A control method and system for reducing power consumption after screen-off of a mobile phone

Abstract

This invention proposes a method and system for reducing power consumption when a mobile phone screen is off. The method includes: acquiring users' daily behavior data and performing preliminary analysis on the fused time series of the behavior data; constructing a social behavior model and a shopping behavior model, and training them separately to predict users' behavioral trends and purchasing intentions; dynamically setting the priority of background tasks based on the prediction results; dynamically allocating resources based on priority; designing a low-power preloading mechanism to preload data and quickly access certain high-priority tasks when the user's screen is off, reducing the waiting time after the screen is turned on; and optimizing the prediction combination and resource allocation strategy of the social behavior model and the shopping behavior model based on changes in user habits to adapt to users' long-term needs. This invention overcomes several shortcomings of existing technologies in power management, especially in dynamically adapting to user needs, achieving more intelligent and efficient power management in the screen-off state.

Description

Technical Field

[0001] This invention belongs to the field of mobile phone power consumption control, and particularly relates to a control method and system for reducing power consumption when the mobile phone screen is off. Background Technology

[0002] With the rapid proliferation of mobile devices, smartphones have become indispensable tools in daily life. Smartphones offer increasingly diverse functions, ranging from basic communication and social interaction to online shopping, media consumption, and advertising, significantly increasing user dependence. However, with the increase in device functionality and diversified application demands, battery life has become a prominent issue. Even when the screen is off, numerous background tasks continue to run, especially applications involving social interaction, shopping recommendations, and advertising that frequently reside in the background. The continuous operation of these applications poses a significant challenge to battery life. Traditional power management strategies primarily rely on background task optimization schemes provided by application developers or unified scheduling by the operating system, typically based on simple "fixed rules" to pause or reduce background activity. However, this static rule-based management approach often fails to dynamically adapt to users' personalized needs and cannot effectively distinguish users' actual usage intentions and requirements after the screen is off.

[0003] For example, a user might be engaged in an important social interaction or shopping decision before the screen goes off. Existing background process management systems cannot intelligently recognize these scenarios, sometimes causing users to miss crucial push notifications or promotional information. Furthermore, when users don't need certain functions, such as in inactive social or shopping apps, the system fails to promptly stop background processes, further increasing battery waste. In addition, apps that push personalized ads and shopping recommendations frequently make network data requests even when the screen is off, continuing to consume power even if these notifications are useless for a short time after the screen is off. This "one-size-fits-all" background management mechanism fails to effectively adjust to user behavior patterns, resulting in a significant conflict between meeting user needs and optimizing power consumption.

[0004] While existing technologies attempt to optimize resource allocation through intelligent power management functions (such as using sensors to identify environmental conditions and freezing low-priority applications when the screen is off), most of these solutions remain at the level of simple "scene detection," failing to deeply analyze user behavior patterns and personalized needs. This results in insufficiently intelligent power optimization in real-world usage scenarios. Especially when applications are closely related to social or shopping behaviors, existing technologies have failed to effectively handle these complex dynamic needs, unable to achieve truly intelligent power management when the screen is off. The complexity and diversity of user behavior mean that traditional power optimization methods cannot meet the precise and dynamic power management demands of the intelligent era. Existing power management methods are too simplistic and lack flexibility, thus remaining significantly inadequate in meeting user experience requirements and extending battery life. Summary of the Invention

[0005] The purpose of this invention is to propose a control method and system for reducing power consumption when the mobile phone screen is off. By intelligently analyzing the user's social and shopping behavior, a personalized background task management and push control strategy is designed, which overcomes several defects in power consumption management of the prior art. In particular, in terms of dynamically adapting to user needs, it achieves more intelligent and efficient power consumption management in the screen-off state.

[0006] To achieve the above objectives, a method for controlling power consumption reduction after a mobile phone screen is turned off is provided in a first aspect of the present invention, the method comprising:

[0007] S1. Obtain users' daily behavior data, including social interaction data and shopping behavior data, and conduct preliminary analysis of the time series of behavior data fusion;

[0008] S2. Based on the preliminary analysis results, construct a social behavior model and a shopping behavior model, and train them respectively to predict user behavior trends and purchase intentions; wherein, the social behavior model and the shopping behavior model adopt a recurrent neural network structure for time series modeling;

[0009] The predictions made by the social behavior model are as follows:

[0010]

[0011] in, Let be the predicted social behavior value at time t, representing the predicted social activity level; The social behavior characteristics of the user at time t-1; This is a social behavior fluctuation adjustment item used to handle situations where users suddenly use or stop using social applications frequently. To represent the short-term trend of social behavior, used to capture rapid changes in user social behavior; α1 and β1 are weighting coefficients, representing the impact of the adjustment term on the overall prediction result; Wsocial The weight matrix is ​​obtained through model training; b social is the bias term, which controls the overall translation of the model; f is the activation function, and the nonlinear function ReLU or LeakyReLU is selected to increase the complexity of the prediction;

[0012] The shopping behavior model predicts the following:

[0013]

[0014] in, This is the predicted value of shopping behavior at time t; The characteristics of a user's shopping behavior at time t-1; This is a purchase intent adjustment item that reflects short-term changes in users' shopping interests; α² represents the short-term trend term of shopping behavior; α² and β² are the corresponding weighting parameters, controlling the contribution of the moderating term; W shop λ is the weight matrix of the model, serving as the core parameter for model learning; λ is the regularization parameter, used to limit the complexity of the model and prevent overfitting; g is the activation function, used to output non-linear changes, commonly LeakyReLU; ||W shop || 2 This is a regularization term used to prevent overfitting of the weight matrix;

[0015] S3. Dynamically set the priority of background tasks based on the prediction results; wherein the priority is designed as follows:

[0016]

[0017] in, The priority of the i-th background task determines the priority order of its resource allocation; Let be the predicted value of social behavior related to the i-th task; The predicted value of shopping behavior related to the i-th task; and α1 and α2 are the power consumption coefficients for the social and shopping activities of the i-th task, respectively, reflecting the impact of this task on battery consumption; α1 and α2 are the weight parameters for social and shopping behaviors, representing the influence of the weights of different types of tasks in priority calculation; U t This represents the system's current remaining battery percentage. When the battery is low, task priorities are more strictly controlled; E t γ represents the current load level of the system, used to adaptively adjust task priorities; γ is the system load and power consumption adjustment coefficient, used to adjust the trade-off between power consumption and load under different scenarios. This is a behavioral fluctuation item used to measure the volatility and suddenness of user behavior in order to adjust priorities;

[0018] Then, a resource allocation matrix is ​​introduced to allocate specific system resources to background tasks according to their priority, as shown below:

[0019]

[0020] in, The resource allocation for the i-th task; R total N represents the total available system resources; N represents the total number of currently running background tasks.

[0021] S4. Combine priority with dynamic resource allocation to ensure that high-priority tasks receive sufficient resource support, while controlling the resource consumption of low-priority tasks and optimizing power consumption.

[0022] S5. Design a low-power preloading mechanism to preload data in advance when the user turns off the screen to quickly access certain high-priority tasks and reduce the waiting time after the screen is turned on.

[0023] S6. Based on changes in user habits, optimize the prediction and resource allocation strategies of social behavior models and shopping behavior models to adapt to users' long-term needs.

[0024] Preferably, step S1 further includes preprocessing the behavioral data; the preliminary analysis of the fused time series of the behavioral data specifically includes:

[0025] The preprocessed data is used for feature mapping to quantify the activity level of user behavior, including:

[0026] Regarding social interaction data:

[0027]

[0028] in, w represents the user's social activity at time t. i Assigning weights to different dimensions of social data, u i The collected user social behaviors, where n is the total number of users interacting with social interactions, and i is the number of users interacting with social interactions.

