A device power consumption optimization method and device, electronic device, and storage medium

Abstract

This application provides a device power consumption optimization method and apparatus, electronic device, and storage medium, belonging to the field of power management technology. The method includes: constructing a real-time component coupling strength matrix based on the device's real-time hardware electrical parameters, real-time software operating parameters, and real-time data interaction volume; obtaining a dynamic coupling weight matrix based on the real-time component coupling strength matrix and historical component coupling strength matrices; determining component coupling constraints based on the dynamic coupling weight matrix and determining power consumption parameter boundary conditions based on device scenario parameters; generating an initial particle swarm based on the initial power consumption parameters of all components and the hard constraint range of each component's parameters; updating the initial particle swarm based on the component coupling constraints and power consumption parameter boundary conditions to obtain a first particle swarm; updating the first particle swarm, and using the particle with the lowest fitness value during the update process as the device's component power consumption optimization parameter. This application can achieve stable and accurate power consumption optimization adaptable to multiple scenarios.

Description

Technical Field

[0001] This application belongs to the field of power management technology, and more specifically, relates to a device power consumption optimization method and apparatus, electronic equipment, and storage medium. Background Technology

[0002] As electronic devices become more multifunctional and integrated, power consumption optimization has become a core requirement for improving battery life, reducing operating costs, and ensuring stability. Devices typically consist of multiple cooperating components. Current power consumption optimization methods often focus on adjusting the parameters of individual components, which can easily lead to parameter mismatches, resulting in increased overall power consumption or decreased performance. Alternatively, they may calculate total power consumption based on a fixed model and employ a uniform optimization strategy; however, the accuracy of power consumption calculations is low under complex external scenarios, making it difficult to achieve optimal global power consumption in complex environments and failing to meet the high-efficiency energy management needs of devices operating in dynamic environments.

[0003] Therefore, there is an urgent need for a precise power consumption optimization method that can adapt to multiple scenarios and take stability into account. Summary of the Invention

[0004] The purpose of this application is to provide a device, electronic device, and storage medium for optimizing device power consumption, so as to achieve accurate power consumption optimization that is adaptable to multiple scenarios and takes into account stability.

[0005] A first aspect of this application provides a device power consumption optimization method, including:

[0006] A real-time component coupling strength matrix is ​​constructed based on the device's real-time hardware electrical parameters, real-time software operating parameters, and real-time data interaction volume; the component coupling strength matrix includes real-time component coupling vectors between multiple components of the device;

[0007] The historical component coupling strength matrix of the device is obtained, and a dynamic coupling weight matrix is ​​obtained by weighted summation based on the real-time component coupling strength matrix and the historical component coupling strength matrix. The historical component coupling strength matrix is ​​determined based on the historical hardware electrical parameters, historical software operating parameters and historical data interaction volume of the device. The historical component coupling strength matrix includes the historical average component coupling vector among multiple components in the device. The dynamic coupling weight matrix includes the target component coupling vector among multiple components in the device. The target component coupling vector is used to characterize the interaction influence strength of the multiple components in the current operating scenario of the device.

[0008] The component coupling constraints are determined based on the dynamic coupling weight matrix, and the power consumption parameter boundary conditions are determined based on the device's scenario parameters; the device's scenario parameters include the device's ambient temperature, task type, and operating mode.

[0009] An initial particle swarm is randomly generated based on the initial power consumption parameters of all components and the hard constraint range of the parameters of each component. The initial particle swarm includes multiple initial particles, and each initial particle includes the power consumption parameters corresponding to each component.

[0010] The initial particle swarm is updated based on the component coupling constraint and power consumption parameter boundary conditions to obtain the first particle swarm that satisfies the component coupling constraint and power consumption parameter boundary conditions.

[0011] The first particle swarm is iteratively updated until the iterative convergence condition is met. The power consumption parameters of all components corresponding to the particle with the lowest fitness value during the iterative update process are used as the component power consumption optimization parameters of the device; the larger the fitness value, the higher the power consumption.

[0012] A second aspect of this application provides a device power consumption optimization apparatus, comprising:

[0013] The coupling vector acquisition module is used to construct a real-time component coupling strength matrix based on the device's real-time hardware electrical parameters, real-time software operating parameters, and real-time data interaction volume; the component coupling strength matrix includes real-time component coupling vectors between multiple components in the device.

[0014] The dynamic coupling analysis module is used to obtain the historical component coupling strength matrix of the device. The dynamic coupling weight matrix is ​​obtained by weighted summation of the real-time component coupling strength matrix and the historical component coupling strength matrix. The historical component coupling strength matrix is ​​determined based on the historical hardware electrical parameters, historical software operating parameters and historical data interaction of the device. The historical component coupling strength matrix includes the historical average component coupling vector among multiple components in the device. The dynamic coupling weight matrix includes the target component coupling vector among multiple components in the device. The target component coupling vector is used to characterize the interaction influence strength of multiple components in the current operating scenario of the device.

[0015] The constraint analysis module is used to determine component coupling constraints based on the dynamic coupling weight matrix and to determine power consumption parameter boundary conditions based on the device's scenario parameters, including the device's ambient temperature, task type, and operating mode.

[0016] The initialization module is used to randomly generate an initial particle swarm based on the initial power consumption parameters of all components and the hard constraint range of the parameters of each component. The initial particle swarm includes multiple initial particles, and each initial particle includes the power consumption parameters corresponding to each component.

[0017] The condition constraint module is used to update the initial particle swarm based on the component coupling constraint condition and the power consumption parameter boundary condition to obtain the first particle swarm that satisfies the component coupling constraint condition and the power consumption parameter boundary condition.

[0018] The power consumption optimization module is used to iteratively update the first particle swarm until the iterative convergence condition is met. The power consumption parameters of all components corresponding to the particle with the lowest fitness value during the iterative update process are used as the component power consumption optimization parameters of the device; the higher the fitness value, the higher the power consumption.

[0019] A third aspect of this application provides an electronic device, including a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor executes the computer program to implement the steps of the device power consumption optimization method described above.

[0020] A fourth aspect of this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the above-described device power consumption optimization method.

[0021] The beneficial effects of the device power consumption optimization method and apparatus, electronic device, and storage medium provided in this application embodiment are as follows:

[0022] This application's embodiments accurately capture the coupling and interaction effects between components through a dynamic coupling weight matrix. This matrix is ​​generated based on real-time and historical data of the device, including hardware electrical parameters and software operating parameters, and reflects the collaborative relationship between different components in the current scenario. The component coupling constraints determined in this way can prevent the optimization of parameters of a single component from becoming disconnected from other components, ensuring that the parameters of each component are mutually adapted, thereby reducing power consumption as a whole and avoiding performance fluctuations.

[0023] This application's embodiments introduce scenario parameters such as ambient temperature and task type, and set boundary conditions for power consumption parameters accordingly, ensuring that the optimization process closely matches the actual operating scenario of the device. Then, a particle swarm optimization algorithm is used to iteratively optimize within the constraints, ultimately obtaining parameters that accurately adapt to the current scenario requirements, solving the problem of poor adaptability of a unified optimization strategy.

