Chromatographic injection control method and system based on dynamic programming

Abstract

The application provides a chromatographic sampling control method and system based on dynamic programming, which comprises the following steps: collecting sample properties and constructing a characteristic data set of the sample; designing a comprehensive optimization objective function according to a low-dimensional feature representation set and the most suitable device operating parameters of the current partition for intra-partition dynamic programming; performing global resource scheduling according to the optimal detection sequence of each partition combined with the most suitable device operating parameters of the current partition to generate a global resource plan; collecting real-time operating states and environmental parameters of the device, calculating device state deviation and environmental state deviation; performing sample detection in batches according to the optimized global resource plan, collecting feedback data of the device, and analyzing and optimizing performance evaluation results of all batches. The application provides a comprehensive and efficient solution for chromatographic sampling optimization. This method not only has a significant advantage in large-scale sample detection tasks, but also can adapt to diverse device and environmental requirements.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of chromatography sample control, and particularly relates to a chromatography sample control method and system based on dynamic programming. BACKGROUND

[0002] In the field of modern chromatography analysis, the sample injection order directly affects the detection efficiency and the accuracy of the results. Traditional chromatography analysis methods mostly rely on manual experience or simple rules, such as injecting samples according to the order or category of sample submission. However, with the improvement of the performance of chromatography analysis equipment and the increase in the complexity of samples, these traditional methods have exposed a series of limitations.

[0003] Firstly, the chemical and physical properties of samples are different, and different samples have different requirements for equipment parameters (such as column temperature, flow rate, etc.). This diversity makes it impossible for simple sorting methods to meet the optimization requirements. For example, continuous injection of samples with high retention time and samples with low retention time may lead to a decrease in equipment performance or cross-contamination. Secondly, there are certain resource limitations (such as maximum load, cleaning time) in the operation of the equipment, and the traditional sorting methods do not effectively consider these constraints, resulting in low resource utilization and even exacerbating equipment wear and tear. In addition, with the increase in analysis requirements, the detection task of large-scale samples is becoming more and more common, which further increases the sorting complexity. Although static optimization methods can improve efficiency to some extent, they usually rely on fixed initial conditions and are difficult to dynamically respond to real-time changes in the environment (such as temperature and pressure fluctuations) or equipment status (such as operation failure or overload).

[0004] Existing optimization techniques, such as sorting methods based on simple heuristic algorithms or optimization methods based on single dynamic programming models, also have deficiencies in application. On the one hand, heuristic algorithms are usually designed for specific scenarios and are difficult to take into account the diverse characteristics of samples and equipment parameters; on the other hand, although single dynamic programming methods have high theoretical optimality, they have a large increase in computational complexity due to the explosion of the state space when facing large-scale sample injection problems, making it difficult to achieve real-time optimization. At the same time, the lack of adaptive ability to the running state of the equipment and the external environment makes these methods often perform poorly in practical applications.

[0005] In summary, the current chromatography sample optimization techniques have the following outstanding problems:

[0006] Insufficient adaptability to sample diversity: unable to fully utilize sample characteristic information, resulting in suboptimal or even failed sorting results.

[0007] Low efficiency of equipment resource utilization: unable to effectively consider equipment constraints and batch scheduling strategies, resulting in serious waste of equipment resources.

[0008] Lack of dynamic adaptive capability: unable to respond to device state and environmental changes in real time, poor robustness of optimization results.

[0009] In view of the above problems, there is an urgent need for a new technical solution that can comprehensively optimize chromatography sample injection sequencing based on sample characteristics, device constraints and real-time environmental changes, thereby improving detection efficiency and device utilization, and ensuring the accuracy and robustness of detection results. SUMMARY

[0010] The purpose of the present application is to provide a dynamic programming-based chromatography sample injection control method and system, which effectively solves the problems of insufficient adaptability of sample diversity, low utilization efficiency of device resources and lack of dynamic adaptive capability, and provides a comprehensive and efficient solution for chromatography sample optimization. This method not only has significant advantages in large-scale sample detection tasks, but also can adapt to diversified device and environmental requirements.

[0011] In order to achieve the above purpose, in the first aspect of the present application, a dynamic programming-based chromatography sample injection control method is provided, which comprises:

[0012] S1, collecting sample attributes and constructing a characteristic data set of samples, wherein each sample feature in the characteristic data set contains sample attributes, and then performing dimensionless processing on the attributes of each sample according to the following standardization processing according to the different dimensions of sample characteristic values, mapping the dimensionless processed characteristic data to a low-dimensional feature space to obtain a low-dimensional feature representation of the current sample, and then performing partition calculation based on a dynamic clustering algorithm according to the low-dimensional feature representation to obtain k partitions, and the sample characteristics in each partition are highly similar, and then for each partition, according to the characteristic distribution of the samples in the partition, the device operating parameters most suitable for the current partition are allocated; wherein the sample attributes include retention time, separation degree and detection sensitivity;

[0013] S2, designing a comprehensive optimization objective function according to the low-dimensional feature representation set and the device operating parameters most suitable for the current partition for dynamic programming within the partition to minimize the comprehensive cost of each pair of samples, and generating an optimal detection sequence for each partition; wherein the device operating parameters include optimal column temperature of the partition device, maximum load of the partition device and optimal flow rate of the partition device;

[0014] S3, performing global resource scheduling according to the optimal detection sequence of each partition combined with the device operating parameters most suitable for the current partition to generate a global resource plan, including a sample list and order of each batch; wherein the global resource plan satisfies the constraints of the maximum load of the partition device and the constraint of minimizing the batch switching cost;

[0015] S4, collect the real-time running state and environmental parameters of the equipment, calculate the equipment state deviation and the environmental state deviation, respectively measure the deviation degree between the current running state and the required state of the scheduling plan, if the current equipment state deviation or the environmental state deviation exceeds the preset threshold value respectively, trigger dynamic optimization, and generate an optimized global resource plan; wherein the dynamic optimization specifically comprises:

[0016] design a new target optimization function of the global resource plan according to the current equipment state deviation and the environmental state deviation;

[0017] recalculate the batch switching cost according to the real-time running state of the equipment;

[0018] determine a batch adjustment window, including the current batch and the subsequent h batches;

[0019] use a rolling optimization algorithm to reorder the batch sequence in the window to minimize the target optimization function of the new global resource plan;

[0020] output the optimized sub-plan;

[0021] replace the optimized sub-plan into the global resource plan to form an optimized global resource plan: if the batch covered by the window has been completed, monitor the equipment state of the subsequent batch; otherwise, the dynamic optimization continues to be executed in a rolling manner;

[0022] S5, execute sample detection batch by batch according to the optimized global resource plan, collect feedback data of the equipment, calculate the deviation between the actual result and the scheduling plan target, analyze the optimization performance evaluation result by comprehensively analyzing the deviation of all batches, and adjust the target optimization function and the scheduling strategy according to the optimization performance evaluation result.