[0029] Shopping behavior data:

[0030]

[0031] Among them, w j Assigning weights to various dimensions of shopping behavior, v j Here is the user's shopping data, where m is the total number of shopping users and j is the shopping user j;

[0032] Processed behavioral characteristics and The time series data is combined and represented and stored as follows: Where T is the total number of time segments collected. and Social and shopping behavior characteristics at each moment.

[0033] Preferably, the social behavior fluctuation adjustment item The definition is as follows:

[0034]

[0035] Where, μ social is the mean of social behavioral characteristics; N is the size of the backtracking time window, used to determine the historical length of the evaluation;

[0036] Purchase Intent Adjustment Item The definition is as follows:

[0037]

[0038] in, This refers to the user's recent shopping behavior data; i The weights represent the different behavior types, indicating the importance of different shopping behaviors; M is the size of the most recent time window.

[0039] Preferably, the behavioral fluctuation term The priority is fine-tuned based on the rate of change in recent user behavior, as follows:

[0040]

[0041] in, Let be the behavioral characteristic value of the i-th task at time t; M is the length of the backtracking time window, used to evaluate the historical cycle of behavioral fluctuations;

[0042] The power consumption coefficients for the social and shopping aspects of the i-th task and Among them, the power consumption coefficient of social tasks The power consumption of the shopping task reflects the power consumption of the i-th social task. This indicates the battery consumption for shopping-related tasks. The power consumption coefficient of the social task is dynamically adjusted based on its active and passive consumption behavior; wherein, the power consumption coefficient of the social task is dynamically adjusted based on its active and passive consumption behavior. The dynamic adjustment is expressed as:

[0043]

[0044] in This represents the active consumption coefficient of the i-th task, i.e., the power consumption when the task is active. This represents the passive consumption coefficient of the task, such as the power consumption during data synchronization; ω social A weighting parameter used to adjust the balance between active and passive power consumption.

[0045] Preferably, the objective function of S4 is:

[0046]

[0047] in, The amount of resources allocated to the i-th task; Let be the priority of the i-th task, reflecting its resource urgency in the current system; N is the total number of current background tasks; R total R represents the total amount of resources currently available to the system. battery U represents the total amount of resources adjusted based on battery level, indicating the amount of resources the system can allocate when the battery level is low. t E represents the current percentage of the system's remaining battery capacity. t The current system load status represents the overall busy level of the system; η is a weighting coefficient used to balance the total resource R. total and resource R based on battery status adjustment battery ;

[0048] Simultaneously, a power consumption limit for the task is designed, and the power consumption threshold for the task is dynamically set based on battery status and system load, expressed as:

[0049]

[0050] in, P represents the power consumption limit for the i-th task, indicating the maximum power consumption that the task can consume; max U represents the maximum power consumption limit for the task. t E represents the current remaining battery level of the system. t This represents the current system load status, indicating whether the system is operating under high load; λ power This is a power consumption adjustment factor that determines the degree of power consumption adjustment under different battery and load conditions. When the battery is in good condition, the task receives resources close to the maximum power consumption, while when the battery is at its lowest or the system load is at its highest, the power consumption limit will be reduced to the minimum to ensure that unnecessary tasks do not waste power.

[0051] Preferably, step S4 further includes designing a task freezing mechanism, which freezes resources for low-priority tasks and releases more resources for high-priority tasks when the system is in an extreme situation, as expressed as:

[0052]

[0053] in, θ represents the frozen state of the i-th task, where 1 indicates frozen and 0 indicates unfrozen; freeze This is the priority threshold for task freezing; tasks below this threshold will be frozen when battery power is low. min The minimum battery level threshold set for the system; when the battery level falls below this threshold, some low-priority tasks need to be frozen.

[0054] Preferably, step S5 further includes designing an adaptive adjustment factor, adjusting it based on the system's historical load and power status, and continuously optimizing the resource and power management strategy, expressed as:

[0055]

[0056] Where, λ adaptive λ is a dynamic adjustment factor used to adjust the sensitivity to resource and power management; initial δ is the initially set adjustment factor value; δ is the adjustment increment coefficient, which controls the adjustment range of the adjustment factor; U t E represents the battery charge at time t. t The system load status at time t; T represents the system historical data over the past T time points.

[0057] Preferably, S5 specifically includes:

[0058] A low-power preloading mechanism is designed so that when the device enters screen-off mode, although the user will not immediately use the device, background tasks still need to handle critical operations, as shown below:

[0059]

[0060]

[0061] in, Let be the preload coefficient for the i-th task, representing the amount of preload allowed for that task in low-power mode; The priority of the i-th task comes from the previous steps and is calculated based on social and shopping behavior prediction and power consumption model. The resource allocation for the i-th task reflects the proportion of the task in the system resources; The maximum preload power consumption threshold set for the system is used to limit the total power consumption of preload tasks and ensure that the system does not consume excessive power due to preloading. Let be the scheduling coefficient for the i-th task, representing the resource scheduling priority of the task in the screen-off state; This is the preloading coefficient for the i-th task, reflecting the preloading requirements of the task; λ represents the resource allocation ratio for the i-th task; N is the total number of active background tasks in the current system; θ is the balancing coefficient, controlling the scheduling ratio of preloaded tasks and actually running tasks; preload λ is the power consumption-preload adjustment factor, representing the power consumption sensitivity of the control system to the preload task in a low-power state; base U represents the initial preload power sensitivity; α is the adjustment increment, reflecting the system's power control over the preload task based on load and power conditions; t E represents the current system's remaining battery percentage. t This represents the current system load status, indicating the overall busyness of the system.

[0062] Preferably, S6 specifically includes:

[0063] An adaptive learning mechanism is designed to learn from the actual resource consumption and power consumption data of tasks through feedback loops, and to dynamically optimize task scheduling and power management strategies.

[0064] In another aspect of the invention, a control system for reducing power consumption when a mobile phone screen is off is provided, the system comprising:

[0065] The daily behavior data acquisition unit is used to acquire users' daily behavior data, including social interaction data and shopping behavior data, and to perform preliminary analysis on the fusion time series of the behavior data;

[0066] The model building unit is used to construct a social behavior model and a shopping behavior model based on the results of the preliminary analysis, and train them respectively to predict users' behavioral trends and purchase intentions; wherein, the social behavior model and the shopping behavior model adopt a recurrent neural network structure for time series modeling;

[0067] The predictions made by the social behavior model are as follows:

[0068]

[0069] in, Let be the predicted social behavior value at time t, representing the predicted social activity level; The social behavior characteristics of the user at time t-1; This is a social behavior fluctuation adjustment item used to handle situations where users suddenly use or stop using social applications frequently. To represent the short-term trend of social behavior, used to capture rapid changes in user social behavior; α1 and β1 are weighting coefficients, representing the impact of the adjustment term on the overall prediction result; W social The weight matrix is ​​obtained through model training; b socialis the bias term, which controls the overall translation of the model; f is the activation function, and the nonlinear function ReLU or LeakyReLU is selected to increase the complexity of the prediction;

[0070] The shopping behavior model predicts the following:

[0071]

[0072] in, This is the predicted value of shopping behavior at time t; The characteristics of a user's shopping behavior at time t-1; This is a purchase intent adjustment item that reflects short-term changes in users' shopping interests; α² represents the short-term trend term of shopping behavior; α² and β² are the corresponding weighting parameters, controlling the contribution of the moderating term; W shop λ is the weight matrix of the model, serving as the core parameter for model learning; λ is the regularization parameter, used to limit the complexity of the model and prevent overfitting; g is the activation function, used to output non-linear changes, commonly LeakyReLU; ||W shop || 2 This is a regularization term used to prevent overfitting of the weight matrix;

