A printing device dynamic load balancing distribution method based on real-time order flow

Abstract

The application belongs to the technical field of load balancing, and discloses a printing equipment dynamic load balancing distribution method based on real-time order flow; the method comprises the following steps: collecting order submission data streams distributed in multiple business areas in real time, and constructing an order flow space-time heat matrix; obtaining cumulative running time records, historical fault maintenance logs and periodic maintenance records of each printing equipment, and dynamically evaluating a residual health life index; obtaining geographic coordinates of each business area and each printing equipment, and forming a time delay cost matrix; constructing an equipment load protection weight vector based on the residual health life index of each printing equipment, and performing dynamic fusion operation with the time delay cost matrix to generate a scheduling priority score matrix; generating a final printing task distribution result table according to the scheduling priority score matrix and performing task scheduling; and the application can guarantee optimal scheduling effect under complex and changeable business environment, thereby improving printing service quality.

Description

Technical Field

[0001] This invention relates to the field of load balancing technology, and more specifically, to a dynamic load balancing distribution method for printing equipment based on real-time order traffic. Background Technology

[0002] With the continuous growth of enterprises' demand for digital office solutions, the volume of printing orders is showing a trend of large-scale and high-frequency development. How to achieve efficient task allocation among multiple printing devices has become a key issue in improving the quality of printing services. Existing printing device load balancing methods usually adopt round-robin scheduling or dynamic allocation strategies based on the current queue depth of the device. That is, according to the backlog of the real-time task queue of each printing device, the newly arrived printing orders are allocated to the device with the lightest queue load to achieve a balance of task load among devices.

[0003] However, the existing technologies mentioned above have the following shortcomings: On the one hand, the existing methods only focus on the real-time load status of the equipment and fail to analyze the regional distribution characteristics and temporal fluctuation patterns of order traffic from a spatiotemporal perspective, making it difficult to effectively optimize the transmission latency and delivery costs of cross-regional orders; on the other hand, the existing methods ignore the differences in the cumulative operating wear and remaining health life of each printing device, and simply pursuing short-term load balancing can easily lead to overuse of some devices and accelerated aging, making it difficult to achieve long-term sustainable operation and maintenance of the equipment group, and the economic efficiency of overall resource allocation needs to be improved.

[0004] In view of this, the present invention proposes a dynamic load balancing distribution method for printing equipment based on real-time order traffic to solve the above problems. Summary of the Invention

[0005] To overcome the aforementioned deficiencies of the prior art and achieve the above objectives, the present invention provides the following technical solution: a dynamic load balancing distribution method for printing equipment based on real-time order traffic, comprising:

[0006] Step S1: Collect order submission data streams from order entry nodes distributed across multiple business regions in real time, extract the business region and submission time of each order in the order submission data stream, and construct a spatiotemporal heat matrix of order traffic through spatiotemporal aggregation calculation;

[0007] Step S2: Obtain the cumulative runtime record, historical fault repair log and periodic maintenance record of each printing device, and combine them with the predefined standard life decay curve model to dynamically evaluate the remaining health life index of each printing device at the current moment.

[0008] Step S3: Obtain the geographical coordinates of each business area and each printing device in the order traffic spatiotemporal heat matrix, and calculate the network transmission latency and logistics delivery latency of order transmission from each business area to each printing device based on the geographical coordinates to form a latency cost matrix;

[0009] Step S4: Based on the remaining health life index of each printing device, construct the device load protection weight vector, and perform dynamic fusion calculation on the device load protection weight vector and the latency cost matrix to generate the scheduling priority scoring matrix;

[0010] Step S5: Based on the scheduling priority scoring matrix, the adaptive weighted round-robin scheduling algorithm is used to match the orders to be assigned to the target printing devices in each spatiotemporal unit of the order traffic spatiotemporal heat matrix, generate the final printing task allocation result table, and perform task scheduling based on the final printing task allocation result table.

[0011] Furthermore, methods for constructing a spatiotemporal heat matrix of order traffic include:

[0012] Calculate the time offset of each order relative to midnight of the day and mark it as the intraday time offset; preset time granularity parameters and business cycle, and divide the business cycle into multiple consecutive time windows according to the time granularity parameters; determine the time window to which each order belongs based on the intraday time offset and time granularity parameters of each order; initialize the order traffic spatiotemporal heat matrix according to the total number of time windows and the total number of business regions; where each element in the traffic spatiotemporal heat matrix corresponds to a spatiotemporal unit;

[0013] The spatiotemporal aggregation operation process is defined as follows: Iterate through each order submission message in the order submission data stream, and locate the corresponding spatiotemporal unit in the order traffic spatiotemporal heat matrix based on the business region and time window of each order submission message. The order submission data stream includes the order submission message for each order, which contains basic order information and order attribute information. Simultaneously, whenever an order is located in a spatiotemporal unit, perform a traffic accumulation operation on the corresponding spatiotemporal unit to obtain the order count value and cumulative weighted traffic value for that unit. Finally, perform a heat conversion on the cumulative weighted traffic value of each spatiotemporal unit to obtain the traffic heat value for that unit.

[0014] The order count and traffic heat value of each spatiotemporal unit are integrated to form a spatiotemporal heat record for each spatiotemporal unit; the spatiotemporal heat records of all spatiotemporal units are then sequentially filled into the order traffic spatiotemporal heat matrix to complete the construction of the order traffic spatiotemporal heat matrix.

[0015] Furthermore, methods for dynamically assessing the remaining healthy life expectancy index include:

[0016] Based on the standard life decay curve model, the corresponding basic decay value, fault decay factor, and maintenance recovery value are calculated sequentially for each printing device. The standard life decay curve model includes a basic decay function, a fault accelerated decay function, and a maintenance recovery function. The basic decay value is calculated using the basic decay function and accumulated runtime records; the fault decay factor is calculated using the fault accelerated decay function and historical fault repair logs; and the maintenance recovery value is calculated using the maintenance recovery function and periodic maintenance records. Based on the basic decay value, fault decay factor, and maintenance recovery value of each printing device, the original health index of each printing device is calculated. Boundary constraints are applied to the original health indices of all printing devices sequentially, and the original health index after boundary constraint processing is used as the remaining healthy life index of each printing device.

[0017] Furthermore, methods for forming the time delay cost matrix include:

[0018] Obtain the network path information corresponding to the network path between the order entry node and each printing device in each business area, and calculate the basic path latency between each business area and each printing device based on the network path information; obtain the current network load rate between the order entry node and each printing device in each business area, and calculate the network congestion coefficient based on the current network load rate; calculate the product of the basic path latency and the network congestion coefficient to obtain the network transmission latency between each business area and each printing device; integrate all network transmission latencies to construct a network latency matrix;

[0019] Based on the geographical coordinates of each business area and each printing device, calculate the straight-line distance between each business area and each printing device; based on the delivery area where each business area and each printing device are located, determine the corresponding average delivery speed and regional delivery coefficient; based on the straight-line distance and average delivery speed, calculate the basic delivery time; calculate the product of the basic delivery time and the regional delivery coefficient to obtain the logistics delivery delay between each business area and each printing device; integrate all logistics delivery delays to construct a logistics delay matrix.

[0020] Corresponding delay weights are set for network transmission delay and logistics delivery delay, and the network delay matrix and logistics delay matrix are weighted and merged based on the delay weights to form a delay cost matrix.

[0021] Furthermore, methods for constructing the equipment load protection weight vector include:

[0022] Based on the remaining health life index of each printing device, the health status of each printing device is classified into four levels: healthy, sub-healthy, critical, and dangerous. Based on the health status of each printing device, a corresponding basic protection coefficient is set. Based on the remaining health life index and the basic protection coefficient, the load protection weight of each printing device is calculated through a nonlinear protection weight mapping function.

[0023] Obtain real-time task queue information and rated processing capacity for each printing device; the real-time task queue information includes the estimated execution time and task load of each printing task in the current task queue; sum the estimated execution time and task load of all printing tasks in the current task queue for the same printing device to obtain the queue backlog time and current total load for each printing device; calculate the ratio of the current total load to the rated processing capacity to obtain the load saturation of each printing device;

[0024] Calculate the load adjustment factor for each printing device based on its load saturation; calculate the dynamic protection weight for each printing device based on the load protection weight and the load adjustment factor; integrate the dynamic protection weights of all printing devices to form a device load protection weight vector.

[0025] Furthermore, methods for generating the scheduling priority scoring matrix include:

[0026] Based on the total number of business areas and the total number of printing devices, an initial scheduling priority scoring matrix is ​​established; each element in the scheduling priority scoring matrix corresponds to a priority scoring unit; dynamic fusion operation rules are designed: corresponding fusion weights are set for the comprehensive latency cost and dynamic protection weight, and the fusion weights include latency fusion weights and protection fusion weights; the sum of the latency fusion weights and protection fusion weights is equal to 1; an adaptive fusion adjustment mechanism is introduced: the mean of all dynamic protection weights in the device load protection weight vector is calculated to obtain the average protection weight; the average protection weight is compared with the preset protection weight benchmark value, and the protection optimization weight and latency optimization weight are calculated based on the comparison result;

[0027] Traverse each priority scoring unit in the scheduling priority scoring matrix, and obtain the comprehensive latency cost of the corresponding position in the latency cost matrix, as well as the dynamic protection weight of the corresponding printing device in the device load protection weight vector. Based on the latency optimization weight and the protection optimization weight, perform a fusion calculation on the comprehensive latency cost and dynamic protection weight of the same priority scoring unit to obtain the original priority score of each priority scoring unit. Perform reverse transformation on each original priority score in turn to obtain the scheduling priority score of each priority scoring unit. Fill the scheduling priority score of each priority scoring unit into the scheduling priority scoring matrix in turn to complete the construction of the scheduling priority scoring matrix.

