the fitness of each individual in the initial population is obtained; text{Fitness} = w_1 \times \text{order completion time} + w_2 \times \text{laboratory load balancing} + w_3 \times \text{laboratory equipment utilization} + w_4 \times \text{penalty for retest / delay} (1) Where (w_1, w_2, w_3, w_4) are weights, the order completion time is the sum of the order start time and the order detection time, the load balancing of the experimenters is obtained based on the work efficiency of the experimenters, and the penalty for retesting is determined based on the retesting probability. One or more of the following parameters—order testing time, retest probability, lab technician efficiency, and lab equipment utilization rate—are predicted based on the first AI model.

[0014] By integrating multiple evaluation parameters, the fitness function can comprehensively assess the merits of different scheduling schemes, ensuring the comprehensiveness and balance of the optimization objectives. Furthermore, the weights can be adjusted according to the actual situation, further improving the flexibility and adaptability of the fitness function.

[0015] In some embodiments, iterative evolution of the initial population includes: Based on fitness, select the corresponding individuals from the initial population for reproduction; Based on the current crossover rate and the current mutation rate, perform the corresponding iterative calculations; In this model, one or more genetic parameters, including the weight coefficients, current crossover rate, and current mutation rate, in the current fitness function are predicted by the second AI model. This second AI model is trained based on collected historical genetic parameters using a machine learning or deep learning model.

[0016] It is evident that predicting the parameters of the genetic algorithm through an AI model and dynamically adjusting the crossover and mutation rates improves the performance and convergence speed of the genetic algorithm. Furthermore, the dynamic adjustment of genetic parameters enables the genetic algorithm to better adapt to different problem sizes and complexities, thereby improving the stability and reliability of the queue optimization.

[0017] In some embodiments, satisfying the defined genetic termination conditions includes: If the current iteration number is equal to the set iteration number, determine that the set genetic termination condition is met; If the absolute difference between the current fitness of the current individual and the previous fitness in the current iteration population is less than a set threshold when the current fitness function evaluates, the recorded duration is incremented by 1. If the updated current duration reaches the set number, the set genetic termination condition is determined to be met.

[0018] Multiple termination conditions are provided to ensure that the genetic algorithm stops when it reaches a preset number of iterations or when the fitness converges, thus improving the efficiency and practicality of the algorithm.

[0019] In some embodiments, the device includes: The learning encoding module is configured to encode the current laboratory order information, current experimental personnel information, and current experimental equipment information into chromosomes based on the current ordering strategy generated by the reinforcement learning model, so as to obtain the corresponding initial population. The genetic iteration module is configured to evaluate the fitness of individuals in the initial population based on the current fitness function, and iteratively evolve the initial population by selecting and improving chromosome crossover and mutation operations, wherein the current fitness function is obtained based on the examination parameters of the laboratory order predicted by the first AI model; The termination module is configured to obtain the optimal laboratory order scheduling scheme when the set genetic termination conditions are met.

[0020] In some embodiments, the apparatus includes a processor and a memory storing program instructions, wherein the processor is configured to execute the above-described method for laboratory order scheduling when executing the program instructions.

[0021] In some embodiments, the device includes a device body; the aforementioned apparatus for laboratory order scheduling is installed on the device body.

[0022] In some embodiments, the storage medium stores program instructions that, when executed, perform the above-described method for laboratory order scheduling.

[0023] The method, apparatus, equipment, and storage medium for laboratory order scheduling provided in this disclosure can achieve the following technical effects: By using reinforcement learning models, the scheduling strategy can be dynamically adjusted to adapt to the dynamic changes in the laboratory, improving the flexibility of laboratory order scheduling. Furthermore, by using AI models to predict the examination parameters of laboratory orders and combining the global search capabilities of genetic algorithms, the optimal scheduling scheme can be found, achieving comprehensive optimization of laboratory order scheduling and improving scheduling efficiency and resource utilization.

[0024] The above general description and the description below are exemplary and illustrative only and are not intended to limit this application. Attached Figure Description

[0025] One or more embodiments are illustrated by way of example with reference to the accompanying drawings. These illustrations and drawings do not constitute a limitation on the embodiments. Elements having the same reference numerals in the drawings are shown as similar elements. The drawings are not to be scaled. And wherein: Figure 1 This is a schematic flowchart of a laboratory order scheduling method provided in an embodiment of this disclosure; Figure 2 This is a schematic diagram of a process for obtaining an AI model provided in an embodiment of this disclosure; Figure 3 This is a schematic flowchart of a laboratory order scheduling method provided in an embodiment of this disclosure; Figure 4 This is a schematic diagram of a laboratory order scheduling device provided in an embodiment of this disclosure; Figure 5 This is a schematic diagram of a laboratory order scheduling device provided in an embodiment of this disclosure; Figure 6 This is a schematic diagram of a laboratory order scheduling device provided in an embodiment of this disclosure; Figure 7 This is a schematic diagram of the structure of a device provided in an embodiment of this disclosure. Detailed Implementation

[0026] To provide a more detailed understanding of the features and technical content of the embodiments of this disclosure, the implementation of the embodiments of this disclosure will be described in detail below with reference to the accompanying drawings. The accompanying drawings are for illustrative purposes only and are not intended to limit the embodiments of this disclosure. In the following technical description, for ease of explanation, several details are used to provide a full understanding of the disclosed embodiments. However, one or more embodiments may still be implemented without these details. In other cases, well-known structures and devices may be simplified in their depiction to simplify the drawings.

[0027] The terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this disclosure are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate for the embodiments of this disclosure described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion.

[0028] Unless otherwise stated, the term "multiple" means two or more.

[0029] In this embodiment of the disclosure, the character " / " indicates that the objects before and after it are in an "or" relationship. For example, A / B means: A or B.

[0030] The term "and / or" describes an association between objects, indicating that three relationships can exist. For example, A and / or B means: A or B, or A and B.

