Intelligent decision-making method and system based on computer network integrated industrial PLC controller

Abstract

The invention relates to the technical field of intelligent decision making, and discloses an intelligent decision making method and system based on a computing network integrated industrial PLC, and the method comprises the steps: carrying out the time slice distribution of control parameters of water supply equipment, and dividing the decision period of the PLC, and obtaining an initial control parameter queue; determining occurrence probability distribution of a normal operation state, an efficient control state and a decision conflict state to obtain a water supply parameter conflict probability table; redistributing the parameter control time resource to obtain a water supply parameter optimization control time sequence; converting the parameter groups which still have conflicts into mutually independent control variables, and respectively executing optimization calculation to obtain a multi-parameter parallel control instruction of the water supply equipment; and carrying out dynamic resource scheduling on the multi-parameter parallel control instruction of the water supply equipment to generate an energy-saving control scheme of the water supply equipment. The method effectively optimizes the dynamic distribution of control resources, remarkably shortens the response time of control parameters, and improves the overall throughput of the system.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent decision-making, and particularly to an intelligent decision-making method and system based on a computing-network integrated industrial PLC controller. Background Art

[0002] Traditional PLC controllers use fixed priority or polling methods for resource scheduling. When multiple water supply control parameters (such as water pressure, flow rate, water level, and water quality) compete for limited controller processing resources simultaneously, it often leads to decision conflicts and control delays. Especially in the actual operation of the water supply system, the request frequencies and priorities of various control parameters change dynamically, and the fixed scheduling strategy cannot adapt to such a complex working condition environment, often resulting in high-priority parameters not being processed in a timely manner or low-priority parameters not being responded to for a long time.

[0003] The computing-network integration technology provides a new idea for solving the above problems, but there are still many challenges in practical applications. There are complex correlations among the parameters in the water supply equipment control system, which makes the traditional independent parameter optimization method unable to effectively handle the conflicts and coordination problems among the parameters; secondly, the water supply equipment is usually distributed in different locations and has different energy states, and how to minimize the system energy consumption while ensuring the control performance has become a difficult problem to be solved; furthermore, when the system scale expands, the number of control parameters increases sharply, and how to achieve efficient multi-parameter parallel decision-making with limited computing resources is also a bottleneck of traditional methods. Summary of the Invention

[0004] The present invention provides an intelligent decision-making method and system based on a computing-network integrated industrial PLC controller. The present invention effectively optimizes the dynamic allocation of control resources, significantly shortens the response time of control parameters, and improves the overall throughput of the system.

[0005] In the first aspect, the present invention provides an intelligent decision-making method based on a computing-network integrated industrial PLC controller. The intelligent decision-making method based on a computing-network integrated industrial PLC controller includes: Performing time slice allocation for the control parameters of the water supply equipment and dividing the decision-making cycle of the PLC controller to obtain an initial control parameter queue; Determining the probability distribution of the occurrence of the normal operation state, the efficient control state, and the decision conflict state according to the initial control parameter queue to obtain a water supply parameter conflict probability table; Reallocating the parameter control time resources based on the water supply parameter conflict probability table to obtain an optimized control time sequence for the water supply parameters; Converting the parameter groups still having conflicts in the optimized control time sequence of the water supply parameters into independent control variables, and respectively performing optimization calculations on each independent control variable to obtain multi-parameter parallel control instructions for the water supply equipment; Perform dynamic resource scheduling on the multi-parameter parallel control instructions for the water supply equipment to generate an energy-saving control scheme for the water supply equipment.

[0006] In a second aspect, the present invention provides an intelligent decision-making system based on an arithmetic-network integrated industrial PLC controller. The intelligent decision-making system based on the arithmetic-network integrated industrial PLC controller includes: A first allocation module, configured to perform time slice allocation on the control parameters of the water supply equipment and divide the decision-making cycle of the PLC controller to obtain an initial control parameter queue; An analysis module, configured to determine the probability distribution of the occurrence of the normal operation state, the efficient control state, and the decision conflict state according to the initial control parameter queue to obtain a water supply parameter conflict probability table; A second allocation module, configured to re-allocate the parameter control time resources based on the water supply parameter conflict probability table to obtain an optimized control time sequence for the water supply parameters; An optimization calculation module, configured to convert the parameter groups still having conflicts in the optimized control time sequence of the water supply parameters into independent control variables, and perform optimization calculations on each independent control variable respectively to obtain multi-parameter parallel control instructions for the water supply equipment; A resource scheduling module, configured to perform dynamic resource scheduling on the multi-parameter parallel control instructions for the water supply equipment to generate an energy-saving control scheme for the water supply equipment.

[0007] In the technical solution provided by the present invention, through probability modeling based on the Bayesian and Poisson distribution models for precise time slice allocation, the dynamic allocation of control resources is effectively optimized, the response time of the control parameters is significantly shortened, and the overall throughput of the system is improved; the FastICA independent component analysis technology is used to convert the conflict parameters into independent control variables, realizing the parallel optimization processing of multiple conflict parameters within the same time slice, greatly reducing the parameter processing conflict rate, and improving the resource utilization efficiency of the PLC controller; combined with the dynamic resource scheduling mechanism of the energy state of the water supply equipment, multi-objective optimization of control performance and energy consumption is achieved through the Q-learning reinforcement learning algorithm, reducing the total system energy consumption on the premise of ensuring control accuracy and extending the service life of the water supply equipment; the time resource re-allocation strategy based on parameter correlation analysis and conflict sensitivity index enhances the adaptability of the system to parameter request fluctuations, and can still maintain stable operation even in the case of a sudden increase in the parameter request frequency; through the task allocation mechanism corrected by the energy factor, the calculation load balance considering the remaining energy state of the equipment is realized, effectively avoiding the problem of single-point equipment resource overload; the parameter grouping strategy and multi-parameter parallel control architecture based on the graph coloring algorithm improve the number of water supply equipment that can be managed by a single arithmetic-network integrated PLC controller at the same time, and significantly enhance the expansion ability of the system.

[0008] Other features and advantages of the present invention will be set forth in the following description, and in part will be obvious from the description, or may be learned by practice of the present invention. The objectives and other advantages of the present invention are realized and attained by the structure particularly pointed out in the description, claims, and drawings.

[0009] To make the above objectives, features, and advantages of the present invention more obvious and understandable, the following specific preferred embodiments are given in conjunction with the accompanying drawings and are described in detail as follows. Brief Description of the Drawings

[0010] Figure 1 It is a schematic diagram of an embodiment of the intelligent decision-making method based on the computing-network integrated industrial PLC controller in an embodiment of the present invention; Figure 2 It is a schematic diagram of an embodiment of the intelligent decision-making system based on the computing-network integrated industrial PLC controller in an embodiment of the present invention. Detailed Embodiment

[0011] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0012] The terms "including" and "having" and any variations thereof mentioned in the embodiments of the present invention are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device including a series of steps or units is not limited to the listed steps or units, but optionally further includes other unlisted steps or units, or optionally further includes other steps or units inherent to these processes, methods, products, or devices.

[0013] For ease of understanding of this embodiment, first, a detailed introduction is given to an intelligent decision-making method based on the computing-network integrated industrial PLC controller disclosed in the embodiments of the present invention. As Figure 1 shown, the method includes the following steps: 101. Perform time slice allocation and PLC controller decision cycle division on the control parameters of the water supply equipment to obtain an initial control parameter queue; It can be understood that the execution subject of the present invention can be an intelligent decision-making system based on the computing-network integrated industrial PLC controller, or a terminal or a server. Specifically, it is not limited here. In the embodiments of the present invention, the server is taken as the execution subject for illustration.