[0023] Further, the dynamic clustering algorithm-based partition calculation is performed according to the low-dimensional feature representation to obtain k partitions, and the sample characteristics in each partition are highly similar, specifically as follows:

[0024] Randomly select k samples from the low-dimensional characteristic data set Z as initial cluster centers {c1, c2, …, ck}; k};

[0025] Assign the sample to the nearest cluster center:

[0026]

[0027] Wherein, Cluster(i) represents the cluster to which the sample i belongs, c j is the center of cluster j, and z i is the low-dimensional feature representation;

[0028] For each cluster, recalculate the center as the average value of the samples assigned to the cluster:

[0029]

[0030] where N j is the number of samples in cluster j, Cluster(j) is the cluster to which sample j belongs;

[0031] The process of sample assignment and center update is repeated until the center position no longer changes;

[0032] Finally, k partitions are formed, and the samples within each partition are highly similar in characteristics.

[0033] Further, the comprehensive optimization objective function F j is expressed as:

[0034]

[0035] where T(i k ,i k+1 ) is the detection time cost, measuring the detection time between samples i k and i k+1 ; P(i k ,i k+1 ) is the cross-contamination cost, measuring the cross-contamination risk between samples i k and i k+1 ; R(i k ,i k+1 ) is the uncertainty regularization term, penalizing the uncertainty of the order between samples i k and i k+1 ; and α and λ are the weight coefficients of the cross-contamination and uncertainty regularization terms, respectively, used to balance the influence of different optimization objectives;

[0036] The detection time cost is determined by the sample residence time difference and the device flow rate , and the formula is:

[0037]

[0038] where and are the residence times of samples i k and i k+1 , respectively; is the device flow rate, representing the adjustment factor of the detection speed;

[0039] The cross-contamination cost is calculated by the difference between the sample separation degree r i and the sensitivity e i , and an innovative weight term is added to penalize extreme contamination, and the formula is:

[0040]

[0041] wherein, is the separation degree of sample i k , indicating the separation effect of the sample from other samples in the device, the higher the separation degree, the smaller the mutual interference between samples; is the separation degree of sample i k+1 , indicating the separation degree of the next sample from other samples; is the sensitivity of sample i k , indicating that the sample has the highest response in the device detection, the highest sensitivity, and the strongest detection signal of the sample; is the sensitivity of sample i k+1 , β is a dynamic adjustment weight, which is generated based on the separation degree distribution of the sample population in the partition:

[0042] β = 1 + exp(-σ r )

[0043] wherein, σ r is the standard deviation of the separation degree of the partition sample, used to adjust the influence of extreme pollution;

[0044] The uncertainty regularization term introduces the constraint of sample sequence fluctuation, which is used to reduce the large change of the detection sequence:

[0045]

[0046] wherein, γ is a regularization strength factor, controlling the smoothness of the detection sequence; is the absolute difference of sample residence time, and a larger difference will increase the regularization penalty.

[0047] Further, the minimum comprehensive cost of each pair of samples is represented as:

[0048] The state transition formula of dynamic programming is constructed:

[0049]

[0050] wherein, f(i,S) represents the optimal detection cost of the remaining sample set S from sample i; j represents sample j, T(i,j) represents the detection time cost between samples i and j, P(i,j) represents the cross-contamination cost between samples i and j, α represents the weight coefficient of the cross-contamination cost, λ represents the weight coefficient of the uncertainty regularization term, and f(j,S\{j}) represents the optimal cost after removing j from sample j;

[0051] The initial state represents that all samples have been detected; wherein, represents that there is no remaining sample;

[0052] The optimal detection sequence o is generated by backtracking from the final state using a dynamic programming table to record the optimal path of each step j = {i1, i2, …, i m}, and the optimal detection sequence {o1, o2, …, o j , o k} of each partition is obtained.

[0053] Further, the global resource scheduling is performed according to the optimal detection sequence of each partition and the most suitable device running parameter of the current partition, and specifically includes:

[0054] According to the maximum load of the partition device, the optimal detection sequence o of the partition is sequentially allocated to the batch j in sequence.

[0055] When it is found that the maximum load of the current partition device has reached, the current generation detection sample needs to be switched to the device with the minimum batch switching cost, so as to control the running load balance of the device.

[0056] An optimal batch sequence is solved by using a recursive optimization algorithm:

[0057] The initial state indicates that the scheduling of all batches is completed.

[0058] The transition formula is:

[0059]

[0060] wherein, is the kth batch in the batch scheduling;

[0061] A global scheduling plan is generated by backtracking in reverse

[0062] Further, the batch switching cost is composed of a batch-to-batch temperature adjustment time and a cross-contamination cost.

[0063] An optimization objective function G is designed to minimize the switching cost between all batches while controlling the running load balance of the device, and is expressed as:

[0064]

[0065] wherein, λ is a weight coefficient of load balance; m represents the total number of batches, represents the batch switching cost, represents a batch-to-batch switching cost matrix, represents the jth batch, represents the j+1th batch; the second term By controlling the deviation of batch load, the efficiency of equipment operation is improved; Load avg represents the average load of the equipment, represents the load of the jth batch.

[0066] Further, the target optimization function G' of the new global resource plan is represented as:

[0067]

[0068] wherein, is the adjusted batch switching cost, which is recalculated based on real-time state; and η are weight coefficients, respectively controlling the influence degree of equipment deviation and environmental adaptability, Δr is the equipment state deviation, and Δe is the environmental state deviation;

[0069] The adjusted batch switching cost is represented as:

[0070]

[0071] wherein, represents the current temperature T real and the adjustment time of the target temperature ; is the updated cross-contamination cost, which is corrected in combination with real-time sensitivity deviation:

[0072]

[0073] wherein, Δe k = |e k -e real | represents the real-time deviation of sample sensitivity; represents; r i represents the separation degree of sample i, r k represents the separation degree of sample k, e i represents the sensitivity of sample i, e k represents the sensitivity of sample k, and ω represents the weight coefficient of cross-contamination cost, which is used to adjust the influence of sensitivity difference on contamination, Δe k = |e k -e real | represents the real-time deviation of sample sensitivity, and e real represents the real-time sample sensitivity.

