Intelligent reservation management system for laboratory equipment

Abstract

The invention relates to the technical field of laboratory management, and discloses an intelligent reservation management system for laboratory equipment, and the system comprises a data collection and processing module which is used for collecting equipment state data, user reservation request data and historical use data in real time, and carrying out the cleaning, standardization and storage of the collected data; the appointment scheduling engine module is used for carrying out appointment scheduling on the equipment according to a multi-objective optimization algorithm, maximizing the utilization rate of the equipment, minimizing appointment conflicts and optimizing time distribution; and the game theory decision module is used for simulating a plurality of users to play a game in the appointment process. By introducing deep learning, reinforcement learning, the game theory and a real-time feedback mechanism, the problems of insufficient dynamic adjustment, unreasonable resource allocation, inaccurate demand prediction and frequent reservation conflicts in a traditional laboratory equipment reservation system are solved, and the equipment utilization rate, the fairness and the scheduling response capability are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of laboratory management, and particularly to an intelligent reservation management system for laboratory equipment. Background Art

[0002] In the field of laboratory equipment management, traditional reservation systems adopt fixed rules or static scheduling methods, and rely on simple priority setting or first-come, first-served models to complete equipment allocation. These systems are mostly used in scenarios with low equipment demand and limited user numbers, and can meet basic equipment reservation needs. However, with the increase in the number of users and the complexity of reservation requirements, traditional systems gradually expose many limitations and are difficult to adapt to complex environments with multi-user and multi-equipment sharing. Especially in peak periods of equipment demand or in cases of resource shortages, existing systems are difficult to ensure efficient and reasonable allocation of equipment resources.

[0003] The structure of the prior art includes an equipment status acquisition module, a user reservation module, and a simple scheduling module, but lacks dynamic adjustment and intelligent optimization capabilities. Specifically, the following problems exist in the prior art: First, static rules cannot adjust the scheduling strategy according to real-time status, resulting in low resource utilization; Second, reservation conflicts among multiple users are often not effectively resolved, affecting the user experience; Third, prediction methods are mostly based on simple statistical analysis and are difficult to accurately predict future equipment demand, and the scheduling lacks foresight; Fourth, there is a lack of a real-time feedback mechanism and it is unable to quickly respond when equipment fails or demand changes, and the overall flexibility and reliability of the system are poor. Summary of the Invention

[0004] In view of the deficiencies of the prior art, the present invention provides an intelligent reservation management system for laboratory equipment, and the present invention solves the problems of low utilization rate of equipment resources, frequent user reservation conflicts, and lagged scheduling response caused by the lack of dynamic adjustment, intelligent optimization, demand prediction, and real-time feedback capabilities in the prior art.

[0005] To achieve the above objectives, the present invention is realized through the following technical solutions: An intelligent reservation management system for laboratory equipment, including: A data acquisition and processing module, which is used to collect equipment status data, user reservation request data, and historical usage data in real time, and clean, standardize, and store the collected data; A reservation scheduling engine module, which is used to perform reservation scheduling on equipment according to a multi-objective optimization algorithm, maximize the utilization rate of equipment, minimize reservation conflicts, and optimize time allocation; A game theory decision-making module, which is used to simulate the game among multiple users during the reservation process, calculate the optimal reservation plan according to the Nash equilibrium theory, avoid conflicts, and improve the efficiency of resource allocation; A deep learning prediction module, which is used to train and generate a demand prediction model based on historical data, predict future device usage demands, and provide dynamic adjustment information for the scheduling engine module; A reinforcement learning dynamic scheduling module, which is used to automatically adjust the reservation strategy of devices according to real-time data and system feedback, optimize resource allocation and adapt to changes in user demands; A real-time feedback and dynamic adjustment module, which is used to monitor the system operation status, obtain real-time changes in user demands and device statuses, and dynamically adjust the scheduling strategy according to the feedback.

[0006] Preferably, the data collection and processing module collects device status data in real time through Internet of Things technology, and the device status data includes the current availability of the device, fault records, maintenance status, and device usage duration; the user reservation request data includes information such as the reservation period of the user, device type, reservation duration, and user priority.

[0007] Preferably, the reservation scheduling engine module performs reservation scheduling on devices based on a multi-objective constraint optimization algorithm, and the optimization objectives include but are not limited to resource utilization rate, service quality, time fairness, and the number of reservation conflicts. The constraint conditions include device capacity, user reservation demands, and available time periods of the devices.

[0008] Preferably, the game theory decision-making module is used to calculate the utility value of each user under a given resource state, and determine the optimal reservation strategy according to the Nash equilibrium principle. The utility function includes factors such as the availability of the device, the user's reservation waiting time, cost, and priority.

[0009] Preferably, the deep learning prediction module adopts a long short-term memory network or a gated recurrent unit model to generate prediction results of future device demands based on historical reservation data, and the prediction results are provided to the reservation scheduling engine module to optimize device reservation scheduling.

[0010] Preferably, the reinforcement learning dynamic scheduling module is based on the Q-learning algorithm. During the peak period of device reservation or when demands fluctuate, it continuously adjusts the scheduling strategy through real-time data and system feedback to optimize the device utilization rate and resource allocation. The strategy update is based on the feedback information of the real-time system state and user reservation behavior.

[0011] Preferably, the real-time feedback and dynamic adjustment module can collect real-time fault information of the device, changes in user demands, actual utilization rates of reservation periods, and changes in system status, and dynamically adjust the scheduling strategy of the scheduling engine according to this information to ensure efficient allocation of resources.

