Printer energy consumption optimization control system and method based on artificial intelligence

Abstract

The invention discloses an artificial intelligence-based printer energy consumption optimization control system and method, and relates to the technical field of printer energy consumption optimization control, and the artificial intelligence-based printer energy consumption optimization control method specifically comprises the following steps: 1, collecting multi-source data of a laser printer, the method comprises the steps of 1, acquiring original data, transmitting the acquired original data to a cloud AI analysis platform or a local edge calculation program for processing, 2, collecting historical data, quantifying the relevance between the comprehensive environmental comfort index and the power consumption of the laser printer, and defining different task processing time periods, and 3, predicting the historical data by using an LSTM prediction model, and according to a prediction result, calculating the power consumption of the laser printer according to the prediction result. According to the scheme, through multi-source data fusion and AI prediction, a weighted comprehensive environment comfort index is provided, multi-parameter cooperative influence is quantized, the future task probability and the environment trend are predicted based on historical data, and the working mode is dynamically adjusted to achieve balance of energy consumption and efficiency.

Description

Technical Field

[0001] This invention belongs to the field of printer energy consumption optimization technology, specifically relating to an artificial intelligence-based printer energy consumption optimization control system and method. Background Technology

[0002] As a core device in office environments, the energy management and environmental adaptability of laser printers directly impact operating costs, equipment lifespan, and user experience. However, traditional laser printers have significant shortcomings in the following aspects: Existing devices typically reduce power consumption using simple sleep / wake-up strategies (such as entering sleep mode after a period of inactivity), lacking fine-grained control over the dynamic correlation between environmental parameters (temperature, humidity, light intensity) and task priorities. For example, high-temperature environments may cause the fixing components to operate under continuous high loads, while traditional solutions cannot proactively adjust their operating modes to adapt to environmental changes. Print jobs are typically processed in the order they are submitted, without prioritizing them (e.g., urgent documents versus regular documents) or considering environmental factors (e.g., high humidity may cause paper jams). Low-priority jobs may increase overall energy consumption because they are not properly allocated to low-power periods. The device cannot predict the task load and environmental changes in the future, resulting in a delay in switching working modes (such as starting warm-up only when suddenly entering a high-load printing task, increasing instantaneous power consumption). Therefore, there is an urgent need for an AI-based printer energy consumption optimization and control method to solve the above problems. Summary of the Invention

[0003] The purpose of this invention is to provide an artificial intelligence-based printer energy consumption optimization control system and method to solve the technical problems of crude energy consumption management, lack of intelligence in task scheduling, and sluggish switching of working modes in the prior art.

[0004] To achieve the above objectives, the present invention adopts the following technical solution: An AI-based printer energy consumption optimization control system and method, including: Step 1: Collect multi-source data from the laser printer and transmit the collected raw data to a cloud-based AI analysis platform or a local edge computing program for processing; By extracting data features from multi-source data of laser printers, the comprehensive environmental comfort index and idle time are obtained; Step 2: Collect historical data, quantify the correlation between the overall environmental comfort index and the power consumption of the laser printer, and define different task processing time periods; Determine the correlation between the comprehensive environmental comfort index and average power consumption, and define different task processing time periods based on the weight of time period type, average power consumption, and comprehensive environmental comfort index; Step 3: Use the LSTM prediction model to predict historical data, and adjust the laser printer decision based on the prediction results.

[0005] Furthermore, multi-source data is collected from the laser printer, specifically through the following methods: The laser printer uses an external environmental sensor to collect real-time environmental parameters around the laser printer and its own operating status data to classify the laser printer's working modes, including sleep, warm-up, printing, and deep sleep. The system records the time point at which each user submits a print job. Print jobs are divided into high-priority and low-priority jobs. The print time order of different priority print jobs is obtained by sorting them according to the time point. The number of jobs to be processed is determined by the network connection status of the laser printer. Print job data for each time period is collected anonymously by logging the server or local logs connected to the laser printer.

[0006] Furthermore, by extracting data features from multi-source data of the laser printer, the comprehensive environmental comfort index and idle time are obtained. The specific method is as follows: Divide the time period into t-hour intervals, determine the average value and maximum difference of environmental parameters for each time period, set different optimal working environment parameter values ​​for different types of environmental parameters, and the single type of environmental comfort index is equal to the difference between the preset constant value and the deviation value of the environmental parameter of that type. Set different weight coefficients for different types of environmental comfort indices, and obtain the comprehensive environmental comfort index by weighted summation. The absolute value of the difference between the average value of the environmental parameters and the optimal working environment parameter value is divided by the maximum difference of the environmental parameters to obtain the environmental parameter deviation value. The idle time is calculated based on the difference between the current time point and the time point of the last print or user operation.