[0014] Specifically, the control parameters in the water supply system are collected. The initial collection process includes core parameters such as water pressure, flow rate, water level, and water quality that affect the system operation, forming an initial water supply parameter set. For each parameter in this set, a parameter priority value is assigned according to its importance in system control. This priority value is quantified in the form of an integer from 1 to 10, and the higher the value, the higher the urgency or importance of the control. At the same time, a maximum allowable delay time value for each parameter is set, which is in milliseconds and is used to restrict the longest acceptable response time between the request and execution of the parameter. To more accurately reflect the computing load of the PLC controller when processing different parameters, a computing resource occupancy rate value is assigned to each parameter, which is represented by the percentage of CPU resources required by the parameter within the control cycle. These three types of data together constitute the parameter attribute matrix. Each row in the matrix represents a parameter, and each column corresponds to the priority value, the maximum allowable delay time, and the computing resource occupancy rate respectively. Based on the parameter attribute matrix, the initial time slice allocation for each parameter is completed through a refined calculation formula. This calculation process performs a weighted processing on the priority value and the resource occupancy rate, then multiplies by the length of the entire PLC control cycle, and combines with the system adjustment coefficient for normalization processing to ensure that the time slice allocation can reflect the importance of the parameter without exceeding the total scheduling cycle. After the calculation is completed, the parameter time slice allocation result is obtained. The initial water supply parameter set is sorted according to the parameter priority value to generate an initial control instruction set. Each instruction set entry contains the control parameter identifier, the corresponding time slice length, and its control logic command set. At the same time, a system state matrix is established to record the operating parameter values, current energy state, and communication load information of each water supply device. The matrix structure is in the form of a two-dimensional array, where the rows represent the devices and the columns represent the device parameters, such as operating frequency, motor power, pump status, etc. All elements are collected and updated in real time through the industrial fieldbus to ensure the timeliness and accuracy of the matrix data. The parameter time slice allocation result, the initial control instruction set, and the device operating state matrix are stored in the shared memory area of the PLC controller to obtain the initial control parameter queue.

[0015] 102. Determine the probability distribution of the occurrence of the normal operation state, the high-efficiency control state, and the decision conflict state according to the initial control parameter queue to obtain the water supply parameter conflict probability table; Specifically, based on the initial control parameter queue, through the unique identifier of the control parameter, match the data recorded in previous control cycles in the database, and extract the historical data sets including the normal operation state, the efficient control state, and the state of scheduling conflicts. These data reflect the true relationship between the system states of each parameter in actual operation and their corresponding control behaviors. Analyze the historical data sets, and use the conditional probability statistical method to evaluate the possibility of each control parameter appearing in different system states, and reveal the correlation between the parameter and the system state. At the same time, according to the request frequency of each parameter in the initial control parameter queue in the time slice, construct a probability distribution model of different numbers of requests occurring per unit time, reflecting the intensity and fluctuation trend of the parameter scheduling requests, so that the system can predict the possibility of multiple parameters concentrating on requesting control resources in a specific time slice. Based on the conditional probability value of the parameter state and the probability distribution of the parameter request frequency, calculate the probability value of decision conflict occurring in each time slice. According to the distribution characteristics of the current parameter queue, combined with historical conflict cases, judge which time slices are more likely to have control conflicts between parameters. When there are multiple high-priority or high-request-frequency parameters in a certain time slice, and there have been multiple resource competitions between these parameters in history, then this time slice is included in the conflict time slice sequence and used as the key monitoring object for scheduling. In order to reveal the conflict relationship between parameters, based on the resource allocation situation in the time slice, construct a parameter conflict probability matrix. In this matrix, the element value between each pair of parameters represents the possible degree of their control conflict in the same time slice. On this basis, calculate the conflict sensitivity of each control parameter, which reflects the tendency of a parameter to generate conflicts in the entire scheduling system. Integrate the conflict sensitivity value of each parameter with the conflict probability matrix to form a water supply parameter conflict probability table.

[0016] In this embodiment, the actual request times of each control parameter within different time slices are extracted from the historical dataset. Based on a large number of operation records accumulated during the long-term operation of the water supply control system, by analyzing the time points when each parameter is scheduled and requested in each control cycle, the request distribution characteristics of the parameter under different time slices are restored, forming a preliminary parameter request frequency dataset, which reflects the real-time requirements and call density of the parameter. Pattern recognition processing is performed on the above parameter request frequencies, the request frequencies are clustered and grouped according to time slices, and the change trends presented under different time periods and different operation scenarios are extracted to obtain the request frequency change pattern. Based on the request frequency change pattern, the Poisson distribution parameter λ value of each parameter under different operation scenarios is determined, and this parameter is used to describe the average number of times a parameter request occurs per unit time. Through this λ value, the probability of k requests arriving simultaneously in each time slice is quantitatively calculated, thereby generating a parameter request quantity probability matrix. Each element of this matrix represents the probability that a specific parameter request reaches a certain frequency within a certain time slice. The time slices and parameter combinations with probability values exceeding the set threshold in the parameter request quantity probability matrix are marked, and the combinations identified as having potential high conflict risks in scheduling are summarized into a high conflict risk time slice list. This list serves as a key monitoring area for the operation scheduling of the control system, reflecting the critical time windows for problems such as concentrated requests, priority conflicts, or response delays in the short term. Based on this list, the control parameter priority information is introduced, and the intersection structure of different priority parameters in the same time slice is modeled to construct a time slice resource competition distribution map. This distribution map represents the scheduling resource contention relationship in a graph structure, where its nodes represent parameters and the edge weights represent the conflict probability or competition intensity, which helps to identify the core nodes and critical paths most prone to contention. The time slice resource competition distribution map and the parameter request quantity probability matrix are fused, and through the joint evaluation of the node priority weights in the graph and the request probabilities in the matrix, a parameter request frequency probability distribution is formed.

[0017] 103. Reallocate the parameter control time resources based on the water supply parameter conflict probability table to obtain an optimized control time sequence for the water supply parameters; Specifically, extract the parameter conflict matrix and parameter priority values from the water supply parameter conflict probability table. By analyzing the conflict probability between each parameter and other parameters in the conflict matrix, and performing weighted aggregation of the conflict values and the priority values of the corresponding parameters, calculate the conflict sensitivity index of each parameter, and obtain the conflict sensitivity index set of the entire control parameter set. Calculate the deviation rate from the average conflict sensitivity index based on the parameter conflict sensitivity index set to quantify the relative prominence of each parameter in terms of conflict risk. To enable the resource adjustment to have flexible control capabilities, multiply each deviation rate by a preset adjustment intensity factor, which is set according to the requirements of the control system for dynamic adjustment sensitivity and fluctuates within a reasonable range. The product result is the time slice adjustment coefficient for this parameter, indicating the ratio by which its current time slice should be enlarged or reduced. Using the time slice adjustment coefficient, correct the time slice values allocated according to priority and resource requirements in the initial stage to obtain the adjusted time slice allocation values for each parameter. The adjustment process ensures that parameters with high conflict sensitivity obtain more control resources to avoid scheduling jams, and ensures that the total amount of time slice allocation does not exceed the upper limit of the decision-making cycle, maintaining the system scheduling balance. This correction result establishes a dynamic mapping relationship between the time slice resources and the parameter conflict risk. Based on the data in the water supply parameter conflict probability table, construct a parameter relationship topology graph, with each control parameter as a node in the graph. If there is a conflict relationship between any two parameters, establish a connecting edge between the nodes and set the weight of the edge to the corresponding conflict probability value. Perform conflict avoidance processing on this topology graph. According to the high and low of the edge weights, identify parameter pairs with conflict probabilities higher than the preset threshold, and use the forced grouping strategy to allocate these parameters to different time slice groups to ensure that they do not request resources simultaneously in the same time slice, reducing the probability of scheduling conflicts. Through the dual mechanisms of time slice correction and parameter topology reconstruction, output the optimized control timing of the water supply parameters.

[0018] 104. Convert the parameter groups still having conflicts in the optimized control timing of the water supply parameters into independent control variables, and perform optimization calculations on each independent control variable respectively to obtain the multi-parameter parallel control instructions for the water supply equipment; Specifically, parameter groups with a significant conflict probability still existing within the same time slice are identified from the optimized control timing of water supply parameters. These parameter groups could not be fully decoupled during previous optimized scheduling due to their complex coupling and high-frequency request behaviors, and are marked as conflict parameter sets. Based on this, the correlation data between parameters is extracted. Based on historical operation records and actual sampling results, a parameter correlation coefficient matrix is constructed. This matrix quantifies the intensity and direction of co-variation between parameters and is an important basis for judging whether there is potential control interference. Rearrange the historical observation values of the conflict parameter groups in the format of rows for parameters and columns for time points to form an original observation data matrix. This matrix records the specific behavioral patterns of each parameter over time within multiple control cycles. On this basis, an independent component analysis algorithm is introduced to process this matrix, and a set of mutually independent control variables is extracted from the mixed observation data. By converting the original highly correlated parameter set into a set of statistically independent control factors, the coupling structure between parameters is broken. Based on each independent control variable, a weighted objective function including a control delay term and an energy consumption term is established. The control delay term reflects the time required from parameter request to execution completion, and the energy consumption term quantifies the energy load caused by control operations. The weights of the two are set according to the sensitivity to response speed and energy consumption in actual applications. To effectively solve the objective function, the gradient descent method is used to iteratively optimize each independent control variable. By continuously adjusting the value direction and step size of the control variable, it gradually converges to the optimal solution of the objective function to obtain the optimized values of the independent control variables. Convert the optimized values of the independent control variables back to the parameter space through inverse transformation using the mixing matrix established by the previous independent component analysis to ensure that the control behavior of each independent control factor on the original parameters is restored to an executable command form. Based on the restored parameter control quantities, multi-parameter control instructions for water supply equipment that can be executed in parallel within the same time slice are generated.

