An intelligent scheduling system and method based on laboratory automation

Abstract

This invention discloses an intelligent scheduling system and method based on laboratory automation, belonging to the field of laboratory intelligent scheduling technology. The invention includes a multi-source event monitoring module, an intelligent decision-making module, an instruction execution module, and a local database. It comprehensively integrates multi-source heterogeneous information such as sample priority, equipment operating status, and consumable inventory, accurately captures core data to construct a standardized event set, clearly depicts the logical sequence of tasks and the correlation between resource usage based on dynamic graphs, and scientifically balances clinical urgency and drug development urgency with resource competition intensity using dual-value assessment. It generates global scheduling instructions, adapts them to heterogeneous equipment characteristics via equipment protocol, converts them into executable sequences, and synchronizes equipment execution status to achieve real-time closed-loop feedback. This effectively solves the pain points of traditional scheduling, such as scattered information, single decision-making, and difficulty in adapting to heterogeneous equipment, achieving dynamic optimal matching of tasks and resources, and significantly improving the efficiency and reliability of laboratory automation processes and scheduling.

Description

Technical Field

[0001] This invention relates to the field of intelligent laboratory scheduling technology, and specifically to an intelligent scheduling system and method based on laboratory automation. Background Technology

[0002] The development of laboratory automation and intelligent scheduling systems has always revolved around process efficiency and resource collaboration. Early automation was applied to repetitive operations in the form of stand-alone machines, reducing human error based on standardized logic. Subsequently, multiple types of equipment were integrated through interfaces to form automated workstations for local processes. With the combination of information technology and management needs, a basic platform supporting data acquisition and task recording was established. The scheduling mode shifted from manual planning to information-based coordination. Current technological evolution focuses on intelligent scheduling, which achieves dynamic resource allocation and real-time process optimization by integrating sensing technology and data analysis algorithms, driving laboratories towards intelligent operation of the entire process. This trend is significantly reflected in laboratory scenarios with complex processes from clinical testing to drug development. Drug development involves a multi-stage task chain of synthesis, screening, and testing, and its urgent requirements for resource coordination and intelligent scheduling further highlight the necessity and broad prospects of developing intelligent scheduling systems for laboratory automation.

[0003] Existing technologies, such as the invention patent applications related to laboratory automation scheduling disclosed in announcement numbers CN110275032B and CN119398461A, show that current laboratory automation scheduling solutions have sample information, equipment information, and consumable information scattered across different systems. This leads to scheduling conflicts such as mismatches between urgent and regular samples, and tasks remaining in queues when equipment malfunctions. Data such as the working status of automated analysis equipment, consumable inventory, and task progress cannot be synchronized in real time, resulting in inefficient scenarios such as automated analysis equipment running out of consumables still being assigned tasks, high-urgency samples queuing backlogs, and faulty equipment still being assigned tasks. Furthermore, scheduling relies on only a single priority factor, often making decisions solely based on task priority, without considering resource competition factors such as equipment queuing load, causing resource allocation imbalances. At the same time, there is poor compatibility with heterogeneous equipment from different manufacturers, and the lack of a unified standard for instruction conversion severely restricts the overall efficiency and reliability of laboratory automation processes. Summary of the Invention

[0004] In view of the above-mentioned technical shortcomings, the purpose of this invention is to provide an intelligent scheduling system and method based on laboratory automation.

[0005] To solve the above-mentioned technical problems, the present invention adopts the following technical solution: The first aspect of the present invention provides an intelligent scheduling system based on laboratory automation, including a multi-source event monitoring module: used to obtain sample information from clinical and drug development laboratory information systems, the sample information including sample number, test item type, clinical and drug development priority information; to obtain the operating status information of automated analysis equipment and the inventory and availability status of key consumables from the equipment monitoring system, the operating status information of the automated analysis equipment including working status and progress of the current task; and to identify and construct a standardized event set including emergency sample testing, equipment fault alarm and material shortage warning.

[0006] Intelligent decision-making module: When a standardized event is received, it constructs a dynamic graph that reflects the logical order and resource consumption relationship between detection tasks, performs a first value assessment and a second value assessment based on the dynamic graph, and generates a global scheduling instruction.

[0007] Instruction execution module: It is used to parse global scheduling instructions and convert them into executable operation sequences that match various automated analysis devices in clinical and drug development laboratories, and distribute the executable operation sequences to the corresponding automated analysis devices for execution.

[0008] A second aspect of the present invention provides a method for an intelligent scheduling system based on laboratory automation, comprising the following steps: Step 1. Multi-source event monitoring: obtaining sample information from clinical and drug development laboratory information systems, the sample information including sample number, test item type, clinical and drug development priority information; obtaining the operating status information of automated analysis equipment and the inventory and availability status of key consumables from equipment monitoring systems, the operating status information of automated analysis equipment including working status and progress of current tasks; and identifying and constructing a standardized event set including emergency sample testing, equipment fault alarm, and material shortage warning.

[0009] Step 2. Intelligent Decision-Making: When a standardized event is received, a dynamic graph reflecting the logical order and resource consumption relationship between detection tasks is constructed. Based on the dynamic graph, a first value assessment and a second value assessment are performed, and a global scheduling instruction is generated.

[0010] Step 3. Instruction Execution: Parse the global scheduling instructions and convert them into executable operation sequences that match various automated analysis devices in clinical and drug development laboratories, and distribute the executable operation sequences to the corresponding automated analysis devices for execution.