[0012] Preferably, the management system works through the following steps: S1. Collect device status data and user reservation request data in real time through the data collection and processing module, and perform data cleaning, standardization, and storage. S2. Through the reservation scheduling engine module, perform reservation scheduling for the device based on the multi-objective optimization algorithm and constraint conditions. S3. Through the game theory decision-making module, simulate the reservation behaviors among users and calculate the optimal reservation plan. S4. Generate a prediction of the future usage demand of the device through the deep learning prediction module and feedback it to the reservation scheduling engine module. S5. Optimize the scheduling strategy of the system through the reinforcement learning module so that the reservation strategy can be adjusted when the device demand fluctuates. S6. Process the system feedback information through the real-time feedback and dynamic adjustment module and dynamically adjust the scheduling plan.

[0013] Preferably, the device status data includes the current availability, maintenance records, fault logs, and device usage duration of the device, and the user reservation request data includes information such as the reservation period, reservation duration, type of reserved device, and user priority.

[0014] Preferably, when there is a conflict in the user reservation requests, the system can adjust the reservation plan through the game theory decision-making module and the reinforcement learning module to avoid conflicts and ensure fair distribution of the user reservation periods.

[0015] The present invention provides an intelligent reservation management system for laboratory equipment, which has the following beneficial effects: 1. The present invention adopts a technical solution that combines deep learning and reinforcement learning. Through learning the historical data of device usage and the feedback of real-time status, it achieves the effects of accurate prediction of device demand and rapid response of resource allocation. Compared with the technical solutions that only rely on static rules or preset scheduling strategies in the prior art, it solves the problems of slow response and rigid scheduling in the face of complex reservation requirements, and significantly improves the flexibility and practicality of the system.

[0016] 2. By introducing the game theory decision-making module, the present invention models the multi-user competition environment using the user utility function and the Nash equilibrium model, achieving the technical effects of fairness in user reservation and rationality in resource allocation. Compared with the technical solutions in the prior art that lack effective means to solve the conflicts among user reservations, it overcomes the problem of decreased user satisfaction caused by resource contention, and at the same time optimizes the utilization efficiency of resources.

[0017] 3. The present invention constructs a reservation scheduling engine based on a multi-objective optimization algorithm, and combines a real-time feedback and dynamic adjustment module, achieving the technical effects of maximizing equipment utilization rate and minimizing conflicts. Compared with the scheduling method of fixed resource allocation in the prior art, it solves the problems of resource waste and high equipment idle rate. At the same time, through the real-time feedback mechanism, it enhances the system's rapid response ability to sudden demands.

[0018] 4. The present invention closely combines the data acquisition and processing module with the deep learning prediction module, and adopts multi-dimensional data analysis technology to achieve the advance perception of future equipment demand trends. Compared with the technical solutions in the prior art that only rely on artificial experience or fixed usage patterns for prediction, it overcomes the defects of low prediction accuracy and decision-making lag, providing a scientific basis for the overall optimal scheduling of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 It is a system architecture diagram of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0020] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0021] Please refer to the attached Figure 1 , the embodiments of the present invention provide an intelligent reservation management system for laboratory equipment, including: A data acquisition and processing module, configured to collect equipment status data, user reservation request data, and historical usage data in real time, and clean, standardize, and store the collected data; Implementation manner of the data acquisition and processing module First of all, the data acquisition and processing module is a crucial component in the intelligent reservation management system for laboratory equipment of the present invention. It plays a core role in the entire system architecture, responsible for collecting various information from laboratory equipment and user terminals, and converting this information into standardized data for subsequent modules to use. This process provides the necessary basic data support for the accurate decision-making and optimal scheduling of the system. This module ensures that the system can obtain the status of equipment and the reservation requirements of users in real time, and process these data, providing high-quality input data for subsequent modules such as reservation scheduling, game theory decision-making, deep learning prediction, and reinforcement learning dynamic scheduling.

[0022] In this embodiment, the data acquisition and processing module mainly includes two sub-modules: device status data acquisition and user reservation request data acquisition. These two sub-modules are respectively responsible for collecting data from the device and the user side, and performing necessary processing and standardization on these data.

[0023] Device status data acquisition In this embodiment, the device status data acquisition module monitors the status of all devices in the laboratory in real time through Internet of Things technology and sensor technology. Specifically, the status of the device includes information such as device availability, fault records, maintenance status, and usage duration. These data can be collected in real time through sensors integrated on the device or intelligent devices connected to the device management system. The status data of each device is transmitted to the system in real time, and the system determines whether the device is available for reservation or needs to be repaired based on these data.

[0024] For example, assume device E j The status data includes availability (whether the device is online), fault records (whether a fault occurs), maintenance status (whether it is in the maintenance period), and usage duration (the total running time of the device since the last maintenance). These data can be expressed as: A j : Availability status of device E j If the device is available, then A j = 1, otherwise A j = 0; F j : Fault status of device E j If a fault occurs in the device, then F j = 1, otherwise F j = 0; M j : Maintenance status of device E j If the device is under maintenance, then M j = 1, otherwise M j = 1; U j : Usage duration of device E j

[0025] On this basis, the real-time status of the device can be expressed by the following formula: S j = A j ·(1 - F j )·(1 - M j ) Where S j is the status value of device E j , indicating whether the device is available for reservation at this moment. If the device is available, not faulty, and not under maintenance, then S j= 1; Otherwise, S j = 0.

[0026] User Appointment Request Data Collection On the user side, the data collection module is responsible for collecting the user's appointment request information, including the appointment device type, appointment time period, appointment duration, user priority, etc. When the user submits an appointment request through the system front-end, the system will record the specific requirements of each user.

[0027] For example, assume that the appointment request submitted by user i includes the appointment device type T i , the appointment time period P i and the appointment duration D i . Where: T i is the device type reserved by user i, which may include different device types, such as laboratory instruments, computers, laboratory spaces, etc.