[0007] Furthermore, historical data is collected, specifically through the following methods: Collect historical data for duration H, determine the laser printer power consumption P at each time point based on the historical data, and calculate the average power consumption Pc of the laser printer in each time period. Set a preset threshold O for power consumption variation, and iterate through the relationship between P and Pc at each time point within the time period, removing... The environmental parameters at the corresponding time point are replaced by the average value of the environmental parameters at other time points within the same time period.

[0008] Furthermore, the correlation between the overall environmental comfort index and the power consumption of laser printers was quantified using the following method: By calculating the Pearson correlation coefficient between the comprehensive environmental comfort index and average power consumption at different historical time periods, it was determined that there is a linear relationship between the two. When the comprehensive environmental comfort index increases, the average power consumption during that period decreases, indicating that the more comfortable the environment, the lower the power consumption. By integrating the baseline power consumption, power consumption influence coefficient, and adjustment coefficient, a linear regression model between the comprehensive environmental comfort index and average power consumption was established. Historical data is used to fit the baseline power consumption, power consumption influence coefficient, and adjustment coefficient. Historical real values ​​are used to verify the accuracy of the model. The model is judged to be accurate when the error is less than or equal to the preset error threshold.

[0009] Furthermore, different task processing time periods are defined, specifically as follows: Determine the time period type, preset the type weight of each time period, define different task processing time periods based on the time period type weight, average power consumption and comprehensive environmental comfort index, preset the comprehensive environmental comfort index threshold and power consumption threshold, calculate the product of the type weight of different time periods and the average power consumption of that time period, and record the result as the power consumption level value. When the overall environmental comfort index is greater than or equal to the overall environmental comfort index threshold and the power consumption level is less than the power consumption threshold, the time period is defined as a priority processing period. When the overall environmental comfort index is less than the overall environmental comfort index threshold and the power consumption level is greater than the power consumption threshold during a certain period, the period is defined as an avoidance processing period. Time periods that are neither priority processing periods nor avoidance processing periods are designated as regular processing periods.

[0010] Furthermore, the LSTM prediction model is used to predict historical data. The specific method is as follows: Using J as the period, historical data from each time period within the historical time of the laser printer are input into the trained LSTM prediction model to predict historical data in different future time periods. By inputting the historical comprehensive environmental comfort index and historical average power consumption into the LSTM prediction model, the comprehensive environmental comfort index and average power consumption for future time periods can be obtained. The number of printing tasks, the time point when printing is triggered, the idle time, and the average processing time of historical tasks are aggregated into a feature vector and input into the LSTM prediction model. The LSTM gating mechanism captures the processing time points of laser printer tasks. The input feature sequence is regarded as a time-dependent input. Based on the changes in the amount of tasks processed in the historical time period, the probability of tasks at each time point in the future time period is output. The output layer compresses the prediction results to the 0-1 interval by using the Sigmoid activation function to represent the probability of tasks occurring at different time points. The model is supervised by the real task probability labels in the historical data.

[0011] Furthermore, based on the prediction results, the decision regarding the laser printer is adjusted, specifically as follows: Predict the task probability value q at the next time point, preset the task probability threshold Q. If the laser printer's continuous idle time at the current time point is greater than the preset time threshold Y, and q is less than or equal to Q, change the current working state of the laser printer to sleep state. If q is greater than Q, change the current working state of the laser printer to warm-up state. If q is less than or equal to Q for L consecutive time points, change the current working state of the laser printer to deep sleep state. If the continuous idle time of laser printing at the current moment is less than or equal to the preset time threshold Y, the working status of the laser printer will not be changed; Identify all pending tasks at the current time point, and prioritize printing the highest priority tasks in the next time period according to their printing time order. Identify all low-priority tasks in the pending tasks and pre-set a task completion time limit K for each low-priority task; Determine the priority processing period, the avoidance processing period, and the normal processing period within the future K-time period; If a priority processing period exists within the next K time period, all low-priority tasks will be assigned to the priority processing period for processing. If no priority processing period exists within the next K time period, all low-priority tasks will be assigned to the regular processing period for processing. If neither a priority processing period nor a regular processing period exists within the next K time period, low-priority tasks will be processed in chronological order. If any low-priority task is not completed within K time periods, it will be automatically converted into a high-priority task for processing.