[0028] Furthermore, methods for generating the final print job assignment result table include:

[0029] Step S501: Determine the active spatiotemporal units in the order flow spatiotemporal heat matrix, obtain the orders to be assigned in each active spatiotemporal unit, and combine them to form a queue of orders to be assigned;

[0030] Step S502: For the orders to be assigned in the order queue corresponding to each active spatiotemporal unit, perform intelligent task slicing in sequence to obtain the print task queue for each active spatiotemporal unit.

[0031] Step S503: Based on the adaptive weighted round-robin scheduling algorithm, target devices are matched for the print tasks in each print task queue, and the final print task allocation result table is generated.

[0032] Methods for task scheduling based on the final print task allocation result table include:

[0033] Step S504: Schedule all print tasks according to the final print task allocation result table, and perform emergency order preemptive scheduling during the task scheduling process;

[0034] Step S505: After the task scheduling is completed, perform scheduling effect feedback and parameter adaptive learning.

[0035] Furthermore, methods for obtaining the queue of tasks to be printed include:

[0036] For each order in the queue to be assigned, order sharding identification is performed sequentially to divide all orders into standard orders and large orders; for each large order, intelligent sharding operation is performed sequentially to obtain the order shards corresponding to each large order;

[0037] Each large order's corresponding order fragment and each standard order are treated as printable tasks. Based on the standard orders and order fragments corresponding to each active spatiotemporal unit, printable tasks for each active spatiotemporal unit are obtained. Printable tasks for the same active spatiotemporal unit are aggregated to form a printable task queue for each active spatiotemporal unit.

[0038] Furthermore, methods for generating the final print job assignment result table include:

[0039] Based on the scheduling priority scoring matrix, initialize the device scheduling weight table and design a dynamic weight decay mechanism;

[0040] Define a single-round scheduling process: Sort all active spatiotemporal units in descending order based on the traffic heat value corresponding to each active spatiotemporal unit; based on the sorting result of the active spatiotemporal units, select the first task in the queue of tasks to be printed corresponding to each active spatiotemporal unit as the current task.

[0041] For the current task, determine the task type and perform the corresponding device matching operation to obtain the target printing device for the current task. Specifically, if the current task is a standard order, perform a single device matching operation directly. If the current task is the first order segment of a large order, perform a multi-device collaborative scheduling decision to determine the corresponding scheduling mode and perform the corresponding device matching operation according to the scheduling mode. If the current task is not the first order segment of a large order, perform the corresponding device matching operation according to the scheduling mode of the corresponding large order. The scheduling modes include collaborative scheduling mode and serial scheduling mode.

[0042] Based on the dynamic weight decay mechanism, the target printing device of the current task is dynamically updated, and the current task is removed from the corresponding queue of tasks to be printed.

[0043] Repeat the single-round scheduling process until all active spatiotemporal units have empty print task queues; combine each print task with its corresponding target printing device to obtain a task allocation combination; summarize all task allocation combinations to generate the final print task allocation result table.

[0044] Furthermore, methods for feedback on scheduling effects include:

[0045] After task scheduling is completed, real-time scheduling effect feedback data is collected. The scheduling effect feedback data includes the actual allocation time of each printing task and the remaining health life index of the corresponding target printing device. Based on the scheduling effect feedback data, scheduling effect evaluation indicators are calculated. Among them, scheduling effect evaluation indicators include load balancing index, device health protection index and latency optimization index.

[0046] The load balancing index, equipment health protection index, and latency optimization index are assigned corresponding evaluation weights, and a weighted summation is performed based on the evaluation weights to obtain a comprehensive scheduling effect score. The comprehensive scheduling effect score is compared with the preset effect benchmark value, and the parameter adaptive learning is determined based on the comparison result.

[0047] The technical effects and advantages of the dynamic load balancing distribution method for printing equipment based on real-time order traffic in this invention are as follows:

[0048] By deploying order entry nodes in various business areas to collect order submission data streams in real time, and constructing a spatiotemporal heat matrix of order traffic based on time granularity parameters and business area divisions, combined with spatiotemporal aggregation operations of order count accumulation and weighted traffic accumulation, the system achieves accurate perception and quantitative analysis of the regional distribution characteristics and temporal fluctuation patterns of order traffic. This effectively overcomes the shortcomings of existing technologies that only focus on immediate load status while ignoring the spatiotemporal distribution characteristics of order traffic. By acquiring the cumulative runtime records, historical fault repair logs, and periodic maintenance records of each printing device, and combining them with a predefined standard lifespan decay curve model to dynamically evaluate the remaining health life index of each printing device, the system can comprehensively consider the combined effects of operating losses, fault damage, and maintenance recovery on the health status of the equipment. This effectively solves the problem of existing technologies neglecting the difference between cumulative operating losses and remaining health life, leading to the overuse and accelerated aging of some equipment. By calculating the network transmission time between each business area and each printing device... This paper proposes a method to optimize the transmission latency and delivery cost of cross-regional orders by constructing a latency cost matrix, comprehensively considering the impact of network topology, network load status, geographical distance, and regional delivery complexity on transmission and delivery efficiency. It also proposes to construct a device load protection weight vector and introduce a nonlinear protection weight mapping function and an adaptive fusion adjustment mechanism to generate a scheduling priority scoring matrix, achieving intelligent balanced scheduling for device protection and latency optimization. Furthermore, it utilizes intelligent order task fragmentation and a multi-device collaborative scheduling decision mechanism to fully leverage the parallel processing capabilities of multiple printing devices, improving the overall processing efficiency of large orders. An emergency order preemptive scheduling mechanism ensures rapid response and timely processing of urgent priority orders. Finally, it employs scheduling effect feedback and parameter adaptive learning mechanisms to continuously optimize and improve scheduling strategies, ensuring optimal scheduling performance in complex and ever-changing business environments and significantly improving the economic efficiency of overall resource allocation and the quality of printing services. Attached Figure Description

[0049] Figure 1 This is a flowchart of a dynamic load balancing distribution method for printing equipment based on real-time order traffic, according to Embodiment 1 of the present invention. Detailed Implementation

[0050] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0051] Example 1

[0052] Please see Figure 1As shown in this embodiment, a dynamic load balancing distribution method for printing equipment based on real-time order traffic includes:

[0053] Step S1: Collect order submission data streams from order entry nodes distributed across multiple business regions in real time, extract the business region and submission time of each order in the order submission data stream, and construct a spatiotemporal heat matrix of order traffic through spatiotemporal aggregation operations.

[0054] Methods for collecting order submission data streams include:

[0055] Order entry nodes are deployed in each business area. These entry nodes are specifically edge computing gateway devices capable of receiving and forwarding orders. Each entry node is configured with a unique node identifier and pre-associated with the corresponding business area's region code and geographic coordinates. A business area refers to a geographic region unit divided according to preset rules, used to identify the spatial affiliation of an order. The region code is a pre-defined unique identifier for the business area, and the geographic coordinates of the business area include the longitude and latitude values ​​of the business area's center point. The node identifier, region code, and geographic coordinates are all pre-configured by those skilled in the art based on the actual business deployment.

[0056] The message queue middleware (a data communication middleware component used for asynchronous message transmission and aggregation between order entry nodes and the backend processing system) subscribes to order submission messages pushed by each order entry node in real time and aggregates all order submission messages to form an order submission data stream. Each order submission message in the order submission data stream contains basic order information and order attribute information. The basic order information includes a unique order identifier, order submission time, order source node, and order file size. The order attribute information includes the number of pages to print, print specification type, and print priority. The print specification type includes single-sided printing, double-sided printing, color printing, and black and white printing. The print priority flags include normal priority, rush priority, and express priority.

[0057] Methods for extracting the business region and submission time of each order include:

[0058] Obtain the order submission information for each order in the order submission data stream; obtain the order entry node for each order based on the order source node in the order submission information; obtain the business region to which each order belongs based on the region code corresponding to each order entry node; obtain the submission time of each order based on the order submission time in the order submission information.

[0059] Methods for constructing a spatiotemporal heat matrix of order traffic include:

[0060] Calculate the time offset of each order relative to midnight of the day (i.e., the starting time of the date corresponding to the order submission time) and mark it as the intra-day time offset; specifically, the intra-day time offset is equal to the time difference between the order submission time and midnight of the day.