[0073] A priority design unit is used to dynamically set the priority of background tasks based on the prediction results; wherein, the priority design is as follows:

[0074]

[0075] in, The priority of the i-th background task determines the priority order of its resource allocation; Let be the predicted value of social behavior related to the i-th task; The predicted value of shopping behavior related to the i-th task; and α1 and α2 are the power consumption coefficients for the social and shopping activities of the i-th task, respectively, reflecting the impact of this task on battery consumption; α1 and α2 are the weight parameters for social and shopping behaviors, representing the influence of the weights of different types of tasks in priority calculation; U t This represents the system's current remaining battery percentage. When the battery is low, task priorities are more strictly controlled; E t γ represents the current load level of the system, used to adaptively adjust task priorities; γ is the system load and power consumption adjustment coefficient, used to adjust the trade-off between power consumption and load under different scenarios. This is a behavioral fluctuation item used to measure the volatility and suddenness of user behavior in order to adjust priorities;

[0076] Then, a resource allocation matrix is ​​introduced to allocate specific system resources to background tasks according to their priority, as shown below:

[0077]

[0078] in, The resource allocation for the i-th task; R total N represents the total available system resources; N represents the total number of currently running background tasks.

[0079] The resource allocation unit is used to dynamically allocate resources based on priority, ensuring that high-priority tasks receive sufficient resource support, while controlling the resource consumption of low-priority tasks and optimizing power consumption.

[0080] The low-power loading unit is used to design a low-power preloading mechanism. When the user turns off the screen, the data is preloaded in advance to quickly access certain high-priority tasks, reducing the waiting time after the screen is turned on.

[0081] The system optimization unit is used to optimize the prediction and resource allocation strategies of social behavior models and shopping behavior models in light of changes in user habits, so as to adapt to users' long-term needs.

[0082] The beneficial technical effects of the present invention are at least as follows:

[0083] To address the shortcomings of existing technologies in power management during screen-off mode, this invention proposes an intelligent screen-off power management system based on social and shopping behavior analysis. This system performs in-depth analysis of users' social interactions and e-shopping behavior, combines it with users' personalized usage patterns, dynamically adjusts the priority and resource allocation of background tasks, optimizes power management, and achieves a balance between user needs and battery life.

[0084] First, this invention introduces a social behavior analysis module, which intelligently determines a user's social needs by collecting real-time data on social activities before and after the screen is turned off. If a user is frequently interacting with specific contacts or has unread important messages before the screen is turned off, this module dynamically manages the background processes of social applications after the screen is turned off, ensuring that the user does not miss important message notifications. For social behaviors with low activity levels, the system automatically reduces the background priority of related applications, thereby avoiding unnecessary resource consumption and power waste. This module overcomes the shortcomings of existing technologies that cannot effectively distinguish between the intensity of user interaction and social needs, achieving personalized power consumption optimization for social behaviors.

[0085] Secondly, this invention also includes a shopping behavior analysis module. This module intelligently predicts a user's shopping needs when the screen is off by analyzing their browsing history, favorites, and purchase history. For users with potential shopping behavior, the system will continue to actively push shopping information, ensuring that important promotional information and product recommendations are delivered to the user in a timely manner. For users with unclear shopping needs, the system will delay or stop the background shopping application from running, reducing unnecessary network data requests and resource consumption, thereby effectively reducing battery consumption. This function solves the problem of power wastage caused by the inability to intelligently determine shopping needs and the blind push of information in existing technologies.

[0086] The third key innovation of this invention is a personalized power consumption optimization engine. This engine combines users' social and shopping behaviors with an adaptive learning mechanism. Through in-depth learning of users' long-term usage patterns, it dynamically adjusts the resource allocation and network request frequency of background applications. The system can continuously optimize power consumption management strategies based on users' historical usage data, gradually improving its responsiveness to users' personalized needs. This engine overcomes the shortcomings of rigid power consumption management strategies and lack of personalized adjustment in existing technologies, enabling the system to intelligently adapt to different user scenarios and further improve battery life.

[0087] In summary, this invention, through intelligent analysis of users' social and shopping behaviors, designs personalized background task management and push control strategies, overcoming several shortcomings of existing technologies in power consumption management. In particular, it achieves more intelligent and efficient power consumption management in the screen-off state by dynamically adapting to user needs. Attached Figure Description

[0088] The present invention will be further described with reference to the accompanying drawings, but the embodiments in the drawings do not constitute any limitation on the present invention. For those skilled in the art, other drawings can be obtained based on the following drawings without creative effort.

[0089] Figure 1 This is a flowchart of a method for controlling power consumption of a mobile phone after the screen is off, according to an embodiment of the present invention.

[0090] Figure 2 This is a framework diagram of a control system for reducing power consumption when a mobile phone screen is off, according to an embodiment of the present invention. Detailed Implementation

[0091] Embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0092] like Figure 1 As shown in the figure, an embodiment of the present invention provides a method for reducing power consumption when a mobile phone screen is off, the method comprising the following steps S1-S6:

[0093] S1. Obtain users' daily behavior data, including social interaction data and shopping behavior data, and conduct preliminary analysis of the time series of behavior data fusion;

[0094] Specifically, the user behavior dataset will contain two main categories of data: social data and shopping data, denoted as D. raw Among them, D raw This is the raw behavior dataset. The initial input data is obtained from sources such as the phone's sensors, application logs, and network interaction records. This data will be used for subsequent model training and behavior prediction.

[0095] D raw The structure is as follows:

[0096] This is a vector of social behavior data collected at time t, which includes multiple dimensions such as user actions in social applications, number of message interactions, unread message status, and usage duration.

[0097] This is a vector of shopping behavior data collected at time t, which includes the user's browsing history on the e-commerce platform, items in the shopping cart, favorites, and ad click data.

[0098] Specific data example: Let... Where u n This provides data for the user's nth social application, such as the number of message interactions and unread messages for each application.

[0099] Similarly, Where v m This indicates a user's actions on a particular e-commerce platform, such as the number of clicks or the number of product categories viewed.

[0100] Data collection and The initial input dataset D was formed. raw This data is used for subsequent analysis. The time dimension of this data is necessary because behavioral analysis relies on trends of change over time.

[0101] Furthermore, the original behavioral data D raw Data often contains noisy or invalid data, such as incomplete log records or blank sensor inputs, which can mislead subsequent models. Therefore, this invention requires a unique preprocessing step tailored to this scenario to remove noise and correct the data.

[0102] Further, preprocessing steps:

[0103] This invention ensures that social data is collected in each time segment t. Shopping data The timestamps are aligned, ensuring that all dimensions of data are successfully collected at the same time.

[0104] When certain time segments contain missing data (such as...) or (Some values ​​in the data are empty), and linear interpolation is used to fill in the missing data. Let the missing data be denoted as . Calculate using data from adjacent time points:

[0105]

[0106] in, For the complete data at time t, S t-1 With S t+1 The data are for the preceding and following times respectively.

[0107] Furthermore, raw data often has excessively high dimensionality and contains redundant information. Therefore, it is necessary to extract features that effectively represent user behavior and reduce the dimensionality of the data. The steps are as follows:

[0108] A frequency- and intensity-based behavioral feature mapping method is designed to quantify the activity level of user behavior. For example, the frequency of social application use and the intensity of message interaction can be obtained using the following formula:

[0109]

[0110] in, w represents the user's social activity at time t. i To assign weights to different dimensions of social data (e.g., message unread status is more important than browsing time), u i For the collected user social behavior.