[0011] The beneficial effects of the present invention are as follows: (1) The first part of the present invention: multi-source information integration and standardized event set architecture, solves the scheduling conflict caused by the dispersion of sample and automated analysis equipment and consumable information, effectively avoids scheduling misjudgment caused by information lag or fragmentation, greatly improves the pre-judgment capability and response timeliness of scheduling decision, lays a solid data foundation for intelligent scheduling of the whole process, and realizes unified adaptation and event-based management of data from different systems.

[0012] (2) The second part of the present invention: The construction of the dynamic map realizes the visualization and dynamic presentation of task logic and resource status. The dual value assessment mechanism takes into account the urgency of clinical and drug development and the intensity of resource competition, gets rid of the limitations of single priority decision, ensures the priority processing of high-value tasks, and avoids the overall inefficiency caused by resource congestion. Through iterative allocation and real-time updates, it can flexibly adapt to dynamic changes such as equipment failure and consumable shortage, realize the optimal allocation of resources and efficient task flow, and significantly improve the scientific nature and global adaptability of scheduling decisions.

[0013] (3) The third part of the present invention: Based on the equipment protocol library and standardized template, the unified adaptation of heterogeneous automated analysis equipment is realized, which solves the industry pain point of incompatibility between equipment instructions from different manufacturers, greatly reduces the docking cost, and the instruction conversion and closed-loop feedback mechanism ensures the accurate parsing and reliable execution of scheduling instructions. At the same time, the operating status of automated analysis equipment is transmitted back in real time, forming an adaptive closed loop of decision-execution-feedback. This not only improves the accuracy and stability of operation execution, but also enhances the scalability and maintainability of the system, ensuring the smooth operation of the automated process. Attached Figure Description

[0014] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0015] Figure 1 This is a schematic diagram of the system modules of the present invention.

[0016] Figure 2 This is a schematic diagram of the method flow of the present invention. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] Reference Figure 1 As shown, the present invention provides an intelligent scheduling system based on laboratory automation, including a multi-source event monitoring module, an intelligent decision-making module, an instruction execution module, and a local database.

[0019] It should be noted that the multi-source event monitoring module is connected to the intelligent decision-making module, the intelligent decision-making module is connected to the instruction execution module, and the local database is connected to the multi-source event monitoring module, the intelligent decision-making module, and the instruction execution module.

[0020] The multi-source event monitoring module is used to obtain sample information from clinical and drug development laboratory information systems, including sample number, test item type, and priority information for clinical and drug development; to obtain the operating status information of automated analysis equipment and the inventory and availability of key consumables from the equipment monitoring system, including working status and progress of the current task; and to identify and construct a standardized event set including emergency sample testing, equipment fault alarms, and material shortage warnings.

[0021] For example, testing items such as specific tests like complete blood count, blood glucose, liver function, and myocardial enzyme profiles, as well as research and development projects such as compound screening and toxicity testing in drug development; clinical and drug development priority information such as classification labels indicating the urgency of samples as routine, urgent, and critical values; automated analytical equipment such as fully automated biochemical analyzers, fully automated blood analyzers, chemiluminescence immunoassay analyzers, high-throughput screening systems, and automated liquid chromatographs; the working status of automated analytical equipment such as idle, running, paused, alarmed, offline, or under maintenance; the progress of the currently executed task such as the number of the currently processed sample, the percentage of steps completed, or the estimated remaining completion time; and the availability status of critical consumables such as the remaining quantity or availability of reaction cups, sampling needles, dilution cups, and printing paper.

[0022] In a specific embodiment of the present invention, the method for identifying and constructing a standardized event set including emergency sample detection, equipment failure alarm and material shortage warning is as follows: obtaining the clinical and drug development priority of the sample, the working status of the automated analysis equipment and the current inventory of key consumables; when the clinical and drug development priority tag of a sample is urgent, the sample is determined to be an emergency sample and an emergency sample detection event is generated.

[0023] The system matches the operating status of a certain automated analysis device with a predefined database of device malfunctions. When the matched operating status indicates that the device is malfunctioning, the system determines that the automated analysis device is malfunctioning and generates a device malfunction alarm event.

[0024] When the current stock of critical consumables for a certain automated analysis device is less than the preset safety stock threshold, it is determined that the current stock of critical consumables is insufficient and a material shortage warning event is generated.

[0025] The generated emergency sample detection events, equipment failure alarm events, and material shortage early warning events are encapsulated into a standardized event set with a unified format.

[0026] In one specific embodiment, the generated emergency sample detection events, equipment failure alarm events, and material shortage early warning events are encapsulated into a standardized event set with a unified format. The specific method is as follows: a standardized data record is created for each generated event. Each record contains the following predefined data fields: event type, event unique identifier, event generation timestamp, associated resource identifier, and event details. The information corresponding to each event is filled into the corresponding fields of the standardized data record. All the filled standardized data records are organized in sequence into an event set to constitute a standardized event set.

[0027] The intelligent decision-making module is used to construct a dynamic graph reflecting the logical order and resource usage relationship between detection tasks when a standardized event is received, perform a first value assessment and a second value assessment based on the dynamic graph, and generate a global scheduling instruction.