[0019] 105. Perform dynamic resource scheduling on the multi-parameter parallel control instructions for water supply equipment to generate an energy-saving control scheme for water supply equipment.

[0020] Specifically, construct a system state space that describes the current control environment. This state space includes the set of parameters currently in the active state, the priority values of each parameter, the specific allocation of time slices, the energy states of each water supply device, and the communication load information of each node in the system. By constructing the state space, comprehensively perceive the dynamic changes of resources involved in the control process; on this basis, define an action space, which includes a series of alternative resource scheduling operation forms, such as dynamic adjustment of resource allocation ratios, optimal selection of communication channels, temporary changes in the priorities of control parameters, etc. By mapping the system state and action content, a system state-action mapping table is formed. Real-time monitor the current energy reserve of each water supply device, and define the ratio of its current remaining energy value to the full charge state value as the energy factor. The introduction of the energy factor enables the system to dynamically adjust its ability to participate in control calculations according to the available energy of each device. Therefore, combine the energy factor of each device to correct its original basic calculation load, and obtain a calculation task allocation volume that more conforms to the actual state of the device. Construct a reward function based on the average response delay, total energy consumption, and control accuracy of the water supply device, and calculate the comprehensive performance evaluation index through the reward function. Execute the Q-learning reinforcement learning algorithm based on the system state-action mapping table, the device calculation task allocation volume, and the comprehensive performance evaluation index. During the training process, the system continuously establishes an optimal mapping relationship between states and actions through trial-and-error learning, and gradually optimizes the Q-value table. By learning historical experience and environmental feedback, obtain the optimal control strategy. According to the optimal control strategy allocation, perform device and communication channel allocation for each control task in the multi-parameter parallel control instruction of the water supply device, and form a structured task allocation plan. Integrate the task allocation plan with the original multi-parameter parallel control instruction to form an energy-saving control plan for the water supply device.

[0021] In this embodiment, the continuous system state space and action space are discretized. State parameters originally containing a large number of floating-point numbers or continuous values, such as device energy level, communication load, control response delay, etc., are divided into several finite levels or intervals, such as high, medium, and low levels, to facilitate the encoding and indexing of state variables. Similarly, all optional actions are discretized. For example, the resource allocation ratio is set to several predefined gears, or the communication channel selection is limited to a finite number of paths, forming a two-dimensional table structure composed of states and actions. Based on this structure, a Q-value table, that is, a Q matrix, is initialized. Each row represents a system state, and each column represents a possible action. Each element in the table is initially set to the same value, which is used to record the expected reward obtained by taking a specific action in different states. When the system receives a control instruction generated by the multi-parameter parallel control module, it identifies the corresponding system state according to the current control situation, device state, and time-slot resource configuration, and combines the execution result of the control instruction with the pre-constructed comprehensive performance evaluation index system to calculate the immediate reward value. This reward value comprehensively considers the average response time, energy consumption level, and control accuracy of the system, reflects the actual benefit brought by the selected action in the current state, and serves as an important basis for updating the Q value. The system locates the corresponding state-action pair in the Q-value table corresponding to the current state and assigns this immediate reward to it. The action selection strategy is executed, and the ε-greedy method is used, that is, in some cases, the optimal action with the largest Q value in the current state is selected with a higher probability, and in other cases, a random selection is made with a lower probability to achieve a balance between exploration and exploitation. By dynamically switching between maximum value selection and random exploration, the system makes full use of the existing learning results and can avoid falling into local optima and missing better solutions. After the action is selected, the system executes this action in the actual control process and calculates the Q-value update amount based on the execution feedback, such as the transition to the new state and the new reward value. This update amount is comprehensively determined by the old Q value in the current state, the immediate reward value, and the maximum Q value among all actions in the new state, and is weighted and averaged in combination with the learning rate and the discount factor to replace the corresponding position in the old Q-value table. The entire reinforcement learning process consists of four links: action selection, state observation, immediate reward evaluation, and Q-value iterative update, and is continuously looped and executed during the system operation. As experience accumulates, the values in the Q-value table gradually converge and approach the long-term expected return in the real environment. After the Q value converges, the action with the largest Q value is extracted from the Q-value table for each discretized system state as the optimal operation selection in this state. The set of optimal actions in all states constitutes the output of the optimal control strategy.

[0022] In the embodiments of the present invention, through probability modeling based on Bayesian and Poisson distribution models for precise time slice allocation, the dynamic allocation of control resources is effectively optimized, the response time of control parameters is significantly shortened, and the overall system throughput is improved; the FastICA independent component analysis technology is used to transform conflict parameters into independent control variables, realizing parallel optimization processing of multiple conflict parameters within the same time slice, greatly reducing the parameter processing conflict rate, and improving the resource utilization efficiency of the PLC controller; combined with the dynamic resource scheduling mechanism of the water supply equipment energy state, multi-objective optimization of control performance and energy consumption is achieved through the Q-learning reinforcement learning algorithm, reducing the total system energy consumption while ensuring control accuracy, and extending the service life of the water supply equipment; based on the time resource reallocation strategy of parameter correlation analysis and conflict sensitivity index, the adaptability of the system to parameter request fluctuations is enhanced, and it can still operate stably even when the parameter request frequency suddenly increases; through the task allocation mechanism corrected by the energy factor, the computing load balance considering the remaining energy state of the equipment is realized, effectively avoiding the problem of single-point equipment resource overload; based on the parameter grouping strategy and multi-parameter parallel control architecture of the graph coloring algorithm, the number of water supply equipment that can be simultaneously managed by a single computing and networking integrated PLC controller is increased, significantly enhancing the expansion ability of the system.

[0023] In a specific embodiment, the process of executing step 101 may specifically include the following steps: Collect the control parameters in the water supply system to obtain the initial water supply parameter set, and assign a parameter priority value, a maximum allowable delay time value, and a computing resource occupancy rate value to each parameter in the initial water supply parameter set to obtain the parameter attribute matrix; Calculate the initial time slice allocation amount of each parameter according to the parameter attribute matrix to obtain the parameter time slice allocation result; Sort the initial water supply parameter set according to the parameter priority value to obtain the initial control instruction set; Establish a system state matrix including the operation parameter values, energy state values, and communication load values of each water supply equipment, and collect and update the values of each matrix element in real time through the industrial field bus to obtain the equipment operation state information; Store the parameter time slice allocation result, the initial control instruction set, and the equipment operation state information in the shared memory area of the PLC controller to obtain the initial control parameter queue.