[0061] A preset time granularity parameter and business cycle are defined, whereby the time granularity parameter defines the smallest unit of division for the time dimension. Based on the time granularity parameter, the business cycle is divided into multiple consecutive time windows. Both the time granularity parameter and the business cycle are preset by those skilled in the art according to the real-time requirements of the business. The ratio of the business cycle to the time granularity parameter is calculated to obtain the total number of time windows. For example, with a complete business cycle of 24 hours (i.e., 1440 minutes) and the time granularity parameter set to 15 minutes, the total number of time windows is... The following are, in order: ;

[0062] Based on the intraday time offset and time granularity parameters of each order, the time window to which each order belongs is determined. Specifically, each time window is assigned an incrementing numerical label, which is then marked as the window label. The value range of the window label is as follows: , The total number of time windows; calculate the ratio of intraday time offset to time granularity parameter, and round down to obtain the time index value; the time window corresponding to the window label of the time index value is used as the time window to which the corresponding order belongs;

[0063] Count the number of all business regions to obtain the total number of business regions; initialize the order traffic spatiotemporal heat matrix based on the total number of time windows and the total number of business regions; specifically, create a two-dimensional matrix, where the number of rows equals the total number of business regions and the number of columns equals the total number of time windows; each element in the matrix corresponds to a spatiotemporal unit, which is used to store the order traffic heat information of a specific business region within a specific time window; set the initial value of all elements in the matrix to zero.

[0064] The spatiotemporal aggregation operation process is defined as follows: traverse each order submission message in the order submission data stream, and locate the corresponding spatiotemporal unit in the order traffic spatiotemporal heat matrix based on the business region and time window of each order submission message; simultaneously, whenever an order is located in a spatiotemporal unit, perform a traffic accumulation operation on the corresponding spatiotemporal unit to obtain the order count value and the cumulative weighted traffic value of the corresponding spatiotemporal unit; the traffic accumulation operation includes order count accumulation and weighted traffic accumulation.

[0065] Specifically, the order count accumulation method is as follows: The order quantity count value corresponding to the spatiotemporal unit is incremented by one to obtain the updated order quantity count value; where the order quantity count value reflects the total number of orders within the spatiotemporal unit. The weighted traffic accumulation method is as follows: Priority weights are determined based on the printing priority of the orders; where the priority weight of urgent priority is greater than that of expedited priority, and the priority weight of expedited priority is greater than that of normal priority; each priority weight is preset by those skilled in the art based on actual business needs; corresponding load weights are set for the number of printed pages and the order file size, and the number of printed pages and the order file size are weighted and summed based on the load weights to obtain the task load of the order; where each load weight is preset by those skilled in the art based on actual conditions; the product of the task load and the priority weight is calculated to obtain the weighted task load; the cumulative weighted traffic value corresponding to the spatiotemporal unit is added to the weighted task load to obtain the updated cumulative weighted traffic value; where the cumulative weighted traffic value reflects the comprehensive task load of all orders within the spatiotemporal unit.

[0066] The cumulative weighted traffic value of each spatiotemporal unit in the order traffic spatiotemporal heat matrix is ​​converted to obtain the traffic heat value of each spatiotemporal unit. Specifically, the cumulative weighted traffic values ​​of all spatiotemporal units in the order traffic spatiotemporal heat matrix are obtained and compared. The cumulative weighted traffic value with the largest value is marked as the maximum weighted traffic, and the cumulative weighted traffic value with the smallest value is marked as the minimum weighted traffic. The difference between the maximum weighted traffic and the minimum weighted traffic is calculated to obtain the weighted traffic difference. The difference between the cumulative weighted traffic value and the minimum weighted traffic of each spatiotemporal unit is calculated and then divided by the weighted traffic difference to obtain the traffic heat value of each spatiotemporal unit.

[0067] The order count and traffic heat value of each spatiotemporal unit are integrated to form a spatiotemporal heat record for each spatiotemporal unit; the spatiotemporal heat records of all spatiotemporal units are then sequentially filled into the order traffic spatiotemporal heat matrix to complete the construction of the order traffic spatiotemporal heat matrix.

[0068] Step S2: Obtain the cumulative runtime record, historical fault repair log and periodic maintenance record of each printing device, and combine them with the predefined standard life decay curve model to dynamically evaluate the remaining health life index of each printing device at the current moment.

[0069] The cumulative runtime records of each printing device are obtained from the printing device monitoring system (i.e., an IoT monitoring platform used to collect and manage the operating status data of printing devices in real time). Each printing device in the printing device monitoring system is equipped with a unique device identifier and a rated lifespan, which is the standard usage time specified by the printing device manufacturer. The cumulative runtime record records the total runtime information of the printing device from its first use to the current moment, specifically including the device identifier and the cumulative runtime. The cumulative runtime is calculated as follows: all runtime segment records of the printing device from its first use to the current moment are obtained, and the runtime segment records include the start time and end time of each run; the difference between the end time and start time in each runtime segment record is calculated to obtain the single runtime of the corresponding runtime segment record; the summation of all single runtimes of the same printing device is calculated to obtain the cumulative runtime of the corresponding printing device.

[0070] The historical fault repair logs of each printing device are obtained from the equipment maintenance management system (an information management platform used to record and manage fault repair and maintenance information of printing equipment). These historical fault repair logs record all fault and repair information of the printing equipment from its initial use to the present moment, specifically including multiple fault repair records. Each fault repair record includes the equipment identifier, fault occurrence time, fault type, fault level, and repair time. Fault types include mechanical faults, electrical faults, software faults, and consumable faults; fault levels include minor faults, general faults, and serious faults; and repair time is the time interval from the occurrence of the fault to the completion of the repair.

[0071] Retrieve periodic maintenance records for each printing device from the equipment maintenance management system. These periodic maintenance records document all maintenance and upkeep information for the printing devices from their initial use to the present moment, and include multiple records. Each record contains a device identifier, maintenance execution time, maintenance type, and maintenance assessment level. Maintenance types include routine maintenance, periodic maintenance, and deep maintenance. The maintenance assessment level is the evaluation result of the device's condition after maintenance, categorized as excellent, good, and satisfactory.

[0072] Methods for predefined standard lifetime decay curve models include:

[0073] The standard lifespan decay curve model includes a basic decay function, a fault-accelerated decay function, and a maintenance recovery function, used to describe the dynamic changes in the health status of the printing equipment with runtime, accumulated faults, and maintenance effectiveness. Specifically: the basic decay function describes the natural decay process of the printing equipment's health status with accumulated runtime; the fault-accelerated decay function describes the negative impact of accumulated faults on the printing equipment's health status; and the maintenance recovery function describes the positive recovery effect of accumulated maintenance on the printing equipment's health status.

[0074] The expression for the fundamental decay function is: In the formula, Based on the attenuation value, For cumulative runtime, For the rated service life, Here, the attenuation curvature parameter controls the shape of the attenuation curve. It exhibits a decay characteristic of being slow at first and then fast. It exhibits a decay characteristic of being fast at first and then slowing down. It exhibits a linear attenuation characteristic; the attenuation curvature parameter is preset by those skilled in the art based on the attenuation characteristics of the device type;

[0075] The expression for the fault acceleration decay function is: In the formula, As the fault attenuation factor, This is the fault impact coefficient. This is the weighted cumulative fault value. Mean time between failures (MTBF) It is an exponential function; where the fault impact coefficient is used to adjust the intensity of the impact of the fault on the equipment health, and is preset by those skilled in the art according to the actual situation;

[0076] The expression for the maintenance and recovery function is: In the formula, To maintain the recovery value, To maintain the efficiency conversion coefficient, To accumulate maintenance efficiency value, To restore the decay rate, The maintenance interval duration; wherein, the maintenance efficiency conversion coefficient is used to convert the cumulative maintenance efficiency into health recovery amount, and the recovery decay rate is used to describe the rate at which the maintenance effect decays over time; both the maintenance efficiency conversion coefficient and the recovery decay rate are preset by those skilled in the art according to the actual situation.

[0077] Methods for dynamically assessing the remaining healthy life expectancy index include:

[0078] Based on the standard life decay curve model, the corresponding basic decay value, fault decay factor, and maintenance recovery value are calculated sequentially for each printing device. Specifically, the cumulative running time and rated life of each printing device are substituted into the basic decay function to calculate the basic decay value of each printing device; the weighted cumulative fault value and average fault interval duration of each printing device are substituted into the fault acceleration decay function to calculate the fault decay factor of each printing device; and the cumulative maintenance efficiency value and maintenance interval duration of each printing device are substituted into the maintenance recovery function to calculate the maintenance recovery value of each printing device.

[0079] The original health index of each printing device is calculated based on its base attenuation value, fault attenuation factor, and maintenance recovery value. Specifically, the product of the base attenuation value and fault attenuation factor for the same printing device is calculated, and the corresponding maintenance recovery value is added to the result to obtain the original health index of each printing device. The original health index comprehensively reflects the combined effect of three factors—operating losses, fault damage, and maintenance recovery—on the health status of the equipment.

[0080] The original health index of all printing devices is sequentially subjected to boundary constraint processing, and the original health index after boundary constraint processing is used as the remaining health life index of each printing device. Specifically, if the original health index is greater than 1, the original health index is set to 1; if the original health index is less than 0, the original health index is set to 0; if the original health index is greater than or equal to 0 and less than or equal to 1, the original health index is not adjusted.

[0081] The weighted cumulative fault value is calculated as follows: based on the fault level in the fault repair record, a corresponding fault weight is set for each fault repair record; the fault weight for a serious fault is greater than that for a general fault, and the fault weight for a general fault is greater than that for a minor fault; each fault weight is preset by those skilled in the art based on the degree of impact of the fault on the equipment's lifespan; the fault weights of all fault repair records corresponding to the printing equipment are summed to obtain the weighted cumulative fault value of the printing equipment; the weighted cumulative fault value is used to reflect the cumulative lifespan damage caused by multiple faults of different fault levels in the printing equipment.