[0028] In a specific embodiment of the present invention, the method for constructing a dynamic graph reflecting the logical order and resource occupancy relationship between detection tasks is as follows: based on sample information and the operating status information of automated analysis equipment, each detection task to be executed is taken as a task node in the graph, and corresponding resource nodes are created for the automated analysis equipment and key consumables. According to the process dependency relationship between each detection task to be executed, directed edges representing the logical execution order are established between the task nodes. According to the actual occupancy requirements of each detection task to be executed for the automated analysis equipment and key consumables, edges representing the resource occupancy relationship are established between the corresponding task nodes and resource nodes. By introducing a time dimension and a dynamic update mechanism, the relationship between node attributes and edges in the graph is updated in real time, thereby constructing a dynamic graph reflecting the logical order and resource occupancy relationship between each detection task to be executed.

[0029] It should be noted that the resource nodes are divided into automated analysis equipment resource nodes and critical consumable resource nodes. The automated analysis equipment resource nodes include equipment identification, current working status, task queue status, estimated standard processing time, and progress of the current task. The critical consumable resource nodes include consumable type, identification, current inventory, and availability status.

[0030] It should be noted that pending testing tasks refer to all testing requests that have arrived at the laboratory but have not yet been executed by automated analysis equipment. Corresponding resource nodes are created for automated analysis equipment and key consumables, including resources reserved for pending testing tasks and resources actually used by ongoing testing tasks. For example, the status of an automated analysis device that is currently running (busy), its estimated completion time for the current task, and the real-time inventory of key consumables are all derived from monitoring information of all tasks, including those currently being executed. Therefore, the resource status in the graph is global and updated in real time, providing a basis for accurately assessing the second value degree.

[0031] It should be noted that the process dependency between the various testing tasks to be executed refers to the logical association and resource constraints between different testing tasks in a laboratory automation scenario that determine their execution order. This dependency is not limited to the operation steps within a single testing task, but focuses on the timing and resource mutual exclusion relationships between tasks.

[0032] For example, the process dependencies between various testing tasks to be performed include: sequential execution dependency, that is, when the same sample needs to undergo multiple testing items with a fixed process sequence, the start of subsequent tasks must wait for the completion of the preceding tasks; resource mutual exclusion dependency, that is, when multiple testing tasks to be performed need to use the same automated analysis equipment or the same type of key limited resources, the execution order constraints formed by resource competition between tasks; data result dependency, that is, the output data of one testing task to be performed is a necessary input for another testing task to be performed, thus forming a sequential association; and logical process dependency, that is, according to the provisions of the laboratory standardized operating procedures, some testing items must be performed only after other items are completed, even if these tasks are for different samples, they still constitute cross-sample logical order constraints.

[0033] It should be noted that introducing a time dimension and a dynamic update mechanism means endowing nodes and edges in the dynamic graph with two key capabilities when constructing and maintaining the graph: first, a time dimension for representing and calculating temporal states; and second, an event-driven update capability to respond to environmental changes. The time dimension is manifested in adding sample arrival timestamps, cumulative waiting times, and estimated processing times to task nodes; recording the start and estimated completion times of current tasks for resource nodes; and assigning conflict heat based on real-time queue calculations and measured in time to resource-occupying edges. The dynamic update mechanism refers to the graph being driven by a standardized event set for real-time updates. As the system evolves, when the multi-source event monitoring module identifies events such as emergency sample testing, equipment failure, reagent shortage, or changes in task status, it automatically locates and modifies the attributes of affected nodes and edges in the graph based on standardized event types and associated resource identifiers. This includes adding task nodes, updating resource status, and recalculating conflict heat. This mechanism ensures that the graph is always an accurate mapping of the laboratory's real-time status, enabling graph-based value assessment and global scheduling decisions to be continuously optimized based on the latest information. This achieves a leap from static planning to dynamic real-time scheduling, forming the core foundation for the system's intelligence and adaptability.

[0034] For example, a dynamic graph reflecting the logical order and resource occupancy relationship between detection tasks is constructed. For instance, if a laboratory simultaneously receives task T1 (a routine blood test requiring equipment D1 and reagent R) and task T2 (preliminary compound activity screening requiring equipment D2 and reagent R), the system constructs an initial graph, creates task nodes T1 and T2, resource nodes D1, D2, and R, and establishes resource occupancy edges between T1 and D1 and R, and between T2 and D2 and R. When a material shortage warning event for reagent R is received, the dynamic graph update mechanism is immediately activated, updating the current stock attribute value of resource node R and simultaneously increasing the conflict heat weight of all resource occupancy edges connected to reagent R. This achieves real-time updates of graph node attributes and edge relationships based on event changes.

[0035] In a specific embodiment of the present invention, the first value assessment and the second value assessment based on the dynamic graph are as follows: The first value assessment: For each detection task node to be performed in the dynamic graph, the urgency of the clinical and drug development value as a function of time delay is calculated based on its clinical and drug development priority and the type of detection item.

[0036] The second value assessment: For each edge in the dynamic graph that connects the node to be executed detection task and the resource node, calculate the conflict heat, which represents the intensity of the queuing competition for each node to be executed detection task to be assigned to this automated analysis device.