[0024] Specifically, the control parameters of the water supply system are collected to form an initial set of water supply parameters. Through data communication with various levels of equipment in the water supply system and with the help of various sensors deployed at key positions, the system operating status is monitored with fine granularity. The collected parameters cover various key control variables that affect water supply stability and energy efficiency performance, such as water pressure, water level, instantaneous flow rate, cumulative flow rate, pump speed, valve opening, water quality indicators, as well as environmental temperature and humidity, motor temperature rise, and power consumption. All the collected raw data is formatted, denoised, and anomaly-processed by the preprocessing module to form the initial set of water supply parameters. Three types of attribute values are assigned to each control parameter in the initial set of water supply parameters, namely the parameter priority value, the maximum allowable delay time value, and the computing resource occupancy rate value. The parameter priority value is set according to the importance of the control task or automatically learned from historical control behaviors. It is graded in integer type, and the higher the value, the more urgent the control requirement for this parameter. The maximum allowable delay time value reflects the longest response time allowed between the trigger request and the execution of this parameter, and this value determines whether it is suitable for parallel scheduling with other parameters. The computing resource occupancy rate value represents the percentage of processing resources required by the PLC controller when processing the control task of this parameter, considering multiple factors including CPU time, storage access load, and communication bandwidth. By uniformly organizing the above three types of attribute values into a structured data table form, a parameter attribute matrix is constructed. Each row in this matrix corresponds to a parameter, and each column corresponds to an attribute item. The time slice allocation for all control parameters is calculated based on the parameter attribute matrix. A time slice is the operation and response time window allocated by the PLC controller for a parameter in each control cycle, and its length determines the order of parameter scheduling and the resource allocation density. During the time slice allocation process, the weighted ratio of priority and resource occupancy rate is introduced for normalization to ensure that parameters with high priority and moderate resource occupancy obtain more control time, while parameters with low priority or high resource occupancy will have their time slices moderately compressed to achieve the balance of the overall scheduling load. This process dynamically allocates the time slice length of each parameter on the premise of ensuring that the total time of the control cycle remains unchanged, forming the result of parameter time slice allocation, which is used to guide the periodic scheduling frequency and specific response time of the parameters. The initial set of water supply parameters is sorted according to the parameter priority value, and the parameters with higher priority values are ranked in the front, so that the control tasks of high-urgency parameters are preferentially processed during the scheduling process. The sorting result serves as the basis for the initial control instruction set. Each control instruction includes a parameter identifier, priority information, and also includes the allocated time slice length and a preset control strategy template, forming the core unit of the scheduling queue. This instruction set is used as the primary basis for task distribution in each control cycle and is read and executed by the PLC controller one by one for the corresponding control commands. At the same time, a system state matrix describing the current operating status of the water supply equipment is established, and the construction of this matrix depends on the data collected by the industrial field bus and the internal sensing units of each device.The rows of the matrix correspond to specific devices or functional units in the water supply system, such as water pumps, motors, water storage tanks, valve control modules, etc., and the columns represent various operating parameters, such as voltage, current, rotational speed, temperature, vibration value, communication load, remaining power, etc. Each matrix element represents the real-time value of an operating index of a device. This matrix is continuously updated through a high-speed communication interface with the PLC controller to ensure that the scheduling algorithm obtains the latest operating information. The time slice allocation result, the initial control instruction set, and the device operating status information in the system state matrix are stored in the shared memory area of the PLC controller. The shared memory serves as a common data interface among the multi-threaded scheduling module, the system status monitoring module, and the control execution module. Its structure is designed for high-speed and low-latency access and has a synchronization lock mechanism to prevent data read-write conflicts. The time slice allocation result provides a time window reference as the basic time structure for the parameter scheduling process, the initial control instruction set provides the scheduling order and specific control action instructions, and the dynamic data in the system state matrix helps to determine whether certain parameters need to be delayed, skipped, or weighted. The initial control parameter queue formed by combining the three is gradually executed by the scheduling core module according to the time trigger method.

[0025] In a specific embodiment, the process of executing step 102 may specifically include the following steps: According to the initial control parameter queue, extract the historical data set including the normal operating state, the efficient control state, and the decision conflict state from the historical database of the arithmetic-network integrated industrial PLC controller; Perform conditional probability distribution calculation on the historical data set to obtain the parameter state conditional probability value; According to the parameter request frequency within each time slice, calculate the occurrence probability of different request quantities per unit time to obtain the parameter request frequency probability distribution; Based on the parameter state conditional probability value and the parameter request frequency probability distribution, calculate the probability value of decision conflict occurring within each time slice to obtain the conflict time slice sequence; Construct a parameter conflict probability matrix, calculate the probability value of decision conflict occurring between each pair of parameters within the same time slice to obtain the parameter conflict relationship within the time slice; Calculate the conflict sensitivity value for each parameter in the conflict time slice sequence to obtain the parameter conflict sensitivity sequence, and merge the parameter conflict sensitivity sequence with the parameter conflict relationship within the time slice into the water supply parameter conflict probability table.

[0026] Specifically, by controlling all the parameter identifiers included in the parameter queue, the operation record data related to them is extracted from the historical database embedded in the computing-network integrated industrial PLC controller, and the extracted data is classified according to the operation result tags marked in the records. A large amount of status data from the actual operation process has been accumulated in this database, and the system response results under each control cycle are archived, including three categories: normal operation status, efficient control status, and decision conflict status, corresponding to the situations where the system successfully completes the control according to the plan, realizes efficient resource allocation under the action of the optimization strategy, and the response is abnormal due to parameter concurrent conflicts or control resource competition respectively. By screening and matching the data of these different states according to the parameters, a multi-state historical data set covering all key parameters is obtained. Analyze the distribution of each parameter in the data set under different system states, and use the conditional probability calculation method to calculate the state probability values of each parameter under specific state conditions. By analyzing the relationship between the historical state of the parameter and its operating environment, clarify the probability distribution of it being in the normal, efficient, or conflict state. This analysis takes the parameter as the known condition and the system state as the result event, and examines the frequency of different operating states of the system under the given parameter value or scheduling condition. A set of conditional probability values formed reflects the influence degree of each parameter on the operation stability of the system. At the same time, based on the request frequency of each parameter in each time slice, deduce the occurrence probability of the request quantity per unit time. Call the data of the parameter request scheduling record of the PLC controller in each control cycle, count how many parameters request control resources simultaneously in each discrete time slice, and record the frequency distribution of this request quantity. Through statistical analysis, form a panoramic map of the scheduling density within the time slice, that is, depict the probability trend of concurrent requests under different control load situations. Convert these statistical results into a parameter request frequency probability distribution function, which reflects the relationship between the concurrent request quantity per unit time and its occurrence probability. Combine the conditional probability values of the parameter state with the probability distribution of the parameter request frequency to calculate the overall probability of potential decision conflicts in each time slice. Taking the time slice as the unit, based on the analysis of the parameter request frequency within this time period, introduce the parameter state characteristics to judge the possibility of their joint triggering of conflicts. When high-frequency requests and high-conflict sensitive states appear simultaneously, this time slice is regarded as a key scheduling window with high conflict risk. Through this mechanism, a sequence of conflict time slices is generated. This sequence lists all the time slice indices with high conflict probabilities, and marks the main parameters involved in the conflict in each time slice. In order to quantify the probability relationship of scheduling conflicts between parameters within a specific time slice, construct a parameter conflict probability matrix based on the data in the conflict time slices. In this matrix, the association between each pair of parameters is quantified as the historical probability of jointly triggering control conflicts in the same time slice. The higher the value of the matrix element, the higher the frequency of mutually exclusive or competitive behaviors of this parameter pair in the scheduling.This matrix has the characteristics of symmetry and sparsity. Based on this, a conflict graph of parameters is established for subsequent graph coloring optimization and parallel scheduling strategy design. For each participating parameter in the conflict time slice sequence, its conflict sensitivity value is calculated. This value is obtained by statistically calculating the conflict probability with other parameters while introducing the importance of the parameter itself, that is, the priority factor, as a weight factor for weighted summation. The resulting value reflects the centrality degree of this parameter in controlling resource contention. The higher the conflict sensitivity value, the more likely it is that this parameter not only easily forms conflicts with other parameters, but also has a higher priority in the scheduling queue. Therefore, higher attention and scheduling priority guarantee are required during the resource scheduling process. By calculating the sensitivity values for all parameters, an ordered sequence of parameter conflict sensitivities is obtained. The conflict sensitivity sequence is fused with the parameter conflict probability matrix to form a water supply parameter conflict probability table.

[0027] In a specific embodiment, the process of performing the step of calculating the occurrence probability of different request quantities per unit time according to the parameter request frequency within each time slice to obtain the parameter request frequency probability distribution may specifically include the following steps: Count the request times of each control parameter in different time slices from the historical dataset to obtain the parameter request frequency; Group the parameter request frequencies to obtain the request frequency change pattern; Based on the request frequency change pattern, determine the Poisson distribution parameter λ value of each parameter in different operating scenarios, and substitute the Poisson distribution parameter λ value into the Poisson distribution formula to calculate the probability distribution of k requests arriving simultaneously to obtain the parameter request quantity probability matrix; Mark the time slices and parameter combinations in the parameter request quantity probability matrix whose probability values exceed the set threshold to obtain a list of high-conflict-risk time slices; Use the list of high-conflict-risk time slices and the parameter priority values to calculate the time slice resource competition distribution map; Combine the time slice resource competition distribution map with the parameter request quantity probability matrix and output the parameter request frequency probability distribution.