[0111] Correspondingly, the user's shopping activity level is:

[0112]

[0113] Where w j The weighting of various dimensions of shopping behavior (such as the importance of page views and product clicks), v j For users' shopping data.

[0114] Furthermore, feature selection techniques are applied to transform high-dimensional data into low-dimensional behavioral feature representations. and This method can be used to transform the original high-dimensional dataset. and The data is compressed into a pair of time series, representing users' social activity and shopping behavior activity respectively, thus simplifying the subsequent analysis and modeling process.

[0115] Furthermore, the behavioral characteristics after treatment and It needs to be stored along the time dimension for subsequent model training and prediction. By using time serialization, this invention can better capture the changing trends of user behavior over time, thereby providing input for predictive models of social and shopping behaviors.

[0116] Furthermore, this invention represents the above two behavioral characteristics as time series:

[0117]

[0118] Where T is the total number of time segments collected. and Social and shopping behavior characteristics at each moment.

[0119] Furthermore, time series data and It is stored in a structured database for model training and behavior prediction. This time series not only preserves the dynamic changes in user behavior, but also simplifies the subsequent analysis process, making behavioral characteristics at each point in time traceable.

[0120] S2. Based on the results of the preliminary analysis, construct a social behavior model and a shopping behavior model, and train them respectively to predict users' behavioral trends and purchase intentions; wherein, the social behavior model and the shopping behavior model adopt a recurrent neural network structure for time series modeling.

[0121] Specifically, in predicting social and shopping behaviors, this invention designs a User Behavior Dynamic Prediction Model (UBDP), which combines two information sources: 1) historical user behavior data; and 2) short-term fluctuation trends in user behavior. To capture these dynamic behaviors, the model adds adaptively adjusted regularization terms and weight adjustment terms related to user behavior characteristics to the traditional recurrent neural network, adapting to the specific needs of this invention.

[0122] Model structure: M social and M shop Both methods employ recurrent neural network (RNN) structures for time series modeling to predict future behavioral demands. Furthermore, behavioral stability regularization and personalized adjustment terms are introduced into the loss function, dynamically optimizing it based on user behavior characteristics.

[0123] Furthermore, the social behavior model (M) of the present invention socialThis invention aims to predict users' future social activity levels. Based on the autoregressive mechanism of recurrent neural networks (RNNs), it can capture users' long-term behavioral dependencies. To adapt to the randomness and volatility of social behavior, this invention innovatively introduces a social behavior volatility adjustment term to smooth the frequency of users' abnormal behaviors. Formula definition:

[0124]

[0125] in, This represents the predicted social behavior value at time t, and represents the predicted social activity level. This represents the user's social behavior characteristics at time t-1; This refers to the social behavior fluctuation adjustment item, which is used to handle situations where users suddenly use or stop using social applications frequently; The term represents the short-term trend of social behavior, used to capture rapid changes in users' social behavior; α1 and β1 are weighting coefficients, representing the impact of the adjustment term on the overall prediction result; W social The weight matrix is ​​obtained through model training; b social represents the bias term, which controls the overall translation of the model; f represents the activation function, where the nonlinear function ReLU or LeakyReLU is chosen to increase the complexity of the prediction.

[0126] Furthermore, this is the innovative part of the model, designed to address situations where users' social behavior suddenly increases or decreases significantly over a period of time. This invention is defined based on the variance of user behavior:

[0127]

[0128] Where, μ social The mean of the social behavior characteristics is represented by ; N represents the size of the backtracking time window, used to determine the historical length of the evaluation. This adjustment term can prevent the model from being interfered with by excessively frequent or infrequent social behaviors by using historical fluctuation information, thus maintaining a reasonable prediction of the user's overall social behavior.

[0129] Furthermore, the core of shopping behavior prediction lies in capturing users' short-term purchase desires and long-term shopping trends. This invention introduces a user purchase intention adjustment term, which is corrected based on changes in users' short-term shopping behavior, and combines regularization to limit model complexity. The model adopts a recurrent neural network architecture and introduces an innovative weight adjustment mechanism for shopping behavior. Formula definition:

[0130]

[0131] in, This represents the predicted shopping behavior value at time t; This represents the characteristics of a user's shopping behavior at time t-1; This indicates a purchase intent adjustment item, reflecting short-term changes in users' shopping interests; This represents the short-term trend term of shopping behavior; α² and β² represent the corresponding weighting parameters, controlling the contribution of the moderating term; W shop The weight matrix represents the model, serving as the core parameters for model learning; λ represents the regularization parameter, used to limit the model's complexity and prevent overfitting; g represents the activation function, used to output non-linear results, commonly LeakyReLU; ||W shop || 2 This represents a regularization term used to prevent overfitting of the weight matrix.

[0132] Furthermore, this adjustment is defined based on the frequency of the user's recent shopping behavior, and the specific calculation formula is as follows:

[0133]

[0134] in, This is represented by the user's recent shopping behavior data (e.g., number of clicks, browsing time); c i The weights represent the different behavior types, indicating the importance of different shopping behaviors; M represents the size of the most recent time window. This adjustment term allows the model to capture users' short-term shopping intentions and identify sudden displays of purchase interest, thereby improving the accuracy of shopping behavior predictions.

[0135] Furthermore, the output result is: P social This represents the prediction results of users' future social behavior; P shop These two results represent predictions of the user's future shopping behavior. They are core inputs for setting background task priorities in subsequent steps. The model's predicted social and shopping behaviors can help the system dynamically adjust task priorities, thereby more intelligently managing power consumption after the screen is off.

[0136] Understandably, this invention introduces a social behavior fluctuation modifier. Smoothing out the frequency of users' abnormal behavior improves the robustness of social behavior prediction. A purchase intent adjustment item was designed. It can sensitively capture fluctuations in users' short-term shopping interests, enhancing the accuracy of shopping behavior prediction. A regularization term λ is used in the shopping behavior model to prevent overfitting, and the model's performance in shopping scenarios is optimized by considering the combination of behavioral trends and purchase intentions.

[0137] S3. Dynamically set the priority of background tasks based on the prediction results;

[0138] Specifically, the priority calculation model proposed in this invention aims to dynamically allocate task priorities based on user behavior predictions, while introducing some innovative terms specific to the scenario to enhance the model's adaptability to special situations. These include: power consumption coefficients based on task type, behavioral fluctuation terms, and adaptive adjustment terms for system state.

[0139] Priority calculation formula:

[0140]

[0141] in, This indicates the priority of the i-th background task, which determines the priority order of its resource allocation; This represents the predicted value of social behavior related to the i-th task, derived from M. social Model; This represents the predicted shopping behavior value related to the i-th task, derived from M. shop Model; and α1 and α2 represent the power consumption coefficients for the social and shopping activities of the i-th task, reflecting the impact of this task on battery consumption; α1 and α2 represent the weight parameters for social and shopping behaviors, indicating the influence of the weights of different task types on priority calculation; U t This indicates the current remaining battery percentage of the system. When the battery is low, task priorities are more strictly controlled; E t This indicates the current load level of the system (total number of tasks and system resource usage), used to adaptively adjust task priorities; γ represents the system load and power consumption adjustment coefficient, used to adjust the trade-off between power consumption and load under different scenarios; This represents the behavioral fluctuation item of innovative design, used to measure the volatility and suddenness of user behavior in order to adjust priorities.

[0142] Among them, behavioral fluctuation term This indicates that the priority is fine-tuned based on the rate of change in the user's recent behavior. For example, if a user's recent behavior changes frequently (such as suddenly becoming active on social media), then such tasks should receive higher priority to ensure that key interactions are not missed. Behavioral fluctuation term formula:

[0143]

[0144] in, This represents the behavioral characteristic value of the i-th task at time t (such as social activity or shopping interest); M represents the length of the backtracking time window, used to evaluate the historical cycle of behavioral fluctuations. This behavioral fluctuation term allows the system to make smarter priority adjustments based on the suddenness of user behavior. For example, a sudden surge in user social interaction should increase the priority of related tasks, while stable behavioral patterns will maintain the original priority.