[0037] It should be noted that the first and second value assessments are complementary rather than contradictory dimensions. For example, if task A has high urgency and task B has medium urgency, and both only have the same resource available (i.e., both have high conflict heat), the system will still prioritize A. However, if the only path for task A with high urgency has extremely high conflict heat (e.g., equipment failure requiring 2 hours of repair), while task B with medium urgency has an available path with low conflict heat, the system will temporarily allow task B to execute first, while simultaneously looking for alternative solutions for task A, such as activating backup equipment or adjusting other tasks to free up resources, rather than letting all tasks wait idly in front of the faulty equipment, thus preventing decisions that are locally optimal but globally inefficient.

[0038] In a specific embodiment of the present invention, the first value assessment is specifically assessed by: identifying each node of the test task to be performed based on a dynamic graph, and obtaining the clinical and drug development priority information, test item type and sample arrival timestamp corresponding to each node of the test task to be performed.

[0039] Based on the clinical and drug development priority information of each test task node to be performed, the position of the task node in the preset priority ranking sequence is determined, wherein the priority ranking sequence is in descending order of the urgency of clinical and drug development: samples with critical value indicators, samples with emergency indicators, and samples with routine indicators.

[0040] Based on the detection item type of each pending detection task node, obtain the estimated standard processing time required to complete the detection of that item type from the local database.

[0041] Based on the sample arrival timestamp and current timestamp of each pending detection task node, the cumulative waiting time of each pending detection task node is calculated.

[0042] Based on the position of each pending detection task in the priority ranking sequence, the estimated standard processing time, and the cumulative waiting time, the urgency of each pending detection task node is obtained through a preset urgency calculation function, and this urgency is used as the first value.

[0043] It should be noted that the urgency calculation function takes priority weight, cumulative waiting time and estimated standard processing time as input. The priority weight is determined according to the position of each detection task node to be executed in the preset priority sorting sequence. The earlier the sorting position, the greater the priority weight.

[0044] For example, a preset urgency calculation function, such as: urgency is... ,in The priority weight is assigned to each node of the detection task to be executed based on its position in the priority sorting sequence, with higher positions receiving greater weight. The larger the value, To accumulate waiting time, To estimate the standard processing time, priority weights are used in the preset urgency calculation function. The cumulative waiting time is directly proportional to the score. pass The item's score exhibits a monotonically increasing but marginally decreasing trend with its increase, estimating the standard processing time. The urgency level is inversely proportional to the score, thus the urgency level output by this function can quantify the nonlinear decay of the clinical and drug development value of the detection task over time.

[0045] It should be noted that the urgency calculation function satisfies the following conditions: when the cumulative waiting time and the estimated standard processing time are fixed, the greater the priority weight, the greater the urgency; when the priority weight and the estimated standard processing time are fixed, the greater the cumulative waiting time, the greater the urgency, and the increase in score gradually decreases as the cumulative waiting time increases; when the priority weight and the cumulative waiting time are fixed, the smaller the estimated standard processing time, the greater the urgency.

[0046] In a specific embodiment of the present invention, the second value assessment is specifically assessed by: identifying the occupancy relationship edges connecting each node of the detection task to be executed and the resource node in the dynamic graph, and obtaining the current task queue status of each resource node connected to each node of the detection task to be executed. The task queue status includes at least the number of tasks currently in the queue waiting state of the resource node.

[0047] For each node of the detection task to be executed and its corresponding occupancy relationship edge, the number of queued tasks of the resource node connected to the occupancy relationship edge is multiplied by the estimated standard processing time of the resource node to be executed, so as to obtain the expected waiting time for the resource node to be executed after it is added to the queue of the resource node.

[0048] The expected waiting time for resource occupancy is normalized and used as the conflict heat corresponding to the occupancy relationship edge. This conflict heat is then used as a second value measure to characterize the queuing competition intensity between each pending detection task node and its corresponding resource node.

[0049] It should be noted that the task queue status also includes the queue task sequence list and the attribute information of the tasks in the queue.

[0050] The number of tasks in the queue specifically refers to the detection tasks that have been scheduled to the automated analysis device but have not yet started execution, excluding tasks that are currently being executed.

[0051] For example, the calculation process for the second value assessment is as follows: Assume there are two task nodes to be executed in the dynamic graph, namely task node T1 representing a blood routine test and task node T2 representing a biochemical test. There are also two automated analysis equipment resource nodes, device D1 being a blood cell analyzer and device D2 being a biochemical analyzer. Task node T1 is connected to device D1 through an occupancy relationship edge, and task node T2 is connected to device D2 through an occupancy relationship edge. For task node T1, the current task queue status of device D1 connected to its occupancy relationship edge is obtained, where the number of tasks in the queue is 2. The estimated standard processing time for each task corresponding to device D1 is 5 minutes. Multiplying the number of queued tasks (2) by the estimated standard processing time (5 minutes) yields an estimated waiting resource occupancy time of 10 minutes. For task node T2, the occupancy relationship edge connected to its occupancy relationship edge is obtained... The current task queue status of device D2 has 1 task in the queue waiting state. The estimated standard processing time for each task is 8 minutes. Multiplying the number of queued tasks (1) by the estimated standard processing time (8 minutes) gives the estimated waiting resource occupation time as 8 minutes. This time is then normalized. For example, using the maximum-minimum value normalization method, the maximum original time of all occupation relationship edges in the current system is 15 minutes, and the minimum is 0 minutes. Therefore, the normalized conflict heat of the T1-D1 edge is (10-0) / (15-0)≈0.67, which is the conflict heat corresponding to this occupation relationship edge and also the second value of task node T1 to device D1. The normalized conflict heat of the T2-D2 edge is (8-0) / (15-0)≈0.53, which is the conflict heat corresponding to this occupation relationship edge and also the second value of task node T2 to device D2.