[0074] Further, the deviation analysis optimization performance evaluation result of all batches is integrated, and the target optimization function and the scheduling strategy are adjusted according to the optimization performance evaluation result, specifically including:

[0075] The deviation of all batches is integrated, and the global optimization performance index G is definedperf , reflecting the overall adaptability and scheduling efficiency of plan execution:

[0076]

[0077] where the first term represents the average deviation of all batches; Δ j represents the comprehensive deviation index; the second term represents the scheduling efficiency, measured by the ratio of idle time T j idle between batches to the total execution time; ω is the efficiency weight, dynamically adjusted to adapt to the specific device requirements;

[0078] The comprehensive deviation index Δ j is represented as:

[0079] Δ j = α T · |T j exec -T j plan | + α P · |P j exec -P j plan | + α E · Δe

[0080] where T j plan and P j plan are the corresponding target detection time and pollution level in the scheduling plan respectively; Δe is the deviation degree of environmental parameters, used to evaluate the influence of environmental changes on the execution results; the weights α T , α P , α E are used to balance the contributions of time, pollution and environment, P j exec represents the actual pollution level of batch j, P j plan represents the target pollution level of batch j in the scheduling plan;

[0081] Adjust the target optimization function and scheduling strategy according to the optimization performance evaluation results:

[0082] If the global optimization performance index G perf exceeds the target range, dynamically adjust the target function weight α T , α P , α E of the scheduling plan to balance each optimization target;

[0083] Update the batch switching cost matrix In particular, the highest environmental fluctuations and device aging effects.

[0084] In another aspect of the application, a dynamic programming-based chromatography injection control system is provided, the system comprising:

[0085] a sample attribute acquisition unit configured to acquire sample attributes and construct a sample characteristic data set, wherein each sample characteristic in the sample characteristic data set comprises a sample attribute, and to perform dimensionless processing on each sample attribute according to the following standardization processing according to different dimensions of the sample characteristic values, map the dimensionless processed characteristic data to a low-dimensional feature space to obtain a low-dimensional feature representation of the current sample, and perform dynamic clustering algorithm-based partition calculation according to the low-dimensional feature representation to obtain k partitions, and the sample characteristics in each partition are highly similar, and for each partition, the device operating parameters most suitable for the current partition are assigned according to the characteristic distribution of the samples in the partition; wherein the sample attributes include retention time, separation degree and detection sensitivity;

[0086] a detection sequence generation unit configured to design a comprehensive optimization objective function according to the low-dimensional feature representation set and the device operating parameters most suitable for the current partition to perform dynamic programming within the partition to minimize the comprehensive cost of each pair of samples, and generate an optimal detection sequence for each partition; wherein the device operating parameters include an optimal column temperature of the partition device, a maximum load of the partition device and an optimal flow rate of the partition device;

[0087] a global detection sequence generation unit configured to perform global resource scheduling according to the optimal detection sequence of each partition in combination with the device operating parameters most suitable for the current partition to generate a global resource plan, including a sample list and order of each batch; wherein the global resource plan satisfies the constraint of the maximum load of the partition device and the constraint of minimizing batch switching cost;

[0088] a global detection sequence optimization unit configured to acquire real-time operating states and environmental parameters of the device, calculate device state deviation and environmental state deviation to respectively measure the degree of deviation between the current operating state and the required state of the scheduling plan, and if the current device state deviation or the environmental state deviation exceeds the respective preset threshold, trigger dynamic optimization to generate an optimized global resource plan; wherein the dynamic optimization specifically includes:

[0089] designing a new target optimization function of the global resource plan according to the current device state deviation and the environmental state deviation;

[0090] recalculating the batch switching cost according to the real-time operating state of the device;

[0091] determining a batch adjustment window containing the current batch and the subsequent h batches;

[0092] using a rolling optimization algorithm, reordering the batch sequence within the window to minimize the target optimization function of the new global resource plan;

[0093] outputting the optimized sub-plan;

[0094] replacing the optimized sub-plan into the global resource plan to form an optimized global resource plan: if the batches covered by the window have been completed, monitoring the device state of subsequent batches; otherwise, the dynamic optimization continues to be executed in a rolling manner;

[0095] a sample injection optimization unit, configured to execute sample detection batch by batch according to the optimized global resource plan, collect feedback data of the device, calculate the deviation between the actual result and the target of the scheduling plan, analyze the optimization performance evaluation result by synthesizing the deviations of all batches, and adjust the target optimization function and the scheduling strategy according to the optimization performance evaluation result.

[0096] The beneficial technical effects of the present application are at least as follows:

[0097] (1) The present application innovatively introduces a sample partitioning method, which dynamically partitions the partitions according to the chemical properties (such as retention time, separation degree) and physical properties (such as detection sensitivity) of the samples. By partitioning the samples, the complexity problem caused by the diversity of the samples is effectively solved, and the efficiency and accuracy of the optimization within the partitions are ensured. A multi-dimensional dynamic programming model is used to optimize the sample injection sequence within each partition, and device parameters (such as flow rate, column temperature) are combined to achieve local optimization.

[0098] (2) To solve the problem of low utilization of device resources, the present application designs an intelligent shunting algorithm to divide the samples into batches between partitions, dynamically schedules the batch sequence to balance the device load, and maximizes the resource utilization efficiency.

[0099] Batch optimization not only reduces the idle time of the device, but also reduces the overall complexity of sample injection sequencing, realizing the synergistic work of local optimization within the partitions and global optimization between the partitions.

[0100] (3) The present application increases a real-time feedback module to monitor the device state (such as load, temperature, flow rate) and environmental changes (such as temperature fluctuations) in real time.

[0101] During operation, the partitioning and batch optimization strategies are dynamically adjusted to improve the robustness and flexibility of the system, ensuring that the optimization scheme always adapts to the actual running situation.

[0102] (4) The present application effectively solves the problems of insufficient adaptability of sample diversity, low utilization efficiency of device resources, and lack of dynamic self-adaptive ability, and provides a comprehensive and efficient solution for chromatographic sample injection optimization. This method not only has significant advantages in large-scale sample detection tasks, but also can adapt to diversified device and environmental demands. BRIEF DESCRIPTION OF THE DRAWINGS

[0103] The present invention is further described with reference to the accompanying drawings. However, the embodiments in the accompanying drawings do not constitute any limitation to the present invention. A person skilled in the art can obtain other drawings based on the following drawings without creative effort.