[0028] P i is the time period reserved by user i, which can be expressed as a set of time period sets, such as P i = [t 1 , t 2 ,..., t n , indicating the reserved time period range.

[0029] D i is the duration reserved by user i, and the unit can be hours or days.

[0030] This information can be recorded in data in the following format: R i = (T i , P i , D i ) Here, R i is the appointment request data submitted by user i.

[0031] Data Cleaning and Standardization Once the device and user data are collected, this module will clean and standardize this data. The main task of data cleaning is to remove invalid or duplicate data to ensure that all input data is consistent and valid. Data standardization is to convert data from different sources into the same format to ensure data consistency.

[0032] For example, in the user reservation request data, there may be invalid data caused by inconsistent formats or input errors, such as incorrect time formats, incorrect device types, etc. Through data cleaning, the system can eliminate this invalid data and only retain correct and valid information. The standardization operation ensures that all data can be stored in a unified format so that subsequent modules can process it.

[0033] Data Storage and Transmission The cleaned and standardized data will be stored in the system's database. The data storage scheme usually uses a relational database or a distributed database. Through database storage, the system can ensure the persistence and efficient access of data, and at the same time it is also convenient for subsequent modules of the system to query and operate on the data.

[0034] Data Interaction and Module Collaboration In the entire system, the interaction between the data collection and processing module and other modules is very crucial. All the collected data will play an important role in the subsequent calculations and decisions of the system.

[0035] Reservation Scheduling Engine: The scheduling engine module will use the device status data to determine whether the device is available and allocate device resources based on the user's reservation request data. The scheduling engine module will consider multiple objective functions, such as maximizing device utilization rate, minimizing conflicts, etc., and perform scheduling based on these data.

[0036] Game Theory Decision Module: The game theory decision module needs to rely on the user's reservation request data, calculate the utility of each user and make game decisions to ensure that each user can obtain a reservation opportunity within a reasonable range.

[0037] Deep Learning Prediction Module: The deep learning prediction module will use historical device status data and user reservation request data to train a prediction model, predict future device usage requirements, and provide prediction information for the scheduling engine.

[0038] Reinforcement Learning Dynamic Scheduling Module: The reinforcement learning module obtains data from devices and users in real time according to real-time feedback and dynamically adjusts the scheduling strategy.

[0039] The data collection and processing module is the basic module of the intelligent reservation management system for laboratory equipment of the present invention. Through this module, the system can accurately collect and process key information such as device status and user reservation requests, and provide high-quality data support for subsequent modules. The cleaning, standardization, and storage operations ensure the consistency and effectiveness of the data, providing strong data guarantee for subsequent multi-objective optimization scheduling, game decision-making, demand prediction, etc.

[0040] Reservation Scheduling Engine Module, which is used to perform reservation scheduling on devices according to the multi-objective optimization algorithm, maximize the utilization rate of devices, minimize reservation conflicts, and optimize time allocation; In this embodiment, the main task of the scheduling engine module is to perform multi-objective optimization scheduling based on the device status data and user reservation request data obtained from the data collection and processing module, in combination with the objective function and constraint conditions set by the system. This process depends on the defined optimization objectives, including maximizing the device usage rate, minimizing reservation conflicts, ensuring time fairness, and enhancing the user experience, etc.

[0041] Multi-objective Optimization Algorithm for Device Scheduling In this embodiment, the scheduling engine adopts a multi-objective optimization algorithm, considering multiple objective functions and making trade-offs during the device reservation process. These objectives include but are not limited to: Resource Utilization Rate (R): That is, the utilization frequency of the device, aiming to ensure that the device is utilized as highly as possible during the reservation process.

[0042] Service Quality (S): Reflects the user's satisfaction, mainly considering the success rate of reservation, timeliness of confirmation, etc.

[0043] Time Fairness (T): Ensures fairness among different users during the reservation process, avoiding that some users cannot reserve devices during peak periods.

[0044] Number of Reservation Conflicts (C): Reduces conflicts when multiple users reserve the same device time slot, ensuring the reasonable allocation of resources.

[0045] The scheduling engine schedules devices through the following multi-objective optimization formula: Z = α 1 ·R + α 2 ·S + α 3 ·T - α 4 ·C Where: R: Resource Utilization Rate, indicating the usage frequency of the device. Specifically, the calculation formula for the resource utilization rate is as follows: Where, x i,j Indicates whether user i reserves device j at a specific time slot, I represents the total number of users, and J represents the total number of devices.

[0046] S: Service Quality, reflecting the user's satisfaction, usually an indicator of waiting time and reservation success. It can be calculated through the following formula: Where, w iDenote the waiting time of user i, c i is the confirmation delay of user i.

[0047] T: Time fairness, which represents the fairness of device reservations among all users. Specifically, time fairness can be calculated by the following formula: where, x i,j represents whether user i has reserved a time slot of device j.

[0048] C: Number of reservation conflicts, which is used to measure the conflict situation of multiple users reserving the same device. The formula is as follows: where, x i,j and x k,j represent the reservation status of user i and user k for device j respectively.

[0049] Constraints of device scheduling To ensure the rationality and fairness of device reservations, in addition to the optimization objectives, the scheduling engine also considers some constraints. These constraints mainly include the capacity of the device, the required duration of the user, the availability of the device, etc. Specifically, the constraints include the following items: Device capacity constraint: The number of reservations for each device in each time period cannot exceed the maximum capacity of the device. Assume the capacity of device j is c j , then for each time period t, it is necessary to satisfy: where, x i,j,t represents whether user i has reserved device j in time period t.