[0031] Before products are developed or shipped, they may need to undergo laboratory testing. Therefore, laboratories may need to process a large number of testing orders daily, with different priorities, deadlines, and testing times. In this embodiment, the genetic algorithm-based scheduling system can optimize order scheduling, improving flexibility, complexity, and adaptability. Furthermore, using a genetic algorithm combined with an AI model can further optimize scheduling strategies and dynamically adjust the overall allocation of laboratory equipment.

[0032] Figure 1 This is a flowchart illustrating a laboratory order scheduling method provided in an embodiment of this disclosure, as shown below. Figure 1 As shown, the order scheduling process for this laboratory includes: Step 101: Based on the current order scheduling strategy generated by the reinforcement learning model, chromosome encoding is performed on the current laboratory order information, current experimental personnel information, and current experimental equipment information to obtain the corresponding initial population.

[0033] In this embodiment, the optimal laboratory order scheduling scheme is determined based on iterative evolution using a genetic algorithm. Therefore, it is necessary to encode certain order-related factors within the laboratory using chromosomes to obtain the corresponding initial population. These order-related factors may include: laboratory order information, laboratory personnel information, and laboratory equipment information, etc. Specifically, laboratory order information includes one or more of the following: order submission time, deadline, testing time, whether retesting is required, etc., related to product testing. Laboratory personnel information includes one or more of the following: workload of laboratory personnel, available time, etc. Laboratory equipment information includes one or more of the following: equipment usage status, equipment utilization rate, etc. At the current time of scheduling, the corresponding laboratory order information, laboratory personnel information, and laboratory equipment information can be defined as the current laboratory order information, current laboratory personnel information, and current laboratory equipment information.

[0034] In some embodiments, the current laboratory order information, current personnel information, and current equipment information can be randomly encoded using chromosomes to obtain the corresponding initial population. For example: the current order information includes: Order A: Submission time 8:00, deadline 12:00, testing time 2 hours, requires retesting; Order B: Submission time 9:00, deadline 13:00, testing time 1 hour, does not require retesting. The current personnel information includes: Lab Technician 1: Workload 50%, available time 8:00 - 12:00; Lab Technician 2: Workload 30%, available time 9:00 - 13:00. The current equipment information includes: Station 1: Usage 40%, available time 8:00 - 12:00; Station 2: Usage 60%, available time 9:00 - 13:00. Randomly encoding these order-related factors using chromosomes to obtain the corresponding initial population may include: Chromosome 1: A 1 1,B 2 2. Order A is assigned to lab technician 1 and workstation 1, and order B is assigned to lab technician 2 and workstation 2. Chromosome 2: A 2 2,B 1 1. Order A is assigned to lab technician 2 and workstation 2, and order B is assigned to lab technician 1 and workstation 1. Chromosome 3: A 1 2,B 2 1. Order A is assigned to lab technician 1 and workstation 2, and order B is assigned to lab technician 2 and workstation 1. Chromosome 4: A 2 1,B 1 2. That is, order A is assigned to lab technician 2 and station 1, and order B is assigned to lab technician 1 and station 2.

[0035] In some embodiments, orders can be allocated based on the learning results of the reinforcement learning model. Specifically, based on the current order allocation strategy generated by the reinforcement learning model, chromosome encoding is performed on the current laboratory order information, current laboratory personnel information, and current experimental equipment information to obtain the corresponding initial population. Therefore, generating the current order allocation strategy based on the reinforcement learning model may include: defining a state space including the current laboratory order information, current laboratory personnel information, and current experimental equipment information; and defining an action space for allocating orders to designated laboratory personnel and designated experimental equipment. Based on the reward function and the reinforcement learning model, the optimal order allocation strategy matching the state space and action space is gradually learned and determined as the current order allocation strategy. The reward function is determined based on factors considered in laboratory order evaluation, including one or more of the following: order completion status, laboratory personnel load balancing, and experimental equipment utilization rate.

[0036] The system allows for the definition of a state space, which includes information about laboratory orders, personnel, and equipment. This includes details such as order submission time, deadline, testing time, whether retesting is required, the workload and available time of the lab technicians, and the usage and utilization rate of the equipment. An action space can also be defined, which assigns orders to specific lab technicians and equipment. After defining the state and action spaces, a reinforcement learning model, such as Q-learning or DQN algorithms, can be trained to interact with the environment and learn the optimal order scheduling strategy.

[0037] Evaluation factors typically refer to various factors that need to be considered during strategy development or evaluation. Evaluation factors for laboratory orders may include one or more of the following: order completion status, lab technician workload balance, and lab equipment utilization rate. Therefore, a reward function is determined based on the evaluation factors for laboratory orders. For example, the reward for timely order completion is 1, the reward for increased lab technician workload balance is 0.6, and the reward for increased lab station utilization rate is 0.6, etc.

[0038] Therefore, based on the reward function and a reinforcement learning model, the optimal scheduling strategy that matches the state space and action space is gradually learned. For example, the factors considered include: order completion status, experimenter load balancing, and experimental equipment utilization. The corresponding weights are 0.4, 0.3, and 0.3. Then, a scheduling strategy that assigns order A to experimenter 1 and station 1, and order B to experimenter 2 and station 2, corresponds to an increase in experimenter 1's workload to 70% and an increase in station 1's utilization to 60%. Thus, the reward for timely completion of order A is 1, the reward for balanced experimenter load is 0.5, and the reward for station utilization is 0.6, for a total reward R = 0.4 × 1 + 0.3 × 0.5 + 0.3 × 0.6 = 0.73. The total reward for each scheduling strategy can then be obtained. Therefore, among all possible scheduling strategies, the strategy with the highest total reward is selected as the current scheduling strategy.

[0039] After generating the current order scheduling strategy based on the reinforcement learning model, chromosome encoding can be performed on the laboratory order information, experimental personnel information, and experimental equipment information according to the current order scheduling strategy to obtain the corresponding initial population.