[0024] In summary, the embodiments of this application can enable the device to operate stably with low power consumption in different scenarios, significantly improving battery life and energy management efficiency. Attached Figure Description

[0025] To more clearly illustrate the technical solutions in the embodiments of this application, 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 this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0026] Figure 1 A flowchart illustrating a device power consumption optimization method provided in an embodiment of this application;

[0027] Figure 2 A structural block diagram of a device power consumption optimization apparatus provided in an embodiment of this application;

[0028] Figure 3 This is a schematic block diagram of an electronic device provided in an embodiment of this application. Detailed Implementation

[0029] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0030] To make the objectives, technical solutions, and advantages of this application clearer, the following description will be provided in conjunction with the accompanying drawings and specific embodiments.

[0031] Please refer to Figure 1 , Figure 1 This is a flowchart illustrating a device power consumption optimization method according to an embodiment of this application. The method can be executed by an electronic device, and specifically, the method may include S101 to S106.

[0032] S101: Construct a real-time component coupling strength matrix based on the device's real-time hardware electrical parameters, real-time software operating parameters, and real-time data interaction volume; the component coupling strength matrix includes real-time component coupling vectors between multiple components in the device.

[0033] In this embodiment, "device" refers to a device that can interact with a user, such as smart glasses or a smartphone. The real-time component coupling strength matrix is ​​a matrix constructed based on three types of real-time parameters of the device, containing real-time component coupling vectors between multiple components, used to characterize the interaction characteristics between components in the current operating state. The real-time component coupling vector refers to the vector in the real-time component coupling strength matrix that quantifies the real-time interaction impact between components, integrating real-time multi-dimensional correlation features. Real-time hardware electrical parameters refer to the hardware electrical data currently running on the device, reflecting the current electrical state of the components. Real-time software operating parameters refer to the current software operating data of the device, reflecting the current software load of the components. Real-time data interaction volume refers to the current data transmission volume between components, reflecting the immediate strength of the current data interaction.

[0034] For example, smart glasses can collect three types of real-time parameters at 500ms intervals. These include real-time hardware electrical parameters such as processor operating voltage and display module backlight current; real-time software operating parameters such as AR navigation application process utilization and sensor data processing task cycle; and real-time data interaction volume such as the amount of image data transmitted from the camera to the processor and the amount of attitude data transmitted from the sensors to the processor. This embodiment can calculate the real-time multi-dimensional correlation between components based on these parameters, integrate them to obtain a real-time component coupling vector, and construct a real-time component coupling strength matrix with the core components as rows and columns.

[0035] S102: Obtain the historical component coupling strength matrix of the device, and obtain the dynamic coupling weight matrix by weighted summation based on the real-time component coupling strength matrix and the historical component coupling strength matrix; the historical component coupling strength matrix is ​​determined based on the historical hardware electrical parameters, historical software operating parameters and historical data interaction volume of the device, and includes the historical average component coupling vector among multiple components in the device; the dynamic coupling weight matrix includes the target component coupling vector among multiple components in the device, and the target component coupling vector is used to characterize the interaction influence strength of multiple components in the current operating scenario of the device.

[0036] In this embodiment, the historical component coupling strength matrix refers to a matrix determined based on three types of historical parameters of the device, containing historical average component coupling vectors among multiple components. The historical average component coupling vector is a vector obtained by statistically averaging the historical component coupling vectors, reflecting the long-term stable characteristics of the interaction between components. The target component coupling vector is a vector obtained by weighted summation of the target component coupling strength matrix with the historical component coupling strength matrix in real time, used to accurately characterize the interaction influence strength between components in the current scenario.

[0037] In this embodiment, the interaction strength between components is dynamically influenced by the current operating state, yet also exhibits stable patterns based on historical data. Relying solely on real-time data is susceptible to instantaneous fluctuations, leading to deviations in coupling representation; relying solely on historical data cannot adapt to the dynamic changes of the current scenario. By constructing two types of component coupling strength matrices—real-time and historical—and merging their advantages through weighted summation, a balance between dynamism and stability can be achieved. The dynamic coupling weight matrix composed of the target component coupling vectors can accurately match the current operating scenario, providing reliable data support for subsequent determination of component coupling constraints and power consumption optimization, avoiding poor optimization results caused by a single data dimension.

[0038] For example, smart glasses can extract three types of historical parameters from the database over the past week, calculate the historical component coupling vectors between each component, obtain the historical average component coupling vector through arithmetic averaging, and construct a historical component coupling strength matrix. This matrix contains the historical average interaction characteristics between the processor and components such as the display module and camera. Alternatively, the component coupling strength matrix constructed daily over the past week can be directly extracted, and the historical component coupling strength matrix can be obtained by averaging all historical component coupling strength matrices.

[0039] Assuming the current scenario is AR navigation, the smart glasses can set the weight of the real-time component coupling strength matrix to 0.6 and the weight of the historical component coupling strength matrix to 0.4. For each pair of components, the real-time component coupling vector and the historical average component coupling vector are weighted and summed to obtain the target component coupling vector. For example, the target component coupling vector between the processor and the camera is real-time vector × 0.6 + historical average vector × 0.4 = target component coupling vector. All target component coupling vectors are filled into the matrix to form a dynamic coupling weight matrix.

[0040] S103: Determine component coupling constraints based on dynamic coupling weight matrix, and determine power consumption parameter boundary conditions based on device scenario parameters; device scenario parameters include device ambient temperature, task type, and operating mode.

[0041] In this embodiment, component coupling constraints refer to rules that limit the interaction relationship of power consumption parameters of each component. These rules ensure that adjustments to component parameters conform to the current intensity of interaction, preventing parameter conflicts that could lead to increased power consumption or abnormal performance. Power consumption parameter boundary conditions refer to the allowed range of power consumption parameters for each component, defining the upper and lower limits of parameter adjustments. Ambient temperature refers to the real-time temperature of the environment in which the device operates, affecting component heat dissipation and power consumption tolerance. Task type refers to the current function category being performed by the device, such as AR navigation or voice calls for smart glasses. Operating mode refers to the device's preset operating state, such as standby mode or high-performance mode.

[0042] In this embodiment, the intensity of interaction between components changes dynamically with the operating state. If fixed component coupling constraints are used, they are easily decoupled from the current interaction requirements, resulting in poor component parameter coordination. Therefore, this embodiment determines constraints based on a dynamic coupling weight matrix, allowing constraints to accurately match the real-time interaction relationships of components and ensuring coordination during parameter adjustments. Scene parameters directly determine the power consumption characteristics of components. For example, components need to reduce power consumption to prevent overheating at high temperatures. Different task types have different performance and power consumption requirements for components. If the power consumption parameter boundaries are fixed, they are likely to exceed the tolerance or requirement range of components in the current scene. This embodiment determines boundary conditions based on scene parameters, which can define a reasonable range for component parameter adjustments, providing a feasible framework for subsequent particle swarm optimization, and ensuring that the optimization results meet power consumption requirements and are adapted to the actual scene.