[0082] The method for calculating the average interval between failures is as follows: count the number of failure repair records in the historical failure repair log of the printing equipment to obtain the cumulative number of failures of the printing equipment; calculate the ratio of the cumulative running time of the printing equipment to the cumulative number of failures to obtain the average interval between failures of the printing equipment; wherein, the average interval between failures is used to reflect the operational reliability of the printing equipment.

[0083] The method for calculating the cumulative maintenance efficiency value is as follows: First, count the number of maintenance records in the periodic maintenance records corresponding to the printing equipment to obtain the cumulative number of maintenance visits. Second, based on the maintenance type in the maintenance records, set corresponding maintenance recovery coefficients for each maintenance record. The maintenance recovery coefficient for deep maintenance is greater than that for periodic maintenance, and the maintenance recovery coefficient for periodic maintenance is greater than that for routine maintenance. Each maintenance recovery coefficient is preset by a person skilled in the art based on the restoration effect of maintenance on equipment lifespan. Third, based on the maintenance assessment level in the maintenance records, set corresponding maintenance quality coefficients for each maintenance record. The maintenance quality coefficient for excellent maintenance is greater than that for good maintenance, and the maintenance quality coefficient for good maintenance is greater than that for acceptable maintenance. Each maintenance quality coefficient is preset by a person skilled in the art based on actual conditions. Fourth, calculate the product of the maintenance recovery coefficient and the maintenance quality coefficient corresponding to each maintenance record to obtain the comprehensive maintenance efficiency of the maintenance record. Fifth, sum all the comprehensive maintenance efficiencies corresponding to the printing equipment to obtain the cumulative maintenance efficiency value of the printing equipment. The cumulative maintenance efficiency value reflects the cumulative effect of lifespan restoration and health improvement achieved by the printing equipment through multiple maintenance visits during use.

[0084] The maintenance interval duration is calculated as follows: obtain the maintenance execution time of the most recent maintenance record (i.e., the maintenance record with the smallest time difference between the maintenance execution time and the current time) in the periodic maintenance records of the printing equipment, and mark it as the most recent maintenance time; calculate the difference between the current time and the most recent maintenance time to obtain the maintenance interval duration of the printing equipment; wherein, the maintenance interval duration is used to reflect the running time of the printing equipment since the most recent maintenance.

[0085] Step S3: Obtain the geographical coordinates of each business area and each printing device in the order traffic spatiotemporal heat matrix, and calculate the network transmission latency and logistics delivery latency of order transmission from each business area to each printing device based on the geographical coordinates to form a latency cost matrix.

[0086] Methods for obtaining the geographic coordinates of each business area and each printing device include:

[0087] For each spatiotemporal unit in the order traffic spatiotemporal heat matrix, the business area corresponding to each order is obtained sequentially; the geographical coordinates associated with each order entry node are used as the geographical coordinates of the business area corresponding to the order; the geographical coordinates of each printing device are obtained from the device registration information of the printing device monitoring system; wherein, the geographical coordinates of the printing device include the longitude and latitude values ​​of the location of the printing device, which are pre-configured by those skilled in the art when the printing device is deployed.

[0088] Methods for calculating network transmission latency include:

[0089] Based on the network topology between each business area and each printing device, the network transmission latency of orders from each business area to each printing device is calculated. This network transmission latency reflects the network communication time required for order data to be transmitted from the order entry node of the business area to the printing device. The network transmission latency between all business areas and printing devices is integrated to construct a network latency matrix. This matrix is ​​a two-dimensional matrix, with the number of rows equal to the total number of business areas and the number of columns equal to the total number of printing devices (i.e., the total number of printing devices). Each element in the network latency matrix represents the network transmission latency of orders from the corresponding business area to the corresponding printing device.

[0090] Specifically, network path information corresponding to the network paths between the order entry nodes of each business area and each printing device is obtained from the network management system (i.e., the network operation and maintenance platform used to monitor and manage the network connection status between business areas and printing devices). Here, the network path refers to the data transmission path formed by the hop nodes and network links between each hop node as order data is transmitted from the order entry node to the printing device, according to the network topology. The network path information includes the number of network hops and the baseline transmission delay of each hop node. The number of network hops is the number of hop nodes that order data passes through when transmitted from the order entry node to the printing device. A hop node is a network node located between the order entry node and the printing device, used for order data processing and forwarding. The baseline transmission delay is the standard time required for each hop node to process and forward order data. The baseline transmission delays of all hop nodes corresponding to the same network path are summed to obtain the basic path delay between each business area and each printing device.

[0091] From the network management system, obtain the current network load rate between the order entry node of each business area and each printing device. The current network load rate is the ratio of the bandwidth used on the network path to the real-time network bandwidth, where real-time network bandwidth is the maximum data transmission capacity of the network path. Calculate the difference between the current network load rate and the current network load rate to obtain the network idle rate. Calculate the ratio of the current network load rate to the network idle rate to obtain the network congestion coefficient. The network congestion coefficient reflects the amplification effect of network congestion on transmission delay; the higher the current network load rate, the larger the network congestion coefficient. Calculate the product of the path base delay and the network congestion coefficient to obtain the network transmission delay between each business area and each printing device.

[0092] Methods for calculating logistics delivery delays include:

[0093] Based on the geographical relationship between each business area and each printing device, the logistics delivery delay of orders from each business area to each printing device is calculated. This logistics delivery delay reflects the transportation time required for the printed product to be transported from the location of the printing device back to the business area. The logistics delivery delays between all business areas and all printing devices are integrated to construct a logistics delay matrix. This matrix is ​​a two-dimensional matrix, with the number of rows equal to the total number of business areas and the number of columns equal to the total number of printing devices. Each element in the logistics delay matrix represents the logistics delivery delay of orders from the corresponding business area to the corresponding printing device.

[0094] Specifically, based on the geographical coordinates of each business area and each printing device, the Haversine formula is used to calculate the straight-line geographical distance between each business area and each printing device. The Haversine formula is a well-known technique in the field, and the specific calculation process will not be elaborated upon here. The average delivery speed and regional delivery coefficient for each delivery area are obtained from the logistics management system (i.e., the logistics operation platform used to manage and schedule order delivery tasks). The average delivery speed is the standard driving speed of logistics vehicles within the corresponding delivery area. The regional delivery coefficient reflects the impact of road conditions, traffic congestion, and delivery complexity on delivery time in different delivery areas; the regional delivery coefficient is preset by those skilled in the art based on historical delivery data.

[0095] Based on the delivery areas of each business area and each printing device, the corresponding average delivery speed and regional delivery coefficient are obtained; the ratio of the geographical straight-line distance to the average delivery speed is calculated to obtain the basic delivery time; the product of the basic delivery time and the regional delivery coefficient is calculated to obtain the logistics delivery delay between each business area and each printing device; wherein, the delivery areas of each business area and each printing device are determined by those skilled in the art based on the geographical coordinates of each business area and each printing device.

[0096] Methods for generating the time delay cost matrix include:

[0097] Corresponding delay weights are set for network transmission delay and logistics delivery delay, and the network delay matrix and logistics delay matrix are weighted and fused based on the delay weights to form a delay cost matrix; wherein, each delay weight is preset by those skilled in the art according to the degree of emphasis of the business on network transmission efficiency and logistics delivery efficiency.

[0098] Specifically, the network transmission delays in the network delay matrix are normalized to obtain standard network delays; all standard network delays are then integrated to construct a standard network delay matrix. The normalization method is as follows: all network transmission delays in the network delay matrix are obtained and compared; the network transmission delay with the largest value is marked as the maximum network delay, and the network transmission delay with the smallest value is marked as the minimum network delay; the difference between the maximum and minimum network delays is calculated to obtain the network delay difference; the difference between each network transmission delay and the minimum network delay is calculated and then divided by the network delay difference to obtain the corresponding standard network delay. Similarly, the logistics delivery delays in the logistics delay matrix are normalized to obtain standard logistics delays; all standard logistics delays are then integrated to construct a standard logistics delay matrix. The method for normalizing logistics delivery delays is the same as the method for normalizing network transmission delays.

[0099] Initialize the latency cost matrix. The number of rows in the latency cost matrix equals the total number of business areas, and the number of columns equals the total number of printing devices. Each element in the latency cost matrix corresponds to a latency cost unit, which stores the comprehensive latency cost of order allocation from a specific business area to a specific printing device. Traverse each latency cost unit in the latency cost matrix and obtain the standard network latency in the standard network latency matrix and the standard logistics latency in the standard logistics latency matrix for the corresponding position. Calculate the weighted sum of the standard network latency and standard logistics latency of the same latency cost unit based on latency weights to obtain the comprehensive latency cost of each latency cost unit. Fill the latency cost matrix with the comprehensive latency cost of each latency cost unit in sequence to complete the construction of the latency cost matrix. The smaller the comprehensive latency cost in the latency cost matrix, the higher the transmission and delivery efficiency of order allocation from the corresponding business area to the corresponding printing device.

[0100] Step S4: Based on the remaining health life index of each printing device, construct the device load protection weight vector, and perform dynamic fusion calculation on the device load protection weight vector and the latency cost matrix to generate the scheduling priority scoring matrix.