[0040] For example, considering the order-related factors in the lab as described above, based on the current scheduling strategy, order A should be assigned to lab technician 1 and station 1 because the available time of lab technician 1 and station 1 best matches the detection time of order A. Order B should be assigned to lab technician 2 and station 2 because the available time of lab technician 2 and station 2 best matches the detection time of order B. Therefore, the initial population based on the reinforcement learning model could include: Chromosome 1: A 1 1,B 2 2. Order A is assigned to lab technician 1 and workstation 1, and order B is assigned to lab technician 2 and workstation 2. Chromosome 2: A 1 1,B 2 2. (0.5 hours earlier): Order A is assigned to lab technician 1 and station 1, and order B is assigned to lab technician 2 and station 2, but the testing time for order B is advanced by 0.5 hours. Chromosome 3: A-1-1, B-2-2 (Laborer 2's workload +10%), order A is assigned to lab technician 1 and station 1, and order B is assigned to lab technician 2 and station 2, but lab technician 2's workload increases by 10%. Chromosome 4: A-1-1, B-2-2 (Station 2 utilization +10%), order A is assigned to lab technician 1 and station 1, and order B is assigned to lab technician 2 and station 2, but station 2's utilization increases by 10%.

[0041] Step 102: Evaluate the fitness of individuals in the initial population based on the current fitness function, and iteratively evolve the initial population by selecting and improving chromosome crossover and mutation operations. The current fitness function is obtained based on the examination parameters of the laboratory order predicted by the first AI model.

[0042] In this embodiment of the disclosure, the first AI model can predict the evaluation parameters of laboratory orders. The evaluation parameters include specific values ​​predicted or calculated by the model. Therefore, the evaluation parameters of laboratory orders may include one or more of the following: order testing time, retest probability, and lab technician efficiency. It is evident that in some embodiments, some evaluation parameters correspond to evaluation factors, which may be the numerical values ​​corresponding to those factors.

[0043] Therefore, in some embodiments of this disclosure, it is also necessary to train a first AI model, which may include: collecting historical laboratory order information, historical laboratory personnel information, historical laboratory equipment information, and historical assessment parameters, and extracting corresponding feature information, wherein the feature information includes one or more of the following: order submission time, deadline, testing time, whether retesting is required, laboratory personnel's workload and available time, and laboratory equipment usage information; training a model based on machine learning algorithms or deep learning algorithms, and training a first AI model to predict the assessment parameters of laboratory orders based on the feature information, wherein the assessment parameters include one or more of the following: order testing time, retesting probability, laboratory personnel's work efficiency, and laboratory equipment utilization rate.

[0044] Machine learning or deep learning algorithms can be used to train the model. Machine learning algorithms can include random forests, support vector machines, etc., while deep learning algorithms can include neural networks, etc. Based on historical laboratory order information, historical personnel information, historical equipment information, and historical environmental feedback information (including historical assessment parameters), the first AI model trained using machine learning or deep learning algorithms can predict one or more assessment parameters for an order, such as the detection time, retest probability, operator efficiency, and equipment utilization rate. For example, based on current laboratory order information, current personnel information, and current equipment information related to chromosome coding, the first AI model predicts that order A will have a detection time of 2.1 hours and a retest probability of 0.3. Order B will have a detection time of 1.05 hours and a retest probability of 0.1.

[0045] After the examination parameters of the laboratory orders are predicted by the first AI model, the fitness of individuals in the initial population can be evaluated based on the current fitness function. In some embodiments, evaluating the fitness of individuals in the initial population based on the current fitness function includes: obtaining the fitness of each individual in the initial population according to formula (1); text{Fitness} = w_1 \times \text{order completion time} + w_2 \times \text{laboratory load balancing} + w_3 \times \text{laboratory equipment utilization} + w_4 \times \text{penalty for retest / delay} (1) Where (w_1, w_2, w_3, w_4) are weights, the order completion time is the sum of the order start time and the order detection time, the load balance of the experimenters is obtained based on the work efficiency of the experimenters, the penalty for retesting is determined based on the retesting probability, and one or more of the following are predicted based on the first AI model: order detection time, retesting probability, experimenter work efficiency, and experimental equipment utilization rate.

[0046] For example: w_1=0.4, w_2=0.3, w_3=0.2, w_4=0.1, the order completion time in formula (1) can be the sum of the order start time and the order detection time predicted by the first AI model; after the first AI model predicts the retest probability, the corresponding retest penalty can be obtained. Of course, the work efficiency of the experimenters and the utilization rate of the experimental equipment can also be predicted by the AI ​​model, or obtained through other means. Among them, the closer the order completion time is to the deadline, the higher the corresponding fitness. The more balanced the workload of the experimenters, the higher the corresponding fitness. The higher the utilization rate of the laboratory station, the higher the corresponding fitness. The processing time of retest orders and delayed orders increases, and the corresponding fitness decreases.

[0047] If, after using the first AI model, the predicted detection time for order A is 2.1 hours, and the retest probability is 0.3, then the completion time for order A is 10:10 (8:00 + 2.1 hours), close to the deadline of 12:00, corresponding to a relatively high fitness of 0.9. Experimenter 1's workload is 70%, and the fitness corresponding to balanced workload for experimenters is 0.5. The utilization rate of station 1 is 60%, and the fitness of station utilization is 0.6. With a retest probability of 0.3 for order A, the corresponding fitness is 0.3.

[0048] Therefore, the fitness of this individual chromosome is text{Fitness} = 0.4×0.9+0.3×0.5+0.2×0.6+0.1×0.3=0.68.

[0049] The fitness of each individual in the population is determined, and the initial population is iteratively evolved based on the fitness. The iterative evolution of the genetic algorithm includes operations such as selection, crossover, and mutation. Specifically, iterative evolution of the initial population may include: selecting corresponding individuals from the initial population for reproduction based on fitness; and performing corresponding iterative calculations based on the current crossover rate and mutation rate.