[0052] In a specific embodiment of the present invention, the method for generating global scheduling instructions is as follows: based on a dynamic graph, all available resource nodes of each node to be executed for detection tasks are selected, along with the conflict heat corresponding to the occupancy relationship edges between the node and all available resource nodes.

[0053] All pending detection task nodes are sorted in descending order according to their urgency. In this order, each pending detection task node is assigned a resource node with the lowest conflict heat from all its available resource nodes. If the resource node is already occupied, the node with the lowest conflict heat is reselected from the remaining available resource nodes. The queue status and conflict heat of the relevant resource nodes are updated in real time. By iteratively executing this allocation and update process, resource nodes and their execution order are allocated to all pending detection task nodes, forming a global task execution sequence.

[0054] The global task execution sequence is encoded into a global scheduling instruction, which includes the sample number corresponding to each detection task to be executed, the identifier of the allocated resource node, and the execution order on that resource node.

[0055] It should be noted that iteratively performing this allocation and update means recalculating the conflict heat of all occupancy relationship edges connected to the resource node based on the updated queue size. This ensures that the competition intensity evaluated when allocating subsequent tasks is based on the latest queue situation that includes newly allocated tasks.

[0056] It should be noted that the global task execution sequence is an ordered list, where each item corresponds to a scheduled detection task and includes the sample number, the assigned automated analysis device identifier, and the identifiers and parameter information of the key consumables required for the corresponding detection task. It fully defines which sample, on which device, in which order of execution, and which key consumables are used.

[0057] For example, the process of generating a global scheduling instruction is as follows: Assume there are two detection task nodes T1 and T2 in the dynamic graph, with urgency levels of 85 and 70 respectively. Both can use automated analysis device resource nodes D1 or D2. Initially, the conflict heat of the edges connecting T1 and D1, T2 and D1, and T2 and D2 are 10 minutes, 12 minutes, and 8 minutes respectively. First, the task nodes are sorted in descending order according to urgency, resulting in the order T1, T2. Resources are allocated to T1 first: the available automated analysis device resource nodes D1 have a conflict heat of 10 minutes and D2 have a conflict heat of 15 minutes. D1 with the lowest conflict heat is selected and marked as occupied. Then, the resource status is updated in real time: because D1 is allocated to T1, its queue... As the load increases, the conflict heat of all edges connected to D1 is updated. For example, the conflict heat of the T2-D1 edge is updated to 22 minutes. Then, resources are allocated to T2: at this time, the available resources for T2 are D1 with a conflict heat of 22 minutes and D2 with a conflict heat of 8 minutes. D2 with the lowest conflict heat is selected for allocation. After all the pending detection tasks are allocated, a global task execution sequence is formed: sample S001 corresponds to T1, the automated analysis device resource node is selected as D1, and the execution order is the first; sample S002 corresponds to T2, the automated analysis device resource node is selected as D2, and the execution order is the first. This sequence is encoded into a specific global scheduling instruction, which includes: automated analysis device D1 executes sample S001, and automated analysis device D2 executes sample S002.

[0058] The instruction execution module is used to parse global scheduling instructions and convert them into executable operation sequences that match various automated analysis devices in clinical and drug development laboratories, and to distribute the executable operation sequences to the corresponding automated analysis devices for execution.

[0059] In a specific embodiment of the present invention, the method for parsing and converting the global scheduling instructions into executable operation sequences matching various automated analysis devices in clinical and drug development laboratories is as follows: Identify the sample number, the allocated automated analysis device resource node identifier, and the execution order on the resource node corresponding to each test task to be executed in the global scheduling instructions; query a pre-set device protocol library based on the allocated automated analysis device resource node identifier to obtain the communication protocol and standard operation instruction template corresponding to the automated analysis device resource node; obtain the physical location of the sample and the specific parameters required for the test item from the clinical and drug development laboratory information system based on the sample number corresponding to each test task to be executed; determine the identifiers and parameters of one or more key consumables necessary for executing each test task based on the occupancy relationship between each test task and key consumable resource nodes in the dynamic graph; and integrate the sample physical location, specific parameters of the test item, identifiers and parameters of key consumables, and execution order corresponding to each test task to be executed, based on the standard operation instruction template of the automated analysis device resource node identifier allocated to each test task to be executed and the instruction format and data encapsulation method specified by the communication protocol, to generate an executable operation sequence for the target automated analysis device of each test task to be executed.

[0060] It should be noted that the device protocol library is a mapping database pre-installed in the system. Its core function is to decouple and adapt the scheduling system to diverse physical devices. Since automated analysis devices in clinical and drug development laboratories often come from different manufacturers, have different models, and have incompatible communication interfaces and instruction sets, the device protocol library identifies and associates each type or device resource with its dedicated communication protocol and standard operation instruction template. This allows the unified scheduling instructions at the upper layer to automatically match the correct underlying driver rules based on the target device identifier.