[0104] Figure 1 This is a flow chart of a chromatographic injection control method based on dynamic programming according to an embodiment of the present invention.

[0105] Figure 2 This is a framework diagram of a chromatographic injection control system based on dynamic programming according to an embodiment of the present invention. DETAILED DESCRIPTION

[0106] The following describes embodiments of the present invention in detail. Examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended only to explain the present invention and are not to be construed as limiting the present invention.

[0107] like Figure 1 As shown, the chromatographic injection control method based on dynamic programming provided by an embodiment of the present invention includes the following steps S1-S5:

[0108] S1. Collect sample attributes and construct a characteristic data set of the sample, wherein each sample feature in the characteristic data set includes sample attributes, and then use the following standardization processing to non-dimensionalize the attributes of each sample according to the different dimensions of the sample characteristic values, map the dimensionless characteristic data to a low-dimensional feature space to obtain a low-dimensional feature representation of the current sample, and then perform partition calculation based on a dynamic clustering algorithm based on the low-dimensional feature representation to obtain k partitions, and the sample characteristics in each partition are highly similar, and then for each partition, according to the characteristic distribution of the samples in the partition, allocate the equipment operating parameters that are most suitable for the current partition; wherein, the sample attributes include retention time, separation degree and detection sensitivity.

[0109] Specifically, the characteristic data set D of the received sample is {d1, d2, ..., d n}, where each sample d i Contains three properties:

[0110] t i : Retention time (RetentionTime), in seconds;

[0111] r i : Separation Factor, dimensionless value;

[0112] e i Sensitivity, a dimensionless value in the range [0, 1].

[0113] Further, for different dimensions of sample characteristics, the following normalization formula is used to normalize the attributes of each sample:

[0114]

[0115] where μ t ,μ r ,μ e are the mean values of the residence time, separation degree and detection sensitivity, respectively; σ t ,σ r ,σ e are the standard deviations of the corresponding attributes. After normalization, the characteristic data of all samples form a set D′ = {d1′, d2′, … d i ′, d n ′}, and the attributes of each sample are normalized to a dimensionless value.

[0116] Further, the normalized characteristic data D′ is mapped to a low-dimensional feature space (usually two-dimensional or three-dimensional) to more compactly express the sample characteristics. The mapping formula is:

[0117] z i = W·d i ′ (2)

[0118] where W is the weight matrix, the initial value is calculated by singular value decomposition (SVD) to ensure that the main information between samples is preserved; z i is the low-dimensional feature representation (such as a two-dimensional or three-dimensional vector) of sample i.

[0119] The mapped low-dimensional characteristic data set is Z = {z1, z2, …, z n}, which facilitates subsequent partitioning operations.

[0120] Further, a dynamic clustering algorithm is used to partition the low-dimensional characteristics Z of the samples, obtaining k partitions Each partition contains samples with similar characteristics. The clustering process includes the following steps:

[0121] Initialization: randomly select k samples from Z as initial cluster centers {c1, c2, …, c k};

[0122] Assign samples: assign samples to the nearest cluster center:

[0123]

[0124] where Cluster(i) represents the cluster that sample i belongs to, c j is the center of cluster j;

[0125] Further, update the cluster center: for each cluster, recalculate the center as the average of the samples assigned to this cluster:

[0126]

[0127] where N j is the number of samples in cluster j.

[0128] Iteration: repeat the process of sample assignment and center update until the center position no longer changes.

[0129] Finally, k partitions are formed, and the samples within each partition have highly similar characteristics.

[0130] Further, for each partition According to the characteristic distribution of the samples within the partition, assign the equipment operating parameters p j that are suitable for this partition, including:

[0131] T j opt : the optimal column temperature of partition j, calculated by the weighted average of the retention time distribution of the samples within the partition;

[0132] The maximum load of partition j, associated with the total number of samples in the partition;

[0133] The optimal flow rate of partition j, determined according to the sensitivity distribution of the samples in the partition.

[0134] It can be understood that the output partition result contains a list of samples with similar characteristics within each partition; at the same time, the corresponding equipment operating parameters {p1, p2, …, p k} are output, which will be directly input for subsequent dynamic programming optimization.

[0135] S2, according to the low-dimensional feature representation set and the equipment operating parameters most suitable for the current partition, design a comprehensive optimization objective function for intra-partition dynamic programming to minimize the comprehensive cost of each pair of samples, and generate the optimal detection sequence for each partition.

[0136] Specifically, from the results of step 1, obtain the data and equipment parameters of each partition:

[0137] The sample set within partition j, each sample d i = [t i , r i , ei ] contains the retention time t i , resolution r i , and sensitivity e i ;

[0138] device parameters of partition j, including optimal column temperature T j opt , maximum load , and optimal flow rate

[0139] Objective: Optimize the sample detection order o j = {i1, i2, …, i m} for each partition to minimize the comprehensive cost of detection time and cross-contamination, while considering the uncertainty of detection order and device constraints.

[0140] Further, a comprehensive optimization objective function F j is constructed, with the specific formula as follows:

[0141]

[0142] wherein,

[0143] T(i k ,i k+1 ): detection time cost, measuring the detection time between sample i k and i k+1 ;

[0144] P(i k ,i k+1 ): cross-contamination cost, measuring the cross-contamination risk between samples;

[0145] R(i k ,i k+1 ): newly introduced uncertainty regularization term, penalizing the uncertainty of detection order;

[0146] α and λ: weight coefficients of cross-contamination and uncertainty regularization terms, respectively, used to balance the influence of different optimization objectives.

[0147] Further, the detection time cost T(i k ,i k+1 ) is jointly determined by the sample retention time difference and the device flow rate , with the formula as follows:

[0148]

[0149] wherein, and are the sample retention time difference and the device flow rate k and ik+1 the residence time of the sample; is the flow rate of the device, representing the adjustment factor of the detection speed.

[0150] Further, the cross-contamination cost P(i k ,i k+1 ) is calculated by the difference of sample separation r i and sensitivity e i , while an innovative weight term is added to penalize extreme contamination:

[0151]

[0152] where β: dynamic adjustment weight, is generated based on the separation distribution of the sample population within the partition:

[0153] β = 1 + exp(-σ r ) (8)

[0154] where σ r is the standard deviation of sample separation in the partition, used to adjust the impact of extreme contamination.