[0043] For example, the target coupling vector values ​​of core component pairs are extracted from the current dynamic coupling weight matrix of the smart glasses, such as a coupling value of 0.75 (high coupling) between the processor and the display module, and a coupling value of 0.4 (low coupling) between the processor and the gyroscope sensor. Highly coupled component pairs require strict coordination; for example, the backlight current of the display module can be adjusted synchronously within ±8% for every 100MHz increase in the processor's operating frequency. Lowly coupled component pairs have looser constraints; when the sensor sampling rate changes, the processor resource utilization fluctuation should not exceed 15%. This embodiment can generate a lookup table of component pair parameter adjustment constraints, which serves as the basis for component coupling constraints in subsequent initial particle swarm updates.

[0044] If the current scene parameters for the smart glasses are an ambient temperature of 32℃, a task type of AR fitness course, and a high-performance operating mode, this embodiment can determine the boundary based on component characteristics: at an ambient temperature of 32℃, the processor's maximum operating frequency should not exceed 2.2GHz to avoid overheating damage; the AR fitness course requires continuous image processing, so the processor's power consumption is set to an upper limit of 5W to ensure image processing capabilities; in high-performance mode, the lower limit of the display module's backlight current is 180mA to ensure clear display of fitness movements, while the upper limit of the sensor sampling rate is set to 120Hz to balance motion data acquisition accuracy and power consumption. This embodiment can store the determined component coupling constraints and power consumption parameter boundary conditions in the smart glasses' constraint management module, providing a basis for subsequent initial particle swarm updates and ensuring that the generated first particle swarm meets the current component interaction requirements and the parameter limitations of the scene.

[0045] S104: Randomly generate an initial particle swarm based on the initial power consumption parameters of all components and the hard constraint range of the parameters of each component. The initial particle swarm includes multiple initial particles, and each initial particle includes the power consumption parameters corresponding to each component.

[0046] In this embodiment, the initial power consumption parameters refer to the power consumption control parameters initially set for each component before power consumption optimization, such as the initial operating voltage of the processor and the initial backlight current of the display module. These serve as the basic benchmark data for generating the initial particle swarm. The parameter hard constraint range refers to the extreme range of power consumption parameters determined by the physical characteristics of the component hardware or safe operation requirements. This is the bottom line that the parameters cannot be exceeded; exceeding it will lead to hardware damage or functional abnormalities. The initial particle swarm refers to a set of multiple initial particles randomly generated based on the initial power consumption parameters and the parameter hard constraint range. This serves as the starting solution set for subsequent power consumption optimization iterations. Each initial particle is a single element in the initial particle swarm, and each particle completely contains the power consumption parameters corresponding to all components, representing a set of potential component power consumption parameter combinations.

[0047] In this embodiment, the particle swarm optimization algorithm starts with multiple combinations of potential parameters. The initial particle swarm provides diverse initial solutions, preventing the optimization process from getting trapped in local optima too early and laying the foundation for finding the globally optimal power consumption parameters. The initial power consumption parameters closely resemble the normal operating state of the components. Generating particles based on this reduces the deviation between the initial and optimal solutions, lowers the difficulty of iterative convergence, and improves optimization efficiency. The hard constraint range of parameters is the core guarantee for the safe operation of hardware. Generating particles based on this eliminates the risk of hardware damage caused by parameter exceeding limits. Each particle contains all component parameters, fully reflecting the correlation between parameters between components, meeting the needs of subsequent coupling constraint-based optimization, and ensuring that the initial solution has the feasibility of global optimization.

[0048] For example, this embodiment can extract the initial power consumption parameters of each core component from the system configuration file of the smart glasses, such as the initial operating voltage of the processor, the initial backlight current of the display module, the initial sampling rate of the gyroscope sensor, and the initial transmit power of the Bluetooth communication module. This embodiment can define the hard constraints of each component parameter according to the smart glasses hardware specifications. For example, the hard constraint range for the processor operating voltage is 0.7V-1.3V (below 0.7V will result in insufficient computing power, and above 1.3V will burn out the chip), the hard constraint range for the display module backlight current is 70mA-260mA (below 70mA will result in a dim display, and above 260mA will damage the backlight element), the hard constraint range for the sensor sampling rate is 10Hz-180Hz, and the hard constraint range for the communication module transmit power is 2dBm-7dBm.

[0049] In this embodiment, the initial particle swarm can be set to contain 60 initial particles. For each initial particle, within the hard constraints of each component parameter, the corresponding parameters are randomly adjusted and generated around the initial power consumption parameter. For example, in a certain initial particle, the processor voltage is 1.08V, the display module current is 155mA, the sensor sampling rate is 65Hz, and the communication module power is 4.2dBm. In this way, 60 initial particles containing all component parameters are generated sequentially to form the initial particle swarm.

[0050] S105: Update the initial particle swarm based on the component coupling constraint and power consumption parameter boundary conditions to obtain the first particle swarm that satisfies the component coupling constraint and power consumption parameter boundary conditions.

[0051] In this embodiment, "updating" refers to filtering or adjusting the particles in the initial particle swarm so that the power consumption parameters corresponding to the particles meet the component coupling constraints and power consumption parameter boundary conditions. The first particle swarm refers to the set of particles that, after the update, all satisfy the component coupling constraints and power consumption parameter boundary conditions.

[0052] In this embodiment, although the initial particle swarm is generated based on hard parameter constraints, there will still be particles that do not meet the boundary conditions of component coupling constraints or scene adaptation. This embodiment ensures that the starting solutions of subsequent iterations are all feasible solutions by updating, removing or correcting these particles, avoiding invalid calculations, and laying the foundation for accurate optimization.

[0053] For example, this embodiment can traverse each particle in the initial particle swarm and extract the component power consumption parameters contained in each particle. This embodiment can check the power consumption parameter boundary conditions; if the processor voltage in a particle is 1.35V (exceeding the 0.7V-1.3V hard constraint), it is adjusted to 1.3V; if the display module current is 65mA (below the 70mA boundary), it is adjusted to 70mA. This embodiment can also check the component coupling constraint conditions; if the ratio deviation between the processor frequency and the display module current in a particle reaches 15% (exceeding the 10% constraint), the display module current is proportionally corrected to reduce the deviation to 8%. This embodiment can summarize all particles that meet both conditions after checking and adjustment to form the first particle swarm.

[0054] S106: Iterate and update the first particle swarm until the iterative convergence condition is met. Use the power consumption parameters of all components corresponding to the particle with the lowest fitness value during the iterative update process as the component power consumption optimization parameters of the device; the larger the fitness value, the higher the power consumption.

[0055] In this embodiment, the first particle swarm is iteratively updated until the iterative convergence condition is met, specifically including:

[0056] The first fitness value of each particle in the first particle swarm is calculated based on the fitness function.