[0061] It should be noted that a standard operating instruction template is a structured operational blueprint that defines the sequence of actions and parameter placeholders that a certain type of automated equipment must follow to complete a typical test. For example, a template for a biochemical analyzer includes steps such as initializing the instrument, moving the sample from a certain position {location information} to the detection position, injecting {reagent parameters}, starting optical measurement, and returning the results to the information system of clinical and drug development laboratories. Integrating the physical location of the sample corresponding to each test task, the specific parameters of the test item, the identification and parameters of key consumables, and the execution order means filling the parsed specific sample information into the corresponding placeholders of the template in sequence, and encoding the filled step sequence into a binary instruction stream or standard message format that the equipment can directly parse, according to the provisions of the supporting communication protocol. This process transforms the logical scheduling task into a physical operation sequence that the equipment can execute.

[0062] It should be noted that the generated executable operation sequence is a set of low-level control instructions with clear timing and capable of directly driving device hardware actions. Its technical effect lies in realizing unambiguous conversion from the decision-making layer to the execution layer, ensuring the accurate physical execution of intelligent scheduling instructions. Through this conversion mechanism based on protocol library and template, this system can uniformly manage heterogeneous devices without customizing independent scheduling logic for each device, significantly improving the system's scalability, maintainability, and overall reliability of the automation process.

[0063] For example, the process of parsing and converting a global scheduling instruction into an executable operation sequence is as follows: Suppose the global scheduling instruction received by the system contains a specific scheduling content: instructing the automated analysis device D1 to be the first to perform the detection of sample S001. Parsing this instruction yields three key elements: sample number S001, device identifier D1, and execution order as first. Based on the device identifier D1, a pre-set device protocol library is queried to determine that D1 is a biochemical analyzer that follows a specific serial communication protocol. Its standard operation instruction template is defined as an operation framework that sequentially includes five steps: [initialization] [grab sample, parameter: position] [inject reagent, parameter: reagent compartment identifier and volume] [start detection] [upload results]. Based on the sample number S001, the physical location of the sample is found in the clinical and drug development laboratory information system as sample rack A-01 well, and the specific parameters required for the liver function test are: reagent type: R1, volume: 200μL, and based on the occupancy relationship between the detection task node and the key consumable resource node in the dynamic spectrum, it is determined that the current location of the injection reagent to be used to perform this task is identified as J01, and then the integration process begins: it fills the physical location sample rack A-01 hole into the parameter placeholder of the second step of the template, fills the specific parameter J01 and 200μL into the parameter placeholder of the third step, and adds the highest priority identifier to the entire step sequence according to the execution order, and encodes each step operation in the filled step sequence into a set of hexadecimal control commands that the device can recognize according to the serial communication protocol defined for D1, for example, the sample grabbing step is encoded as 0xAA0x010xA00x010x0D, and finally generates a complete, sequentially arranged hexadecimal command sequence, which is the executable operation sequence that can be directly sent to the biochemical analyzer D1 to drive it to perform sample loading, reagent injection, detection and result reporting in sequence.

[0064] In a specific embodiment of the present invention, the method for distributing the executable operation sequence to the corresponding automated analysis device for execution is as follows: Based on the automated analysis device resource node identifier associated with the executable operation sequence, the automated analysis device corresponding to the executable operation sequence is determined; a data transmission link is established through the physical communication interface corresponding to the target automated analysis device; according to the communication protocol for acquiring the automated analysis device, the instruction units in the executable operation sequence are sequentially encapsulated into data packets conforming to the protocol specifications; and these packets are sent to the automated analysis device through the data transmission link. During the transmission process, the response signals returned by the automated analysis device are continuously received and parsed; the instruction execution status is confirmed according to the instruction response rules defined in the communication protocol; and the instruction distribution progress and the automated analysis device execution feedback status are updated in real time to the status attributes of the corresponding automated analysis device resource node in the dynamic graph.

[0065] The instruction response rules defined in the communication protocol are part of the technical standards provided by the equipment manufacturer. They specify the format and meaning of the specific response signal that the equipment must return after receiving each valid instruction. The status confirmation process according to this specification includes: after sending an instruction to the automated analysis equipment, it waits for and listens for feedback from the communication interface. Once the data is received, it is parsed according to the protocol. If a predefined execution success confirmation code is parsed, such as hexadecimal code 0x06 or the string ACK, it is determined that the instruction has been received by the equipment and is about to be executed. If an error or warning code is parsed, such as ERR_101, it is determined that the instruction execution has failed, and the corresponding processing flow is triggered according to the error type. If no response is received within the timeout period specified in the protocol, it is determined that the communication has timed out or the equipment is unresponsive. This confirmation mechanism transforms instruction distribution from a one-way open-loop process into a monitorable and feedback-enabled closed-loop control process, ensuring that the scheduling instructions are executed by the physical equipment in a real and reliable manner, and feeding back the final execution status to the dynamic graph, thereby achieving system status synchronization.

[0066] It should be noted that when the instruction distribution progress and device execution feedback status are updated to the dynamic graph in real time, the instruction execution status, in particular referring to execution failure, timeout, or device alarm status, is encapsulated as a standardized device execution feedback event and added to the standardized event set. This will trigger the multi-source event monitoring module to send a new event notification to the intelligent decision-making module, which may initiate a new round of dynamic graph updates, value assessment, and scheduling instruction generation, forming a complete adaptive closed loop from decision-making, execution to feedback and re-decision-making.