[0101] Methods for constructing device load protection weight vectors include:

[0102] Based on the remaining health lifespan index of each printing device, the health status of each printing device is classified. Specifically, a set of preset health classification thresholds is used, including a first health threshold, a second health threshold, and a third health threshold. The first health threshold is greater than the second health threshold, and the second health threshold is greater than the third health threshold. Each health threshold is preset by those skilled in the art according to the device management strategy. The remaining health lifespan index of each printing device is compared with each health threshold in the set of health classification thresholds. If the remaining health lifespan index is greater than or equal to the first health threshold, the corresponding printing device is in a healthy state. If the remaining health lifespan index is less than the first health threshold but greater than or equal to the second health threshold, the corresponding printing device is in a sub-healthy state. If the remaining health lifespan index is less than the second health threshold but greater than or equal to the third health threshold, the corresponding printing device is in a critical state. If the remaining health lifespan index is less than the third health threshold, the corresponding printing device is in a dangerous state.

[0103] Based on the health status of each printing device, corresponding basic protection coefficients are set; wherein, the basic protection coefficient for the dangerous level is greater than that for the critical level, the basic protection coefficient for the critical level is greater than that for the sub-healthy level, and the basic protection coefficient for the sub-healthy level is greater than that for the healthy level; each basic protection coefficient is preset by those skilled in the art according to the device protection strategy; for example, the basic protection coefficient for the healthy level is 1.0, the basic protection coefficient for the sub-healthy level is 1.5, the basic protection coefficient for the critical level is 2.5, and the basic protection coefficient for the dangerous level is 4.0;

[0104] Based on the remaining health life index and the basic protection coefficient, the load protection weight of each printing device is calculated using a nonlinear protection weight mapping function; the expression of the nonlinear protection weight mapping function is as follows: In the formula, For load protection weight, Basic protection coefficient, To protect sensitivity parameters, The remaining health life index is used to control the sensitivity of the load protection weight to changes in the remaining health life index. This parameter is preset by those skilled in the art based on actual conditions. The load protection weight reflects the load protection requirement of each printing device. The larger the load protection weight, the higher the load protection requirement of the corresponding printing device, and the fewer order tasks should be assigned to the corresponding printing device.

[0105] The real-time task queue information and rated processing capacity of each printing device are obtained from the printing device monitoring system. The real-time task queue information includes the estimated execution time and task load of each printing task in the current task queue. The estimated execution time indicates the expected processing time required to complete the corresponding printing task on the corresponding printing device, and the task load indicates the degree of resource consumption of the corresponding printing task on the corresponding printing device. The rated processing capacity is the standard task load that the corresponding printing device can handle per unit time, pre-configured by those skilled in the art based on the device's performance parameters. The estimated execution times of all printing tasks in the current task queue corresponding to the same printing device are summed to obtain the queue backlog time for each printing device. The task loads of all printing tasks in the current task queue corresponding to the same printing device are summed to obtain the current total load of each printing device. The ratio of the current total load to the rated processing capacity is calculated to obtain the load saturation of each printing device. The load saturation reflects the current task load pressure of the printing device; a higher load saturation indicates a greater current task pressure on the device.

[0106] Based on the load saturation of each printing device, a load adjustment factor for each printing device is calculated. Specifically, a load saturation threshold is preset, which is pre-set by those skilled in the art according to the device load management strategy. The load saturation of each printing device is compared with the load saturation threshold. If the load saturation is less than or equal to the load saturation threshold, the load adjustment factor for the corresponding printing device is set to 1. If the load saturation is greater than the load saturation threshold, the difference between the load saturation and the load saturation threshold is calculated to obtain the overload amount. The product of the overload amount and the preset adjustment coefficient is calculated, and then 1 is added to obtain the load adjustment factor for the corresponding printing device. The adjustment coefficient is pre-set by those skilled in the art according to the actual situation.

[0107] The load protection weight of each printing device is multiplied by the corresponding load adjustment factor to obtain the dynamic protection weight of each printing device. The dynamic protection weights of all printing devices are integrated to form a device load protection weight vector. The device load protection weight vector is a one-dimensional vector with a length equal to the total number of printing devices. Each element in the device load protection weight vector represents the dynamic protection weight of the corresponding printing device.

[0108] Methods for generating scheduling priority score matrices include:

[0109] Initialize the scheduling priority scoring matrix based on the total number of business areas and the total number of printing devices. Specifically, create a two-dimensional matrix with the number of rows equal to the total number of business areas and the number of columns equal to the total number of printing devices. Each element in the matrix corresponds to a priority scoring unit, which is used to store the scheduling priority score for allocating orders from a specific business area to a specific printing device.

[0110] Design dynamic fusion operation rules: Set corresponding fusion weights for comprehensive latency cost and dynamic protection weight respectively. The fusion weights include latency fusion weight and protection fusion weight; wherein, the sum of latency fusion weight and protection fusion weight is equal to 1; each fusion weight is preset by those skilled in the art according to the emphasis of the service on transmission efficiency and equipment protection.

[0111] An adaptive fusion adjustment mechanism is introduced: The mean of all dynamic protection weights in the equipment load protection weight vector is calculated to obtain the average protection weight; the average protection weight is compared with a preset protection weight benchmark value, which is preset by those skilled in the art based on actual conditions; if the average protection weight is greater than the protection weight benchmark value, the ratio of the average protection weight to the protection weight benchmark value is calculated to obtain the protection weight deviation; the product of the protection fusion weight and the protection weight deviation is calculated to obtain the protection optimization weight; the difference between 1 and the protection optimization weight is calculated to obtain the time delay optimization weight; if the average protection weight is less than or equal to the protection weight benchmark value, the time delay fusion weight is used as the time delay optimization weight, and the protection fusion weight is used as the protection optimization weight; it should be noted that the adaptive fusion adjustment mechanism is used to automatically increase the weight ratio of equipment protection when the overall health of the equipment is poor.

[0112] The process iterates through each priority scoring unit in the scheduling priority scoring matrix, obtaining the comprehensive delay cost of the corresponding position in the delay cost matrix and the dynamic protection weight of the corresponding printing device in the device load protection weight vector. The comprehensive delay cost and dynamic protection weight of the same priority scoring unit are then fused to obtain the original priority score for each unit. Specifically, the product of the comprehensive delay cost and the delay optimization weight is calculated to obtain the weighted delay cost. The dynamic protection weight is normalized to obtain the standard protection weight; the method for normalizing the dynamic protection weight is the same as the method for normalizing network transmission delay. The product of the standard protection weight and the protection optimization weight is calculated to obtain the weighted protection weight. The weighted delay cost and the weighted protection weight are then added to obtain the original priority score.

[0113] Each original priority score is reverse-converted sequentially to obtain the scheduling priority score for each priority score unit. The scheduling priority scores of each priority score unit are then sequentially filled into the scheduling priority score matrix to complete the construction of the scheduling priority score matrix. Specifically, the difference between 1 and the original priority score is calculated to obtain the scheduling priority score. The higher the scheduling priority score, the higher the overall priority of the orders in the corresponding business area to the corresponding printing equipment, meaning that the corresponding printing equipment is more suitable to handle the orders in the corresponding business area.

[0114] Step S5: Based on the scheduling priority scoring matrix, the adaptive weighted round-robin scheduling algorithm is used to match the orders to be assigned to the target printing devices in each spatiotemporal unit of the order traffic spatiotemporal heat matrix, generate the final printing task allocation result table, and perform task scheduling based on the final printing task allocation result table.

[0115] Methods for generating the final print job assignment result table include:

[0116] Step S501: Determine the active spatiotemporal units in the order flow spatiotemporal heat matrix, obtain the orders to be assigned in each active spatiotemporal unit, and combine them to form a queue of orders to be assigned;

[0117] Step S502: For the orders to be assigned in the order queue corresponding to each active spatiotemporal unit, perform intelligent task slicing in sequence to obtain the print task queue for each active spatiotemporal unit.

[0118] Step S503: Based on the adaptive weighted round-robin scheduling algorithm, target devices are matched for the print tasks in each print task queue, and the final print task allocation result table is generated.

[0119] In step S501 above, the method for forming the queue of orders to be assigned includes:

[0120] Traverse each spatiotemporal unit in the order traffic spatiotemporal heat matrix and obtain the order count value corresponding to each spatiotemporal unit; if the order count value is greater than zero, mark the corresponding spatiotemporal unit as an active spatiotemporal unit; for each active spatiotemporal unit, filter out all orders belonging to the corresponding business area and time window from the order submission data stream and treat them as orders to be assigned; summarize all the orders to be assigned corresponding to each active spatiotemporal unit to form the set of orders to be assigned for each active spatiotemporal unit;

[0121] The orders to be assigned in each set of orders to be assigned are prioritized to form a queue of orders to be assigned for each active spatiotemporal unit. Specifically, the priority weight of each order to be assigned in the set of orders to be assigned is obtained; the time difference between the current time and the submission time of each order to be assigned is calculated to obtain the order waiting time of each order to be assigned; corresponding sorting weights are set for the priority weight and the order waiting time, and the priority weight and the order waiting time of the same order to be assigned are weighted and summed based on the sorting weights to obtain the comprehensive sorting score of each order to be assigned; wherein, each sorting weight is preset by those skilled in the art according to the actual situation; all orders to be assigned in the same set of orders to be assigned are sorted in descending order according to the comprehensive sorting score to form a queue of orders to be assigned for each active spatiotemporal unit.