[0028] Specifically, by traversing the scheduling records of each time slice within the historical control period, information on whether control requests are initiated for each parameter within each time slice is extracted and archived according to the time slice, forming a three-dimensional data set containing control parameters, time slice indices, and request counts. The statistical process considers the scheduling logs of the control system under different operating loads, external supply and demand conditions, and device states. Therefore, the request frequencies of the same parameter in different time periods are accumulated and normalized to obtain the request frequency of each parameter within different time slices, that is, the average number of times the parameter initiates a scheduling request within a unit time slice. The parameter request frequencies are grouped to identify the change patterns of parameter request behaviors. This grouping process is classified according to the request frequency magnitude of the parameter, the time slice index, and the operating environment state, forming several representative request frequency pattern clusters. Through clustering algorithms or threshold segmentation methods, parameters with obvious change characteristics are classified into the same request frequency change pattern, and each pattern represents the typical behavior characteristics of a certain type of parameter under specific operating scenarios. Based on the statistical characteristics within each frequency change pattern, the Poisson distribution parameter λ value under different operating scenarios is calculated. The Poisson distribution, as a statistical tool for describing the probability of discrete events occurring within a unit time, is suitable for simulating the behavior pattern of control parameters initiating requests within a unit time slice. The mean value of the time slice request data included in each type of frequency change pattern is calculated as the basis for estimating the λ value, and the expected value of each parameter initiating a request within each time slice under a specific operating state is obtained. These λ values are substituted into the standard Poisson distribution expression to solve the occurrence probabilities of different request quantities k, generating a parameter request quantity probability matrix. The matrix has the time slice as the horizontal axis and the request quantity k as the vertical axis, and the value within each matrix cell represents the probability estimate value that the parameter request quantity reaches k times within that time slice. The time slices and parameter combinations with probability values exceeding the set threshold in the parameter request quantity probability matrix are marked. When the request quantity probability value corresponding to a certain time slice in the matrix exceeds the threshold, it indicates that within this time slice, multiple parameters initiate scheduling requests intensively, thereby exacerbating resource competition and causing scheduling delays or response failures. The system marks all time slices and their associated parameters that meet this condition and summarizes them into a high-conflict risk time slice list. Using the high-conflict risk time slice list and parameter priority values, a scheduling competition relationship graph is constructed. The participating parameters in each high-risk time slice are extracted and weighted according to their priority levels, and on this basis, a time slice resource competition distribution graph is constructed. In this graph, each node represents a control parameter, and the edges between the nodes represent the conflict relationships where the parameters may simultaneously request the same scheduling resource. The weight of the edge is calculated according to the request overlap degree and priority level difference of the parameters. Through this graph, it is reflected which parameters compete most fiercely for resources and which are the key control factors of the system response bottleneck within a certain time slice.Fuse the time slice resource competition distribution map with the parameter request quantity probability matrix, and perform weighted processing through the correlation degree between the estimated value of the request quantity probability within each time slice and the resource competition intensity to correct the original probability matrix, so that it can not only reflect the frequency of scheduling requests, but also express the conflict relationship between requests and the severity of resource contention, and obtain a parameter request frequency probability distribution model.

[0029] In a specific embodiment, the process of executing step 103 may specifically include the following steps: Extract the parameter conflict matrix and the parameter priority value according to the water supply parameter conflict probability table to obtain a parameter conflict sensitivity index set; Calculate the deviation rate from the average conflict sensitivity index based on the parameter conflict sensitivity index set, and multiply the deviation rate by a preset adjustment intensity factor to obtain a time slice adjustment coefficient; Correct the initially allocated time slice value according to the time slice adjustment coefficient to obtain an adjusted time slice allocation value; Construct a parameter relationship topology graph according to the water supply parameter conflict probability table, and according to the parameter relationship topology graph, force the parameters with a conflict probability higher than a preset threshold to be assigned to different time slice groups to obtain an optimized control time sequence for the water supply parameters.

[0030] Specifically, extract the parameter conflict matrix and the priority values of each control parameter from the conflict probability table. The parameter conflict matrix is a symmetric structure, and its elements represent the probability of scheduling conflicts between any two control parameters within the same time slice; the priority value is the control importance weight assigned to each parameter in advance and is the key basis for the system to make scheduling decisions and resource allocations. By using these two pieces of information in combination, calculate the conflict sensitivity index for each parameter. This index not only considers the frequency or intensity of conflicts between this parameter and other parameters, but also considers the control priority of the other parameters that conflict with it in the system. By weighted accumulation of each conflict edge and the priority of the opposite-end parameter, the conflict core degree of the parameter in the scheduling structure, that is, the conflict sensitivity index, is obtained. Calculate the deviation rate from the average conflict sensitivity index based on the set of parameter conflict sensitivity indices to measure the degree of deviation of each parameter from the average conflict sensitivity. Calculate the average value of the conflict sensitivity values of all parameters, and then calculate the relative deviation rate between each parameter and the average value, which reflects the degree of conflict intensity and the relative position of resource contention of this parameter in the overall scheduling structure. To ensure that this deviation rate plays an appropriate role in the time slice adjustment process, introduce a preset adjustment intensity factor. This factor is set according to the requirements of the actual control system for the sensitivity of resource allocation. The larger its value, the more the system tends to make large adjustments to the time slice according to the sensitivity difference; conversely, it indicates that the system pursues a relatively stable scheduling structure. Multiply the adjustment intensity factor by the deviation rate to obtain the time slice adjustment coefficient. This coefficient provides a specific time slice adjustment ratio for each parameter and determines the control window that should be expanded or reduced in the next round of resource allocation. Correct the initial time slice allocation value based on the time slice adjustment coefficient. On the premise that the total control period remains unchanged, appropriately increase the time slice quota for parameters with high sensitivity and high conflict risk to relieve their scheduling pressure. At the same time, appropriately reduce the time slices of parameters with low sensitivity or those in non-core conflict chains to release scheduling resources for key parameters. During the execution of this correction process, ensure that the sum of the time slices of all adjusted parameters still meets the set upper limit of the control period T through linear normalization or proportional balance algorithms, so as to avoid system beat imbalance or scheduling overload caused by excessive adjustment. Construct a parameter relationship topology graph based on the water supply parameter conflict probability table. In this topology graph, each node represents a control parameter, and an edge is used to connect any two parameters with a scheduling conflict relationship, and the weight of this edge is set to the conflict probability value between them. The structure of the entire graph reflects the conflict propagation path, scheduling coupling relationship, and resource competition tightness among the parameters in the water supply control system. Through this graph, identify the conflict hot spots, conflict clusters, and scheduling bottleneck paths.To achieve more thorough conflict isolation, the topological graph is subjected to structured coloring processing. All parameter pairs in the graph with a conflict probability exceeding a preset threshold are forcibly assigned to different time slice groups to ensure that they do not request control resources simultaneously in the same time slice during the scheduling process in a physically isolated manner. This operation is equivalent to performing a round of grouping processing on the control parameter set based on the conflict probability. The parameters within a group are relatively independent, and the conflict between groups is minimized, thus forming a grouped scheduling structure with clear hierarchy and non-interfering tasks in the time slice structure. The corrected time slice allocation values are integrated with the parameter grouping information obtained from the topological graph coloring to form the optimized control timing of the water supply parameters.

[0031] In a specific embodiment, the process of executing step 104 may specifically include the following steps: Extract the parameter groups marked as conflicts from the optimized control timing of the water supply parameters to obtain the parameter correlation coefficient matrix; Arrange the historical observations of the conflict parameter groups corresponding to the parameter correlation coefficient matrix with the row parameters and column time points to construct the original observation data matrix; Perform independent component analysis on the original observation data matrix to obtain mutually independent control variables, and establish weighted objective functions containing control delay terms and energy consumption terms based on each independent control variable; Perform gradient descent optimization calculation on each weighted objective function to obtain the optimized values of the independent control variables; Convert the optimized values of the independent control variables back to the parameter space to generate multi-parameter parallel control instructions for water supply equipment that can be executed in parallel within the same time slice.