[0057] The fitness function is:

[0058] ;

[0059] ;

[0060] ;

[0061] ;

[0062] ;

[0063] in, Let n be the set of power consumption parameters for all components. Let be the parameters of component i, s be the scene parameters of the device, and W be the dynamic coupling weight matrix. Let i be the coupling strength between component i and component j. >0, All are weighting coefficients. , This is the total power consumption item. For the power consumption model of component i, This is a coupling penalty term. Used to quantify the additional power consumption caused by coupling between components. A scene adaptation penalty term is used to characterize the degree to which parameters deviate from the ideal range of the scene. Let i be the ideal parameter range for component i in scenario S. This is a robustness penalty term used to simulate disturbances to power consumption parameters during actual operation. The fluctuation range is , To preset the fluctuation threshold, This represents the fluctuation value of the power consumption parameter of component i during actual operation. To introduce the power consumption model corresponding to component i after the disturbance;

[0064] The first particle swarm is iteratively updated based on the first fitness value of all particles in the first particle swarm until the iteration stopping condition is met. The iteration stopping condition is that the current iteration number reaches the preset maximum iteration number, the change in the fitness value of the globally optimal particle in K consecutive iterations is not greater than the preset precision threshold, or the standard deviation of the fitness values ​​of all particles in the particle swarm is not greater than the preset diversity threshold. The globally optimal particle refers to the particle with the lowest fitness value.

[0065] In this embodiment, iterative update refers to the process of continuously adjusting the component power consumption parameters of particles in the first particle swarm according to preset rules, and gradually optimizing the fitness value of the corresponding particles. The fitness function is an evaluation function used to quantify the merits of combinations of power consumption parameters corresponding to particles, and the result is obtained through multi-dimensional weighted calculation. The total power consumption term refers to the sum of the power consumption of all components calculated based on the power consumption models of each component; the coupling penalty term refers to an evaluation term that quantifies the additional power loss caused by the coupling relationship between components; the scene adaptation penalty term refers to an evaluation term that characterizes the degree to which the component power consumption parameters deviate from the ideal range of the current scene; and the robustness penalty term refers to an evaluation term that simulates the impact on power consumption when parameters are disturbed during actual operation. The weight coefficients are coefficients used to balance the importance of each component of the fitness function, and their sum is 1. The iterative convergence condition refers to the criteria for determining whether to stop the iterative update. The preset maximum number of iterations refers to the maximum number of times the iterative update can be executed in advance; the globally optimal particle refers to the particle with the lowest fitness value (i.e., the lowest power consumption) during the iteration process; the preset accuracy threshold refers to the critical value for judging whether the fitness value of the globally optimal particle is stable; and the preset diversity threshold refers to the critical value for judging whether the particle swarm is overly concentrated.

[0066] In this embodiment, component power consumption optimization requires continuous iteration to gradually approach the optimal parameter combination, with iterative updates providing a path for optimization. The fitness function integrates the combined effects of total power consumption and scenario adaptation, as well as multi-dimensional robustness evaluation, avoiding practical application defects caused by single-dimensional optimization and ensuring that the optimization results fully adapt to the device's operational requirements. The weight coefficient settings can flexibly balance the importance of each evaluation dimension, adapting to the optimization emphasis in different scenarios. Multiple judgment criteria are set for iterative convergence conditions, controlling computational resource consumption by presetting the maximum number of iterations, and ensuring the stability and global optimality of the optimization results through precision thresholds and diversity thresholds, avoiding getting trapped in local optima or infinite iterations. Finally, the lowest power consumption parameter combination is locked through the global optimal particle, achieving the precise power consumption optimization goal.

[0067] In this embodiment, the core principle of the fitness function is to comprehensively quantify the advantages and disadvantages of power consumption parameter combinations from multiple dimensions. By integrating the core influencing factors in actual device operation, a comprehensive evaluation system is constructed to provide accurate judgment criteria for particle swarm optimization, ultimately selecting low-power, highly adaptable, and highly stable component parameter combinations. The fitness function is based on the total power consumption term, directly quantifying the sum of the basic power consumption of all components and anchoring the core optimization goal of low power consumption. The fitness function incorporates a coupling penalty term to respond to the dynamic interaction characteristics between components, avoiding the imbalance of inter-component coordination caused by the optimization of a single component parameter, resulting in additional power loss. The fitness function adds a scene adaptation penalty term, associating the current scene parameters of the device (such as ambient temperature and task type) to prevent parameters from deviating from the ideal range of the scene, ensuring that the optimization results are adapted to the specific use case. The fitness function supplements a robustness penalty term to simulate the fluctuation of parameters under hardware errors and environmental interference in actual operation, ensuring that the optimized parameters can still maintain stable low power consumption under dynamic disturbances.

[0068] For example, based on the current operating scenario of the smart glasses being AR navigation, this embodiment can determine the weight coefficients of the fitness function. Assume that the weight coefficient of the total power consumption term is set to 0.4, the coupling penalty term is set to 0.25, the scene adaptation penalty term is set to 0.2, the robustness penalty term is set to 0.15, and the sum of all coefficients is 1.

[0069] This embodiment can extract the component power consumption parameters of 60 particles in the first particle swarm, including processor operating voltage, display module backlight current sensor sampling rate, etc. For each particle, the total power consumption term is calculated through the power consumption model of each component, and the additional loss caused by the coupling between components is calculated by combining the dynamic coupling weight matrix to obtain the coupling penalty term. The scene adaptation penalty term is obtained by comparing with the ideal parameter range of the scene, and the robustness penalty term is obtained by simulating the power consumption change of the parameters within the preset fluctuation range. The first fitness value of each particle is obtained by weighted summation according to the weight coefficients.

[0070] In this embodiment, the maximum number of iterations can be set to 100, the number of consecutive verifications K to 5, the preset accuracy threshold to a minimum value, the preset diversity threshold to a low value, the current iteration count to 1, and the particle with the lowest first fitness value is determined as the initial global optimal particle.

[0071] For each particle, the power consumption parameters of each component are adjusted by combining its own historical best fitness value and the parameters of the globally best particle. For example, if the fitness value of a certain particle is too high, the processor operating voltage is appropriately reduced and the backlight current of the display module is adjusted to ensure that the component coupling constraints and power consumption parameter boundary conditions are met, and a new particle swarm is generated.

[0072] This embodiment can calculate the fitness value of each particle in the new particle swarm, update the global optimal particle, and check whether the current iteration count has reached 100. If not, it calculates the change in the fitness value of the global optimal particle over 5 consecutive iterations. If the change is not greater than a preset precision threshold, or if it calculates the standard deviation of all fitness values ​​in the particle swarm, and if the standard deviation is not greater than a preset diversity threshold, then the iteration convergence condition is met; if neither condition is met, the iteration count is incremented by 1, and the process returns to step four to continue updating.

[0073] When the iteration convergence condition is met, the iteration stops, and all component power consumption parameters corresponding to the globally optimal particle during the iteration process are extracted, including processor optimized operating voltage, display module optimized backlight current, etc. These parameters are used as component power consumption optimization parameters in the AR navigation scenario of smart glasses and applied to the device operation to achieve the goal of low power consumption.

[0074] As can be seen from the above, the embodiments of this application accurately capture the real-time interaction effects between components through a dynamic coupling weight matrix. This matrix is ​​generated based on real-time operating data and historical data of the device, such as hardware electrical parameters and software operating parameters, and can reflect the collaborative relationship of different components in the current scenario. The component coupling constraints determined in this way can avoid the decoupling of individual component parameters from other components, ensure that the parameters of each component are compatible with each other, reduce power consumption as a whole, and avoid performance fluctuations.