[0012] Furthermore, the present invention also provides an artificial intelligence-based printer energy consumption optimization control system, comprising: The multi-source data acquisition and processing module collects multi-source data from the laser printer and transmits the collected raw data to the cloud AI analysis platform or local edge computing program for processing. The task processing time period definition module collects historical data, quantifies the correlation between the comprehensive environmental comfort index and the power consumption of the laser printer, and defines different task processing time periods. The laser printer decision management module uses an LSTM prediction model to predict historical data and adjusts laser printer decisions based on the prediction results.

[0013] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are: 1. This invention defines a single type of environmental comfort index by calculating the average value and maximum difference of environmental parameters, and obtains a comprehensive environmental comfort index by weighted summation. This model can quantify the impact of the environment on equipment performance, reflect the environmental suitability in real time, guide the equipment to adjust its working mode, calculate the idle time based on the time difference between the current time point and the last operation, and dynamically adjust the equipment status. 2. This invention verifies that the comprehensive environmental comfort index and average power consumption are linearly negatively correlated by calculating the Pearson correlation coefficient between the two. By setting a power consumption change threshold, it identifies and removes power consumption data at abnormal time points (such as sensor failure or instantaneous power consumption surge caused by sudden high load tasks), and uses the average environmental parameters at other times during the same period to replace them, thus avoiding interference from outliers on the accuracy of the model. It also sets the weight of time period type (such as the weight can be increased if the task density is high on weekdays) to further refine the division of time periods. 3. This invention uses an LSTM model to predict the future comprehensive environmental comfort index and average power consumption, thereby sensing environmental changes and power consumption trends in advance and providing a basis for adjusting the device state. Based on features such as the number of historical tasks, trigger time points, and idle time, it predicts the probability of tasks occurring at future time points, dynamically adjusts the device state, reduces unnecessary energy consumption, and reduces frequent state switching caused by short-term fluctuations through a time threshold Y and a continuous judgment mechanism, thereby improving device stability. It prioritizes high-priority tasks and intelligently allocates low-priority tasks, thereby improving overall efficiency. Attached Figure Description

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

[0015] Figure 1 The diagram illustrates the steps of the AI-based laser printer energy consumption optimization control method of the present invention. Figure 2 The flowchart of the laser printer energy consumption optimization control method based on artificial intelligence of the present invention is shown; Figure 3 The diagram shows a block diagram of the laser printer energy consumption optimization control system based on artificial intelligence according to the present invention. Detailed Implementation

[0016] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0017] Example 1, such as Figure 1 The energy consumption optimization and control method for laser printers based on artificial intelligence, as shown, specifically includes the following steps: Step 1: Collect multi-source data from the laser printer and transmit the collected raw data to a cloud-based AI analysis platform or a local edge computing program for processing.

[0018] The laser printer uses external environmental sensors (such as temperature and humidity sensors, and light sensors) to collect real-time environmental parameters (e.g., temperature, humidity, and light intensity) around the printer. It also collects real-time operating status data of the laser printer itself, including the current power consumption measured by the built-in current / power sensor, and the processor temperature and fuser temperature obtained by the built-in temperature sensor. The laser printer's working modes are divided into sleep (the display is turned off, the scanner head is turned off, and the CPU is reduced to a low-speed mode), warm-up (the fuser is heated to the target temperature, the printhead is cleaned and the ink is preheated, and other modules are in low-power standby mode), printing, and deep sleep (only the motion sensor is powered). Record the moment when a user submits a print job. Print jobs are divided into high-priority and low-priority jobs. Sort them according to the time point to obtain the print time order of different priority print jobs. Determine the number of jobs to be processed by the network connection status of the laser printer. Collect print job data for each time period, including print job time and print job number, by logging the server or local logs connected to the laser printer. The collected raw data is transmitted to a cloud AI analysis platform or a local edge computing program (such as the AI ​​acceleration chip built into the laser printer) through the network interface (wired / wireless) of the laser printer. The data is then cleaned and standardized to extract the data characteristics. Specifically, the data is divided into time periods of t, and the average value and maximum difference of environmental parameters in each time period are determined (the difference between the maximum and minimum values ​​of environmental parameters within the time period). Different optimal working environment parameter values ​​are set for different types of environmental parameters (such as optimal temperature, optimal humidity, and optimal light intensity). The comfort index of a single type of environment (such as temperature comfort index, humidity comfort index, and light comfort index) is equal to the difference between the preset constant value and the deviation value of that type of environmental parameter. Different weighting coefficients are set for different types of environmental comfort indices, and the weighted sum is used to obtain the comprehensive environmental comfort index. The absolute value of the difference between the average value of the environmental parameters and the optimal working environment parameter value is divided by the maximum difference of the environmental parameters to obtain the environmental parameter deviation value. The idle time is calculated based on the difference between the current time point and the time point of the last print or user operation.