[0032] Specifically, identify the parameter groups that still exhibit resource contention or control interference phenomena during actual operation from the optimized control timing. These parameters are core competing objects with each other in the resource scheduling structure, having high request frequencies, close priorities, and a certain functional dependence relationship between control targets. Cluster these parameter groups through the conflict probability marking mechanism, and extract the conflict parameter groups as the processing targets for subsequent parallel deconstruction. Quantify the linear correlation degree between these parameters in the identified conflict parameter groups to determine whether there is a mathematical basis for independent decomposition. Based on the multi-round operation records of this parameter group in the historical control cycle, construct a correlation coefficient matrix between the parameters. In this matrix, each element represents the degree of change consistency between a certain two parameters during operation, with a value range from -1 to 1. A positive value indicates positive correlation, a negative value indicates negative correlation, and the larger the absolute value, the stronger the correlation. If the correlation between the conflict parameter groups is relatively high, it indicates that there are problems such as information overlap, repeated control logic, or cross-over of resource allocation paths during scheduling execution. Modifying the time slice directly cannot completely decouple the conflict. Therefore, perform a spatial transformation on the highly correlated parameters, and re-express the parameters with multi-dimensional coupling characteristics in the original parameter space as a set of statistically independent control variables to achieve a fundamental improvement in parallel scheduling capabilities. To achieve this transformation, rearrange the original historical observation data to construct a standardized observation data matrix. Arrange the historical observation values of each parameter in the conflict parameter group by row, with each row corresponding to a parameter and each column representing different sampling time points, forming a two-dimensional observation matrix with parameters as rows and time points as columns. In this matrix, each data element represents the sampling value of a certain control parameter at a specific time point, and these values constitute the dynamic observation trajectory of the behavior of this parameter during the past operation of the system. Input this matrix into the independent component analysis algorithm for processing. Independent component analysis is an algorithm that extracts the latent factors hidden behind the observation data and statistically independent of each other from multivariate statistical data. Through this algorithm, the original parameter matrix is decomposed into two sub-matrices: one is the mixing matrix, which is used to describe how the independent variables linearly combine to generate the original parameters, and the other is the independent variable matrix, that is, the decoupled control factors. Based on each independent control variable, establish a weighted objective function that includes a control delay term and an energy consumption term. The control delay term is used to characterize the response time offset caused by the scheduling behavior corresponding to the variable during execution, and the energy consumption term reflects the consumption level of the device's electrical energy or communication resources during the execution of this control behavior. According to the operation scenario and priority strategy, set reasonable weights for the two index terms to construct the weighted objective function. Perform gradient descent optimization calculation on each weighted objective function. During the optimization process, starting from the initial estimated value of the current variable, iterate and update along the direction of the gradient descent of the objective function. At each step, make a small adjustment to the variable according to the current gradient value, and gradually approach the optimal solution of the objective function. Determine the termination of the optimization process according to the set convergence condition or the maximum number of iterations.For each independent control variable, obtain its optimized numerical expression, which represents the optimal behavior or scheduling response strategy that the control variable should adopt within the current control period. Convert the optimized independent control variable values back to the parameter space to generate specific executable parallel control instructions. Perform an inverse mapping operation through the mixing matrix recorded in independent component analysis, and re-convert each optimized independent variable into the respective parameter values in the original parameter set according to its weight in the linear combination, restoring the specific control amount of each original control parameter within the current scheduling period. Integrate the reconstructed control parameters within the same time slice to generate multi-parameter parallel control instructions for water supply equipment with a reasonable structure, independent instructions, and non-interfering resources.

[0033] In a specific embodiment, the process of executing step 105 may specifically include the following steps: Define a system state space that includes the current active parameter set, parameter priorities, time slice allocation, device energy status, and communication load, construct an action space that includes resource allocation ratio adjustment, communication channel selection, and temporary modification of parameter priorities, and obtain a system state-action mapping table; Calculate the ratio of the current energy value of each water supply device to the full charge as the energy factor, and correct the basic calculation load according to the energy factor to obtain the device calculation task allocation; Construct a reward function based on the average response delay, total energy consumption, and control accuracy of the water supply device, and calculate the comprehensive performance evaluation index through the reward function; Execute the Q-learning reinforcement learning algorithm based on the system state-action mapping table, device calculation task allocation, and comprehensive performance evaluation index to obtain the optimal control strategy; According to the optimal control strategy allocation, perform device and communication channel allocation for each control task in the multi-parameter parallel control instructions for the water supply device to obtain a task allocation plan; Integrate the task allocation plan with the multi-parameter parallel control instructions for the water supply device to generate an energy-saving control plan for the water supply device.

[0034] Specifically, define the system state space that describes the current operating state of the water supply control system, which includes the set of currently active control parameters, i.e., the parameter index set participating in the control task in the current scheduling period; the control priority values of the parameters, which represent the importance ranking in the control target sequence; the time slice allocation of the parameters, which is used to characterize the proportion of computing resources obtained during the control period; the current energy state of each control device in the water supply system, i.e., the current remaining power or battery capacity; and the network load level of the communication module in the current scheduling period, which reflects the bandwidth and delay pressure of the system in information transmission. These state variables together form a high-dimensional state vector, and each state combination represents the specific control and resource allocation environment of the water supply system at a certain moment. Construct the action space, which contains all available resource scheduling operation items. In this control system, the actions include adjusting the control resource allocation ratio, i.e., dynamically increasing or decreasing the time slice quota of the parameters according to the current parameter priority and scheduling requirements within a given period; selecting the communication channel, i.e., choosing the optimal transmission path for data interaction between different devices according to the network load and communication energy consumption; and temporarily modifying the priority of the control parameters, i.e., increasing or decreasing the priority weight of some parameters according to abnormal states or policy feedback in specific operating scenarios to affect their scheduling order. These actions constitute the set of operations executed by the system in each state. The Cartesian product of the action space and the state space forms a state-action mapping table, which is used to record the expected rewards of each action of the current policy in a specific state during the reinforcement learning process, thereby guiding the agent to learn the optimal control path. After completing the state-action modeling, introduce an energy consumption modeling mechanism to dynamically adjust the load capacity of the devices. Calculate the ratio of the current energy state of each water supply device to its rated full-charge state to obtain the energy factor, which represents the proportion of the remaining working capacity of the device. According to the energy factor, correct the basic computing load of the device in task allocation. The more energy-sufficient the device is, the higher the load weight it will obtain in task allocation. Conversely, its task priority will be reduced in scheduling to extend the overall operating time of the system and prevent some devices from interrupting work due to power exhaustion. At the same time, to guide the convergence direction of the reinforcement learning algorithm, design a reward function for evaluating the quality of the scheduling strategy. This function consists of three core indicators: the average response delay, which is used to measure the average time experienced from the initiation to the completion of the control task and reflects the real-time response ability of the system; the total energy consumption, which covers the computing consumption and communication energy consumption during the device operation and is the key objective of the system's energy-saving optimization; and the control accuracy, i.e., the error deviation between the control result and the target parameter, which is an indicator for measuring the reliability of the control effect. Set the weight coefficients of these three indicators according to the requirements of the operating environment, and combine them into a unified comprehensive performance evaluation indicator. This indicator can be calculated in real time after each state-action pair is executed and is used to update the Q-value table, which is the core feedback source in the Q-learning process.Execute the Q-learning reinforcement learning algorithm based on the system state-action mapping table, the device computing task allocation amount, and the comprehensive performance evaluation index to perform policy training. In each control cycle, the algorithm selects the action with the maximum expected return in the current state from the Q-value table according to the current state vector, or performs random exploration with a certain probability according to the ε-greedy policy. After the action is executed, calculate the actual reward value according to the control result, and then update the Q-value of the corresponding state-action pair iteratively according to the Q-learning update formula. By continuously repeating the processes of action selection, feedback acquisition, and Q-value update, the system gradually learns the optimal action combinations in various states and forms an optimal control strategy set. When the Q-value table converges stably, the system extracts its policy output as the actual execution basis for the current scheduling cycle. In the specific scheduling execution, according to the optimal policy output, allocate each control task in the multi-parameter parallel control instruction, select the device that best suits its energy consumption level and scheduling ability to execute the corresponding task, and select the optimal communication path according to the current communication network load and channel energy efficiency to complete data interaction. Integrate the task device allocation information with the original multi-parameter control instruction to form a task allocation plan. Integrate the task allocation plan with the control instruction to construct an energy-saving control plan for the water supply equipment.