[0075] This application's embodiments introduce scenario parameters such as ambient temperature and task type, and set boundary conditions for power consumption parameters accordingly, ensuring that the optimization process closely matches the actual operating scenario of the device. Then, a particle swarm optimization algorithm is used to iteratively optimize within the constraints, ultimately obtaining parameters that accurately adapt to the current scenario requirements, solving the problem of poor adaptability of a unified optimization strategy.

[0076] In summary, the embodiments of this application can enable the device to operate stably with low power consumption in different scenarios, significantly improving battery life and energy management efficiency.

[0077] In one embodiment of this application, a real-time component coupling strength matrix is ​​constructed based on the device's real-time hardware electrical parameters, real-time software operating parameters, and real-time data interaction volume:

[0078] A hardware feature library is constructed based on the real-time hardware electrical parameters of the device; a software feature library is constructed based on the real-time software operation parameters of the device; and an interaction feature library is constructed based on the real-time data interaction volume of the device.

[0079] The hardware feature library includes the hardware features of multiple components; the software feature library includes the task scheduling features between multiple components; the interaction feature library includes the total data transmission volume of each component and the data transmission volume between each component and other components.

[0080] A real-time component coupling strength matrix is ​​generated based on hardware feature libraries, software feature libraries, and interaction feature libraries.

[0081] In this embodiment, a real-time component coupling strength matrix is ​​generated based on a hardware feature library, a software feature library, and an interaction feature library, including:

[0082] Calculate the correlation degree of hardware parameters between every two components in the hardware feature library;

[0083] Calculate the overlap of task scheduling cycles between every two components in the software feature library, and obtain the correlation of software parameters between the two components based on the overlap of task scheduling cycles between the two components.

[0084] For each pair of components in the interaction feature library, the interaction correlation degree between the two components is calculated based on the total amount of data transmission of each component and the amount of data transmission between the two components.

[0085] A real-time component coupling strength matrix is ​​constructed based on the correlation of hardware parameters, software parameters, and interaction between each pair of components.

[0086] In this embodiment, a real-time component coupling strength matrix is ​​constructed based on the hardware parameter correlation, software parameter correlation, and interaction correlation between every two components, including:

[0087] The synergistic influence coefficient between two components is calculated based on the correlation of hardware parameters, software parameters, and interaction between them; a real-time component coupling vector between the two components is constructed based on the correlation of hardware parameters, software parameters, interaction, and synergistic influence coefficient.

[0088] The real-time component coupling strength matrix is ​​constructed by using the real-time component coupling vector between each pair of components as the matrix elements corresponding to those two components.

[0089] In this embodiment, the hardware feature library refers to a collection built based on the real-time hardware electrical parameters of the device, containing the hardware features of multiple components. The software feature library refers to a collection built based on the real-time software operating parameters of the device, containing the task scheduling features between multiple components. The interaction feature library refers to a collection built based on the real-time data interaction volume of the device, containing the total data transmission volume of each component and the data transmission volume between components. Hardware parameter correlation refers to the quantified value of the correlation between the hardware parameters of each two components in the hardware feature library. Task scheduling cycle overlap refers to the quantified value of the overlap between the task scheduling cycles of each two components in the software feature library. Software parameter correlation refers to the quantified value of the correlation at the software level between each two components obtained based on the overlap of task scheduling cycles. Interaction correlation refers to the quantified value of the correlation at the interaction level calculated based on the total data transmission volume of two components and the data transmission volume between components. The synergistic influence coefficient refers to a coefficient calculated based on the hardware, software, and interaction correlation of two components, reflecting the synergistic influence of the three on the component coupling. The component coupling vector refers to a vector that integrates the hardware, software, interaction correlation, and synergistic influence coefficient of two components, used to characterize the coupling characteristics between components.

[0090] In this embodiment, the coupling relationship between components is influenced by multiple dimensions, including hardware parameter correlation, software task scheduling correlation, and data interaction correlation. A single dimension cannot comprehensively and accurately represent the coupling strength. This embodiment establishes three major feature libraries: hardware, software, and interaction. These libraries can systematically store real-time data for each dimension, providing a structured foundation for multi-dimensional correlation calculation and avoiding evaluation bias caused by mixed data. By calculating three types of correlation, this embodiment can capture the correlation characteristics between components from different levels of hardware, software, and interaction. The introduction of a synergistic influence coefficient reflects the synergistic effect of the three, avoiding the one-sidedness of single-dimensional evaluation. The component coupling vector integrates multi-dimensional information and uses it as matrix elements to construct a component coupling strength matrix, which can comprehensively reflect the coupling relationship between components. This provides reliable and accurate data support for subsequent device power consumption optimization, ensuring that the optimization strategy fits the actual interaction state of the components.

[0091] For example, taking smart glasses as an example, the application process of this embodiment may include:

[0092] (1) Construct three feature libraries. This embodiment can extract the hardware features of each component based on the real-time hardware electrical parameters of the smart glasses in the past 2 hours (the specific time can be set according to the needs), such as the processor working voltage, the display module backlight current, the sensor power supply, etc., and construct a hardware feature library. The library contains the hardware feature data of five core components: processor, display module, camera, gyroscope sensor, and Bluetooth communication module.

[0093] This embodiment can extract task scheduling features between components and build a software feature library based on real-time software operating parameters, such as the start time of image processing tasks between the processor and the camera, the execution cycle of attitude data processing tasks between the sensor and the processor, and the scheduling interval of audio data transmission tasks between the communication module and the processor.

[0094] This embodiment can be based on real-time data interaction volume to count the total data transmission volume of each component every 30 minutes and the daily data transmission volume between components, such as the total daily image data transmission volume of the camera, the total daily data transmission volume of the processor, the image data transmission from the camera to the processor, and the attitude data transmission from the sensor to the processor, etc., to build an interaction feature library.

[0095] (2) Calculate the multi-dimensional correlation. In this embodiment, the real-time hardware parameter sequence of the processor and the display module can be extracted from the hardware feature library. By analyzing the consistency of the changing trends of the two, the correlation between the hardware parameters of the processor and the display module can be quantified. In this embodiment, the same method can be used to calculate the correlation between the hardware parameters of the sensor and the processor, and the correlation between the hardware parameters of the communication module and the processor.

[0096] This embodiment can extract the task scheduling cycles of the processor and the camera from the software feature library, count the proportion of overlapping time between the two scheduling cycles, obtain the task scheduling cycle overlap, and obtain the software parameter correlation between the processor and the camera based on the overlap metric; similarly, the software parameter correlation between the sensor and the processor is calculated.

[0097] This embodiment can obtain data from the processor and camera from the interaction feature library, and obtain the interaction correlation between the processor and camera by calculating the proportion of data transmission between the components to the total transmission of the two components; and calculate the interaction correlation between the sensor and the processor.

[0098] (3) Calculate the synergistic influence coefficient and component coupling vector. For each pair of components, such as the processor and the display module, the weights of the hardware parameter correlation, software parameter correlation, and interaction correlation are set to 0.4, 0.3, and 0.3, respectively. In this embodiment, the synergistic influence coefficient can be calculated by weighted summation based on the hardware parameter correlation, software parameter correlation, and interaction correlation of the two components. In this embodiment, the hardware parameter correlation, software parameter correlation, interaction correlation, and synergistic influence coefficient can be integrated to form the component coupling vector of the processor and the display module. In this embodiment, the same method can be used to calculate the synergistic influence coefficient and component coupling vector of all component pairs such as the processor and the camera, and the sensor and the processor.