[0050] User required duration constraint: The reservation duration of each user must match their actual needs. Assume the required reservation duration of user i is d i , then: where, x i,j represents whether user i has reserved device j and meets the required duration of user i.

[0051] Device availability constraint: The availability of the device will change according to device failures, repairs, etc., so the scheduling engine needs to dynamically update the availability status of the device. Assume the availability of device j in time period t is A j,t , then: Among them, A j,t is the available state of device j in time period t. If the device is available, then A j,t = 1, otherwise A j,t = 0.

[0052] Dynamic Scheduling and Adjustment To cope with the real-time changing device states and user demands, the scheduling engine also needs to be able to make dynamic adjustments when the device states change. Specifically, when a device fails or is under temporary maintenance, the scheduling engine can recalculate and adjust the appointment time slots of all affected users. The scheduling engine will obtain real-time device status feedback information and re-optimize the scheduling plan according to the latest device status.

[0053] In this possible implementation, the scheduling engine also adjusts resource allocation by monitoring user appointment behaviors. For example, when multiple users request the same device and the time slots overlap, the scheduling engine will try to recommend appointments at different time periods or reschedule the appointment time slots of other users to ensure that all users can make fair appointments for the device.

[0054] System Interaction and Data Flow The appointment scheduling engine module works closely with other modules. First, the data acquisition and processing module provides device status and user appointment request data. Then, the scheduling engine module performs multi-objective optimization scheduling based on this data to obtain a preliminary appointment plan. Next, the game theory decision-making module further adjusts the appointment time period according to the scheduling plan and user demands to ensure the fairness of resource allocation. Finally, the deep learning prediction module provides prediction information on future device demands to help the scheduling engine better allocate resources.

[0055] The appointment scheduling engine module of the present invention provides an intelligent device appointment scheduling plan through multi-objective optimization algorithms and constraint conditions, combined with device states and user demands. Through real-time data feedback and dynamic adjustment mechanisms, the system can ensure quick response in case of device failures, demand fluctuations, etc., and guarantee the efficient operation of the system. This module not only improves the utilization efficiency of devices, but also provides fair appointment opportunities for users, ensuring the flexibility and stability of the system.

[0056] A game theory decision-making module, used to simulate the game among multiple users during the appointment process, calculate the optimal appointment plan according to the Nash equilibrium theory, avoid conflicts and improve the efficiency of resource allocation; In this embodiment, the game theory decision-making module simulates the reservation decision-making processes of multiple users and calculates the optimal choice for each user according to the user's utility function. This process utilizes the Nash equilibrium principle to ensure that all users reach a "stable" state under the given resources and conditions, that is, no user can unilaterally change their strategy to obtain higher utility. In this case, the resource allocation of the entire system is the most reasonable, and the balance between fairness and efficiency is also achieved.

[0057] Calculation of User Utility Function Generally, when users select a reservation time period, they make decisions based on various factors such as their own needs, the availability of the device, and the reservation cost. The game theory decision-making module simulates this process by defining a utility function for each user. The utility function U i (j) is used to quantify the utility value obtained by the user when selecting a certain reservation period j. This utility function usually includes the following items: Device Availability: Whether the device is in a reservable state.

[0058] Waiting Time: The time from when the user submits a reservation request to when the final confirmation is obtained.

[0059] Reservation Cost: The cost of reserving the device during a specific time period.

[0060] Priority: The reservation priority of the user, which may be determined based on factors such as the user's historical records and credit ratings.

[0061] The utility function of the user can be expressed as: U i (j) = α 1 ·availability j -α 2 ·waiting timej + α 3 ·cost j +α 4 ·priority i Where: availability j represents the availability of device j during reservation period j. If the device is available, then availability j = 1, otherwise it is 0.

[0062] waiting time j represents the reservation waiting time, that is, the time from when the user initiates a reservation to when the final confirmation is obtained.

[0063] cost j is the cost that the user needs to pay for reserving the device during this period.

[0064] priority i Indicates the reservation priority of user i, reflecting the impact of the user's credit or historical behavior on the reservation priority.

[0065] α 1 , α 2 , α 3 , α 4 is the weight coefficient, used to adjust the contribution of various factors to the user's utility.

[0066] Specifically, the design of the utility function takes into account the availability of the device (whether the device is online), the user's waiting time (the longer the waiting time during the reservation process, the lower the user's utility), and the reservation cost (users tend to choose a time period with a lower cost). The priority helps the system decide whether to give the user the opportunity to make a reservation first based on the user's credit or historical behavior.

[0067] Calculation of Nash Equilibrium In a multi-user reservation system, each user's reservation choice will affect the choices of other users. To ensure the reasonable allocation of resources in the competition of multiple demands, the game theory decision module applies the Nash equilibrium theory. Nash equilibrium means that, given the choices of other users, no user can obtain a higher utility by changing their own choice. In other words, each user's decision is optimal provided that the decisions of all other users are known and do not change.

[0068] The Nash equilibrium in game theory can be described by the following conditions: In a system with multiple users, if each user adjusts their strategy according to the choices of other users to maximize their own utility, then the Nash equilibrium is reached. Specifically, for user i, if the reservation time period j selected by it * satisfies the following conditions: then the system is in a Nash equilibrium state.

[0069] In other words, when the choices of all users no longer change, the system reaches a stable state. In this state, no user can improve their own utility by unilaterally changing their reservation strategy (i.e., changing the reservation time period or device). Therefore, the reservation plan reached by the system among multiple users is the optimal resource allocation plan.