[0019] Step 2: Collect historical data, quantify the correlation between the overall environmental comfort index and the power consumption of the laser printer, and define different task processing time periods; Collect historical data for duration H (historical environmental parameters, historical laser printer power consumption, historical print job data), determine the laser printer power consumption P at each time point based on the historical data, and calculate the average power consumption Pc of the laser printer over each time period. Set a preset threshold O for power consumption variation, and iterate through the relationship between P and Pc at each time point within the time period, removing... The environmental parameters at the corresponding time point are replaced by the average value of the environmental parameters at other time points within the same time period; By calculating the Pearson correlation coefficient between the comprehensive environmental comfort index and average power consumption over different historical time periods, a linear relationship was established. When the comprehensive environmental comfort index increases, the average power consumption decreases during that period, indicating that the more comfortable the environment, the lower the power consumption. A linear regression model between the comprehensive environmental comfort index and average power consumption was established, and the specific formula is shown below: ; in, Indicates average power consumption. This indicates the overall environmental comfort index. This indicates the reference power consumption (power consumption without environmental influence). This represents the power consumption impact coefficient, and Less than 0, Indicates the adjustment factor; Fitting using historical data , and The accuracy of the model is verified using historical real values, and the model is judged to be accurate when the error is less than or equal to a preset error threshold.

[0020] Determine the time period type (e.g., weekday time period, holiday time period), preset the type weight of each time period type, and define different task processing time periods based on the time period type weight, average power consumption, and overall environmental comfort index. The specific method is as follows: The comprehensive environmental comfort index threshold and power consumption threshold are preset. The product of the type weight of different time periods and the average power consumption of that time period is calculated, and the result is recorded as the power consumption level value. When the comprehensive environmental comfort index is greater than or equal to the comprehensive environmental comfort index threshold and the power consumption level is less than the power consumption threshold, the time period is defined as a priority processing period. Such periods are suitable for centralized processing of non-urgent tasks to reduce energy consumption. When the overall environmental comfort index is less than the overall environmental comfort index threshold and the power consumption level is greater than the power consumption threshold during a certain period, the period is defined as an avoidance processing period. Time periods that are neither priority processing periods nor avoid processing periods are recorded as regular processing periods; Step 3: Use the LSTM prediction model to predict historical data, and adjust the laser printer decision based on the prediction results.

[0021] Using J as the period (e.g., one day, one week), historical data from various time periods within the historical time of the laser printer are input into the trained LSTM prediction model to predict historical data for different future time periods. By inputting the historical comprehensive environmental comfort index and historical average power consumption into the LSTM prediction model, the comprehensive environmental comfort index and average power consumption for future time periods can be obtained. The number of printing tasks, the time point when printing is triggered, the idle time, and the average processing time of historical tasks are aggregated into a feature vector and input into the LSTM prediction model. The LSTM gating mechanism captures the processing time points of laser printer tasks. The input feature sequence is regarded as a time-dependent input. Based on the changes in the amount of tasks processed in the historical time period, the probability of tasks at each time point in the future time period is output. The output layer compresses the prediction results to the 0-1 interval by using the Sigmoid activation function to represent the probability of tasks occurring at different time points. The model is supervised by the real task probability labels in the historical data (such as whether tasks actually occurred in a certain time period in the past). By using the cross-entropy loss function, the error between the predicted value and the true value is minimized, and the system is retrained periodically with the latest data to adapt to changes in user behavior or environment (such as new device access or seasonal demand changes). Predict the task probability value q at the next time point, preset the task probability threshold Q. If the laser printer's continuous idle time at the current time point is greater than the preset time threshold Y, and q is less than or equal to Q, change the current working state of the laser printer to sleep state. If q is greater than Q, change the current working state of the laser printer to warm-up state. If q is less than or equal to Q for L consecutive time points, change the current working state of the laser printer to deep sleep state. If the continuous idle time of laser printing at the current moment is less than or equal to the preset time threshold Y, the working status of the laser printer will not be changed; Identify all pending tasks at the current time point, and prioritize printing the highest priority tasks in the next time period according to their printing time order. Identify all low-priority tasks in the pending tasks, and pre-set a task completion time limit K for each low-priority task, meaning that the printing task must be completed within K hours after the user submits the printing task. Based on the predicted values ​​of the comprehensive environmental comfort index and average power consumption for future time periods, priority processing periods, avoidance processing periods, and normal processing periods are determined within the future K-time period. If a priority processing period exists within the next K time period, all low-priority tasks will be assigned to that priority period for processing. If no priority processing period exists within the next K time period, all low-priority tasks will be assigned to the regular processing period for processing. If neither a priority processing period nor a regular processing period exists within the next K time period, low-priority tasks will be processed in chronological order. If any low-priority task is not completed within K time period, it will be automatically converted into a high-priority task for processing.