[0155] Further, the uncertainty regularization term R(i k ,i k+1 ) introduces the constraint of sample sequence fluctuation, used to reduce the large variation of detection sequence:

[0156]

[0157] where γ: regularization strength factor, controls the smoothness of the detection sequence; the absolute difference of sample residence time, larger difference will increase the regularization penalty.

[0158] Further, the state transition formula of dynamic programming is constructed:

[0159]

[0160] where f(i,S): represents the optimal detection cost from sample i, in the remaining sample set S;

[0161] The initial state represents that all samples have been detected.

[0162] Further, the dynamic programming table is used to record the optimal path at each step, and the optimal detection sequence o j = {i1,i2,…,i m} is generated by backtracking from the final state.

[0163] Finally, the optimal detection sequence of each partition {o1,o2,…o j ,o k}:the overall cost (including time, contamination, uncertainty) of each pair of samples for visualizing and subsequent scheduling analysis.

[0164] S3, global resource scheduling according to the optimal detection sequence of each partition and the most suitable device operating parameters of the current partition, to generate a global resource plan, including the sample list and order of each batch.

[0165] Specifically, the following input data is obtained from step 2:

[0166] {o1,o2,…,o k}:the optimal detection sequence within the partition, each o j ={i1,i2,…,i m} represents the detection order of partition j;

[0167] {p1,p2,…,p k}:device parameters corresponding to the partition, each represents the column temperature, maximum load, and flow rate of partition j.

[0168] Objective: generate a global scheduling plan S across partitions to minimize batch switching cost while ensuring efficient use of device resources.

[0169] Further, according to the maximum load of the device The partition samples {o j} are sequentially assigned to the batch

[0170]

[0171] Initialization: samples are added to the current batch in the optimal order within the partition

[0172] Check batch load whether it exceeds

[0173]

[0174] where w i is the load weight of sample i, with an initial value of sensitivity e i ; when the load limit is exceeded, a new batch is started

[0175]

[0176] Further, innovative design: add a load balancing term η when calculating the load to punish cases where the load difference between partitions is large, making the batch division more balanced:

[0177]

[0178] Load avg is the average load of all batches so far.

[0179] Further, the batch switching cost is mainly composed of temperature adjustment time between batches and cross-contamination cost:

[0180] Temperature adjustment cost: the difference between the column temperature T j opt and the target temperature T .

[0181] Cross-contamination cost: determined by the difference between the resolution r i i and sensitivity e

[0182] Combining these two parts, the switching cost can be expressed as:

[0183]

[0184] where the cross-contamination cost has been defined in step 2 and will not be repeated here.

[0185] Further, the optimization goal is to minimize the switching cost between all batches while controlling the load balance of the device:

[0186]

[0187] λ is the weight coefficient of load balance;

[0188] The second term further improves the efficiency of device operation by controlling the deviation of batch load.

[0189] Further, a recursive optimization algorithm is used to solve the optimal batch sequence:

[0190] The initial state indicates that all batch scheduling is completed;

[0191] The transition formula is:

[0192]

[0193] Generate the global scheduling plan by reverse backtracking

[0194] It can be understood that the final output global scheduling plan S includes the sample list and sequence of each batch;

[0195] The batch switching cost matrix is used for further visualization and result analysis.

[0196] ​S4, collect real-time running state of equipment and environmental parameters, calculate equipment state deviation and environmental state deviation, respectively measure the deviation between the current running state and the required state of the scheduling plan, if the current equipment state deviation or environmental state deviation exceeds the respective preset threshold, trigger dynamic optimization, generate optimized global resource plan.

[0197] Specifically, obtain the global scheduling plan from step S3 and batch switching cost matrix At the same time, the following states are collected in real time:

[0198] r = [T real , F real , P real ]: real-time running state of equipment, including column temperature T real , flow rate F real and pressure P real ;

[0199] e = [θ temp , θ humidity ]: environmental parameters, including environmental temperature θ temp and humidity θ humidity .

[0200] Objective: based on the real-time collected equipment and environmental data, dynamically adjust the scheduling plan S to minimize the deviation in actual execution and resource waste.

[0201] Further, define equipment state deviation Δr and environmental state deviation Δe to measure the deviation between the current running state and the required state of the scheduling plan:

[0202] Δr = |r-p j |, Δe = |e-e ref | (16)

[0203] Wherein, the equipment running parameters in the scheduling plan for the batch ;

[0204] e ref : nominal working environment parameters of the equipment.

[0205] Wherein, when any index in Δr or Δe exceeds the preset threshold τ r or τ e , trigger dynamic optimization.

[0206] Further, define a new optimization objective function for the adjusted scheduling plan S', considering switching cost, equipment state deviation and environmental adaptability:

[0207]

[0208] in, The adjusted batch switching cost is recalculated based on the real-time status; and η: weight coefficients, which control the influence of equipment deviation and environmental adaptability respectively.

[0209] Furthermore, the batch switching cost The impact of real-time status on temperature adjustment and cross contamination is taken into account:

[0210]

[0211] in,

[0212] Indicates the adjustment time between the current temperature and the target temperature;

[0213] The updated cross-contamination cost, combined with real-time sensitivity bias correction:

[0214]

[0215] Δe k =|e k -e real |Indicates the real-time deviation of sample sensitivity.

[0216] Furthermore, the adjustment window is determined Contains the current batch and the next h batches;

[0217] Use the rolling optimization algorithm to reorder the batch order within the window to minimize the adjusted objective function G′;

[0218] Output optimized sub-plan

[0219] Understandably, the adjusted sub-plan Replace it into the global scheduling plan S to form a new scheduling plan S′;

[0220] If the batch covered by the window has been completed, continue to monitor the equipment status of subsequent batches;

[0221] Otherwise, dynamic optimization continues to execute rollingly.

[0222] Finally, the updated global scheduling plan S′ includes the order adjustment of all batches;

[0223] Record the deviation of equipment status and the number of adjustments during the optimization process for subsequent performance analysis.

[0224] S5. Perform sample testing batch by batch according to the optimized global resource plan, collect feedback data from the equipment, calculate the deviation between the actual results and the scheduling plan target, analyze the optimization performance evaluation results based on the deviation of all batches, and adjust the target optimization function and scheduling strategy based on the optimization performance evaluation results.

[0225] Specifically, the updated global scheduling plan is obtained from step S4 and the adjusted batch switching cost matrix The real-time operating parameters of the experimental equipment r=[T real ,F real ,P real ] and the environmental state e=[θ temp ,θ humidity ] is also used as input for result recording and optimization evaluation.