[0122] In step S502 above, the method for obtaining the queue of tasks to be printed for each active spatiotemporal unit includes:

[0123] For each order in the queue to be assigned, order segmentation is performed sequentially to classify all orders into standard orders and large orders. Specifically, a task segmentation threshold is preset, which is pre-set by those skilled in the art based on the optimal processing load of a single printing device. The task load of each order to be assigned is compared with the task segmentation threshold. If the task load is less than or equal to the task segmentation threshold, the corresponding order is marked as a standard order and no segmentation is performed. If the task load is greater than the task segmentation threshold, the corresponding order is marked as a large order and segmentation is required.

[0124] For each large order, intelligent sharding is performed sequentially to obtain the order shards corresponding to each large order. Specifically, the ratio of the task load of the large order to the task sharding threshold is calculated and rounded up to obtain the target number of shards; the ratio of the number of printed pages of the large order to the target number of shards is calculated, and the base number of pages for each shard is calculated; the number of printed pages of the large order is evenly divided according to the base number of pages to form multiple order shards; each order shard inherits the original large order's unique order identifier, printing priority, and printing specification type, and is assigned a unique shard number; the ratio of the base number of pages of each order shard to the number of printed pages of the corresponding large order is calculated, and the result is multiplied by the task load of the corresponding large order to obtain the shard load of each order shard;

[0125] Each large order's corresponding order fragment and each standard order are treated as printable tasks. Based on the standard orders and order fragments corresponding to each active spatiotemporal unit, printable tasks for each active spatiotemporal unit are obtained. Printable tasks for the same active spatiotemporal unit are aggregated to form a printable task queue for each active spatiotemporal unit.

[0126] In step S503 above, the method for generating the final print job assignment result table includes:

[0127] Based on the scheduling priority scoring matrix, initialize the device scheduling weight table. Specifically, for each printing device, calculate the average scheduling score of each printing device by averaging all scheduling priority scores in the corresponding column of the scheduling priority scoring matrix. Use the average scheduling score of each printing device as the corresponding scheduling weight, and associate the device identifier of each printing device with the scheduling weight to form the device scheduling weight table. At the same time, set a cumulative allocation counter for each printing device and set the corresponding cumulative allocation counter value to zero.

[0128] Design a dynamic weight decay mechanism: preset weight decay coefficient and weight recovery period; wherein, the weight decay coefficient is used to control the decay of the scheduling weight after each order allocation, and the weight recovery period is used to control the time interval for the scheduling weight to recover to the initial state; the weight decay coefficient and weight recovery period are both preset by those skilled in the art according to the load balancing strategy.

[0129] Define a single-round scheduling process: Sort all active spatiotemporal units in descending order based on their traffic heat values; according to the sorting results, select the first task in the print task queue corresponding to each active spatiotemporal unit as the current task; for example, the print task queue of active spatiotemporal unit A contains... The print queue of active spatiotemporal unit B contains Furthermore, active spatiotemporal unit A is ranked before active spatiotemporal unit B, therefore the current task selection order is... ;

[0130] For the current task, determine the task type and perform the corresponding device matching operation to obtain the target printing device for the current task. Specifically, if the current task is a standard order, perform a single device matching operation directly; if the current task is the first order segment of a large order, perform a multi-device collaborative scheduling decision to determine the corresponding scheduling mode, and perform the corresponding device matching operation according to the scheduling mode; if the current task is not the first order segment of a large order, perform the corresponding device matching operation according to the scheduling mode of the corresponding large order.

[0131] Based on the dynamic weight decay mechanism, a dynamic weight update operation is performed on the target printing device of the current task. Specifically, the product of the scheduling weight of the target printing device in the device scheduling weight table and the weight decay coefficient is calculated to obtain the decayed scheduling weight. Based on the decayed scheduling weight, the scheduling weight of the target printing device in the device scheduling weight table is updated. The cumulative allocation count value of the target printing device is incremented by one, and the current task is removed from the corresponding waiting task queue.

[0132] The single-round scheduling process is repeated until the print task queues of all active spatiotemporal units are empty. Each print task is then combined with its corresponding target printing device to obtain a task allocation combination. All task allocation combinations are summarized to generate the final print task allocation result table. It should be noted that, in order to avoid the printing device's scheduling weight remaining low due to continuous selection over a long period, which would affect the fairness of subsequent task allocation and system load balancing, the scheduling weight of all printing devices in the device scheduling weight table is reset to the corresponding average scheduling score after each weight recovery cycle. This allows the printing devices to regain a balance in the next round of scheduling, ensuring that each printing device has a fair possibility of being allocated and maintaining the overall stability of the system.

[0133] The single-device matching operation method is as follows: Based on the business area to which the current task belongs, obtain all scheduling priority scores for the corresponding row from the scheduling priority scoring matrix to form a candidate device scoring sequence; multiply each scheduling priority score in the candidate device scoring sequence by the scheduling weight of the corresponding printing device in the device scheduling weight table to obtain the real-time scheduling score of each printing device; introduce a load balancing constraint mechanism to adjust the real-time scheduling score to obtain the adjusted scheduling score of each printing device; compare the adjusted scheduling scores of all printing devices, and select the printing device with the highest adjusted scheduling score as the target printing device for the current task;

[0134] The load balancing constraint mechanism works as follows: The average cumulative allocation count for all printing devices is calculated to obtain the average allocation quantity; for each printing device, the difference between the corresponding cumulative allocation count and the average allocation quantity is calculated to obtain the allocation deviation; if the allocation deviation is greater than a preset balance deviation threshold, the real-time scheduling score of the corresponding printing device is multiplied by a preset penalty coefficient to obtain an adjustment scheduling score; if the allocation deviation is less than or equal to the balance deviation threshold, the real-time scheduling score is used as the adjustment scheduling score; The balance deviation threshold and penalty coefficient are preset by those skilled in the art based on actual conditions; it should be noted that the purpose of introducing the load balancing constraint mechanism is to avoid excessive concentration of orders on at least a few printing devices.

[0135] The method for multi-device collaborative scheduling decision-making is as follows: Based on the business region to which the large order belongs, obtain all scheduling priority scores for the corresponding row from the scheduling priority scoring matrix and mark them as candidate priority scores; sort all printing devices in descending order according to their corresponding candidate priority scores, and select the top-ranked devices. A set of 1-bit printing devices is used as a collaborative candidate device set; among them... The target number of partitions is defined. Based on the geographical coordinates of each printing device in the collaborative candidate device set, the straight-line distance between any two printing devices is calculated sequentially and marked as the device straight-line distance. The mean of all device straight-line distances is calculated to obtain the average device spacing. The standard deviation of the remaining health life index corresponding to all printing devices in the collaborative candidate device set is calculated to obtain the health index dispersion. Corresponding collaborative weights are set for the average device spacing and the health index dispersion, and the average device spacing and the health index dispersion are weighted and summed based on the collaborative weights to obtain the collaborative cost value. The collaborative cost value is normalized, and the difference between 1 and the normalized collaborative cost value is calculated. The value is used to obtain the collaboration fit degree; if the collaboration fit degree is greater than the preset collaboration threshold, the scheduling mode of the corresponding large order is marked as the collaboration scheduling mode, and a one-to-one correspondence is established between each order segment and each printing device in the collaboration candidate device set arranged in descending order of scheduling priority score, according to the order segment number corresponding to each order segment of the large order; if the collaboration fit degree is less than or equal to the collaboration threshold, the scheduling mode of the corresponding large order is marked as the serial scheduling mode, and the printing device with the highest scheduling priority score in the collaboration candidate device set is used as the designated allocation device for the corresponding large order; wherein, the collaboration threshold is preset by those skilled in the art according to the actual situation;

[0136] The device matching operation in the collaborative scheduling mode is as follows: Based on the fragmentation number of the current task, obtain the corresponding printing device from the collaborative candidate device set and mark it as a candidate device; according to the load balancing constraint mechanism, determine whether the candidate device meets the load balancing constraint; if it meets the load balancing constraint, use the candidate device as the target printing device for the corresponding current task; if it does not meet the load balancing constraint, obtain other available devices, which are all printing devices other than those in the collaborative candidate device set; sort all other available devices in descending order of scheduling priority score, verify whether each other available device meets the load balancing constraint in turn, and use the first other available device that meets the constraint as the target printing device for the current task;

[0137] In serial scheduling mode, the device matching operation method is as follows: the designated device for the large order corresponding to the current task is used as the target printing device for the current task.

[0138] Methods for task scheduling based on the final print task allocation result table include:

[0139] Step S504: Schedule all print tasks according to the final print task allocation result table, and perform emergency order preemptive scheduling during the task scheduling process;

[0140] Step S505: After the task scheduling is completed, perform scheduling effect feedback and parameter adaptive learning.

[0141] In step S504 above, the method for preemptive scheduling of emergency orders includes:

[0142] During task scheduling, the system monitors in real time whether there are any newly submitted urgent priority orders (i.e., orders marked as urgent priority) in the order submission data stream; if there are any newly submitted urgent priority orders, the system performs preemptive scheduling for urgent orders.