[0022] Example 2, as follows Figure 3 The AI-based energy consumption optimization control system for laser printers shown specifically includes: The multi-source data acquisition and processing module collects multi-source data from the laser printer and transmits the collected raw data to the cloud AI analysis platform or local edge computing program for processing. The task processing time period definition module collects historical data, quantifies the correlation between the comprehensive environmental comfort index and the power consumption of the laser printer, and defines different task processing time periods. The laser printer decision management module uses an LSTM prediction model to predict historical data and adjusts laser printer decisions based on the prediction results.

[0023] The above description is only a preferred specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with the technical field, within the technical scope disclosed by the present invention, who makes equivalent replacements or changes based on the technical solution and inventive concept of the present invention, should be covered by the scope of protection of the present invention.

[0024] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to specific implementations. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. A printer energy consumption optimization control method based on artificial intelligence, characterized in that, include: Step 1: Collect multi-source data from the laser printer and transmit the collected raw data to a cloud-based AI analysis platform or a local edge computing program for processing; By extracting data features from multi-source data of laser printers, the comprehensive environmental comfort index and idle time are obtained; Step 2: Collect historical data, quantify the correlation between the overall environmental comfort index and the power consumption of the laser printer, and define different task processing time periods; Determine the correlation between the comprehensive environmental comfort index and average power consumption, and define different task processing time periods based on the weight of time period type, average power consumption, and comprehensive environmental comfort index; Step 3: Use the LSTM prediction model to predict historical data, and adjust the laser printer decision based on the prediction results.

2. The printer energy consumption optimization control method based on artificial intelligence according to claim 1, characterized in that, The specific method for collecting multi-source data from a laser printer is as follows: The laser printer uses an external environmental sensor to collect real-time environmental parameters around the laser printer and its own operating status data to classify the laser printer's working modes, including sleep, warm-up, printing, and deep sleep. The system records the time point at which each user submits a print job. Print jobs are divided into high-priority and low-priority jobs. The print time order of different priority print jobs is obtained by sorting them according to the time point. The number of jobs to be processed is determined by the network connection status of the laser printer. Print job data for each time period is collected anonymously by logging the server or local logs connected to the laser printer.

3. The printer energy consumption optimization control method based on artificial intelligence according to claim 1, characterized in that, By extracting data features from multi-source data of laser printers, the comprehensive environmental comfort index and idle time are obtained. The specific method is as follows: Divide the time period into t-hour intervals, determine the average value and maximum difference of environmental parameters for each time period, set different optimal working environment parameter values ​​for different types of environmental parameters, and the single type of environmental comfort index is equal to the difference between the preset constant value and the deviation value of the environmental parameter of that type. Set different weight coefficients for different types of environmental comfort indices, and obtain the comprehensive environmental comfort index by weighted summation. The absolute value of the difference between the average value of the environmental parameters and the optimal working environment parameter value is divided by the maximum difference of the environmental parameters to obtain the environmental parameter deviation value. The idle time is calculated based on the difference between the current time point and the time point of the last print or user operation.

4. The printer energy consumption optimization control method based on artificial intelligence according to claim 1, characterized in that, The specific methods for collecting historical data are as follows: Collect historical data for duration H, determine the laser printer power consumption P at each time point based on the historical data, and calculate the average power consumption Pc of the laser printer in each time period. Set a preset threshold O for power consumption variation, and iterate through the relationship between P and Pc at each time point within the time period, removing... The environmental parameters at the corresponding time point are replaced by the average value of the environmental parameters at other time points within the same time period.