[0226] Furthermore, sample testing is performed batch by batch according to the scheduling plan S′:

[0227] In each batch Before starting, according to the required equipment parameters Adjust the device status r.

[0228] If Δr=|rp j |Exceeding the set threshold τ r , trigger recalibration of the device status and record the calibration time T calib And dynamically adjust the testing schedule.

[0229] Furthermore, sample testing is completed in batches, and the following indicators are recorded:

[0230] Total batch detection time T j exec ;

[0231] The contamination level P within the batch j exec , measured by the dynamic matching degree of sample characteristics and sensitivity.

[0232] Furthermore, after each batch is executed, the deviation between the actual result and the scheduling plan target is calculated, and a comprehensive deviation index is defined:

[0233] Δ j =α T ·|T j exec -T j plan |+α P ·|P j exec -P j plan |+α E• Δe(20)

[0234] where T j plan and P j plan are the corresponding target detection time and pollution level in the schedule plan respectively; Δe is the deviation of environmental parameters, used to evaluate the impact of environmental changes on the execution results; weights α T , α P , α E are used to balance the contributions of time, pollution, and environment.

[0235] Further, the deviations of all batches are integrated to define a global optimization performance indicator G perf , reflecting the overall fitness and scheduling efficiency of the plan execution:

[0236]

[0237] where the first term represents the average deviation of all batches;

[0238] the second term represents the scheduling efficiency, measured by the ratio of idle time T j idle between batches to the total execution time;

[0239] ω is the efficiency weight, dynamically adjusted to adapt to the specific device requirements.

[0240] Further, according to the optimization performance evaluation results, the optimization model and scheduling strategy are adjusted:

[0241] If G perf exceeds the target range, the objective function weights α T , α P , α E of the scheduling plan are dynamically adjusted to balance each optimization target.

[0242] The batch switching cost matrix is updated , especially considering higher environmental fluctuations and device aging effects.

[0243] Finally, the output results are:

[0244] Execution records: actual detection time, pollution level, device status, and environmental parameters of each batch;

[0245] Optimization evaluation: global optimization performance indicator G perf and key deviation indicators Δ j ;

[0246] Improvement suggestions: including objective function adjustment and plan update strategy, providing reference for the next round of scheduling optimization.

[0247] AsFigure 2 As shown in another embodiment of the present application provides a dynamic programming-based chromatography sample control system, the system comprises:

[0248] The sample attribute acquisition unit 301 is configured to acquire sample attributes and construct a sample characteristic data set, wherein each sample characteristic in the sample characteristic data set contains sample attributes, and the attributes of each sample are processed by the following standardization processing according to different dimensions of sample characteristic values, the characteristic data after the non-dimensional processing is mapped to a low-dimensional feature space to obtain a low-dimensional feature representation of the current sample, and the low-dimensional feature representation is used for partition calculation based on a dynamic clustering algorithm to obtain k partitions, and the sample characteristics in each partition are highly similar, and for each partition, the device operating parameters most suitable for the current partition are allocated according to the characteristic distribution of the samples in the partition; wherein the sample attributes include retention time, separation degree and detection sensitivity;

[0249] The detection sequence generation unit 302 is configured to design a comprehensive optimization objective function according to the low-dimensional feature representation set and the device operating parameters most suitable for the current partition to perform dynamic programming in the partition to minimize the comprehensive cost of each pair of samples, and generate an optimal detection sequence for each partition; wherein the device operating parameters include an optimal column temperature of the partition device, a maximum load of the partition device and an optimal flow rate of the partition device;

[0250] The global detection sequence generation unit 303 is configured to perform global resource scheduling according to the optimal detection sequence of each partition and the device operating parameters most suitable for the current partition to generate a global resource plan, including a sample list and order of each batch; wherein the global resource plan satisfies the constraint of the maximum load of the partition device and the constraint of minimizing the batch switching cost;

[0251] The global detection sequence optimization unit 304 is configured to acquire real-time operating states and environmental parameters of the device, calculate device state deviation and environmental state deviation, and measure the deviation between the current operating state and the state required by the scheduling plan, respectively, if the current device state deviation or the environmental state deviation exceeds the respective preset threshold, a dynamic optimization is triggered to generate an optimized global resource plan; wherein the dynamic optimization specifically includes:

[0252] Designing a new target optimization function of the global resource plan according to the current device state deviation and the environmental state deviation;

[0253] Recalculating the batch switching cost according to the real-time operating state of the device;

[0254] Determining a batch adjustment window containing the current batch and the subsequent h batches;

[0255] using a rolling optimization algorithm, reordering the batch sequence within the window to minimize the target optimization function of the new global resource plan;

[0256] outputting the optimized sub-plan;

[0257] replacing the optimized sub-plan into the global resource plan to form an optimized global resource plan: if the batches covered by the window have been completed, monitoring the device status of subsequent batches; otherwise, the dynamic optimization continues to be executed in a rolling manner;

[0258] the sample injection optimization unit 305 is configured to execute sample detection in batches according to the optimized global resource plan, collect feedback data of the device, calculate the deviation between the actual result and the target of the scheduling plan, analyze the optimization performance evaluation result by synthesizing the deviations of all batches, and adjust the target optimization function and the scheduling strategy according to the optimization performance evaluation result.

[0259] It should be noted that the above-described workflow is merely illustrative and does not limit the protection scope of the present application. In actual applications, a person skilled in the art can select part or all of the above-described workflow to achieve the purpose of the present embodiment, and the present application is not limited in this regard.

[0260] In addition, technical details not described in the present embodiment can be found in the parameter running method provided by any embodiment of the present application, and will not be described here again.

[0261] It should be noted that, in this document, the terms “comprise” “contain” or any other variant thereof are intended to cover non-exclusive inclusion, so that a process, method, article or system including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such a process, method, article or system. Without more limitations, the element defined by the statement “comprises a” does not exclude the presence of another identical element in the process, method, article or system including the element.

[0262] The above-mentioned embodiment numbers of the present application are only for description, and do not represent the advantages and disadvantages of the embodiments.

[0263] Through the above description of the embodiments, those skilled in the art can clearly understand that the above-mentioned example method can be realized by means of software and a necessary general hardware platform, and of course, it can also be realized by hardware, but in many cases, the former is a better embodiment. Based on such understanding, the technical solutions of the present application can be embodied in the form of a software product, which is stored in a storage medium (such as a read-only memory / random access memory, a magnetic disk, or an optical disk) and includes a number of instructions for causing an end device (which can be a mobile phone, a computer, a server, an air conditioner, or a network device) to execute the method described in each embodiment of the present application.