[0070] Adjustment and Feedback Mechanism Generally, the game theory decision-making module will cooperate closely with other modules in the system. Specifically, when the device status changes (for example, when the device fails or needs maintenance), the game theory decision-making module will recalculate the utility functions of all users based on the change in device availability and readjust the reservation periods of users based on the Nash equilibrium. At the same time, when the system obtains prediction data on future device requirements through the deep learning prediction module, the game theory decision-making module will also consider this prediction data to make scheduling preparations in advance.

[0071] For example, assume device E j has a change in availability during a certain period j, and the system recalculates the utility functions of all users through the game theory decision-making module. If the utility function U i (j) of a certain user i changes because the device is unavailable, the game theory module will adjust the reservation period of this user or reallocate resources based on the new device status and the user's utility value. Finally, the system will output a new set of reservation plans that conform to the Nash equilibrium to ensure that the reservation allocation of all users is as fair and efficient as possible.

[0072] Dynamic feedback and adjustment of the system During the operation of the system, the calculation of the game theory decision-making module is dynamic. With the real-time changes in device status and user reservation requests, the game theory decision-making module will continuously calculate and adjust the reservation strategies of users. When multiple users reserve a certain device simultaneously, the game theory module will ensure that the final reservation plan can minimize conflicts and improve resource utilization efficiency by adjusting reservation periods, allocating priorities, adjusting fees, etc.

[0073] In addition, the game theory decision-making module works together with other modules (such as the scheduling engine, deep learning prediction module, etc.). Through cooperation with the scheduling engine module, the game theory decision-making module can perform dynamic adjustments under high load and conflict conditions to ensure that the system always maintains the best resource allocation under complex user requirements and device status conditions.

[0074] The game theory decision-making module ensures the fairness and efficiency of laboratory equipment reservation by defining user utility functions and calculating the Nash equilibrium. Through cooperation with other modules in the system (such as the scheduling engine, deep learning prediction module), the game theory decision-making module can dynamically adjust reservation strategies, avoid conflicts, and optimize the device usage rate. This module can provide the optimal scheduling plan in an environment where multiple users compete for limited resources and is an important part of the intelligent reservation management system for laboratory equipment in the present invention.

[0075] The deep learning prediction module is used to train and generate a demand prediction model based on historical data, predict future device usage requirements, and provide dynamic adjustment information for the scheduling engine module; In this embodiment, the deep learning prediction module adopts a long short-term memory network (LSTM) or a gated recurrent unit (GRU) model. These models are particularly suitable for processing tasks with temporal characteristics because they can consider the dependencies in the time dimension when processing data. The deep learning prediction module is trained on the historical reservation data of the device, learns to identify the device usage patterns and demand fluctuation rules, and then predicts future demands.

[0076] Selection and Application of Deep Learning Models Generally, LSTM models and GRU models are widely used in processing temporal prediction tasks, such as predicting the reservation demands of devices. Specifically, the LSTM model solves the vanishing gradient problem existing in traditional recurrent neural networks (RNNs) through its special gate structures (input gate, forget gate, and output gate), and thus can handle long-term dependencies. Therefore, the LSTM model has obvious advantages in processing long-term dependent data such as device reservations.

[0077] As an alternative, the GRU model is a simplified version of the LSTM, with similar performance but higher computational efficiency. The GRU controls the flow of information through an update gate and a reset gate, reducing the number of parameters compared to the LSTM. Therefore, the GRU model is also a viable alternative in cases where computational resources are limited.

[0078] During the implementation of the present invention, the deep learning prediction module uses the LSTM or GRU model to process the historical usage data of the device (including information such as the reservation frequency of the device, usage time period, device type, user demands, etc.). Through continuous training, the model can gradually learn the rules of device usage and demand fluctuations, and then predict the device demands for a future period of time.

[0079] Model Training and Data Input Specifically, when training the LSTM or GRU model, the system first needs to collect historical reservation data. This data includes users' reservation requests, device usage situations, device failures and repair statuses, etc. Each reservation data can be represented as a time series, where time is one of the input features of the model.

[0080] For example, assume device E j 's reservation data can be represented as: D j =(t 1 ,t 2 ,...,t n ) Where t i represents the reservation time of device E jThe usage status during time period i, including information such as the number of reserved people, equipment type, reserved time period, etc. By inputting this time series data, the LSTM or GRU model will train on it to learn the fluctuation pattern of equipment demand.

[0081] When making a prediction, the model will output the equipment demand for the next m time periods. Assume the predicted equipment demand is: Where is the demand for the prediction time period n + i. This prediction result will be used as the input to the scheduling engine to guide the scheduling decision.

[0082] Specific implementation of deep learning prediction In a possible implementation, the deep learning prediction module first receives the historical equipment usage data transmitted from the data collection and processing module. After this data is cleaned and standardized, it will be input into the LSTM or GRU model. The model adjusts the weights through the backpropagation algorithm and gradually learns the pattern of equipment usage. After the model training is completed, the system can use the trained model to predict the future equipment demand.

[0083] Specifically, the training process of the LSTM model can be described by the following recurrence relation: h t = LSTM(x t , h t-1 , c t-1 ) Where: x t is the input data, representing the equipment status or user reservation information at the current time t; h t is the output of the LSTM cell, representing the hidden state at the current time; c t is the memory state of the LSTM; h t-1 and c t-1 are the hidden state and memory state at the previous time.

[0084] During the training process, the LSTM model learns the relationship between the input data and the target output by optimizing the loss function to minimize the prediction error.

[0085] After the training is completed, the output of the LSTM model is the prediction of the equipment demand for the next several time periods. The prediction result will be transmitted to the reservation scheduling engine module for optimizing the reservation decision and resource scheduling.