[0145] Furthermore, when prioritizing tasks, it's crucial to consider not only user behavior predictions but also the task's impact on system power consumption. Each background task has a corresponding power consumption coefficient, reflecting its energy consumption. Specifically:

[0146] Social task power consumption coefficient This represents the power consumption of the i-th social task. For example, a task that continuously keeps messages synchronized consumes more power than a task that only occasionally checks for new messages. (Shopping task power consumption coefficient) This indicates the power consumption of shopping-related tasks, such as background operations like ad pushes and data synchronization.

[0147] Dynamic adjustment of power consumption factor:

[0148]

[0149] in, Let represent the active consumption coefficient of the i-th task, that is, the power consumption when the task is active. This represents the passive consumption coefficient of the task, such as the power consumption during data synchronization; ω social This represents the weighting parameter used to adjust the balance between active and passive power consumption, depending on the task type.

[0150] Similarly, for shopping tasks, It also dynamically adjusts based on the user's active and passive consumption behavior. This approach ensures that priority settings are not only based on user needs but also balanced according to the actual power consumption of the task, preventing high-power tasks from unnecessarily consuming resources.

[0151] Furthermore, the task priority is calculated. Subsequently, this invention introduces a resource allocation matrix R. allocation This is used to allocate specific system resources (such as CPU, memory, network bandwidth) to background tasks based on their priority. To dynamically adapt to the resource needs of different tasks, resource allocation depends not only on priority but also on balancing the current system load. Resource allocation formula:

[0152]

[0153] in, This represents the resource allocation for the i-th task; The priority of the i-th task is represented by the aforementioned formula; R total N represents the total amount of available system resources (such as allocable CPU, network bandwidth, etc.); N represents the total number of currently running background tasks.

[0154] Understandably, this resource allocation formula ensures that system resources are allocated proportionally according to task priority. Higher priority tasks receive a larger share of resources, while lower priority tasks have their resource usage restricted, or are even frozen when power or system load limits are reached.

[0155] Furthermore, the priority setting scheme of this invention allows the system to adaptively adjust based on fluctuations in user behavior, dynamic changes in power consumption coefficients, and the current state of the system (such as load and battery level). This mechanism ensures that background tasks can operate efficiently in the screen-off state without unnecessarily burdening battery life.

[0156] Furthermore, the task freeze mechanism: when Too low (below a certain set threshold θ) low When a task is frozen, the system will freeze it, stopping all its activities until the system state returns to an acceptable level. Freezing tasks reduces unnecessary power consumption and ensures the normal operation of core tasks under high load or low power conditions.

[0157] This priority setting scheme uses the behavioral prediction value P. social and P shop Power consumption coefficient adjustment and and behavioral fluctuation items Through innovative mechanisms, this solution achieves the goal of intelligent background task priority allocation. Compared to traditional power management methods, this solution can dynamically adapt to changes in user behavior and perform refined resource management based on task power consumption, system load, and battery status.

[0158] Priority calculation incorporates adjustments based on user behavior fluctuations and active and passive power consumption coefficients for tasks, ensuring that background tasks remain active based on actual user needs even when the screen is off, avoiding unnecessary power consumption, and greatly improving the system's responsiveness and battery life in complex scenarios.

[0159] S4. Combine priority with dynamic resource allocation to ensure that high-priority tasks receive sufficient resource support, while controlling the resource consumption of low-priority tasks and optimizing power consumption.

[0160] Specifically, traditional resource allocation is typically based on task priority and total system resources. However, this invention not only dynamically adjusts resource allocation according to priority but also considers current battery status and system load to perform fine-grained resource control, ensuring minimal power consumption. To this end, a new resource allocation and power management strategy is designed. Core formula: Resource Allocation Formula

[0161]

[0162] in, This represents the amount of resources (such as CPU time, memory, network bandwidth, etc.) allocated to the i-th task. R represents the priority of the i-th task, reflecting its resource urgency in the current system. N represents the total number of background tasks. total This represents the total amount of resources currently available to the system (such as CPU and bandwidth). R battery This represents the total amount of resources adjusted based on battery level, indicating the amount of resources the system can allocate when the battery level is low. U t This indicates the current percentage of the system's remaining battery capacity. E t This indicates the current load status of the system, representing the overall busy level of the system. η represents the weighting coefficient, used to balance the total resource R. total and resource R based on battery status adjustment battery .

[0163] This formula calculates resource allocation in two parts: one part is a weighted calculation based on task priorities, and the other part determines the resource allocation ratio based on the system battery status and current load. The weight η controls the balance between system resources and battery status. If the battery is low, the battery adjustment part of the formula will play a greater role, and the resources allocated to tasks will be more strictly controlled, reducing unnecessary resource waste.

[0164] Understandably, traditional resource allocation often relies on task priority, but in the field of power management, this invention must perform more precise dynamic control of power consumption. This formula introduces the remaining battery capacity U... t and load adjustment factor E t This effectively enables dynamic resource regulation based on the real-time status of the system.

[0165] Furthermore, after determining resource allocation, the system also needs to strictly control the power consumption of each task. This invention designs a task power consumption upper limit P. limit The power consumption threshold for tasks is dynamically set based on battery status and system load. Core formula: Power Management Formula

[0166]

[0167] in, P represents the power consumption limit for the i-th task, indicating the maximum power consumption that the task can consume. max This indicates the maximum power consumption limit for the task, which is usually a power consumption limit set by the system. t This indicates the current remaining battery level in the system. E t This indicates the current load status of the system and whether the system is operating under high load. This indicates the task's priority, which determines its importance in power allocation. λ power This represents the power consumption adjustment factor, which determines the degree of power consumption adjustment under different battery and load conditions.

[0168] Understandably, this formula sets a power consumption limit for tasks. When the battery is in good condition, tasks can access resources close to their maximum power consumption. However, when the battery is low or the system load is high, the power consumption limit will be significantly reduced, ensuring that unnecessary tasks do not waste power. This provides further control over system resources and effectively reduces the negative impact of background tasks on battery life.

[0169] Furthermore, an innovative task freezing mechanism is designed: when the system is in extreme conditions (such as extremely low power or excessive system load), this invention designs an automatic task freezing mechanism to freeze resources for low-priority tasks and release more resources for high-priority tasks.

[0170] Task Freeze Formula

[0171]

[0172] in, θ represents the frozen state of the i-th task, where 1 indicates frozen and 0 indicates unfrozen. freeze This indicates the priority threshold for task freezing; tasks below this threshold will be frozen when battery power is low. min This indicates the minimum battery level threshold set by the system. When the battery level falls below this threshold, some low-priority tasks need to be frozen.

[0173] Furthermore, when the system power is lower than the set threshold U min And the task priority is lower than θ freeze When this happens, the system will freeze the task and stop all its background activities. This mechanism not only effectively frees up resources but also ensures that battery power is used to maximize its utilization for more important tasks, extending device battery life.

[0174] Furthermore, a dynamic adjustment mechanism is designed: to better adapt to changes in users' dynamic needs, the system also needs to continuously optimize resource and power consumption management strategies. This invention introduces an adaptive adjustment factor λ. adaptive This factor is adjusted based on the system's historical load and power status to ensure the system can adapt to users' personalized behaviors and habits over long-term use. Adaptive adjustment formula λ adaptive :

[0175]

[0176] Where, λ adaptive This represents a dynamic adjustment factor used to adjust the sensitivity of resource and power management. λ initial This represents the initial set adjustment factor value. δ represents the adjustment increment coefficient, controlling the adjustment range of the adjustment factor. U t E represents the battery charge at time t. t This represents the system load status at time t. T represents the system historical data from the past T times.