[0050] The selection operation can include: a roulette wheel selection strategy, which selects individuals with higher fitness for breeding; or a tournament selection strategy, which randomly selects several individuals for comparison and selects the individual with higher fitness, for example, comparing every three individuals and selecting the individual with the highest fitness. The crossover operation can include: partial mapping crossover, where partial genes are selected from two parent individuals and exchanged to generate offspring individuals, for example: exchanging chromosome A... 1 1,B 2 2 and A 2 2,B 1 1. Crossover occurs, generating a new chromosome A. 1 1,B 1 1. Alternatively, sequential crossover can be used, maintaining the order of the orders while cross-assigning experimental personnel and equipment. Mutation operations can include: random mutation, i.e., randomly changing the allocation of experimental personnel or workstations for certain orders, for example: changing chromosome A... 1 1,B 1 In order B of 1, the experimenter changes from 1 to 2, generating a new chromosome A. 1 1,B 2 1. Alternatively, local search mutation, which involves adjusting the allocation of orders within a local scope to optimize fitness.

[0051] During each genetic iteration, the corresponding current crossover rate and mutation rate can be fixed or dynamic. Therefore, in some embodiments, one or more genetic parameters, including the weight coefficients, current crossover rate, and current mutation rate in the current fitness function, are predicted by a second AI model. This second AI model is trained based on collected historical genetic parameters using a machine learning or deep learning model.

[0052] A first AI model for predicting laboratory order assessment parameters can be obtained by training a machine learning or deep learning model on historical laboratory order information, historical laboratory personnel information, historical laboratory equipment information, and historical assessment parameters. Furthermore, in some embodiments of this disclosure, a second AI model for predicting genetic parameters can be obtained by training a machine learning or deep learning model on collected historical genetic parameters. These genetic parameters include one or more of the following in the fitness function: weight coefficients, crossover rate, and mutation rate.

[0053] For example, if the preset crossover rate is 0.7 and the mutation rate is 0.2, during the genetic iteration process, based on the current laboratory order information, current experimental personnel information, current experimental equipment information, and current environmental feedback information, the genetic parameter AI model predicts that when the crossover rate is 0.8, the convergence speed is faster; and when the mutation rate is 0.3, the quality of the optimal solution is higher. Therefore, the corresponding current crossover rate is 0.8 and the current mutation rate is 0.3.

[0054] Of course, in some embodiments, the weight coefficients in the fitness function can also be predicted by the second AI model. Therefore, during the genetic iteration process, the fitness of individuals in the current population can be determined based on the weight coefficients in the current fitness function predicted by the second AI model. Then, selection, crossover, mutation, and other operations are performed based on the fitness. Since there may be resource changes, such as machine failures or the availability of emergency orders for researchers, the current crossover rate and mutation rate in crossover, mutation, and other operations may also be predicted by the genetic parameter AI model during the genetic iteration process and will change according to these circumstances. This achieves dynamic adjustment of genetic parameters, enabling the genetic algorithm to better adapt to different problem sizes and complexities, and improving the stability and reliability of the order optimization.

[0055] Step 103: Under the condition of meeting the set genetic termination conditions, obtain the optimal laboratory order scheduling plan.

[0056] In some embodiments, satisfying the set genetic termination condition includes: determining that the set genetic termination condition is satisfied when the current iteration number is equal to the set iteration number; incrementing the recorded duration number by 1 when the absolute difference between the current fitness of the current individual in the current iteration population and the previous fitness is less than a set threshold when the current fitness function evaluates the current individual; and determining that the set genetic termination condition is satisfied when the updated current duration number reaches the set number.

[0057] For example, if the number of iterations is set to 100, then if the current iteration reaches 100, the set genetic termination condition can be determined to be met. Alternatively, if the number of iterations is set to 10, and the change in fitness of the current individual in the iterative population is less than the set threshold of 0.03 for 10 consecutive iterations (where the change is the absolute difference between the current fitness and the previous fitness), then the set genetic termination condition can also be determined to be met.

[0058] Once the set genetic termination conditions are met, the individual with the highest fitness in the current population can be output as the optimal laboratory order testing scheme.

[0059] As can be seen, in this embodiment, the reinforcement learning model can dynamically adjust the scheduling strategy to adapt to the dynamic changes in the laboratory, improving the flexibility of laboratory order scheduling. Furthermore, by predicting the examination parameters of laboratory orders through an AI model and combining this with the global search capability of the genetic algorithm, the optimal scheduling scheme is found, achieving comprehensive optimization of laboratory order scheduling and improving scheduling efficiency and resource utilization. Moreover, the AI ​​model can predict the parameters of the genetic algorithm, dynamically adjusting the crossover and mutation rates, improving the performance and convergence speed of the genetic algorithm. The dynamic adjustment of genetic parameters also allows the genetic algorithm to better adapt to different problem sizes and complexities, further improving the stability and reliability of scheduling optimization.

[0060] The following describes the operational process in a specific embodiment, illustrating the laboratory order scheduling process provided by the embodiments of the present invention.

[0061] In this embodiment, the first AI model can predict the examination parameters of laboratory orders, including one or more of the following: order testing time, retest probability, lab technician efficiency, and lab equipment utilization rate. Furthermore, a second AI model can predict genetic parameters, which may include one or more of the following: weight coefficients, crossover rate, and mutation rate in the fitness function. Therefore, the corresponding first AI model and second AI model must be trained first.

[0062] Figure 2 This disclosure provides a schematic flowchart for obtaining an AI model. For example... Figure 2 As shown, the process of obtaining the AI ​​model includes: Step 201: Collect historical laboratory order information, historical laboratory personnel information, historical laboratory equipment information, and historical assessment parameters, and extract the corresponding feature information. The feature information includes one or more of the following: order submission time, deadline, testing time, whether retesting is required, laboratory personnel workload and available time, and laboratory equipment usage information.