[0035] In a specific embodiment, the process of executing the steps based on the system state-action mapping table, the device computing task allocation amount, and the comprehensive performance evaluation index to execute the Q-learning reinforcement learning algorithm to obtain the optimal control strategy may specifically include the following steps: Discretize the system state space and the action space to obtain an initial Q-value table; Calculate the immediate reward value according to the multi-parameter parallel control instruction of the water supply equipment and the comprehensive performance evaluation index to obtain the reward value of the current state-action pair; Switch between greedily selecting the action with the maximum Q value and randomly exploring actions according to the system state-action mapping table to obtain an action selection strategy; Execute the selected action based on the action selection strategy and calculate the Q-value update amount, iteratively update the Q-value update amount, and repeat the reinforcement learning process of action selection, observation feedback, and Q-value update to obtain the converged Q-value distribution; Extract the action with the maximum Q value corresponding to each system state based on the converged Q-value distribution and output the optimal control strategy.

[0036] Specifically, the system state space and action space are discretized to construct a finite state-action pair set suitable for Q-value learning. The system state space consists of multiple continuous variables, such as the priority of control parameters, the energy state of devices, the communication network load, the time slice resource allocation ratio, etc. The continuous variables are divided into discrete levels. For example, the priority is divided into three levels: low, medium, and high; the energy state is divided into three levels: full charge, half charge, and low charge; the communication load is divided into three levels: light, medium, and heavy, etc. In this way, the state space is compressed into a set composed of finite state vectors, and each state vector represents a typical scheduling scenario in which the system is currently located. Similarly, the resource scheduling operations involved in the action space include multiple variables, such as the adjustment direction and amplitude of the resource allocation ratio, the selection of communication channels, the fine-tuning of the priority of control parameters, etc. The parameter ranges of each action are divided, and after numbering the communication channels, the selectable ranges are limited to form a discrete action set. After the discretization of the state space and action space, an initial Q-value table, that is, a Q matrix, is constructed. The number of rows of this Q matrix is equal to the total number of states in the discretized state space, and the number of columns is equal to the number of all available action types. Each cell in the matrix represents the expected reward value of performing a specific action in a specific state. In the initial stage, all Q values are uniformly set to zero or a small constant, indicating that the system has not yet recognized the effects of each state-action pair. As the training progresses, the system continuously corrects the Q values to gradually approach the long-term rewards brought by the optimal strategy. When the system receives a multi-parameter parallel control instruction for a water supply device, it immediately evaluates the execution effect within the current control cycle. The comprehensive performance evaluation index is called, and this index consists of three parts: one is the average delay time of the control response, which measures the real-time performance of the instruction execution; the second is the total energy consumption during the control process, including the device operation energy consumption and communication energy consumption; the third is the control accuracy, that is, the error between the control result and the expected target. The above three indicators are weighted according to the pre-set weights to synthesize the immediate reward value. The higher this reward value is, the better the control effect brought by the current state-action pair is, and the more inclined the system is to repeat this action in a similar state; on the contrary, it is more inclined to try other actions. Through the calculation of the immediate reward, a quantitative feedback is formed for each action decision, so as to achieve step-by-step optimization based on experience. In the reinforcement learning process, the ε-greedy strategy is adopted for action selection, that is, in each state judgment, the system does not always select the action with the largest current Q value, but executes a random action with a certain probability. An ε value is set to represent the exploration probability. In each decision, the action with the largest Q value is selected with a probability of 1 - ε, and other actions are randomly selected with a probability of ε. This strategy effectively avoids the risk of falling into local optimality and enables the system to discover potential better scheduling paths during continuous attempts. As the training progresses, the system gradually reduces the ε value, shifting from exploration-oriented to exploitation-oriented, so that the control strategy tends to be stable and converges to the optimal solution.After the action selection is completed, the action is executed, that is, resources are allocated, priorities are adjusted, or communication paths are switched under the current control state, and the feedback results of the control system are observed. The reward value is recalculated according to the feedback results, and the Q-value of the current state-action pair is corrected according to the update formula of Q-learning. The update of the Q-value takes into account the weighted average between the current immediate reward and the future potential maximum benefit, making the Q-value gradually approach the true return after each decision. The system uses the learning rate to determine the amplitude of the Q-value update and the discount factor to determine the weight of considering future returns. These two parameters jointly affect the convergence speed and stability of the Q-table. By continuously repeating the processes of state recognition, action selection, execution feedback, and Q-value update, the system gradually accumulates experience and forms an optimal response strategy for different control situations. After multiple rounds of iterative training, the Q-value table in the system gradually converges, indicating that the system has mastered the optimal actions to be taken in different states. At this time, the system directly extracts the actions pointed to by the maximum Q-value for each state from the Q-table, which is the optimal control strategy for that state. The set of optimal actions for all states together constitutes a full-coverage strategy mapping table.

[0037] The intelligent decision-making method based on the computing-network integrated industrial PLC controller in the embodiments of the present invention has been described above. Next, the intelligent decision-making system based on the computing-network integrated industrial PLC controller in the embodiments of the present invention will be described. Please refer to Figure 2 , an embodiment of the intelligent decision-making system based on the computing-network integrated industrial PLC controller in the embodiments of the present invention includes: The first allocation module 201 is used to allocate time slices for the control parameters of the water supply equipment and divide the decision-making cycle of the PLC controller to obtain an initial control parameter queue; The analysis module 202 is used to determine the probability distribution of the occurrence of the normal operation state, the efficient control state, and the decision conflict state according to the initial control parameter queue to obtain a water supply parameter conflict probability table; The second allocation module 203 is used to reallocate the parameter control time resources based on the water supply parameter conflict probability table to obtain an optimized control time sequence for the water supply parameters; The optimization calculation module 204 is used to convert the parameter groups still having conflicts in the optimized control time sequence of the water supply parameters into independent control variables, and perform optimization calculations on each independent control variable respectively to obtain multi-parameter parallel control instructions for the water supply equipment; The resource scheduling module 205 is used to perform dynamic resource scheduling on the multi-parameter parallel control instructions for the water supply equipment to generate an energy-saving control scheme for the water supply equipment.

[0038] Through the collaborative cooperation of the above-mentioned various components, through probability modeling based on Bayesian and Poisson distribution models for precise time slice allocation, the dynamic allocation of control resources is effectively optimized, significantly shortening the response time of control parameters and enhancing the overall system throughput; the FastICA independent component analysis technology is adopted to transform conflict parameters into independent control variables, realizing the parallel optimization processing of multiple conflict parameters within the same time slice, greatly reducing the parameter processing conflict rate and improving the resource utilization efficiency of the PLC controller; combined with the dynamic resource scheduling mechanism of the energy state of the water supply equipment, multi-objective optimization of control performance and energy consumption is achieved through the Q-learning reinforcement learning algorithm, reducing the total system energy consumption while ensuring control accuracy and extending the service life of the water supply equipment; the time resource reallocation strategy based on parameter correlation analysis and conflict sensitivity index enhances the system's adaptability to parameter request fluctuations, and can still maintain stable operation even when the parameter request frequency suddenly increases; through the task allocation mechanism corrected by the energy factor, the computing load balance considering the remaining energy state of the equipment is realized, effectively avoiding the problem of single-point equipment resource overload; the parameter grouping strategy and multi-parameter parallel control architecture based on the graph coloring algorithm improve the number of water supply equipment that can be managed simultaneously by a single computing-network integrated PLC controller, significantly enhancing the system's expansion ability.

[0039] Those skilled in the art can clearly understand that for the convenience and simplicity of description, the specific working processes of the above-described systems, systems, and units can refer to the corresponding processes in the foregoing method embodiments and will not be elaborated herein.

[0040] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs that can store program codes.

[0041] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An intelligent decision-making method based on an integrated computing and network industrial PLC controller, characterized in that, Including: Performing time slice allocation on the control parameters of the water supply equipment and dividing the decision-making cycle of the PLC controller to obtain an initial control parameter queue; Determining the probability distribution of the occurrence of normal operation state, high-efficiency control state and decision conflict state according to the initial control parameter queue to obtain a water supply parameter conflict probability table; Reallocating the parameter control time resources based on the water supply parameter conflict probability table to obtain an optimized control time sequence for water supply parameters; Converting the parameter groups still having conflicts in the optimized control time sequence of the water supply parameters into independent control variables, and performing optimization calculations on each independent control variable respectively to obtain a multi-parameter parallel control instruction for the water supply equipment; Performing dynamic resource scheduling on the multi-parameter parallel control instruction of the water supply equipment to generate an energy-saving control scheme for the water supply equipment.