[0177] Furthermore, the adaptive adjustment factor adjusts the system's sensitivity to resources and power consumption based on historical battery and load conditions. For example, when the system load is low and the battery is sufficient, the adjustment factor will loosen the power consumption restrictions and allocate more resources to tasks; while when the system load is too high or the battery is insufficient, the adjustment factor will tighten resource allocation to further reduce unnecessary power consumption.

[0178] Furthermore, this step effectively achieves the objectives of the present invention by combining resource scheduling, power management, and task freezing mechanisms: dynamically optimizing resource allocation in the screen-off state, reducing ineffective power consumption, and improving the system's long-term adaptability to user behavior through adaptive adjustment factors.

[0179] Understandably, unlike traditional power consumption optimization, this solution comprehensively regulates multiple factors such as battery status, system load, and task priority, enabling the system to maintain the normal operation of critical tasks even under resource constraints, thus avoiding the problem of low-priority tasks consuming limited resources. In particular, the introduction of a task freezing mechanism and dynamic power consumption limit setting can efficiently optimize battery life in complex usage scenarios.

[0180] S5. Design a low-power preloading mechanism to preload data in advance when the user turns off the screen to quickly access certain high-priority tasks and reduce the waiting time after the screen is turned on.

[0181] Specifically, when a device enters screen-off mode, the user may not immediately use the device, but background tasks (such as social media messages and shopping recommendations) still need to handle some critical operations. To ensure a good user experience (such as receiving important notifications promptly after the screen is off), this invention requires preloading these high-priority tasks with low power and controlling their behavior in low-power mode. Preloading formula

[0182]

[0183] in, This represents the preload coefficient for the i-th task, indicating the amount of preload allowed for that task in low-power mode. This indicates the priority of the i-th task, derived from previous steps and calculated based on social and shopping behavior predictions and a power consumption model. This represents the resource allocation for the i-th task, reflecting the proportion of the task in the system resources. This indicates the maximum power consumption of the task, preventing the task from consuming too much power during preloading. This indicates the maximum preload power consumption threshold set by the system, which is used to limit the total power consumption of preload tasks and ensure that the system does not consume excessive power due to preloading.

[0184] Among them, the task scheduling formula

[0185]

[0186] in, represents the scheduling coefficient of the i-th task, and represents the resource scheduling priority of the task in the screen-off state. This represents the preloading coefficient for the i-th task, reflecting the preloading requirements of the task. This represents the resource allocation ratio for the i-th task. N represents the total number of active background tasks in the current system. θ represents the balancing coefficient, which controls the scheduling ratio between preloaded tasks and actually running tasks.

[0187] Wherein, the adjustment formula λ preload :

[0188]

[0189] Where, λ preload λ represents the power consumption-preload adjustment factor, indicating the control system's sensitivity to the power consumption of preloaded tasks in a low-power state. base This represents the initial preload power sensitivity setting. α represents the adjustment increment, reflecting the system's power control over the preload task based on load and power consumption. U t This indicates the current remaining battery percentage of the system. Et It indicates the current system load status and the overall busy level of the system.

[0190] S6. Based on changes in user habits, optimize the prediction and resource allocation strategies of social behavior models and shopping behavior models to adapt to users' long-term needs.

[0191] Specifically, user behavior patterns, task power consumption requirements, and resource loads are constantly changing, therefore task scheduling and preloading strategies cannot remain static. This step introduces an adaptive learning mechanism that learns from the actual resource consumption and power consumption data of tasks through a feedback loop, dynamically optimizing task scheduling and power management strategies.

[0192] Furthermore, this invention introduces a learning rate η learn This is used to control how the system adjusts its scheduling and resource allocation strategies for future tasks based on historical behavioral data. The learning rate formula is η. learn :

[0193]

[0194] in, η represents the adaptive learning rate for the i-th task, controlling the step size for learning optimization. base This represents the base learning rate, which sets the initial sensitivity for learning. This represents the amount of resources actually consumed by the i-th task. This represents the amount of resources allocated to the i-th task. This represents the actual power consumption of the i-th task. This represents the upper limit of power consumption for the i-th task. α represents the resource difference weight, controlling the impact of resource differences on the learning rate. β represents the power consumption difference weight, controlling the impact of power consumption differences on the learning rate.

[0195] Among them, adaptive learning rate The core principle lies in dynamically adjusting the learning rate based on the difference between the actual resources and power consumption of each task and the allocated values. When the actual resource consumption and power consumption of a task differ significantly from the preset values, the learning rate increases, allowing the system to adjust its allocation strategy more quickly. Conversely, if the difference is small, the learning rate gradually decreases, reducing the frequency of task scheduling adjustments by the system.

[0196] Furthermore, after collecting resource and power consumption data from historical tasks, the system needs to adjust future task scheduling and power management strategies through an adaptive learning mechanism to better adapt to users' dynamic needs. Optimization Formula

[0197]

[0198] in, Let represent the optimized scheduling coefficient of the i-th task, and let represent the resource scheduling weight of the task after learning. The initial scheduling coefficient for the i-th task is calculated in step 5. This represents the learning rate of the i-th task, controlling the optimization magnitude of task scheduling.

[0199] Understandable, the optimized scheduling coefficient Based on learning rate Dynamic adjustments are made. By increasing the learning rate, the system can quickly respond to the actual needs of tasks, improving the accuracy of resource allocation; when the actual performance of a task is close to the preset value, the adjustment of the scheduling coefficient will become more stable. This approach ensures that the system can perform optimized scheduling in low-power mode and gradually adapt to the user's personalized usage habits over time.

[0200] Furthermore, adaptive learning systems need to continuously acquire actual task execution data and adjust task scheduling and preloading strategies through feedback loops. The specific feedback process includes:

[0201] Actual resource consumption feedback: Data on actual resource consumption collected during task execution. Used to update resource allocation strategies.

[0202] Actual power consumption feedback: Collect actual power consumption data for the task. Used to adjust power consumption limits.

[0203] Task execution result evaluation: Optimize the preloading strategy for future tasks by evaluating the actual execution results of tasks (such as response speed, loading latency, etc.).

[0204] Feedback Formula: Resource and Power Consumption Optimization Updates:

[0205]

[0206] in, This represents the updated resource allocation value for the i-th task. This represents the upper limit of power consumption after the update for the i-th task. This represents the learning rate of the i-th task, which determines the magnitude of resource and power consumption adjustments.

[0207] Understandably, through the adaptive learning process, the resource allocation values ​​and power consumption limits for each task are continuously updated, allowing the system to gradually optimize the resource usage and power consumption performance of each task. The feedback mechanism ensures that each task receives more appropriate resource allocation and power consumption limits in subsequent runs, guaranteeing resource optimization effects under low power consumption conditions.

[0208] The adaptive learning system dynamically adjusts the scheduling and power consumption strategies for future tasks by combining historical task data with preset resource allocations and power consumption limits. Unlike traditional fixed strategies, this system continuously optimizes based on user behavior patterns, system load, and task performance, improving the efficiency of background tasks and battery management capabilities.

[0209] The adaptive learning and optimization mechanism in this step plays a crucial role in this invention: by continuously learning user behavior and task execution data, the system can more effectively schedule resources and manage power consumption in low-power mode. Especially in screen-off mode, the system can intelligently adjust task resource usage and preloading strategies to ensure a good user experience while saving power.