[0063] Historical environmental feedback information includes the examination parameters of historical laboratory orders, i.e., historical examination parameters, such as: one or more of the following: order testing time, retest probability, laboratory technician work efficiency, and laboratory equipment utilization rate.

[0064] Step 202: Train a model based on machine learning algorithms. Based on feature information, train and obtain a first AI model to predict the examination parameters of laboratory orders. The examination parameters include one or more of the following: order detection time, retest probability, laboratory technician work efficiency, and laboratory equipment utilization rate.

[0065] Step 203: Collect historical genetic parameters from population iterations. Genetic parameters may include one or more of the following: weight coefficients in the fitness function, crossover rate, and mutation rate.

[0066] Step 204: Based on historical laboratory order information, historical experimental personnel information, historical experimental equipment information, and historical genetic parameters, train and obtain a second AI model to predict genetic parameters using a neural network.

[0067] Of course, the order in which the first and second AI models are obtained is not limited. In some embodiments, the second AI model can be trained first, followed by the first AI model. Alternatively, in some embodiments, only the first AI model is trained. In this case, the weight coefficients, crossover rate, and mutation rate in the comfort function can be preset. Specific examples are not listed here.

[0068] In this way, by collecting and analyzing historical data, the AI ​​model can predict key assessment parameters, providing more accurate input to the fitness function. The machine learning algorithm improves the accuracy of assessment parameter prediction, thereby enhancing the reliability of the scheduling scheme. Furthermore, the AI ​​model can also predict the parameters of the genetic algorithm, enabling it to better adapt to different problem sizes and complexities, further improving the stability and reliability of scheduling optimization.

[0069] After obtaining the first AI model for predicting laboratory orders and the second AI model for predicting genetic parameters, laboratory order scheduling can begin. Figure 3 This is a schematic flowchart illustrating a laboratory order scheduling method provided in an embodiment of this disclosure. Figure 3 As shown, the laboratory order scheduling process includes: Step 301: The device acquires the current laboratory order information, current experimental personnel information, and current experimental equipment information.

[0070] Step 302: After defining the state space including the current laboratory order information, the current experimental personnel information, and the current experimental equipment information, and defining the action space for assigning orders to designated experimental personnel and designated experimental equipment, the equipment gradually learns the optimal order scheduling strategy that matches the state space and action space based on the reward function and the reinforcement learning model, and determines it as the current order scheduling strategy.

[0071] The reward function is determined based on factors considered in laboratory orders, including one or more of the following: order completion status, lab technician load balancing, and lab equipment utilization. For example, on-time order completion awards 1, increased lab technician load balancing awards 0.6, and increased lab station utilization awards 0.6, etc. Thus, based on the reward function, by training a model using Q-learning or DQN algorithms and gradually learning, an optimal order scheduling strategy that matches the state and action spaces can be obtained.

[0072] Step 303: Based on the current order scheduling strategy, the device performs chromosome encoding on the current laboratory order information, current experimental personnel information, and current experimental equipment information to obtain the corresponding initial population, which is then used as the current population.

[0073] Step 304: The device uses the first AI model to predict the current examination parameters corresponding to the current population; and the device uses the second AI model to predict the current genetic parameters corresponding to the current population.

[0074] The parameters examined include one or more of the following: order testing time, retest probability, lab technician efficiency, and experimental equipment utilization rate. Genetic parameters include one or more of the following: weighting coefficients in the fitness function, crossover rate, and mutation rate.

[0075] Step 305: The device obtains the fitness of each chromosome individual in the current population according to formula (1), and performs selection, crossover and mutation operations based on the fitness to iteratively evolve the current population.

[0076] text{Fitness} = w_1 \times \text{order completion time} + w_2 \times \text{laboratory load balancing} + w_3 \times \text{laboratory equipment utilization} + w_4 \times \text{penalty for retest / delay} (1) Among them, (w_1, w_2, w_3, w_4) may be predicted by the second AI model. The order completion time is the sum of the order start time and the order detection time. The load balancing of the experimenters is obtained based on the work efficiency of the experimenters. The penalty for retesting is determined based on the retesting probability. One or more of the following are based on the order detection time, retesting probability, experimenter work efficiency, and experimental equipment utilization rate: prediction by the first AI model. The current crossover rate and mutation rate for crossover and mutation operations may also be predicted by the second AI model.

[0077] Among them, the selection operation can be a roulette wheel operation or a tournament operation, the crossover operation can be a partial mapping crossover or a sequential crossover, and the mutation operation can be a random mutation or a local search mutation. The specific processes will not be described in detail.

[0078] Step 306: Determine if the current iteration count is equal to 100. If yes, proceed to step 307; otherwise, proceed to step 308.

[0079] Of course, in some embodiments, the determination can also be made based on whether the fitness converges, or based on a combination of the number of iterations and fitness convergence. These will not be elaborated on in detail.

[0080] Step 307: The device determines and outputs the optimal laboratory order scheduling plan based on the individual with the highest fitness in the current population.

[0081] Step 308: The device iterates to obtain the population, identifies it as the current population, and returns to step 304.

[0082] As can be seen, in this embodiment, the reinforcement learning model can dynamically adjust the scheduling strategy to adapt to the dynamic changes in the laboratory, improving the flexibility of laboratory order scheduling. Furthermore, by using an AI model to predict the examination parameters of laboratory orders and combining this with the global search capability of the genetic algorithm, the optimal scheduling scheme is found, achieving comprehensive optimization of laboratory order scheduling and improving scheduling efficiency and resource utilization. Moreover, the AI ​​model can predict the parameters of the genetic algorithm and dynamically adjust the crossover and mutation rates, improving the performance and convergence speed of the genetic algorithm. The dynamic adjustment of genetic parameters allows the genetic algorithm to better adapt to different problem sizes and complexities, further improving the stability and reliability of scheduling optimization.

[0083] Based on the above-mentioned laboratory order scheduling process, a device for laboratory order scheduling can be constructed.