[0143] Specifically, newly submitted priority orders are marked as urgent orders; based on the business region to which the urgent order belongs, all scheduling priority scores for the corresponding row are obtained from the scheduling priority scoring matrix; simultaneously, the load saturation and remaining health life index of all printing devices are obtained; corresponding emergency response weights are set for the scheduling priority score, the reciprocal of the load saturation, and the remaining health life index, and a weighted summation is performed based on the emergency response weights to obtain the emergency response capability score for each printing device; wherein, each emergency response weight is pre-set by those skilled in the art according to the emergency scheduling strategy; the emergency response capability scores of all printing devices are compared, and the printing device with the highest emergency response capability score is selected as the target printing device for the urgent order and marked as an urgent printing device; the current task queue corresponding to the urgent printing device is obtained, and the urgent order is inserted at the beginning of the corresponding current task queue.

[0144] In step S505 above, the methods for scheduling effect feedback and parameter adaptive learning include:

[0145] After task scheduling is completed, real-time scheduling effect feedback data is collected. The scheduling effect feedback data includes the actual allocation time of each print task and the remaining health life index of the corresponding target printing device. The actual allocation time refers to the time from when the print task enters the print task queue to when the target device is matched. Based on the scheduling effect feedback data, scheduling effect evaluation indicators are calculated. The scheduling effect evaluation indicators include load balancing index, device health protection index, and latency optimization index.

[0146] Specifically, the cumulative allocation count of all printing devices is obtained, the standard deviation of all cumulative allocation counts is calculated, and the allocation dispersion is obtained. The allocation dispersion is normalized, and the difference between 1 and the normalized allocation dispersion is calculated to obtain the load balancing index. The higher the load balancing index, the more evenly the printing tasks are distributed.

[0147] The average remaining health lifetime index of all target printing devices is calculated to obtain the average target device health index, which is used as the device health protection index. The higher the device health protection index, the more the scheduling strategy tends to allocate tasks to printing devices with better health status.

[0148] The average allocation time is calculated by averaging all actual allocation times. The average allocation time is then normalized, and the difference between 1 and the normalized average allocation time is calculated to obtain the delay optimization index. The higher the delay optimization index, the better the delay control effect of the scheduling.

[0149] The load balancing index, equipment health protection index, and latency optimization index are each assigned a corresponding evaluation weight, and a weighted sum is calculated based on these weights to obtain a comprehensive scheduling effect score. Each evaluation weight is pre-set by those skilled in the art according to business objectives. The comprehensive scheduling effect score is compared with a preset performance benchmark value. If the comprehensive scheduling effect score is less than the benchmark value, the scheduling effect is deemed unsatisfactory, and parameter adaptive learning is required. If the comprehensive scheduling effect score is greater than or equal to the benchmark value, the scheduling effect is deemed satisfactory, and parameter adaptive learning is not required. The performance benchmark value is pre-set by those skilled in the art based on actual conditions.

[0150] The parameter adaptive learning method is as follows: if the load balancing index is the lowest, the penalty coefficient is increased to strengthen the load balancing constraint; if the equipment health protection index is the lowest, the protection fusion weight is increased to strengthen equipment health protection; if the latency optimization index is the lowest, the latency fusion weight is increased to strengthen latency optimization; the specific adjustment range of each parameter is determined by those skilled in the art based on the degree of deviation of the scheduling effect and the business optimization goal.

[0151] This embodiment collects order submission data streams in real time by deploying order entry nodes in each business area. Based on time granularity parameters and business area divisions, it constructs a spatiotemporal heat matrix of order traffic. Combined with spatiotemporal aggregation operations of order count accumulation and weighted traffic accumulation, it achieves accurate perception and quantitative analysis of the regional distribution characteristics and temporal fluctuation patterns of order traffic. This effectively overcomes the shortcomings of existing technologies that only focus on immediate load status while ignoring the spatiotemporal distribution characteristics of order traffic. By acquiring the cumulative runtime records, historical fault repair logs, and periodic maintenance records of each printing device, and combining them with a predefined standard lifespan decay curve model, it dynamically evaluates the remaining healthy lifespan index of each printing device. This comprehensively considers the combined effects of operating losses, fault damage, and maintenance recovery on the device's health status, effectively solving the problem of existing technologies neglecting the difference between cumulative operating losses and remaining healthy lifespan, leading to overuse and accelerated aging of some devices. Furthermore, by calculating the network transmission between each business area and each printing device… This system constructs a latency cost matrix by combining transmission latency and logistics delivery latency, comprehensively considering the impact of network topology, network load status, geographical distance, and regional delivery complexity on transmission and delivery efficiency. This effectively optimizes the transmission latency and delivery cost of cross-regional orders. By constructing a device load protection weight vector and introducing a nonlinear protection weight mapping function and an adaptive fusion adjustment mechanism to generate a scheduling priority scoring matrix, it achieves intelligent balanced scheduling for device protection and latency optimization. Through intelligent order task fragmentation and a multi-device collaborative scheduling decision mechanism, it fully utilizes the parallel processing capabilities of multiple printing devices to improve the overall processing efficiency of large orders. An emergency order preemptive scheduling mechanism ensures that urgent priority orders receive rapid response and timely processing. Through scheduling effect feedback and parameter adaptive learning mechanisms, it continuously optimizes and improves scheduling strategies, thereby ensuring optimal scheduling performance in complex and ever-changing business environments and significantly improving the economic efficiency of overall resource allocation and the quality of printing services.

[0152] Example 2

[0153] This application also provides an electronic device. The electronic device may include one or more processors and one or more memories. The memories store computer-readable code that, when executed by the one or more processors, can perform a dynamic load balancing distribution method for printing devices based on real-time order traffic, as described above.

[0154] The method according to the embodiments of this application can also be implemented using the architecture of the electronic device shown in this application. The electronic device may include a bus, one or more CPUs, ROM, RAM, a communication port connected to a network, input / output, a hard disk, etc. The storage device in the electronic device, such as ROM or hard disk, may store the dynamic load balancing allocation method for printing devices based on real-time order traffic provided in this application. Furthermore, the electronic device may also include a user interface. Of course, the architecture shown in this application is merely exemplary; when implementing different devices, one or more components in the electronic device shown in this application may be omitted according to actual needs.

[0155] Example 3

[0156] Please refer to the accompanying drawings. One embodiment of this application discloses a computer-readable storage medium. The computer-readable storage medium stores computer-readable instructions. When executed by a processor, the computer-readable instructions can perform a dynamic load balancing allocation method for a printing device based on real-time order traffic, as described in the above-described embodiments of this application. The storage medium includes, but is not limited to, volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and cache memory. Non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc.

[0157] Furthermore, according to embodiments of this application, the processes described in the above-referenced flowcharts can be implemented as computer software programs. For example, this application provides a non-transitory machine-readable storage medium storing machine-readable instructions that can be executed by a processor to perform instructions corresponding to the method steps provided in this application, such as a dynamic load balancing allocation method for printing equipment based on real-time order traffic. When this computer program is executed by a central processing unit (CPU), it performs the functions defined in the method of this application.

[0158] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

[0159] All formulas in this manual are dimensionless and calculated numerically. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters and thresholds in the formulas are set by those skilled in the art according to the actual situation.

[0160] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

Claims

1. A dynamic load balancing method for printing equipment based on real-time order traffic, characterized in that, include: Step S1: Collect order submission data streams from order entry nodes distributed across multiple business regions in real time, extract the business region and submission time of each order in the order submission data stream, and construct a spatiotemporal heat matrix of order traffic through spatiotemporal aggregation calculation; Step S2: Obtain the cumulative runtime record, historical fault repair log and periodic maintenance record of each printing device, and combine them with the predefined standard life decay curve model to dynamically evaluate the remaining health life index of each printing device at the current moment. Step S3: Obtain the geographical coordinates of each business area and each printing device in the order traffic spatiotemporal heat matrix, and calculate the network transmission latency and logistics delivery latency of order transmission from each business area to each printing device based on the geographical coordinates to form a latency cost matrix; Step S4: Based on the remaining health life index of each printing device, construct the device load protection weight vector, and perform dynamic fusion calculation on the device load protection weight vector and the latency cost matrix to generate the scheduling priority scoring matrix; Step S5: Based on the scheduling priority scoring matrix, the adaptive weighted round-robin scheduling algorithm is used to match the orders to be assigned to each spatiotemporal unit in the order traffic spatiotemporal heat matrix with the target printing device, generate the final printing task allocation result table, and perform task scheduling based on the final printing task allocation result table. The method for matching target printing devices is as follows: Initialize the device scheduling weight table based on the scheduling priority scoring matrix; design a dynamic weight decay mechanism: preset weight decay coefficient and weight recovery period; where the weight decay coefficient controls the decay magnitude of the scheduling weight after each order allocation, and the weight recovery period controls the time interval for the scheduling weight to recover to its initial state; both the weight decay coefficient and weight recovery period are preset according to the load balancing strategy; determine the active spatiotemporal units in the order traffic spatiotemporal heat matrix and obtain the print task queues for each active spatiotemporal unit; define a single-round scheduling process: determine the current task in the print task queue; for the current task, determine the task type and perform the corresponding device matching operation to obtain the target printing device for the current task; based on the dynamic weight decay mechanism, perform a dynamic weight update operation on the target printing device for the current task; repeat the single-round scheduling process until the print task queues for all active spatiotemporal units are empty.