[0210] like Figure 2 As shown, in another embodiment of the present invention, a control system for reducing power consumption when a mobile phone screen is off is provided, the system comprising:

[0211] The daily behavior data acquisition unit 3011 is used to acquire users' daily behavior data, including social interaction data and shopping behavior data, and to perform preliminary analysis on the fusion time series of the behavior data;

[0212] The model building unit 3012 is used to build a social behavior model and a shopping behavior model based on the results of the preliminary analysis, and train them respectively to predict users' behavioral trends and purchase intentions; wherein, the social behavior model and the shopping behavior model adopt a recurrent neural network structure for time series modeling.

[0213] The predictions made by the social behavior model are as follows:

[0214]

[0215] in, Let be the predicted social behavior value at time t, representing the predicted social activity level; The social behavior characteristics of the user at time t-1; This is a social behavior fluctuation adjustment item used to handle situations where users suddenly use or stop using social applications frequently. To represent the short-term trend of social behavior, used to capture rapid changes in user social behavior; α1 and β1 are weighting coefficients, representing the impact of the adjustment term on the overall prediction result; W social The weight matrix is ​​obtained through model training; b social is the bias term, which controls the overall translation of the model; f is the activation function, and the nonlinear function ReLU or LeakyReLU is selected to increase the complexity of the prediction;

[0216] The shopping behavior model predicts the following:

[0217]

[0218] in, This is the predicted value of shopping behavior at time t; The characteristics of a user's shopping behavior at time t-1; This is a purchase intent adjustment item that reflects short-term changes in users' shopping interests; α² represents the short-term trend term of shopping behavior; α² and β² are the corresponding weighting parameters, controlling the contribution of the moderating term; W shop λ is the weight matrix of the model, serving as the core parameter for model learning; λ is the regularization parameter, used to limit the complexity of the model and prevent overfitting; g is the activation function, used to output non-linear changes, commonly LeakyReLU; ||W shop || 2 This is a regularization term used to prevent overfitting of the weight matrix;

[0219] The priority design unit 3013 is used to dynamically set the priority of background tasks based on the prediction results; wherein, the priority design is as follows:

[0220]

[0221] in, The priority of the i-th background task determines the priority order of its resource allocation; Let be the predicted value of social behavior related to the i-th task; The predicted value of shopping behavior related to the i-th task; and α1 and α2 are the power consumption coefficients for the social and shopping activities of the i-th task, respectively, reflecting the impact of this task on battery consumption; α1 and α2 are the weight parameters for social and shopping behaviors, representing the influence of the weights of different types of tasks in priority calculation; U t This represents the system's current remaining battery percentage. When the battery is low, task priorities are more strictly controlled; E t γ represents the current load level of the system, used to adaptively adjust task priorities; γ is the system load and power consumption adjustment coefficient, used to adjust the trade-off between power consumption and load under different scenarios. This is a behavioral fluctuation item used to measure the volatility and suddenness of user behavior in order to adjust priorities;

[0222] Then, a resource allocation matrix is ​​introduced to allocate specific system resources to background tasks according to their priority, as shown below:

[0223]

[0224] in, The resource allocation for the i-th task; Rtotal N represents the total available system resources; N represents the total number of currently running background tasks.

[0225] The resource allocation unit 3014 is used to dynamically allocate resources based on priority, ensuring that high-priority tasks receive sufficient resource support, while controlling the resource consumption of low-priority tasks and optimizing power consumption.

[0226] The low-power loading unit 3015 is used to design a low-power preloading mechanism. When the user turns off the screen, the data is preloaded in advance to quickly access certain high-priority tasks, reducing the waiting time after the screen is turned on.

[0227] The system optimization unit 3016 is used to optimize the prediction combination and resource allocation strategy of the social behavior model and shopping behavior model in light of changes in user habits, so as to adapt to the long-term needs of users.

[0228] It should be noted that the workflow described above is merely illustrative and does not limit the scope of protection of this invention. In practical applications, those skilled in the art can select some or all of the workflow to achieve the purpose of this embodiment according to actual needs, and no restrictions are imposed here.

[0229] In addition, for technical details not described in detail in this embodiment, please refer to the parameter operation method provided in any embodiment of the present invention, which will not be repeated here.

[0230] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or system. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or system that includes that element.

[0231] The sequence numbers of the above embodiments of the present invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0232] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as read-only memory / random access memory, magnetic disk, optical disk) and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0233] The above are merely preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Any equivalent structural or procedural transformations made based on the description and drawings of the present invention, or direct or indirect applications in other related technical fields, are similarly included within the scope of protection of the present invention.

Claims

1. A method for controlling power consumption to reduce power consumption when a mobile phone screen is off, characterized in that, The method includes: S1. Obtain users' daily behavior data, including social interaction data and shopping behavior data, and conduct preliminary analysis of the time series fusion of daily behavior data; S2. Based on the preliminary analysis results, construct a social behavior model and a shopping behavior model, and train them respectively to predict user behavior trends and purchase intentions; wherein, the social behavior model and the shopping behavior model adopt a recurrent neural network structure for time series modeling; The predictions made by the social behavior model are as follows: in, Let be the predicted social behavior value at time t, representing the predicted social activity level; The social behavior characteristics of the user at time t-1; This is a social behavior fluctuation adjustment item used to handle situations where users suddenly use or stop using social applications frequently. To represent the short-term trend of social behavior, used to capture rapid changes in user social behavior; α1 and β1 are weighting coefficients, representing the impact of the adjustment term on the overall prediction result; W social The weight matrix is ​​obtained through model training; b social is the bias term, which controls the overall translation of the model; f is the activation function, and the nonlinear function ReLU or LeakyReLU is selected to increase the complexity of the prediction; The shopping behavior model predicts the following: in, This is the predicted value of shopping behavior at time t; The characteristics of a user's shopping behavior at time t-1; This is a purchase intent adjustment item that reflects short-term changes in users' shopping interests; The short-term trend term of shopping behavior is represented by α2 and β2, which are the corresponding weighting parameters; W shop λ is the model's weight matrix; λ is the regularization parameter; g is the activation function used to output nonlinear transformation results; ||W shop || 2 It is a regularization term; S3. Dynamically set the priority of background tasks based on the prediction results; wherein the priority is designed as follows: in, The priority of the i-th background task determines the priority order of its resource allocation; Let be the predicted value of social behavior related to the i-th task; The predicted value of shopping behavior related to the i-th task; and These are the power consumption coefficients for social interaction and shopping in the i-th task, respectively; α1 and α2 are the weight parameters for social interaction and shopping behavior; U t E represents the current remaining battery percentage of the system. t γ represents the current load level of the system; γ is the adjustment coefficient for system load and power consumption. For behavioral fluctuations; Then, a resource allocation matrix is ​​introduced to allocate specific system resources to background tasks according to their priority, as shown below: in, The resource allocation for the i-th task; R total N represents the total amount of available system resources; N represents the total number of currently running background tasks. S4. Combine priority with dynamic resource allocation to ensure that high-priority tasks receive sufficient resource support, while controlling the resource consumption of low-priority tasks and optimizing power consumption. S5. Design a low-power preloading mechanism to preload data and quickly access high-priority tasks after the user turns off the screen, reducing the waiting time after the screen is turned on. S6. Based on changes in user habits, optimize the prediction results and resource allocation strategies of social behavior models and shopping behavior models; The objective function of S4 is: in, The amount of resources allocated to the i-th task; Let be the priority of the i-th task, reflecting its resource urgency in the current system; N is the total number of currently running background tasks; R total R represents the total amount of resources currently available to the system. battery R represents the total resource amount adjusted based on battery power; η is a weighting coefficient used to balance the total resource amount R. total and resource R based on battery status adjustment battery ; Simultaneously, a power consumption limit for the task is designed, and the power consumption threshold for the task is dynamically set based on battery status and system load, expressed as: in, P represents the power consumption limit for the i-th task, indicating the maximum power consumption that the task can consume; max λ represents the maximum power consumption limit for the task. power This is the power consumption adjustment factor, which determines the adjustment range of power consumption under different battery and load conditions; S5 specifically includes: A low-power preloading mechanism is designed so that when the device enters screen-off mode, background tasks need to handle critical operations, represented as follows: in, Let be the preloading coefficient for the i-th task; The priority of the i-th task is calculated based on social and shopping behavior prediction and power consumption model; The amount of resources allocated for the i-th task; The maximum preload power consumption threshold set for the system; λ is the scheduling coefficient for the i-th task, representing the resource scheduling priority of the task in the screen-off state; θ is the balancing coefficient, controlling the scheduling ratio of preloaded tasks and actually running tasks; preload λ is the power consumption-preload adjustment factor, representing the power consumption sensitivity of the control system to the preload task in a low-power state; base α represents the initial preload power sensitivity; α is the adjustment increment, reflecting the system's power control over the preload task based on load and power conditions.