[0099] (4) Construct the component coupling strength matrix. The five core components of the smart glasses (processor, display module, camera, sensor, and communication module) are used as the rows and columns of the matrix to form a 5×5 matrix framework. In this embodiment, the component coupling vector between each pair of components can be filled into the corresponding position of the matrix. For example, the first row (processor) and the second column (display module) of the matrix are filled with the coupling vector between the processor and the display module, the first row and the third column (camera) are filled with the coupling vector between the processor and the camera, the second row and the third column are filled with the coupling vector between the display module and the camera, and so on, to complete the construction of the component coupling strength matrix.

[0100] This embodiment integrates component-related data from multiple dimensions by constructing three major feature libraries: hardware, software, and interaction, avoiding the one-sidedness of coupling relationship evaluation caused by a single data dimension. This embodiment calculates the correlation degree of hardware, software, and interaction, combines it with a synergistic influence coefficient to capture the synergistic effect of the three, and then constructs a matrix using component coupling vectors, achieving a comprehensive and accurate characterization of the coupling relationships between components. The resulting component coupling strength matrix can truly reflect the interaction characteristics of components, providing reliable data support for power consumption optimization, effectively improving the adaptability of device power consumption optimization strategies to actual operating scenarios, and solving the problem of poor optimization results caused by inaccurate characterization of coupling relationships in existing technologies.

[0101] In one embodiment of this application, the data transmission volume between the two components includes a first data transmission volume from the first component to the second component and a second data transmission volume from the second component to the first component; based on the total data transmission volume of each component and the data transmission volume between the two components, the interaction correlation degree between the two components is calculated, including:

[0102] Based on the total data transmission volume and the first data transmission amount of the first component, the first dependency coefficient of the first component on the second component is calculated using the first dependency coefficient calculation formula.

[0103] The formula for calculating the first dependency coefficient is:

[0104] ;

[0105] in, The first dependency coefficient, This is the first data transmission volume. For correction items, This represents the total data transmission volume of the first component.

[0106] Based on the total data transmission volume and the second data transmission amount of the second component, the second dependency coefficient of the second component to the first component is calculated using the second dependency coefficient calculation formula;

[0107] The formula for calculating the second dependency coefficient is:

[0108] ;

[0109] in, The second dependency coefficient, This is the second data transmission volume. This represents the total data transmission volume of the second component.

[0110] The interaction correlation between the two components is calculated based on the first dependency coefficient and the second dependency coefficient.

[0111] In this embodiment, the interaction correlation degree between the two components is calculated based on the first dependency coefficient and the second dependency coefficient. Specifically, this includes: calculating the interaction correlation degree between the two components based on the first dependency coefficient and the second dependency coefficient using the interaction correlation degree calculation formula; the interaction correlation degree calculation formula is:

[0112] ;

[0113] Where C represents the degree of interaction correlation. and All are preset weighting coefficients. This represents the total amount of data transferred across all components.

[0114] In this embodiment, the first data transmission volume refers to the amount of data transmitted from the first component to the second component. The second data transmission volume refers to the amount of data transmitted from the second component to the first component. The first dependency coefficient is a numerical value that quantifies the degree of dependency of the first component on the data transmission of the second component. The second dependency coefficient is a numerical value that quantifies the degree of dependency of the second component on the data transmission of the first component. The correction term is a preset small value used to avoid extreme values ​​or meaningless results in data calculation. The weighting coefficient is a preset coefficient used to balance the importance of different parts in the interaction correlation calculation formula. The total data transmission volume of all components refers to the sum of data transmitted by all components in the device within a specific period.

[0115] In this embodiment, the data interaction between components is bidirectional, and the transmission volume in one direction cannot fully reflect the degree of interaction between the two. This embodiment accurately captures the bidirectional dependency relationship between components by calculating the first and second dependency coefficients separately. The introduction of the correction term avoids the problem of dependency coefficient calculation failure due to the total data transmission volume of a certain component being zero or extremely small, ensuring the effectiveness of the calculation. The interaction correlation degree combines the bidirectional dependency coefficient with the proportion of the transmission volume between components to the total transmission volume, reflecting both the relative dependence strength between components and the overall data transmission background. The setting of the weight coefficient can flexibly adapt to the emphasis requirements of different devices or scenarios on bidirectional dependency and overall proportion, ensuring that the interaction correlation degree calculation is comprehensive and accurate, providing reliable support for the subsequent construction of the coupling strength matrix.

[0116] For example, in this embodiment, the processor of the smart glasses can be selected as the first component, and the camera as the second component, to extract relevant data from the interaction feature library. For instance, the first data transmission volume, i.e., the amount of image data received by the processor from the camera, is 18GB per day; the second data transmission volume, i.e., the amount of control command data received by the camera from the processor, is 0.5GB per day; the total data transmission volume of the first component, i.e., the total daily transmission volume of the processor, is 50GB; the total data transmission volume of the second component, i.e., the total daily transmission volume of the camera, is 20GB; the total data transmission volume of all components, i.e., the total daily transmission volume of all core components of the smart glasses, is 120GB; the preset correction term is a minimum value, the weight coefficient μ is set to 0.6, and σ is set to 0.4.

[0117] In this embodiment, the first dependency coefficient can be obtained by substituting the first data transmission volume of 18GB, the first component data transmission volume of 50GB, and the correction term into the first dependency coefficient calculation formula, thereby quantifying the processor's dependence on the camera.

[0118] In this embodiment, the second dependency coefficient can be obtained by substituting the second data transmission volume of 0.5GB, the total data transmission volume of the second component of 20GB, and the correction term into the second dependency coefficient calculation formula. This second dependency coefficient characterizes the degree of dependence of the camera on the processor.

[0119] In this embodiment, the first dependency coefficient and the second dependency coefficient can be substituted into the interaction correlation degree calculation formula, and the ratio of the sum of the bidirectional transmission volume of the two components (18.5GB) to the total transmission volume of all components (120GB) can be substituted. Combined with the weight coefficients μ and σ, the interaction correlation degree between the processor and the camera can be calculated.

[0120] This embodiment captures the bidirectional characteristics of data interaction between components through a bidirectional dependency coefficient, avoiding the one-sidedness of a single-direction evaluation. The setting of correction terms ensures the stability and effectiveness of the calculation, preventing bias caused by extreme data. The interaction correlation degree integrates bidirectional dependency and overall transmission ratio, combined with adjustable weight coefficients, to adapt to different scenario requirements, resulting in more comprehensive and accurate calculation results. The calculation method of this embodiment provides reliable interaction correlation degree data for the component coupling strength matrix, ensuring that the matrix truly reflects the interaction characteristics between components, thereby improving the accuracy and adaptability of subsequent device power consumption optimization.