[0084] Figure 4 This is a schematic diagram of a laboratory order scheduling device provided in an embodiment of this disclosure. Figure 4As shown, the laboratory order scheduling device 400 includes: a learning encoding module 410, a genetic iteration module 420, and a termination acquisition module 430.

[0085] The learning encoding module 410 is configured to encode the current laboratory order information, current experimental personnel information, and current experimental equipment information into chromosomes based on the current ordering strategy generated by the reinforcement learning model, so as to obtain the corresponding initial population.

[0086] The genetic iteration module 420 is configured to evaluate the fitness of individuals in the initial population based on the current fitness function, and iteratively evolve the initial population by selecting and improving chromosome crossover and mutation operations, wherein the current fitness function is obtained based on the examination parameters of the laboratory order predicted by the first AI model.

[0087] The termination module 430 is configured to obtain the optimal laboratory order scheduling scheme when the set genetic termination conditions are met.

[0088] In some embodiments, the learning encoding module 410 includes: The definition unit is configured to define a state space that includes current laboratory order information, current laboratory personnel information, and current laboratory equipment information, as well as an action space that assigns the order to the designated laboratory personnel and the designated laboratory equipment.

[0089] The learning unit is configured to learn, based on a reinforcement learning model and a reward function, the optimal order scheduling strategy that matches the state space and action space, and determine it as the current order scheduling strategy. The reward function is determined based on one or more of the following factors for evaluating lab orders: order completion status, lab technician load balancing, and lab equipment utilization. In some embodiments, it also includes: The training and generation module is configured to collect historical laboratory order information, historical laboratory personnel information, historical laboratory equipment information, and historical assessment parameters, and extract corresponding feature information. This feature information includes one or more of the following: order submission time, deadline, testing time, whether a retest is required, laboratory personnel's workload and available time, and laboratory equipment usage information. Based on machine learning or deep learning algorithms, a model is trained to predict the assessment parameters of laboratory orders. These assessment parameters include one or more of the following: order testing time, retest probability, laboratory personnel's work efficiency, and laboratory equipment utilization rate.

[0090] In some embodiments, the genetic iteration module 420 includes: The evaluation unit is configured to obtain the fitness of each individual in the initial population according to formula (1); text{Fitness} = w_1 \times \text{order completion time} + w_2 \times \text{laboratory load balancing} + w_3 \times \text{laboratory equipment utilization} + w_4 \times \text{penalty for retest / delay} (1) Where (w_1, w_2, w_3, w_4) are weights, the order completion time is the sum of the order start time and the order detection time, the load balance of the experimenters is obtained based on the work efficiency of the experimenters, and the penalty for retesting is determined based on the retesting probability; one or more of the order detection time, retesting probability, experimenter work efficiency, and experimental equipment utilization rate are predicted based on the first AI model.

[0091] In some embodiments, the genetic iteration module 420 includes: The iterative unit is configured to select corresponding individuals from the initial population for reproduction based on fitness; and to perform corresponding iterative calculations based on the current crossover rate and the current mutation rate. One or more genetic parameters in the current fitness function, including the weight coefficients, current crossover rate, and current mutation rate, are predicted by a second AI model. This second AI model is trained based on collected historical genetic parameters using a machine learning or deep learning model.

[0092] In some embodiments, the termination module 430 includes: The termination unit is configured to determine that the set genetic termination condition is met when the current iteration number is equal to the set iteration number; when the absolute difference between the current fitness of the current individual and the previous fitness in the current iteration population is less than the set threshold, the recorded duration number is incremented by 1; if the updated current duration number reaches the set number, the set genetic termination condition is determined to be met.

[0093] The control process for the laboratory order scheduling device provided in the embodiments of the present invention will be illustrated below with reference to specific examples.

[0094] Figure 5 This is a schematic diagram of a laboratory order scheduling device provided in an embodiment of this disclosure. Figure 5As shown, the laboratory order scheduling device 400 includes: a learning encoding module 410, a genetic iteration module 420, a termination module 430, and a training generation module 440. The learning encoding module 410 includes a definition unit 411 and a learning unit 412. The genetic iteration module 420 includes an evaluation unit 421 and an iteration unit 422. The termination module 430 includes a termination unit 431.

[0095] The training and generation module 440 collects historical laboratory order information, historical laboratory personnel information, historical laboratory equipment information, and historical assessment parameters, and extracts corresponding feature information. This feature information includes one or more of the following: order submission time, deadline, testing time, whether retesting is required, laboratory personnel workload and available time, and laboratory equipment usage information. Based on this feature information, a neural network training model is trained to obtain a first AI model predicting the assessment parameters of laboratory orders. These assessment parameters include one or more of the following: order testing time, retest probability, laboratory personnel efficiency, and laboratory equipment utilization rate. Furthermore, the training and generation module 440 can also collect historical genetic parameters from population iterations. These genetic parameters may include one or more of the following: weight coefficients, crossover rate, and mutation rate in the fitness function. Based on the historical laboratory order information, historical laboratory personnel information, historical laboratory equipment information, and historical genetic parameters, a second AI model predicting genetic parameters is trained using a neural network.

[0096] After the learning encoding module 410 acquires the current laboratory order information, current laboratory personnel information, and current experimental equipment information, the definition unit 411 can define a state space including the current laboratory order information, current laboratory personnel information, and current experimental equipment information, and an action space for allocating orders to designated laboratory personnel and designated experimental equipment. Then, the learning unit 412, based on a reward function and a reinforcement learning model, gradually learns the optimal order scheduling strategy that matches the state space and action space, and determines it as the current order scheduling strategy. The reward function is determined based on factors considered in laboratory orders, including one or more of the following: order completion status, laboratory personnel load balancing, and experimental equipment utilization rate. Therefore, the learning encoding module 410 can perform chromosome encoding on the current laboratory order information, current laboratory personnel information, and current experimental equipment information according to the current order scheduling strategy to obtain the corresponding initial population, and then perform corresponding genetic iterations. The population corresponding to each iteration is the current population.