5. The printer energy consumption optimization control method based on artificial intelligence according to claim 1, characterized in that, The correlation between the overall environmental comfort index and the power consumption of laser printers is quantified using the following method: By calculating the Pearson correlation coefficient between the comprehensive environmental comfort index and average power consumption at different historical time periods, it was determined that there is a linear relationship between the two. When the comprehensive environmental comfort index increases, the average power consumption during that period decreases, indicating that the more comfortable the environment, the lower the power consumption. By integrating the baseline power consumption, power consumption influence coefficient, and adjustment coefficient, a linear regression model between the comprehensive environmental comfort index and average power consumption was established. Historical data is used to fit the baseline power consumption, power consumption influence coefficient, and adjustment coefficient. Historical real values ​​are used to verify the accuracy of the model. The model is judged to be accurate when the error is less than or equal to the preset error threshold.

6. The printer energy consumption optimization control method based on artificial intelligence according to claim 1, characterized in that, Define different task processing time periods, the specific method is as follows: Determine the time period type, preset the type weight of each time period, define different task processing time periods based on the time period type weight, average power consumption and comprehensive environmental comfort index, preset the comprehensive environmental comfort index threshold and power consumption threshold, calculate the product of the type weight of different time periods and the average power consumption of that time period, and record the result as the power consumption level value. When the overall environmental comfort index is greater than or equal to the overall environmental comfort index threshold and the power consumption level is less than the power consumption threshold, the time period is defined as a priority processing period. When the overall environmental comfort index is less than the overall environmental comfort index threshold and the power consumption level is greater than the power consumption threshold during a certain period, the period is defined as an avoidance processing period. Time periods that are neither priority processing periods nor avoidance processing periods are designated as regular processing periods.

7. The printer energy consumption optimization control method based on artificial intelligence according to claim 1, characterized in that, The LSTM prediction model is used to predict historical data. The specific method is as follows: Using J as the period, historical data from each time period within the historical time of the laser printer are input into the trained LSTM prediction model to predict historical data in different future time periods. By inputting the historical comprehensive environmental comfort index and historical average power consumption into the LSTM prediction model, the comprehensive environmental comfort index and average power consumption for future time periods can be obtained. The number of printing tasks, the time point when printing is triggered, the idle time, and the average processing time of historical tasks are aggregated into a feature vector and input into the LSTM prediction model. The LSTM gating mechanism captures the processing time points of laser printer tasks. The input feature sequence is regarded as a time-dependent input. Based on the changes in the amount of tasks processed in the historical time period, the probability of tasks at each time point in the future time period is output. The output layer compresses the prediction results to the 0-1 interval by using the Sigmoid activation function to represent the probability of tasks occurring at different time points. The model is supervised by the real task probability labels in the historical data.

8. The printer energy consumption optimization control method based on artificial intelligence according to claim 1, characterized in that, Based on the prediction results, the decision regarding the laser printer will be adjusted as follows: Predict the task probability value q at the next time point, preset the task probability threshold Q. If the laser printer's continuous idle time at the current time point is greater than the preset time threshold Y, and q is less than or equal to Q, change the current working state of the laser printer to sleep state. If q is greater than Q, change the current working state of the laser printer to warm-up state. If q is less than or equal to Q for L consecutive time points, change the current working state of the laser printer to deep sleep state. If the continuous idle time of laser printing at the current moment is less than or equal to the preset time threshold Y, the working status of the laser printer will not be changed; Identify all pending tasks at the current time point, and prioritize printing the highest priority tasks in the next time period according to their printing time order. Identify all low-priority tasks in the pending tasks and pre-set a task completion time limit K for each low-priority task; Determine the priority processing period, the avoidance processing period, and the normal processing period within the future K-time period; If a priority processing period exists within the next K time period, all low-priority tasks will be assigned to the priority processing period for processing. If no priority processing period exists within the next K time period, all low-priority tasks will be assigned to the regular processing period for processing. If neither a priority processing period nor a regular processing period exists within the next K time period, low-priority tasks will be processed in chronological order. If any low-priority task is not completed within K time periods, it will be automatically converted into a high-priority task for processing.

9. An AI-based printer energy consumption optimization control system, applied to the AI-based printer energy consumption optimization control method according to any one of claims 1-8, characterized in that, include: The multi-source data acquisition and processing module collects multi-source data from the laser printer and transmits the collected raw data to the cloud AI analysis platform or local edge computing program for processing. The task processing time period definition module collects historical data, quantifies the correlation between the comprehensive environmental comfort index and the power consumption of the laser printer, and defines different task processing time periods. The laser printer decision management module uses an LSTM prediction model to predict historical data and adjusts laser printer decisions based on the prediction results.

Images