2. The method for controlling power consumption after a mobile phone screen is off according to claim 1, characterized in that, S1 further includes preprocessing the behavioral data; the preliminary analysis of the fused time series of the behavioral data specifically includes: The preprocessed data is used for feature mapping to quantify the activity level of user behavior, including: Regarding social interaction data: Among them, w i Assigning weights to different dimensions of social data, u i The collected user social behaviors, where n is the nth social application; Regarding shopping behavior data: Among them, w j Assigning weights to various dimensions of shopping behavior, v j Here, m represents the user's shopping data, and m represents the user's actions on the e-commerce platform. Processed behavioral characteristics and The time series data is combined and represented and stored as follows: Where T is the total number of time segments collected. and These are the social and shopping behavior feature values ​​of the user at time t, respectively.

3. The method for controlling power consumption reduction after a mobile phone screen is off according to claim 1, characterized in that, The social behavior fluctuation regulation item The definition is as follows: Where, μ social is the mean of social behavioral characteristics; N' is the size of the backtracking time window, used to determine the historical length of the evaluation; Purchase Intent Adjustment Item The definition is as follows: in, For user shopping behavior data; c i The weights represent the different behavior types, indicating the importance of different shopping behaviors; M' is the size of the most recent time window.

4. The method for controlling power consumption reduction after a mobile phone screen is off according to claim 1, characterized in that, The behavioral fluctuation item The priority is fine-tuned based on the rate of change in recent user behavior, as follows: in, Let be the behavioral characteristic value of the i-th task at time t; M is the length of the backtracking time window, used to evaluate the historical cycle of behavioral fluctuations; The power consumption coefficients for the social and shopping aspects of the i-th task and Among them, the power consumption coefficient of social tasks The power consumption of the shopping task reflects the power consumption of the i-th social task. This indicates the battery consumption for shopping-related tasks. The power consumption coefficient of the social task is dynamically adjusted based on its active and passive consumption behavior; wherein, the power consumption coefficient of the social task is dynamically adjusted based on its active and passive consumption behavior. The dynamic adjustment is expressed as: in This represents the active consumption coefficient of the i-th task, i.e., the power consumption when the task is active. This represents the passive consumption coefficient of the task, including power consumption during data synchronization; ω social A weighting parameter used to adjust the balance between active and passive power consumption.

5. The method for controlling power consumption after a mobile phone screen is off according to claim 1, characterized in that, S4 further includes designing a task freezing mechanism, which freezes resources for low-priority tasks and releases more resources for high-priority tasks when the system is in an extreme situation, as expressed as: in, θ represents the frozen state of the i-th task, where 1 indicates frozen and 0 indicates unfrozen; freeze The priority threshold for task freezing; U min The minimum power threshold set for the system.

6. The method for controlling power consumption reduction after a mobile phone screen is off according to claim 1, characterized in that, S5 further includes designing an adaptive adjustment factor that adjusts based on the system's historical load and power status to continuously optimize resource and power management strategies, expressed as: Where, λ adaptive λ is a dynamic adjustment factor used to adjust the sensitivity to resource and power management; initial δ is the initially set adjustment factor value; δ is the adjustment increment coefficient, which controls the adjustment range of the adjustment factor; U t E represents the battery charge at time t. t The system load status at time t; T is the total number of time segments collected.

7. The method for controlling power consumption reduction after a mobile phone screen is off according to claim 1, characterized in that, S6 specifically includes: An adaptive learning mechanism is designed to learn from the actual resource consumption and power consumption data of tasks through feedback loops, and to dynamically optimize task scheduling and power management strategies.

8. A system for implementing a control method for reducing power consumption after a mobile phone screen is turned off as described in claim 1, characterized in that, The system includes: The daily behavior data acquisition unit is used to acquire users' daily behavior data, including social interaction data and shopping behavior data, and to perform preliminary analysis on the fused time series of daily behavior data; The model building unit is used to construct a social behavior model and a shopping behavior model based on the results of the preliminary analysis, and train them respectively to predict users' behavioral trends and purchase intentions; wherein, the social behavior model and the shopping behavior model adopt a recurrent neural network structure for time series modeling; The predictions made by the social behavior model are as follows: in, Let be the predicted social behavior value at time t, representing the predicted social activity level; The social behavior characteristics of the user at time t-1; This is a social behavior fluctuation adjustment item used to handle situations where users suddenly use or stop using social applications frequently. To represent the short-term trend of social behavior, used to capture rapid changes in user social behavior; α1 and β1 are weighting coefficients, representing the impact of the adjustment term on the overall prediction result; W social The weight matrix is ​​obtained through model training; b social is the bias term, which controls the overall translation of the model; f is the activation function, and the nonlinear function ReLU or LeakyReLU is selected to increase the complexity of the prediction; The shopping behavior model predicts the following: in, This is the predicted value of shopping behavior at time t; The characteristics of a user's shopping behavior at time t-1; This is a purchase intent adjustment item that reflects short-term changes in users' shopping interests; The short-term trend term of shopping behavior is represented by α2 and β2, which are the corresponding weighting parameters; W shop λ is the model's weight matrix; λ is the regularization parameter; g is the activation function used to output nonlinear transformation results; ||W shop || 2 It is a regularization term; A priority design unit is used to dynamically set the priority of background tasks based on the prediction results; wherein, the priority design is as follows: in, The priority of the i-th background task determines the priority order of its resource allocation; Let be the predicted value of social behavior related to the i-th task; The predicted value of shopping behavior related to the i-th task; and These are the power consumption coefficients for social interaction and shopping in the i-th task, respectively; α1 and α2 are the weight parameters for social interaction and shopping behavior; U t E represents the current remaining battery percentage of the system. t γ represents the current load level of the system; γ is the adjustment coefficient for system load and power consumption. For behavioral fluctuations; Then, a resource allocation matrix is ​​introduced to allocate specific system resources to background tasks according to their priority, as shown below: in, The resource allocation for the i-th task; R total N represents the total amount of available system resources; N represents the total number of currently running background tasks. The resource allocation unit is used to dynamically allocate resources based on priority, ensuring that high-priority tasks receive sufficient resource support, while controlling the resource consumption of low-priority tasks and optimizing power consumption. The low-power loading unit is used to design a low-power preloading mechanism. When the user turns off the screen, the data is preloaded in advance to quickly access high-priority tasks and reduce the waiting time after the screen is turned on. The system optimization unit is used to optimize the prediction results and resource allocation strategies of the social behavior model and shopping behavior model based on changes in user habits.

Images