[0086] Collaboration between the model output and the scheduling engine The output of the deep learning prediction module will directly affect the operation of the appointment scheduling engine. Generally, the scheduling engine will adjust the appointment time slots of devices or optimize the allocation of resources according to the predicted device requirements. For example, assume that the model predicts that device E j will have a significant increase in demand in the next time period. The scheduling engine can advance more appointment time slots for device E j or dynamically adjust the appointment arrangements of other devices to avoid device congestion or idleness.

[0087] Specifically, when the prediction result shows that the demand for a device in a certain time period will increase, the scheduling engine can preferentially allocate more appointment time slots for that time period or postpone the appointment time slots of other devices to ensure the rational use of resources. At the same time, the scheduling engine will also dynamically adjust the resource allocation strategy according to the difference between the prediction result and the actual appointment demand.

[0088] Data Update and Real-time Prediction In this embodiment, the deep learning prediction module is not trained once and used for a long time, but has the ability to be updated in real time. As new user appointment data arrives, the model will be continuously updated and the prediction results will be gradually adjusted. This online learning method enables the deep learning model to adapt to rapidly changing appointment demands and maintain a high prediction accuracy.

[0089] Whenever the system receives new appointment requests or device status change data, the deep learning module will retrain the model and perform real-time prediction based on the latest historical data. In this way, the system can make more accurate predictions about future device demands according to the latest demand changes and timely adjust the appointment strategy.

[0090] The deep learning prediction module plays an important role in the present invention. By predicting the device usage requirements through LSTM or GRU models, the system can effectively cope with demand fluctuations, adjust the scheduling strategy in advance, and avoid problems such as device idleness or resource shortage. The deep learning prediction module ensures that the system can make accurate predictions according to the latest data and provides support for the appointment scheduling engine through real-time update and online learning methods. The implementation of this module greatly improves the usage efficiency of devices, optimizes the appointment experience, and enables the system to adapt to rapidly changing demands.

[0091] The reinforcement learning dynamic scheduling module is used to automatically adjust the appointment strategy of devices according to real-time data and system feedback, optimize resource allocation, and adapt to changes in user demands; In this embodiment, the reinforcement learning dynamic scheduling module adopts the Q-learning algorithm or its improved version, such as Deep Q-Learning (DQN). These algorithms gradually optimize the scheduling strategy of the system through interaction with the environment, thereby achieving the rational allocation of resources.

[0092] In general, the task of reinforcement learning can be represented as a Markov Decision Process (MDP), which includes the following five elements: State set S: Represents the state space of the system. Specifically in this system, the state can include the real-time available state of the device, the user's reservation request, historical reservation data, etc.

[0093] Action set A: Represents the set of executable actions of the system. In this system, the actions include adjusting the reservation period, allocating different device resources, canceling conflicting reservations, etc.

[0094] Transition probability P(s,a ′ |s,a ′ ): Represents the probability that the system transfers from state s to state s ′ when the action a is executed.

[0095] Reward function R(s,a): Represents the reward value obtained by the system when executing action a in state s.

[0096] Discount factor γ: Used to weigh the relative importance of short-term rewards and long-term rewards.

[0097] In the present invention, the goal of the system is to maximize the cumulative reward value of the system in long-term operation by learning an optimal policy π * (s).

[0098] State space design of reinforcement learning In this embodiment, the state S of the system includes the following main contents: Device status data: The current availability, fault record, maintenance status, etc. of the device. Assuming s j represents the state of device j, then the device status can be represented as a vector S device =[s 1 ,s 2 ,...,s n .

[0099] User reservation request data: The user reservation requests within the current time period, including the reservation period, required device type, etc.

[0100] Scheduling historical data: The previous device reservation situations, including the allocated devices, remaining available resources, etc.

[0101] Action design of reinforcement learning The action set A of the system includes the following main operations: Adjust the reservation period allocation of the device. For example, adjust the reservation period of a certain device from the peak period to the non-peak period.

[0102] Optimize user appointment allocation. For example, dynamically adjust the user's reserved device or recommend other time slots.

[0103] Reallocate resources. For example, when a device fails, adjust the user appointments originally assigned to that device to other devices.

[0104] Design of the reward function In the reinforcement learning algorithm, the reward function is the core basis for system optimization scheduling. In this embodiment, the design of the reward function R(s,a) considers the following factors: Reward for device utilization: If an action improves the utilization rate of a device, the reward value is positive.

[0105] Reward for user satisfaction: If the system's scheduling meets the appointment needs of more users, the reward value is positive.

[0106] Conflict penalty: If an action causes user appointment conflicts, the system receives a negative reward.

[0107] Reward for time fairness: The system tries to ensure fairness in the user appointment time slots during the scheduling process.

[0108] Specifically, the reward function can be expressed as: R(s,a) = α 1 ·U utilization +α 2 ·U satisfaction -α 3 ·P conflict +α 4 ·F fairness Where: U utilization Represents the reward for the improvement of device utilization; U satisfaction Represents the reward for the improvement of user satisfaction; P conflict Represents the penalty for user appointment conflicts; F fairness Represents the reward for time fairness; α 1 ,α 2 ,α 3 ,α 4 Are weight coefficients that adjust the relative importance of different factors.

[0109] Update process of Q-learning In this embodiment, the core of the Q-learning algorithm is to gradually approximate the optimal policy by iteratively updating the Q value. The Q value update formula is: Wherein: Q(s,a) is the expected cumulative reward for taking action a in the current state s; η is the learning rate, which is used to control the step size of each update; R(s,a) is the immediate reward for the current action; γ is the discount factor, which is used to balance short-term and long-term rewards; max a′ Q(s ′ ,a ′ ) represents the maximum Q value of all possible actions in the next state s ′ next.