[0097] In each iteration, the first AI model predicts the current parameters for the current population, and the second AI model predicts the current genetic parameters for the current population. For example, the first AI model predicts the detection time and retest probability for an individual's order in the current population, while the second AI model predicts the current crossover rate and mutation rate of the current population. Therefore, the weights in the fitness function can be set to specific values.

[0098] Thus, the evaluation unit 421 in the genetic iteration module 420 obtains the fitness of each chromosome individual in the current population according to formula (1), where formula (1) can be applied to the detection time and retest probability of the order predicted by the first AI model. Then, the iteration unit 422 performs selection, crossover and mutation operations based on the fitness to iteratively evolve the current population.

[0099] The current crossover rate and mutation rate corresponding to the crossover and mutation operations can be predicted by the second AI model. During the genetic iteration process, if the absolute difference between the current fitness of the current individual and the previous fitness in the current iteration population is less than a set threshold, the recorded duration is incremented by 1. If the updated current duration reaches the set number, the termination unit 431 in the termination module 430 determines that the set genetic termination condition is met. Furthermore, the termination module 430 can determine and output the optimal laboratory order scheduling plan based on the individual with the highest fitness in the current population. Of course, if the set genetic termination condition is not met, the iteratively evolving population can be determined as the current population, and genetic iteration can continue.

[0100] As can be seen, in this embodiment, the device for laboratory order scheduling can dynamically adjust the scheduling strategy through a reinforcement learning model to adapt to the dynamic changes in the laboratory, thereby improving the flexibility of laboratory order scheduling. Furthermore, by predicting the examination parameters of laboratory orders through an AI model and combining this with the global search capability of a genetic algorithm, the optimal scheduling scheme is found, achieving comprehensive optimization of laboratory order scheduling and improving scheduling efficiency and resource utilization. Moreover, the AI ​​model can predict the parameters of the genetic algorithm and dynamically adjust the crossover and mutation rates, improving the performance and convergence speed of the genetic algorithm. The dynamic adjustment of genetic parameters allows the genetic algorithm to better adapt to different problem sizes and complexities, further improving the stability and reliability of scheduling optimization.

[0101] This disclosure provides an embodiment of a device 600 for laboratory order scheduling, the structure of which is as follows: Figure 6 As shown, it includes: The processor 1000 and memory 1001 may further include a communication interface 1002 and a bus 1003. The processor 1000, communication interface 1002, and memory 1001 can communicate with each other via the bus 1003. The communication interface 1002 can be used for information transmission. The processor 1000 can call logical instructions stored in the memory 1001 to execute the method for laboratory order scheduling described in the above embodiment.

[0102] Furthermore, the logic instructions in the aforementioned memory 1001 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium.

[0103] The memory 1001, as a computer-readable storage medium, can be used to store software programs and computer-executable programs, such as program instructions / modules corresponding to the methods in the embodiments of this disclosure. The processor 1000 executes functional applications and data processing by running the program instructions / modules stored in the memory 1001, that is, it implements the method for laboratory order scheduling in the above method embodiments.

[0104] The memory 1001 may include a program storage area and a data storage area. The program storage area may store the operating system and applications required for at least one function; the data storage area may store data created based on the use of the terminal device. Furthermore, the memory 1001 may include high-speed random access memory and may also include non-volatile memory.

[0105] This disclosure provides an apparatus for laboratory order scheduling, comprising: a processor and a memory storing program instructions, wherein the processor is configured to execute a laboratory order scheduling method when executing the program instructions.

[0106] This disclosure provides an embodiment of a device, such as a terminal device or a server device. Figure 7 As shown, the device includes: a device body 700, and the aforementioned device 400 (600) for laboratory order scheduling, which is installed in the device body 700. The installation relationship described herein is not limited to placement within the product, but also includes installation connections with other components of the product, including but not limited to physical connections, electrical connections, or signal transmission connections. Those skilled in the art will understand that the device 400 (600) for laboratory order scheduling can be adapted to suitable corresponding device bodies to achieve other feasible embodiments.

[0107] This disclosure provides a storage medium storing program instructions that, when executed, perform the method described above for laboratory order scheduling.

[0108] This disclosure provides a computer program product, which includes a computer program stored on a storage medium. The computer program includes program instructions that, when executed by a computer, cause the computer to perform the above-described laboratory order scheduling method.

[0109] The aforementioned storage medium can be a transient computer-readable storage medium or a non-transitory computer-readable storage medium.

[0110] The technical solutions of this disclosure can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes one or more instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in this disclosure. The aforementioned storage medium can be a non-transitory storage medium, including: a USB flash drive, a portable hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk, and other media capable of storing program code; it can also be a transient storage medium.

[0111] The foregoing description and accompanying drawings fully illustrate embodiments of the present disclosure to enable those skilled in the art to practice them. Other embodiments may include structural, logical, electrical, procedural, and other changes. The embodiments represent only possible variations. Individual components and functions are optional unless explicitly required, and the order of operation may vary. Parts and features of some embodiments may be included or replace parts and features of other embodiments. The scope of the embodiments of this disclosure includes the entire scope of the claims and all available equivalents of the claims. While the terms “first,” “second,” etc., may be used in this application to describe elements, these elements should not be limited by these terms. These terms are used only to distinguish one element from another. For example, a first element may be called a second element without changing the meaning of the description, and similarly, a second element may be called a first element, provided that all occurrences of “first element” are consistently renamed and all occurrences of “second element” are consistently renamed. First and second elements are both elements, but may not be the same element. Moreover, the terminology used in this application is only for describing embodiments and is not intended to limit the claims. As used in the description of the embodiments and claims, unless the context clearly indicates otherwise, the singular forms “a,” “an,” and “the” are intended to also include the plural forms. Similarly, the term “and / or” as used herein means including one or more of the associated listed any and all possible combinations. Additionally, when used herein, the terms “comprise” and its variations “comprises” and / or “comprising” refer to the presence of stated features, integrals, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components, and / or groups thereof. Without further limitations, an element defined by the phrase “comprising an…” does not exclude the presence of additional identical elements in the process, method, or apparatus that includes said element. In this document, each embodiment may focus on the differences from other embodiments, and similar or identical parts between embodiments can be referred to mutually. For methods, products, etc., disclosed in the embodiments, if they correspond to the method section disclosed in the embodiments, the relevant parts can be referred to the description of the method section.