[0264] The above is only the preferred embodiment of the present application, and does not limit the patent scope of the present application, and any equivalent structure or equivalent flow transformation made by using the content of the specification and drawings of the present application, or directly or indirectly applied to other related technical fields, are also included in the patent protection scope of the present application.

Claims

1. A method for chromatographic sample control based on dynamic programming, characterized in that The method comprises: S1, collecting sample properties and constructing a sample characteristic data set, wherein each sample feature in the characteristic data set contains sample properties, and then performing non-dimensional processing on the properties of each sample according to the following standardization processing according to different dimensions of sample characteristic values, mapping the characteristic data after non-dimensional processing to a low-dimensional feature space to obtain a low-dimensional feature representation of the current sample, and then performing partition calculation based on a dynamic clustering algorithm according to the low-dimensional feature representation to obtain k partitions, and the sample characteristics in each partition are highly similar, and then for each partition, the device operating parameters most suitable for the current partition are assigned according to the characteristic distribution of the samples in the partition; wherein the sample properties include residence time, separation degree and detection sensitivity; S2, designing a comprehensive optimization objective function according to the low-dimensional feature representation set and the device operating parameters most suitable for the current partition to perform dynamic planning in the partition to minimize the comprehensive cost of each pair of samples, and generating an optimal detection sequence for each partition; wherein the device operating parameters include the optimal column temperature of the partition device, the maximum load of the partition device and the optimal flow rate of the partition device; S3, performing global resource scheduling according to the optimal detection sequence of each partition combined with the device operating parameters most suitable for the current partition to generate a global resource plan, including a sample list and order of each batch; wherein the global resource plan satisfies the constraints of the maximum load of the partition device and the constraint of minimizing the batch switching cost; S4, collecting real-time running states and environmental parameters of the device, calculating device state deviation and environmental state deviation to measure the deviation between the current running state and the state required by the scheduling plan, respectively, if the current device state deviation or environmental state deviation exceeds the respective preset threshold, triggering dynamic optimization to generate an optimized global resource plan; wherein the dynamic optimization specifically comprises: designing a new global resource plan target optimization function according to the current device state deviation and environmental state deviation; recalculating the batch switching cost according to the real-time running state of the device; determining a batch adjustment window containing the current batch and the next h batches; using a rolling optimization algorithm to reorder the batch order in the window to minimize the new global resource plan target optimization function; outputting the optimized sub-plan; replacing the optimized sub-plan into the global resource plan to form an optimized global resource plan: if the batch covered by the window has been completed, monitoring the device state of the subsequent batches; otherwise, the dynamic optimization continues to roll; S5, performing sample detection batch by batch according to the optimized global resource plan, collecting feedback data of the device, calculating the deviation between the actual result and the scheduling plan target, analyzing the optimization performance evaluation result by synthesizing the deviation of all batches, and adjusting the target optimization function and the scheduling strategy according to the optimization performance evaluation result.

2. The dynamic programming based chromatographic injection control method of claim 1, wherein, The dynamic clustering algorithm-based partition calculation according to the low-dimensional feature representation to obtain k partitions, and the sample characteristics in each partition are highly similar, specifically: Randomly select k samples from the low-dimensional characteristic data set Z as initial cluster centers {c1, c2, …, c k} assigning the sample to the nearest cluster center: where Cluster(i) denotes the cluster to which sample i belongs, c j is the center of cluster j, z i is the low-dimensional feature representation; for each cluster, recalculating the center as the average value of the samples assigned to the cluster: where N j is the number of samples in cluster j, Cluster(j) is the cluster to which sample j belongs; The sample allocation and center updating process is repeated until the center position no longer changes; Finally, k partitions are formed, and the sample properties in each partition are highly similar.

3. The dynamic programming based chromatographic sample control method of claim 1, wherein, The integrated optimization objective function F j is expressed as: where T(i k ,i k+1 ) is the detection time cost, measuring the detection time between sample i k and i k+1 ; P(i k ,i k+1 ) is the cross-contamination cost, measuring the cross-contamination risk between sample i k and i k+1 ; R(i k ,i k+1 ) is the uncertainty regularization term, penalizing the uncertainty of the order between sample i k and i k+1 ; and a and l are the weight coefficients of the cross-contamination and uncertainty regularization terms, respectively, used to balance the influence of different optimization objectives. The detection time cost is determined by the sample residence time difference and the device flow rate The joint decision is determined by the formula: wherein, and are the retention times of samples i k and i k+1 respectively; is the equipment flow rate, representing a scaling factor for the detection speed; The cross contamination cost is calculated from the difference in sample resolution r i and sensitivity e i with an innovative weight term added to penalize extreme contamination, expressed as: wherein, is the separation degree of sample i k , which indicates the separation effect of the sample in the device from other samples, and a high separation degree indicates less mutual interference between samples; is the separation degree of sample i k+1 , which indicates the separation degree of the next sample from other samples; is the sensitivity of sample i k , which indicates the highest responsiveness of the sample in the device detection, the highest sensitivity, and the strongest detection signal of the sample; is the sensitivity of sample i k+1 , and β is a dynamic adjustment weight, which is generated based on the separation degree distribution of the total samples in the partition: β = 1 + exp(-σ r ) where σ r is the standard deviation of the partition sample separation, used to adjust for the effect of extreme contamination; The uncertainty regularization term introduces a constraint on the sample sequence fluctuation, which is used to reduce the large changes in the detection sequence: where γ is the regularization strength factor, controlling the stationarity of the detection sequence; is the absolute difference in sample residence time, with larger differences increasing the regularization penalty.

4. The dynamic programming based chromatographic sample control method of claim 3, wherein, The minimum comprehensive cost of each pair of samples is represented as: The state transition formula of dynamic programming is constructed: Where f(i, S) represents the optimal detection cost from sample i in the remaining sample set S; j represents sample j, T(i, j) represents the detection time cost between samples i and j, P(i, j) represents the cross-contamination cost between samples i and j, α represents the weight coefficient of the cross-contamination cost, λ represents the weight coefficient of the uncertainty regularization term, and f(j, S\{j}) represents the optimal cost after removing j from the sample j. wherein the initial state indicates that all samples have been tested; wherein, indicates that there are no samples remaining; The optimal detection sequence o is generated by backtracking from the final state using the dynamic programming table to record the optimal path at each step j = {i1, i2, …, i m}, and the optimal detection sequence {o1, o2, …, o j , o k} of each partition is obtained.