[0067] Reference Figure 2As shown, this invention provides a method for an intelligent scheduling system based on laboratory automation, including step 1. Multi-source event monitoring: obtaining sample information from clinical and drug development laboratory information systems, the sample information including sample number, test item type, clinical and drug development priority information; obtaining the operating status information of automated analysis equipment and the inventory and availability status of key consumables from the equipment monitoring system, the operating status information of the automated analysis equipment including working status and progress of the current task; identifying and constructing a standardized event set including emergency sample testing, equipment fault alarm, and material shortage warning.

[0068] Step 2. Intelligent Decision-Making: When a standardized event is received, a dynamic graph reflecting the logical order and resource consumption relationship between detection tasks is constructed. Based on the dynamic graph, a first value assessment and a second value assessment are performed, and a global scheduling instruction is generated.

[0069] Step 3. Instruction Execution: Parse the global scheduling instructions and convert them into executable operation sequences that match various automated analysis devices in clinical and drug development laboratories, and distribute the executable operation sequences to the corresponding automated analysis devices for execution.

[0070] The examples described in this invention are not limited to the specific embodiments listed above. The examples are merely illustrative to facilitate understanding of the invention and do not constitute a limitation on the scope of protection of this invention. Any modifications, equivalent substitutions, etc., made within the spirit and principles of this invention should be included within the scope of protection.

[0071] The above description is merely an example and illustration of the concept of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described or use similar methods to replace them, as long as they do not deviate from the concept of the invention or exceed the scope defined in this specification, they should all fall within the protection scope of the present invention.

Claims

1. An intelligent scheduling system based on laboratory automation, characterized in that, Includes the following modules: Multi-source event monitoring module: used to obtain sample information from clinical and drug development laboratory information systems, including sample number, test item type, clinical and drug development priority information; obtain the operating status information of automated analysis equipment and the inventory and availability status of key consumables from the equipment monitoring system, including working status and progress of current tasks; identify and construct a standardized event set including emergency sample testing, equipment fault alarm and material shortage warning. Intelligent decision-making module: When a standardized event is received, it constructs a dynamic graph that reflects the logical order and resource consumption relationship between detection tasks, performs a first value assessment and a second value assessment based on the dynamic graph, and generates a global scheduling instruction. Instruction execution module: It is used to parse global scheduling instructions and convert them into executable operation sequences that match various automated analysis devices in clinical and drug development laboratories, and distribute the executable operation sequences to the corresponding automated analysis devices for execution.

2. The intelligent scheduling system based on laboratory automation according to claim 1, characterized in that, The specific method for identifying and constructing a standardized event set, including emergency sample detection, equipment malfunction alarms, and material shortage warnings, is as follows: The system acquires the clinical and drug development priorities of samples, the working status of automated analysis equipment, and the current inventory of key consumables. When the clinical and drug development priority tag of a sample is urgent, the sample is determined to be an urgent sample, and an urgent sample detection event is generated. The working status of a certain automated analysis device is matched with a predefined equipment fault database. When the matched working status indicates that the device function has failed, the automated analysis device is determined to have failed and a device fault alarm event is generated. When the current stock of critical consumables of a certain automated analysis device is less than the preset safety stock threshold, it is determined that the current stock of critical consumables is insufficient and a material shortage warning event is generated. The generated emergency sample detection events, equipment failure alarm events, and material shortage early warning events are encapsulated into a standardized event set with a unified format.

3. The intelligent scheduling system based on laboratory automation according to claim 1, characterized in that, The specific method for constructing the dynamic graph reflecting the logical order and resource consumption relationship between detection tasks is as follows: Based on sample information and the operational status information of automated analysis equipment, each test task to be executed is treated as a task node in the graph, and corresponding resource nodes are created for the automated analysis equipment and key consumables. According to the process dependencies between each test task to be executed, directed edges representing the logical execution order are established between the task nodes. According to the actual occupation requirements of each test task for the automated analysis equipment and key consumables, edges representing the resource occupation relationship are established between the corresponding task nodes and resource nodes. By introducing a time dimension and a dynamic update mechanism, the relationship between node attributes and edges in the graph is updated in real time, thus constructing a dynamic graph that reflects the logical order and resource occupation relationship between each test task to be executed.

4. The intelligent scheduling system based on laboratory automation according to claim 3, characterized in that, The specific content of the first and second value assessments based on dynamic graphs is as follows: The first value assessment: For each detection task node to be performed in the dynamic graph, the urgency of its clinical and drug development value as a function of time delay is calculated based on its clinical and drug development priority and the type of detection item. The second value assessment: For each edge in the dynamic graph that connects the node to be executed detection task and the resource node, calculate the conflict heat, which represents the intensity of the queuing competition for each node to be executed detection task to be assigned to this automated analysis device.

5. The intelligent scheduling system based on laboratory automation according to claim 4, characterized in that, The specific evaluation method for the first value assessment is as follows: Based on dynamic graph identification, each node of the test task to be performed is obtained, along with the clinical and drug development priority information, test item type, and sample arrival timestamp corresponding to each node of the test task to be performed. Based on the clinical and drug development priority information of each test task node to be performed, the position of the task node in the preset priority ranking sequence is determined, wherein the priority ranking sequence is in descending order of the urgency of clinical and drug development: samples with critical value labels, samples with emergency labels, and samples with routine labels. Based on the detection item type of each pending detection task node, obtain the estimated standard processing time required to complete the detection of that item type from the local database; Based on the sample arrival timestamp and current timestamp of each pending detection task node, the cumulative waiting time of each pending detection task node is calculated. Based on the position of each pending detection task in the priority ranking sequence, the estimated standard processing time, and the cumulative waiting time, the urgency of each pending detection task node is obtained through a preset urgency calculation function, and this urgency is used as the first value.