[0112] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the embodiments of this disclosure. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0113] The methods and products disclosed in the embodiments herein (including but not limited to devices and equipment) can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For instance, the division of units may be merely a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed may be through some interfaces, and the indirect coupling or communication connection of devices or units may be electrical, mechanical, or other forms. The units described as separate components may or may not be physically separate. The 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 the units can be selected to implement this embodiment according to actual needs. In addition, the functional units in the embodiments of this disclosure may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0114] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. In some alternative implementations, the functions marked in the blocks may occur in a different order than that shown in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. In the descriptions corresponding to the flowcharts and block diagrams in the accompanying drawings, the operations or steps corresponding to different blocks may also occur in a different order than disclosed in the description, and sometimes there is no specific order between different operations or steps. For example, two consecutive operations or steps may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. Each block in a block diagram and / or flowchart, and combinations of blocks in a block diagram and / or flowchart, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

Claims

1. A method for laboratory order scheduling, characterized in that, include: Based on the current order scheduling strategy generated by the reinforcement learning model, chromosome encoding is performed on the current laboratory order information, current experimental personnel information, and current experimental equipment information to obtain the corresponding initial population; The fitness of individuals in the initial population is evaluated based on the current fitness function, and the initial population is iteratively evolved by selecting and improving chromosome crossover and mutation operations. The current fitness function is obtained based on the examination parameters of the laboratory order predicted by the first AI model. Under the condition of meeting the set genetic termination conditions, the optimal laboratory order scheduling scheme is obtained.

2. The method according to claim 1, characterized in that, The current order placement strategy is generated based on a reinforcement learning model, including: The definition includes a state space containing current laboratory order information, current laboratory personnel information, and current laboratory equipment information, as well as an action space for assigning orders to designated laboratory personnel and designating laboratory equipment. Based on the reward function and the reinforcement learning model, the optimal order scheduling strategy that matches the state space and action space is gradually learned and determined as the current order scheduling strategy. The reward function is determined based on the evaluation factors of laboratory orders, which include one or more of the following: order completion status, experimenter load balancing, and experimental equipment utilization rate.

3. The method according to claim 1, characterized in that, Also includes: Collect historical laboratory order information, historical laboratory personnel information, historical laboratory equipment information, and historical assessment parameters, and extract corresponding feature information. The feature information includes one or more of the following: order submission time, deadline, testing time, whether retesting is required, laboratory personnel workload and available time, and laboratory equipment usage information. The model is trained based on machine learning or deep learning algorithms. Based on feature information, a first AI model is trained to predict the examination parameters of laboratory orders. The examination parameters include one or more of the following: order detection time, retest probability, laboratory technician work efficiency, and laboratory equipment utilization rate.

4. The method according to claim 3, characterized in that, The fitness of individuals in the initial population is evaluated based on the current fitness function, including: According to formula (1), the fitness of each individual in the initial population is obtained; text{Fitness} = w_1 \times \text{order completion time} + w_2 \times \text{laboratory load balancing} + w_3 \times \text{laboratory equipment utilization} + w_4 \times \text{penalty for retest / delay} (1) Where (w_1, w_2, w_3, w_4) are weights, the order completion time is the sum of the order start time and the order detection time, the load balancing of the experimenters is obtained based on the work efficiency of the experimenters, and the penalty for retesting is determined based on the retesting probability. One or more of the following parameters—order testing time, retest probability, lab technician efficiency, and lab equipment utilization rate—are predicted based on the first AI model.

5. The method according to claim 1, characterized in that, The initial population for iterative evolution includes: Based on fitness, select the corresponding individuals from the initial population for reproduction; Based on the current crossover rate and the current mutation rate, perform the corresponding iterative calculations; In this model, one or more genetic parameters, including the weight coefficients, current crossover rate, and current mutation rate, in the current fitness function are predicted by the second AI model. This second AI model is trained based on collected historical genetic parameters using a machine learning or deep learning model.

6. The method according to any one of claims 1-5, characterized in that, The genetic termination conditions are met, including: If the current iteration number is equal to the set iteration number, determine that the set genetic termination condition is met; If the absolute difference between the current fitness of the current individual and the previous fitness in the current iteration population is less than a set threshold when the current fitness function evaluates, the recorded duration is incremented by 1. If the updated current duration reaches the set number, the set genetic termination condition is determined to be met.

7. A device for laboratory order scheduling, characterized in that, include: The learning encoding module is configured to encode the current laboratory order information, current experimental personnel information, and current experimental equipment information into chromosomes based on the current ordering strategy generated by the reinforcement learning model, so as to obtain the corresponding initial population. The genetic iteration module is configured to evaluate the fitness of individuals in the initial population based on the current fitness function, and iteratively evolve the initial population by selecting and improving chromosome crossover and mutation operations, wherein the current fitness function is obtained based on the examination parameters of the laboratory order predicted by the first AI model; The termination module is configured to obtain the optimal laboratory order scheduling scheme when the set genetic termination conditions are met.

8. An apparatus for laboratory order scheduling, the apparatus comprising a processor and a memory storing program instructions, characterized in that, The processor is configured to perform the method for laboratory order scheduling as described in any one of claims 1-6 when executing the program instructions.

9. A device, characterized in that, include: Equipment body; The apparatus for laboratory order scheduling as described in claim 7 or 8 is installed on the main body of the device.

10. A storage medium storing program instructions, characterized in that, When the program instructions are executed, they perform the method for laboratory order scheduling as described in any one of claims 1-6.

Images