[0121] In one embodiment of this application, a device power consumption optimization method further includes:

[0122] Acquire real-time fault data of the equipment, construct a fault feature library based on the real-time fault data, and determine the optimization boundary conditions based on the fault feature library;

[0123] An initial particle swarm is randomly generated based on the initial power consumption parameters of all components and the hard constraint range of the parameters of each component. The initial particle swarm includes multiple initial particles, and each initial particle includes the power consumption parameters corresponding to each component.

[0124] The initial particle swarm is updated based on the component coupling constraint and power consumption parameter boundary conditions to obtain the first particle swarm that satisfies the component coupling constraint and power consumption parameter boundary conditions.

[0125] The first particle swarm is iteratively updated based on the optimized boundary conditions until the iterative convergence condition is met. The power consumption parameters of all components corresponding to the particle with the lowest fitness value during the iterative update process are used as the component power consumption optimization parameters of the device.

[0126] In this embodiment, historical fault data refers to relevant data from past equipment operation when faults occurred, such as component parameters and fault types at the time of the fault. The fault feature library is a collection built upon historical fault data to store characteristic information corresponding to faults. Optimized boundary conditions refer to constraints determined based on the fault feature library, used to prevent component parameters from falling into ranges that are prone to causing faults.

[0127] Considering that relying solely on component coupling constraints and power consumption parameter boundary optimization may overlook the risk of parameter-induced failures, this embodiment introduces historical failure data to construct a failure feature library and determine optimization boundary conditions. This allows for the avoidance of parameters prone to failure during iterative updates, ensuring that the optimization process achieves both low power consumption and reliable equipment operation, and preventing equipment failures caused by pursuing power consumption optimization.

[0128] For example, this embodiment can acquire a fault feature library of smart glasses, filter typical fault types (such as component overheating, function interruption) from the fault feature library, extract the associated parameters and abnormal parameter ranges corresponding to each fault, establish the correspondence between fault type, associated parameter, and abnormal range, and identify the key parameters and failure value ranges that cause the fault. Based on the abnormal parameter range, a preset safety margin (such as 5%-10%) is added to determine the critical threshold. For example, if a fault is caused when the abnormal parameter range of a certain component is ≥1.25V, the critical threshold is set to ≤1.2V after adding the margin to avoid the parameter approaching the fault risk range. The parameters and critical thresholds of each component are organized into structured boundary conditions to form a clear component parameter constraint comparison table.

[0129] This embodiment can generate an initial particle swarm and update it to obtain the first particle swarm. In each iteration, the fitness value of each particle in the first particle swarm is calculated first, and then it is checked whether the parameters of each particle meet the optimization boundary. For example, if a particle's processor voltage is 1.3V, exceeding the boundary, it is adjusted to 1.25V; if a particle displays a module current of 58mA, it is adjusted to 60mA. This embodiment can repeat the above checking and adjustment process until the iteration convergence condition is met, thus determining the component power consumption optimization parameters.

[0130] This embodiment introduces fault-related optimization boundary conditions, so that the final component power consumption optimization parameters not only meet the low power consumption requirements, but also avoid parameter ranges that are prone to causing faults. This avoids equipment failure caused by the optimization process, and ensures the reliability of equipment operation while reducing power consumption, thus improving the practicality and safety of the power consumption optimization scheme.

[0131] A device power consumption optimization method corresponding to the above embodiment, Figure 2 This is a structural block diagram of a device power consumption optimization apparatus according to an embodiment of this application. For ease of explanation, only the parts relevant to the embodiment of this application are shown. References Figure 2 The device power consumption optimization device 20 includes: a coupling vector acquisition module 21, a dynamic coupling analysis module 22, a constraint analysis module 23, an initialization module 24, a condition constraint module 25, and a power consumption optimization module 26.

[0132] Among them, the coupling vector acquisition module 21 is used to construct a real-time component coupling strength matrix based on the real-time hardware electrical parameters, real-time software operating parameters and real-time data interaction volume of the device; the component coupling strength matrix includes real-time component coupling vectors between multiple components in the device;

[0133] The dynamic coupling analysis module 22 is used to obtain the historical component coupling strength matrix of the device, and to obtain the dynamic coupling weight matrix by weighted summation of the real-time component coupling strength matrix and the historical component coupling strength matrix. The historical component coupling strength matrix is ​​determined based on the historical hardware electrical parameters, historical software operating parameters and historical data interaction of the device. The historical component coupling strength matrix includes the historical average component coupling vector among multiple components in the device. The dynamic coupling weight matrix includes the target component coupling vector among multiple components in the device. The target component coupling vector is used to characterize the interaction influence strength of multiple components in the current operating scenario of the device.

[0134] Constraint analysis module 23 is used to determine component coupling constraints based on dynamic coupling weight matrix and power consumption parameter boundary conditions based on device scenario parameters; device scenario parameters include device ambient temperature, task type and operating mode;

[0135] Initialization module 24 is used to randomly generate an initial particle swarm based on the initial power consumption parameters of all components and the hard constraint range of the parameters of each component. The initial particle swarm includes multiple initial particles, and each initial particle includes the power consumption parameters corresponding to each component.

[0136] The condition constraint module 25 is used to update the initial particle swarm based on the component coupling constraint condition and the power consumption parameter boundary condition to obtain the first particle swarm that satisfies the component coupling constraint condition and the power consumption parameter boundary condition.

[0137] The power consumption optimization module 26 is used to iteratively update the first particle swarm until the iterative convergence condition is met. The power consumption parameters of all components corresponding to the particle with the lowest fitness value during the iterative update process are used as the component power consumption optimization parameters of the device; the larger the fitness value, the higher the power consumption.

[0138] In one embodiment of this application, the coupling vector acquisition module 21, when constructing a real-time component coupling strength matrix based on the device's real-time hardware electrical parameters, real-time software operating parameters, and real-time data interaction volume, is specifically used for:

[0139] A hardware feature library is constructed based on the real-time hardware electrical parameters of the device; a software feature library is constructed based on the real-time software operation parameters of the device; and an interaction feature library is constructed based on the real-time data interaction volume of the device.

[0140] The hardware feature library includes the hardware features of multiple components; the software feature library includes the task scheduling features between multiple components; the interaction feature library includes the total data transmission volume of each component and the data transmission volume between each component and other components.

[0141] A real-time component coupling strength matrix is ​​generated based on hardware feature libraries, software feature libraries, and interaction feature libraries.

[0142] In one embodiment of this application, the coupling vector acquisition module 21, when generating a real-time component coupling strength matrix based on the hardware feature library, software feature library, and interaction feature library, is specifically used for:

[0143] Calculate the correlation degree of hardware parameters between every two components in the hardware feature library;

[0144] Calculate the overlap of task scheduling cycles between every two components in the software feature library, and obtain the correlation of software parameters between the two components based on the overlap of task scheduling cycles between the two components.

[0145] For each pair of components in the interaction feature library, the interaction correlation degree between the two components is calculated based on the total amount of data transmission of each component and the amount of data transmission between the two components.

[0146] A real-time component coupling strength matrix is ​​constructed based on the correlation of hardware parameters, software parameters, and interaction between each pair of components.