5. The dynamic programming based chromatographic sample control method of claim 4, wherein, The global resource scheduling is performed according to the optimal detection sequence of each partition and the most suitable device operating parameters of the current partition, specifically including: The optimal detection sequence o of the partition is determined according to the maximum load of the partition device j sequentially assigned to the batch When it is found that the maximum load of the current partition device has reached, the current sample to be detected needs to be switched to the device with the minimum batch switching cost, and the running load of the device is balanced; The recursive optimization algorithm is used to solve the optimal batch order: initial state indicates that all batch scheduling is complete; The transition formula is: wherein, is the kth batch in the batch schedule; Generating global dispatch plans by reverse tracing 6. The dynamic programming based chromatographic sampling control method of claim 1, wherein, The batch switching cost is composed of the temperature adjustment time between batches and the cross-contamination cost; An optimization objective function G is designed to minimize the switching cost between all batches while balancing the running load of the device, and is represented as: where λ is the weight coefficient of load balancing; m represents the total batch number, denotes the batch switching cost, denotes the batch switching cost matrix, denotes the jth batch, denotes the j+1th batch; the second term By controlling the deviation of batch load, the efficiency of equipment operation is improved; Load avg denotes the average load of the equipment, denotes the load of the jth batch.

7. The dynamic programming based chromatographic sampling control method of claim 6, wherein, The target optimization function G' of the new global resource plan is represented as: wherein, is the adjusted batch switching cost, recalculated based on real-time states; and η are weight coefficients, respectively controlling the influence degree of equipment deviation and environmental adaptability, Δr is equipment state deviation, and Δe is environmental state deviation. the adjusted lot change cost is represented as: wherein, represents the current temperature T real and the target temperature adjustment time; is the updated cross-contamination cost, combined with real-time sensitivity bias correction: where Δe k = |e k - e real | represents the real-time deviation of sample sensitivity; r i represents the separation of sample i, r k represents the separation of sample k, e i represents the sensitivity of sample i, e k represents the sensitivity of sample k, ω represents a weight coefficient of cross-contamination cost, used to adjust the influence of sensitivity difference on contamination, Δe k = |e k - e real | represents the real-time deviation of sample sensitivity, e real represents the real-time sample sensitivity.

8. The dynamic programming based chromatographic sampling control method of claim 1, wherein, The deviation analysis optimization performance evaluation results of all batches are integrated, and the target optimization function and the scheduling strategy are adjusted according to the optimization performance evaluation results, specifically including: The global optimization performance indicator G is defined as the sum of the deviations of all batches perf reflecting the overall fitness of the plan execution and scheduling efficiency: where the first term represents the average deviation across all batches; Δ j represents the overall deviation index; the second term represents the scheduling efficiency, measured by the ratio of idle time between batches to the total execution time; ω is the efficiency weight, dynamically adjusted to adapt to the specific device needs; The integrated deviation indicator Δ j is expressed as: wherein, and are the corresponding target detection time and pollution level in the schedule plan, respectively; Δe is the deviation of the environmental parameter, used to assess the impact of environmental changes on the execution result; the weight α T ,α P ,α E is used to balance the contribution of time, pollution and environment, represents the actual pollution level of batch j, represents the target pollution level of batch j in the schedule plan; The target optimization function and the scheduling strategy are adjusted according to the optimization performance evaluation results: If the global optimization performance index G perf If the target range is exceeded, the target function weight a of the scheduling plan is dynamically adjusted T , a P , a E , and the optimization targets are balanced. Update batch switch cost matrix 9. A chromatographic sample introduction control system based on dynamic programming, characterized by, The system includes: A sample attribute acquisition unit is configured to acquire sample attributes and construct a sample characteristic data set, wherein each sample feature in the characteristic data set contains sample attributes, and each sample attribute is subjected to dimensionless processing according to different dimensions of sample characteristic values using the following standardization processing, and the characteristic data after dimensionless processing is mapped to a low-dimensional feature space to obtain a low-dimensional feature representation of the current sample. Then, a dynamic clustering algorithm-based partition calculation is performed according to the low-dimensional feature representation to obtain k partitions, and the sample properties in each partition are highly similar. Then, for each partition, the device operating parameters most suitable for the current partition are allocated according to the characteristic distribution of the samples in the partition. The sample attributes include residence time, separation degree, and detection sensitivity. A detection sequence generation unit is configured to design a comprehensive optimization objective function for dynamic programming in each partition according to the low-dimensional feature representation set and the device operating parameters most suitable for the current partition to minimize the comprehensive cost of each pair of samples and generate an optimal detection sequence for each partition. The device operating parameters include the optimal column temperature of the partition device, the maximum load of the partition device, and the optimal flow rate of the partition device. A global detection sequence generation unit is configured to generate a global resource plan including a sample list and sequence of each batch according to the optimal detection sequence of each partition and the most suitable device operation parameter of the current partition for global resource scheduling, wherein the global resource plan satisfies the constraints of maximum load of the partition device and minimization of batch switching cost; A global detection sequence optimization unit is configured to collect real-time operation state and environmental parameters of the device, calculate device state deviation and environmental state deviation, and measure the deviation between the current operation state and the required state of the scheduling plan, respectively. If the current device state deviation or environmental state deviation exceeds the respective preset threshold, dynamic optimization is triggered to generate an optimized global resource plan. The dynamic optimization specifically includes: designing a new target optimization function of the global resource plan according to the current device state deviation and environmental state deviation; recomputing the batch switching cost according to the real-time operation state of the device; determining a batch adjustment window including the current batch and the subsequent h batches; using a rolling optimization algorithm to reorder the batch sequence in the window to minimize the new target optimization function of the global resource plan; outputting the optimized sub-plan; replacing the optimized sub-plan into the global resource plan to form an optimized global resource plan: if the batch covered by the window has been completed, monitoring the device state of the subsequent batch; otherwise, the dynamic optimization continues to be executed in a rolling manner; A sample injection optimization unit is configured to execute sample detection batch by batch according to the optimized global resource plan, collect feedback data of the device, calculate the deviation between the actual result and the target of the scheduling plan, analyze the optimization performance evaluation result by comprehensively analyzing the deviation of all batches, and adjust the target optimization function and the scheduling strategy according to the optimization performance evaluation result.

Images