[0110] In a possible implementation, the system continuously executes the Q-learning algorithm to dynamically adjust the reservation allocation strategy of the device in real time, so as to maximize the cumulative reward obtained in the long-term operation.

[0111] Real-time dynamic adjustment The reinforcement learning dynamic scheduling module can dynamically adjust the scheduling strategy according to the real-time changes of the device status and user requirements. For example, when a certain device becomes unavailable due to a fault, the system will recalculate the device allocation plan through the reinforcement learning module and adjust the users who originally reserved the device to other available devices. At the same time, the system can also dynamically adjust the allocation of reservation time periods according to the changes in user requirements. For example, more reservation time periods are allocated to popular devices during peak periods.

[0112] Module cooperation and data interaction The reinforcement learning dynamic scheduling module closely cooperates with other modules of the system. The data acquisition and processing module provides real-time device status and user reservation request data; the deep learning prediction module provides prediction information on future device requirements; the preliminary scheduling plan of the reservation scheduling engine module provides a basis for optimization for the reinforcement learning module. Under the combined action of these modules, the reinforcement learning module can optimize the scheduling plan in real time to ensure the efficient operation of the system in a dynamic environment.

[0113] In this embodiment, by introducing the reinforcement learning dynamic scheduling module, the system is equipped with the ability of real-time adaptive scheduling. Through the Q-learning algorithm or its improved version, the system can adjust the reservation strategy according to real-time data, optimize the allocation of device resources, and improve the satisfaction rate of user requirements. The dynamic learning and adaptive ability of this module provide important support for the efficient operation of the laboratory equipment intelligent reservation management system.

[0114] The real-time feedback and dynamic adjustment module is used to monitor the system operation status, obtain the changes in user requirements and device status in real time, and dynamically adjust the scheduling strategy according to the feedback; In this embodiment, the real-time feedback and dynamic adjustment module obtains real-time data through various technical means, and combines the feedback information with the current scheduling plan to dynamically adjust the reservation arrangement of the device. Through this module, the system can handle complex reservation conflicts, equipment failures, and dynamic changes in user requirements, thus ensuring the reasonable utilization of equipment resources and the timely satisfaction of user needs.

[0115] Data Sources and Processing of Real-Time Feedback Generally, the data sources of the real-time feedback and dynamic adjustment module include device status information, user behavior data, and system operation logs. These data are transmitted to this module in real time by the data acquisition and processing module, and after being cleaned and formatted, they are used for further dynamic adjustment calculations.

[0116] As an option, the device status information may include the availability of the device, the fault status, the maintenance record, and the real-time load of the device. For example, assume that the status of device E j is described by the following parameters: A j : The availability of device j. If the device is in a reservable state, then A j = 1, otherwise A j = 0; F j : The fault status of device j. If the device is faulty, then F j = 1, otherwise F j = 0; L j : The real-time load of device j, representing the current utilization rate of the device.

[0117] In a possible implementation, the device status can be comprehensively represented by the following formula: S j = A j ·(1 - F j )·(1 - L j ) Where, S j represents the comprehensive status value of device j, and the system can determine whether the device can continue to accept reservation requests based on this value.

[0118] Implementation of Dynamic Adjustment Specifically, after receiving the status data, the real-time feedback and dynamic adjustment module will compare it with the current reservation scheduling plan to identify the parts that need to be adjusted. The adjustment methods include reallocating reservation resources, adjusting the reservation period, or regenerating the scheduling plan.

[0119] In this embodiment, the calculation process of dynamic adjustment is based on the following main factors: Device Status Change: When a device becomes unavailable due to a fault or maintenance, the system needs to reassign the originally scheduled users to other available devices. For example, if device E j has a fault during time period t, resulting in S j = 0, the system will, through the real-time feedback and dynamic adjustment module, adjust the original reservation assigned to E j to device E k , where E k satisfies the following conditions: where availability k represents the remaining reservable resources of device k.

[0120] User Behavior Feedback: When a user modifies a reservation request (such as changing the reservation time period or canceling the reservation), the system dynamically updates the scheduling plan according to the new user requirements. As a possible implementation, the system can preferentially reassign resources to high-priority users to meet their adjusted reservation needs.

[0121] Peak Demand Adjustment: When the system detects a rapid increase in the reservation demand for a certain device, the real-time feedback and dynamic adjustment module can adjust the reservation arrangement of the device in advance. For example, if the reservation requests for device E j reach a peak within the time period from t to t + n, the system can reassign reservation resources through the following optimization objective: where x i,j,t indicates whether user i reserves device j during time period t, and I and J represent the total number of users and devices respectively.

[0122] Adjustment Mechanism and Feedback Loop In a possible implementation, the real-time feedback and dynamic adjustment module forms a closed-loop feedback mechanism with other modules. Specifically, when the system receives real-time feedback data, the dynamic adjustment module calculates the optimal adjustment plan based on the current state and passes the adjustment result to the reservation scheduling engine module. The updated plan of the scheduling engine module is fed back to the user, and at the same time, the dynamic adjustment module continuously monitors the user feedback and device status to ensure the feasibility and effectiveness of the adjustment plan.

[0123] For example, in some embodiments, when the status of device E j changes from available to unavailable, the real-time feedback module immediately updates the status of the device to S j = 0 and triggers the dynamic adjustment process. The system will first look for an alternative device E kMeet the user's needs and at the same time notify the affected users of the adjusted reservation plan. If the user is not satisfied with the adjustment result, the system will further optimize the allocation plan to ensure the user experience.