2. The intelligent decision-making method based on the computing-network integrated industrial PLC controller according to claim 1, wherein The performing time slice allocation on the control parameters of the water supply equipment and dividing the decision-making cycle of the PLC controller to obtain an initial control parameter queue includes: Collecting the control parameters in the water supply system to obtain an initial water supply parameter set, and assigning a parameter priority value, a maximum allowable delay time value and a computing resource occupancy rate value to each parameter in the initial water supply parameter set to obtain a parameter attribute matrix; Calculating the initial time slice allocation amount of each parameter according to the parameter attribute matrix to obtain a parameter time slice allocation result; Sorting the initial water supply parameter set according to the parameter priority value to obtain an initial control instruction set; Establishing a system state matrix including the operation parameter values, energy state values and communication load values of each water supply equipment, and collecting and updating the values of each matrix element in real time through the industrial field bus to obtain equipment operation state information; Storing the parameter time slice allocation result, the initial control instruction set and the equipment operation state information into the shared memory area of the PLC controller to obtain an initial control parameter queue.

3. The intelligent decision-making method based on the computing-network integrated industrial PLC controller according to claim 1, wherein The determining the probability distribution of the occurrence of normal operation state, high-efficiency control state and decision conflict state according to the initial control parameter queue to obtain a water supply parameter conflict probability table includes: Extracting a historical data set including normal operation state, high-efficiency control state and decision conflict state from the historical database of the computing and networking integrated industrial PLC controller according to the initial control parameter queue; Performing conditional probability distribution calculation on the historical data set to obtain parameter state conditional probability values; Calculating the occurrence probability of different request quantities per unit time according to the parameter request frequency in each time slice to obtain a parameter request frequency probability distribution; Based on the parameter state conditional probability values and the parameter request frequency probability distribution, calculating the probability value of decision conflict occurring in each time slice to obtain a conflict time slice sequence; Constructing a parameter conflict probability matrix and calculating the probability value of decision conflict occurring between each pair of parameters in the same time slice to obtain the parameter conflict relationship within the time slice; Calculating the conflict sensitivity value of each parameter in the conflict time slice sequence to obtain a parameter conflict sensitivity sequence, and merging the parameter conflict sensitivity sequence and the parameter conflict relationship within the time slice into a water supply parameter conflict probability table.

4. The intelligent decision-making method based on the computing-network integrated industrial PLC controller according to claim 3, characterized in that, Calculating the occurrence probability of different request quantities within a unit time according to the parameter request frequency in each time slice to obtain the parameter request frequency probability distribution, including: Counting the request times of each control parameter in different time slices from the historical dataset to obtain the parameter request frequency; Grouping the parameter request frequency to obtain the request frequency change pattern; Based on the request frequency change pattern, determining the Poisson distribution parameter λ value of each parameter under different operating scenarios, and substituting the Poisson distribution parameter λ value into the Poisson distribution formula to calculate the probability distribution of k requests arriving simultaneously, obtaining the parameter request quantity probability matrix; Marking the time slices and parameter combinations with probability values exceeding the set threshold in the parameter request quantity probability matrix to obtain the high-conflict risk time slice list; Using the high-conflict risk time slice list and parameter priority values to calculate the time slice resource competition distribution map; Combining the time slice resource competition distribution map with the parameter request quantity probability matrix to output the parameter request frequency probability distribution.

5. The intelligent decision-making method based on the computing-network integrated industrial PLC controller according to claim 1, wherein, Reallocating the parameter control time resources based on the water supply parameter conflict probability table to obtain the optimized control timing sequence of water supply parameters, including: Extracting the parameter conflict matrix and parameter priority values from the water supply parameter conflict probability table to obtain the parameter conflict sensitivity index set; Calculating the deviation rate from the average conflict sensitivity index based on the parameter conflict sensitivity index set, and multiplying the deviation rate by a preset adjustment intensity factor to obtain the time slice adjustment coefficient; Correcting the initially allocated time slice value according to the time slice adjustment coefficient to obtain the adjusted time slice allocation value; Constructing a parameter relationship topology graph according to the water supply parameter conflict probability table, and according to the parameter relationship topology graph, forcibly allocating the parameters with conflict probability higher than the preset threshold to different time slice groups to obtain the optimized control timing sequence of water supply parameters.

6. The intelligent decision-making method based on the computing-network integrated industrial PLC controller according to claim 1, characterized in that, Converting the parameter groups still in conflict in the optimized control timing sequence of water supply parameters into independent control variables, and performing optimization calculations on each independent control variable respectively to obtain the multi-parameter parallel control instruction for water supply equipment, including: Extracting the parameter groups marked as conflicting from the optimized control timing sequence of water supply parameters to obtain the parameter correlation coefficient matrix; Arranging the historical observation values of the conflicting parameter groups corresponding to the parameter correlation coefficient matrix in rows by parameters and in columns by time points to construct the original observation data matrix; Performing independent component analysis on the original observation data matrix to obtain independent control variables, and respectively establishing weighted objective functions including control delay terms and energy consumption terms based on each independent control variable; Performing gradient descent optimization calculations on each weighted objective function to obtain the optimized independent control variable values; Converting the optimized independent control variable values back to the parameter space to generate the multi-parameter parallel control instruction for water supply equipment that can be executed in parallel within the same time slice.

7. The intelligent decision-making method based on the computing-network integrated industrial PLC controller according to claim 1, wherein Performing dynamic resource scheduling on the multi-parameter parallel control instruction for water supply equipment to generate an energy-saving control scheme for water supply equipment, including: Define the system state space that includes the current active parameter set, parameter priorities, time slice allocation, device energy status, and communication load, construct the action space that includes resource allocation ratio adjustment, communication channel selection, and temporary modification of parameter priorities, and obtain the system state-action mapping table; Calculate the ratio of the current energy value of each water supply device to the full charge as the energy factor, and correct the basic calculation load according to the energy factor to obtain the device calculation task allocation; Construct a reward function based on the average response delay, total energy consumption, and control accuracy of the water supply device, and calculate the comprehensive performance evaluation index through the reward function; Execute the Q-learning reinforcement learning algorithm based on the system state-action mapping table, the device calculation task allocation, and the comprehensive performance evaluation index to obtain the optimal control strategy; According to the optimal control strategy, allocate the execution devices and communication channels for each control task in the multi-parameter parallel control instruction of the water supply device to obtain the task allocation plan; Integrate the task allocation plan with the multi-parameter parallel control instruction of the water supply device to generate an energy-saving control plan for the water supply device.

8. The intelligent decision-making method based on the computing-network integrated industrial PLC controller according to claim 7, characterized in that, The executing the Q-learning reinforcement learning algorithm based on the system state-action mapping table, the device calculation task allocation, and the comprehensive performance evaluation index to obtain the optimal control strategy includes: Discretize the system state space and the action space to obtain the initial Q-value table; Calculate the immediate reward value according to the multi-parameter parallel control instruction of the water supply device and the comprehensive performance evaluation index to obtain the reward value of the current state-action pair; According to the system state-action mapping table, switch between greedily selecting the action with the maximum Q value and randomly exploring actions to obtain the action selection strategy; Execute the selected action based on the action selection strategy and calculate the Q-value update amount, iteratively update the Q-value update amount, and repeat the reinforcement learning process of action selection, observation feedback, and Q-value update to obtain the converged Q-value distribution; Extract the action with the maximum Q value corresponding to each system state based on the converged Q-value distribution and output the optimal control strategy.

9. An intelligent decision-making system based on an arithmetic-network integrated industrial PLC controller, characterized in that, For executing the intelligent decision-making method based on the computing-network integrated industrial PLC controller as described in any one of claims 1-8, the intelligent decision-making system based on the computing-network integrated industrial PLC controller includes: The first allocation module is used to allocate time slices for the control parameters of the water supply device and divide the decision-making cycle of the PLC controller to obtain the initial control parameter queue; The analysis module is used to determine the probability distribution of the occurrence of the normal operation state, high-efficiency control state, and decision conflict state according to the initial control parameter queue to obtain the water supply parameter conflict probability table; The second allocation module is used to reallocate the parameter control time resources based on the water supply parameter conflict probability table to obtain the optimized control timing of the water supply parameters; The optimization calculation module is used to convert the parameter groups still having conflicts in the optimized control timing of the water supply parameters into independent control variables, and perform optimization calculations on each independent control variable respectively to obtain the multi-parameter parallel control instruction of the water supply device; A resource scheduling module for dynamically scheduling resources for the multi-parameter parallel control instruction of the water supply equipment to generate an energy-saving control scheme for the water supply equipment.

Images