6. The intelligent scheduling system based on laboratory automation according to claim 4, characterized in that, The second value assessment, specifically, is conducted using the following method: Identify the occupancy relationship edges connecting each node of the detection task to be executed and the resource node in the dynamic graph, and obtain the current task queue status of each resource node connected to each node of the detection task to be executed. The task queue status includes at least the number of tasks currently in the queue waiting state of the resource node. For each node of the detection task to be executed and its corresponding occupancy relationship edge, the number of queued tasks of the resource node connected by the occupancy relationship edge is multiplied by the estimated standard processing time of the resource node to obtain the expected waiting time for the resource occupancy after the node of the detection task to be executed is added to the queue of the resource node. The expected waiting time for resource occupancy is normalized and used as the conflict heat corresponding to the occupancy relationship edge. This conflict heat is then used as a second value measure to characterize the queuing competition intensity between each pending detection task node and its corresponding resource node.

7. The intelligent scheduling system based on laboratory automation according to claim 4, characterized in that, The specific method for generating the global scheduling instruction is as follows: Based on the dynamic graph, all available resource nodes of each node to be executed detection task are selected, as well as the conflict heat corresponding to the edge of its occupancy relationship with all available resource nodes. All pending detection task nodes are sorted in descending order according to their urgency. In this order, each pending detection task node is assigned a resource node with the lowest conflict heat from all its available resource nodes. If the resource node is already occupied, the node with the lowest conflict heat is reselected from the remaining available resource nodes. The queue status and conflict heat of the relevant resource nodes are updated in real time. By iteratively executing this allocation and update process, resource nodes and their execution order are allocated to all pending detection task nodes, forming a global task execution sequence. The global task execution sequence is encoded into a global scheduling instruction, which includes the sample number corresponding to each detection task to be executed, the identifier of the allocated resource node, and the execution order on that resource node.

8. The intelligent scheduling system based on laboratory automation according to claim 7, characterized in that, The specific method for parsing and converting global scheduling instructions into executable operation sequences that match various automated analysis devices in clinical and drug development laboratories is as follows: The system identifies the sample number, the assigned automated analysis equipment resource node identifier, and the execution order on each pending test task in the global scheduling command. Based on the assigned automated analysis equipment resource node identifier, it queries a pre-set equipment protocol library to obtain the communication protocol and standard operation instruction template corresponding to that automated analysis equipment resource node. Then, based on the sample number corresponding to each pending test task, it retrieves the physical location of the sample and the specific parameters required for the test from the clinical and drug development laboratory information system. According to the occupancy relationship between each pending test task and key consumable resource nodes in the dynamic graph, it determines the identifiers and parameters of one or more key consumables necessary for executing each pending test task. Finally, based on the standard operation instruction template of the automated analysis equipment resource node identifier assigned to each pending test task and the instruction format and data encapsulation method specified by the communication protocol, it integrates the sample physical location, specific parameters of the test, identifiers and parameters of key consumables, and execution order corresponding to each pending test task to generate an executable operation sequence for the target automated analysis equipment of each pending test task.

9. The intelligent scheduling system based on laboratory automation according to claim 8, characterized in that, The specific method for distributing the executable operation sequence to the corresponding automated analysis equipment for execution is as follows: Based on the resource node identifier of the automated analysis device associated with the executable operation sequence, the automated analysis device corresponding to the executable operation sequence is determined. A data transmission link is established through the physical communication interface corresponding to the target automated analysis device. According to the communication protocol for acquiring the automated analysis device, the instruction units in the executable operation sequence are sequentially encapsulated into data packets conforming to the protocol specifications and sent to the automated analysis device through the data transmission link. During the transmission process, the response signals returned by the automated analysis device are continuously received and parsed. The instruction execution status is confirmed according to the instruction response rules defined in the communication protocol. The instruction distribution progress and the execution feedback status of the automated analysis device are updated in real time to the status attributes of the corresponding automated analysis device resource node in the dynamic graph.

10. A method for implementing an intelligent scheduling system based on laboratory automation as described in any one of claims 1-9, characterized in that, include: Step 1. Multi-source event monitoring: Obtain sample information from clinical and drug development laboratory information systems, including sample number, test type, and priority information for clinical and drug development; obtain the operating status information of automated analysis equipment and the inventory and availability of key consumables from the equipment monitoring system, including working status and progress of the current task; identify and construct a standardized event set including emergency sample testing, equipment fault alarm, and material shortage warning. Step 2. Intelligent Decision-Making: When a standardized event is received, a dynamic graph reflecting the logical order and resource consumption relationship between detection tasks is constructed. Based on the dynamic graph, a first value assessment and a second value assessment are performed, and a global scheduling instruction is generated. Step 3. Instruction Execution: Parse the global scheduling instructions and convert them into executable operation sequences that match various automated analysis devices in clinical and drug development laboratories, and distribute the executable operation sequences to the corresponding automated analysis devices for execution.

Images