[0124] Real-time and efficiency of dynamic adjustment In this embodiment, the real-time feedback and dynamic adjustment module can complete the processing of status data and dynamic adjustment calculations within milliseconds. By introducing parallel computing and optimized algorithms, the system can quickly respond to dynamic changes in complex environments, avoiding waste of device resources or excessive waiting time for users.

[0125] As a possible technical extension, this module can also combine the output results of the deep learning prediction module to predict upcoming demand fluctuations. For example, when the prediction results show that the demand for a certain device may exceed the available capacity in the next few time periods, the dynamic adjustment module can optimize the reservation allocation in advance to reduce possible conflicts.

[0126] To ensure the public integrity of the technical solution, all formula parameters involved in this embodiment have been completely defined. Including: S j : The comprehensive status value of device j; A j : The availability of device j; F j : The fault status of device j; L j : The real-time load of device j; x i,j,t : Whether user i reserves device j at time period t; R: The optimization goal of reservation resource allocation.

[0127] Through the real-time feedback and dynamic adjustment module, the intelligent reservation management system for laboratory equipment of the present invention realizes efficient real-time response capabilities and flexible dynamic scheduling strategies. This module can handle device status changes, user demand adjustments, and reservation conflicts during peak periods, and through cooperation with other modules, ensure the efficient operation and stability of the system in complex environments. At the same time, this module combines real-time feedback and prediction information to achieve refined management of resource allocation and continuous optimization of system performance.

[0128] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, it can be understood that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. Laboratory equipment intelligent reservation management system, characterized by: include: Data collection and processing module, used to collect device status data, user reservation request data and historical usage data in real time, and clean, standardize and store the collected data; The appointment scheduling engine module is used to schedule equipment appointments based on a multi-objective optimization algorithm to maximize equipment utilization, minimize appointment conflicts, and optimize time allocation; Game theory decision module, used to simulate multiple users playing games during the reservation process, calculate the optimal reservation plan based on Nash equilibrium theory, avoid conflicts and improve resource allocation efficiency; The deep learning prediction module is used to train and generate demand prediction models based on historical data, predict future equipment usage needs, and provide dynamic adjustment information for the scheduling engine module; Reinforcement learning dynamic scheduling module, which is used to automatically adjust the device reservation strategy based on real-time data and system feedback, optimize resource allocation and adapt to changes in user needs; The real-time feedback and dynamic adjustment module is used to monitor the system operation status, obtain user needs and equipment status changes in real time, and dynamically adjust the scheduling strategy based on feedback.

2. The laboratory equipment intelligent reservation management system according to claim 1 is characterized in that: The data collection and processing module collects device status data in real time through Internet of Things technology. The device status data includes the current availability, fault records, maintenance status and device usage time of the device; the user reservation request data includes the user's reservation period, device type, reservation duration and user priority information.

3. The laboratory equipment intelligent reservation management system according to claim 1 is characterized in that: The reservation scheduling engine module performs reservation scheduling for devices based on a multi-objective constraint optimization algorithm. The optimization objectives include but are not limited to resource utilization, service quality, time fairness and number of reservation conflicts. The constraints include device capacity, user reservation requirements and device available time period.

4. The laboratory equipment intelligent reservation management system according to claim 1 is characterized in that: The game theory decision module is used to calculate the utility value of each user under a given resource state and determine the optimal reservation strategy based on the Nash equilibrium principle. The utility function includes factors such as device availability, user reservation waiting time, cost and priority.

5. The laboratory equipment intelligent reservation management system according to claim 1 is characterized in that: The deep learning prediction module uses a long short-term memory network or a gated recurrent unit model to generate prediction results of future equipment demand based on historical reservation data, and the prediction results are provided to the reservation scheduling engine module to optimize equipment reservation scheduling.

6. The laboratory equipment intelligent reservation management system according to claim 1 is characterized in that: The reinforcement learning dynamic scheduling module is based on the Q-learning algorithm. During the peak period of equipment reservation or when demand fluctuates, the scheduling strategy is continuously adjusted through real-time data and system feedback to optimize equipment utilization and resource allocation. The strategy update is based on real-time system status and feedback information of user reservation behavior.

7. The laboratory equipment intelligent reservation management system according to claim 1 is characterized in that: The real-time feedback and dynamic adjustment module can collect equipment failure information, user demand changes, actual usage rate of reservation time periods and changes in system status in real time, and dynamically adjust the scheduling strategy of the scheduling engine based on this information to ensure efficient allocation of resources.

8. The laboratory equipment intelligent reservation management system according to claim 1 is characterized in that: The management system works through the following steps: S1. Collect device status data and user reservation request data in real time through the data acquisition and processing module, and perform data cleaning, standardization and storage; S2. Schedule the equipment based on the multi-objective optimization algorithm and constraint conditions through the appointment scheduling engine module; S3, through the game theory decision module, simulate the reservation behavior between users and calculate the optimal reservation plan; S4. Generate a forecast of future equipment usage demand through the deep learning prediction module and feed it back to the appointment scheduling engine module; S5. Optimize the system's scheduling strategy through the reinforcement learning module so that the reservation strategy can be adjusted when equipment demand fluctuates; S6. Process the system feedback information through the real-time feedback and dynamic adjustment module and dynamically adjust the scheduling plan.

9. The laboratory equipment intelligent reservation management system according to claim 1, characterized in that: The device status data includes the current availability of the device, maintenance records, fault logs and device usage time, and the user reservation request data includes reservation period, reservation time, type of reserved device and user priority information.

10. The laboratory equipment intelligent reservation management system according to claim 1, characterized in that: The system can adjust the reservation plan through the game theory decision module and the reinforcement learning module when user reservation requests conflict, so as to avoid conflicts and ensure fair allocation of users' reservation time slots.

Images