2. The dynamic load balancing method for printing equipment based on real-time order traffic as described in claim 1, characterized in that, Methods for constructing a spatiotemporal heat matrix of order traffic include: Calculate the time offset of each order relative to midnight of the day and mark it as the intraday time offset; preset time granularity parameters and business cycle, and divide the business cycle into multiple consecutive time windows according to the time granularity parameters; determine the time window to which each order belongs based on the intraday time offset and time granularity parameters of each order; initialize the order traffic spatiotemporal heat matrix according to the total number of time windows and the total number of business regions; where each element in the traffic spatiotemporal heat matrix corresponds to a spatiotemporal unit; The spatiotemporal aggregation operation process is defined as follows: Iterate through each order submission message in the order submission data stream, and locate the corresponding spatiotemporal unit in the order traffic spatiotemporal heat matrix based on the business region and time window of each order submission message. The order submission data stream includes the order submission message for each order, which contains basic order information and order attribute information. Simultaneously, whenever an order is located in a spatiotemporal unit, perform a traffic accumulation operation on the corresponding spatiotemporal unit to obtain the order count value and cumulative weighted traffic value for that unit. Finally, perform a heat conversion on the cumulative weighted traffic value of each spatiotemporal unit to obtain the traffic heat value for that unit. The order count and traffic heat value of each spatiotemporal unit are integrated to form a spatiotemporal heat record for each spatiotemporal unit; the spatiotemporal heat records of all spatiotemporal units are then sequentially filled into the order traffic spatiotemporal heat matrix to complete the construction of the order traffic spatiotemporal heat matrix.

3. The method for dynamic load balancing of printing equipment based on real-time order traffic according to claim 2, characterized in that, Methods for dynamically assessing the remaining healthy life expectancy index include: Based on the standard life decay curve model, the corresponding basic decay value, fault decay factor, and maintenance recovery value are calculated sequentially for each printing device. The standard life decay curve model includes a basic decay function, a fault accelerated decay function, and a maintenance recovery function. The basic decay value is calculated using the basic decay function and accumulated runtime records; the fault decay factor is calculated using the fault accelerated decay function and historical fault repair logs; and the maintenance recovery value is calculated using the maintenance recovery function and periodic maintenance records. Based on the basic decay value, fault decay factor, and maintenance recovery value of each printing device, the original health index of each printing device is calculated. Boundary constraints are applied to the original health indices of all printing devices sequentially, and the original health index after boundary constraint processing is used as the remaining healthy life index of each printing device.

4. The dynamic load balancing method for printing equipment based on real-time order traffic as described in claim 3, characterized in that, Methods for generating the time delay cost matrix include: Obtain the network path information corresponding to the network path between the order entry node and each printing device in each business area, and calculate the basic path latency between each business area and each printing device based on the network path information; obtain the current network load rate between the order entry node and each printing device in each business area, and calculate the network congestion coefficient based on the current network load rate; calculate the product of the basic path latency and the network congestion coefficient to obtain the network transmission latency between each business area and each printing device; integrate all network transmission latencies to construct a network latency matrix; Based on the geographical coordinates of each business area and each printing device, calculate the straight-line distance between each business area and each printing device; based on the delivery area where each business area and each printing device are located, determine the corresponding average delivery speed and regional delivery coefficient; based on the straight-line distance and average delivery speed, calculate the basic delivery time; calculate the product of the basic delivery time and the regional delivery coefficient to obtain the logistics delivery delay between each business area and each printing device; integrate all logistics delivery delays to construct a logistics delay matrix. Corresponding delay weights are set for network transmission delay and logistics delivery delay, and the network delay matrix and logistics delay matrix are weighted and merged based on the delay weights to form a delay cost matrix.

5. A dynamic load balancing distribution method for printing equipment based on real-time order traffic according to claim 4, characterized in that, Methods for constructing device load protection weight vectors include: Based on the remaining health life index of each printing device, the health status of each printing device is classified into four levels: healthy, sub-healthy, critical, and dangerous. Based on the health status of each printing device, a corresponding basic protection coefficient is set. Based on the remaining health life index and the basic protection coefficient, the load protection weight of each printing device is calculated through a nonlinear protection weight mapping function. Obtain real-time task queue information and rated processing capacity for each printing device; the real-time task queue information includes the estimated execution time and task load of each printing task in the current task queue; sum the estimated execution time and task load of all printing tasks in the current task queue for the same printing device to obtain the queue backlog time and current total load for each printing device; calculate the ratio of the current total load to the rated processing capacity to obtain the load saturation of each printing device; Calculate the load adjustment factor for each printing device based on its load saturation; calculate the dynamic protection weight for each printing device based on the load protection weight and the load adjustment factor; integrate the dynamic protection weights of all printing devices to form a device load protection weight vector.

6. The dynamic load balancing method for printing equipment based on real-time order traffic according to claim 5, characterized in that, Methods for generating scheduling priority score matrices include: Based on the total number of business areas and the total number of printing devices, an initial scheduling priority scoring matrix is ​​established; each element in the scheduling priority scoring matrix corresponds to a priority scoring unit; dynamic fusion operation rules are designed: corresponding fusion weights are set for the comprehensive latency cost and dynamic protection weight, and the fusion weights include latency fusion weights and protection fusion weights; the sum of the latency fusion weights and protection fusion weights is equal to 1; an adaptive fusion adjustment mechanism is introduced: the mean of all dynamic protection weights in the device load protection weight vector is calculated to obtain the average protection weight; the average protection weight is compared with the preset protection weight benchmark value, and the protection optimization weight and latency optimization weight are calculated based on the comparison result; Traverse each priority scoring unit in the scheduling priority scoring matrix, and obtain the comprehensive latency cost of the corresponding position in the latency cost matrix, as well as the dynamic protection weight of the corresponding printing device in the device load protection weight vector. Based on the latency optimization weight and the protection optimization weight, perform a fusion calculation on the comprehensive latency cost and dynamic protection weight of the same priority scoring unit to obtain the original priority score of each priority scoring unit. Perform reverse transformation on each original priority score in turn to obtain the scheduling priority score of each priority scoring unit. Fill the scheduling priority score of each priority scoring unit into the scheduling priority scoring matrix in turn to complete the construction of the scheduling priority scoring matrix.

7. The dynamic load balancing method for printing equipment based on real-time order traffic according to claim 6, characterized in that, Methods for generating the final print job assignment result table include: Step S501: Obtain the orders to be assigned from each active spatiotemporal unit and combine them to form a queue of orders to be assigned; Step S502: For the orders to be assigned in the order queue corresponding to each active spatiotemporal unit, perform intelligent task slicing in sequence to obtain the print task queue for each active spatiotemporal unit. Step S503: Based on the adaptive weighted round-robin scheduling algorithm, target devices are matched for the print tasks in each print task queue, and the final print task allocation result table is generated. Methods for task scheduling based on the final print task allocation result table include: Step S504: Schedule all print tasks according to the final print task allocation result table, and perform emergency order preemptive scheduling during the task scheduling process; Step S505: After the task scheduling is completed, perform scheduling effect feedback and parameter adaptive learning.

8. The dynamic load balancing method for printing equipment based on real-time order traffic according to claim 7, characterized in that, Methods for obtaining the queue of tasks to be printed include: For each order in the queue to be assigned, order sharding identification is performed sequentially to divide all orders into standard orders and large orders; for each large order, intelligent sharding operation is performed sequentially to obtain the order shards corresponding to each large order; Each large order's corresponding order fragment and each standard order are treated as printable tasks. Based on the standard orders and order fragments corresponding to each active spatiotemporal unit, printable tasks for each active spatiotemporal unit are obtained. Printable tasks for the same active spatiotemporal unit are aggregated to form a printable task queue for each active spatiotemporal unit.

9. A dynamic load balancing method for printing equipment based on real-time order traffic as described in claim 8, characterized in that, Methods for generating the final print job assignment result table include: Define a single-round scheduling process: Sort all active spatiotemporal units in descending order based on the traffic heat value corresponding to each active spatiotemporal unit; based on the sorting result of the active spatiotemporal units, select the first task in the queue of tasks to be printed corresponding to each active spatiotemporal unit as the current task. For the current task, determine the task type and perform the corresponding device matching operation to obtain the target printing device for the current task. Specifically, if the current task is a standard order, perform a single device matching operation directly. If the current task is the first order segment of a large order, perform a multi-device collaborative scheduling decision to determine the corresponding scheduling mode and perform the corresponding device matching operation according to the scheduling mode. If the current task is not the first order segment of a large order, perform the corresponding device matching operation according to the scheduling mode of the corresponding large order. The scheduling modes include collaborative scheduling mode and serial scheduling mode. Each print job is paired with its corresponding target printing device to obtain a job assignment combination; all job assignment combinations are then summarized to generate the final print job assignment result table.

10. A dynamic load balancing method for printing equipment based on real-time order traffic according to claim 9, characterized in that, Methods for providing feedback on scheduling performance include: After task scheduling is completed, real-time scheduling effect feedback data is collected. The scheduling effect feedback data includes the actual allocation time of each printing task and the remaining health life index of the corresponding target printing device. Based on the scheduling effect feedback data, scheduling effect evaluation indicators are calculated. Among them, scheduling effect evaluation indicators include load balancing index, device health protection index and latency optimization index. The load balancing index, equipment health protection index, and latency optimization index are assigned corresponding evaluation weights, and a weighted summation is performed based on the evaluation weights to obtain a comprehensive scheduling effect score. The comprehensive scheduling effect score is compared with the preset effect benchmark value, and the parameter adaptive learning is determined based on the comparison result.

Images