[0147] In one embodiment of this application, the data transmission volume between the two components includes a first data transmission volume from the first component to the second component and a second data transmission volume from the second component to the first component; the coupling vector acquisition module 21, when calculating the interaction correlation degree between the two components based on the total data transmission volume of each component and the data transmission volume between the two components, is specifically used for:

[0148] Based on the total data transmission volume and the first data transmission amount of the first component, the first dependency coefficient of the first component on the second component is calculated using the first dependency coefficient calculation formula.

[0149] The formula for calculating the first dependency coefficient is:

[0150] ;

[0151] in, The first dependency coefficient, This is the first data transmission volume. For correction items, This represents the total data transmission volume of the first component.

[0152] Based on the total data transmission volume and the second data transmission amount of the second component, the second dependency coefficient of the second component to the first component is calculated using the second dependency coefficient calculation formula;

[0153] The formula for calculating the second dependency coefficient is:

[0154] ;

[0155] in, The second dependency coefficient, This is the second data transmission volume. This represents the total data transmission volume of the second component.

[0156] The interaction correlation between the two components is calculated based on the first dependency coefficient and the second dependency coefficient.

[0157] In one embodiment of this application, when the coupling vector acquisition module 21 calculates the interaction correlation degree between the two components based on the first dependency coefficient and the second dependency coefficient, it is specifically used for:

[0158] Based on the first dependency coefficient and the second dependency coefficient, the interaction degree between the two components is calculated using the interaction degree calculation formula.

[0159] The formula for calculating the degree of interaction relevance is:

[0160] ;

[0161] Where C represents the degree of interaction correlation. and All are preset weighting coefficients. This represents the total amount of data transferred across all components.

[0162] In one embodiment of this application, when constructing a real-time component coupling strength matrix based on the hardware parameter correlation, software parameter correlation, and interaction correlation between each pair of components, the coupling vector acquisition module 21 is specifically used to: calculate the synergistic influence coefficient between the two components based on the hardware parameter correlation, software parameter correlation, and interaction correlation between each pair of components; and construct a real-time component coupling vector between the two components based on the hardware parameter correlation, software parameter correlation, interaction correlation, and synergistic influence coefficient.

[0163] The real-time component coupling strength matrix is ​​constructed by using the real-time component coupling vector between each pair of components as the matrix elements corresponding to those two components.

[0164] In one embodiment of this application, the power consumption optimization module 26, when iteratively updating the first particle swarm until the iterative convergence condition is met, is specifically used for:

[0165] The first fitness value of each particle in the first particle swarm is calculated based on the fitness function.

[0166] The fitness function is:

[0167] ;

[0168] ;

[0169] ;

[0170] ;

[0171] ;

[0172] in, Let n be the set of power consumption parameters for all components. Let be the parameters of component i, s be the scene parameters of the device, and W be the dynamic coupling weight matrix. Let i be the coupling strength between component i and component j. >0, All are weighting coefficients. , This is the total power consumption item. For the power consumption model of component i, This is a coupling penalty term. Used to quantify the additional power consumption caused by coupling between components. A scene adaptation penalty term is used to characterize the degree to which parameters deviate from the ideal range of the scene. Let i be the ideal parameter range for component i in scenario S. This is a robustness penalty term used to simulate disturbances to power consumption parameters during actual operation. The fluctuation range is , To preset the fluctuation threshold, This represents the fluctuation value of the power consumption parameter of component i during actual operation. To introduce the power consumption model corresponding to component i after the disturbance;

[0173] The first particle swarm is iteratively updated based on the first fitness value of all particles in the first particle swarm until the iteration stopping condition is met. The iteration stopping condition is that the current iteration number reaches the preset maximum iteration number, the change in the fitness value of the globally optimal particle in K consecutive iterations is not greater than the preset precision threshold, or the standard deviation of the fitness values ​​of all particles in the particle swarm is not greater than the preset diversity threshold. The globally optimal particle refers to the particle with the lowest fitness value.

[0174] See Figure 3 , Figure 3 This is a schematic block diagram of an electronic device provided according to an embodiment of this application. Figure 3 The electronic device 300 in this embodiment may include one or more processors 301, one or more input devices 302, one or more output devices 303, and one or more memories 304. The processors 301, input devices 302, output devices 303, and memories 304 communicate with each other via a communication bus 305. The memories 304 store computer programs, including program instructions. The processors 301 execute the program instructions stored in the memories 304. Specifically, the processors 301 are configured to invoke the program instructions to perform the functions of the modules in the aforementioned device embodiments, for example... Figure 2 The functions of the coupling vector acquisition module 21, dynamic coupling analysis module 22, constraint analysis module 23, initialization module 24, condition constraint module 25, and power consumption optimization module 26 are shown.

[0175] It should be understood that, in the embodiments of this application, the processor 301 may be a central processing unit (CPU), but it may also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor.

[0176] Input device 302 may include a touchpad, a fingerprint sensor (for collecting the user's fingerprint information and fingerprint orientation information), a microphone, etc., and output device 303 may include a display (LCD, etc.), a speaker, etc.

[0177] The memory 304 may include read-only memory and random access memory, and provides instructions and data to the processor 301. A portion of the memory 304 may also include non-volatile random access memory. For example, the memory 304 may also store device type information.

[0178] In specific implementations, the processor 301, input device 302, and output device 303 described in the embodiments of this application can execute the implementation methods described in the embodiments of the device power consumption optimization method provided in the embodiments of this application, or they can execute the implementation methods of the electronic device 300 described in the embodiments of this application, which will not be repeated here.

[0179] In another embodiment of this application, a computer-readable storage medium is provided. This computer-readable storage medium stores a computer program, which includes program instructions. When executed by a processor, the program instructions implement all or part of the processes in the methods described above. Alternatively, the computer program can instruct related hardware to complete the process. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include any entity or device capable of carrying computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc.

[0180] The computer-readable storage medium can be an internal storage unit of the electronic device in any of the foregoing embodiments, such as a hard disk or memory of the electronic device. The computer-readable storage medium can also be an external storage device of the electronic device, such as a plug-in hard disk, smart media card (SMC), secure digital card (SD) card, flash card, etc., equipped on the electronic device. Furthermore, the computer-readable storage medium can include both internal and external storage units of the electronic device. The computer-readable storage medium is used to store computer programs and other programs and data required by the electronic device. The computer-readable storage medium can also be used to temporarily store data that has been output or will be output.

[0181] Those skilled in the art will recognize that the modules / units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this application.

[0182] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the electronic devices and units described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0183] In the several embodiments provided in this application, it should be understood that the disclosed electronic devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For instance, the division of modules / units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules, units, or components may be combined or integrated into another system, or some features may be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces or modules / units, or it may be an electrical, mechanical, or other form of connection.

[0184] The modules / units described as separate components may or may not be physically separate. Similarly, the components shown as modules / units may or may not be physical modules / units; they may be located in one place or distributed across multiple network modules / units. Some or all of the modules / units can be selected to achieve the purpose of the embodiments in this application, depending on actual needs.

[0185] Furthermore, the functional modules / units in the various embodiments of this application can be integrated into one processing module / unit, or each module / unit can exist physically separately, or two or more modules / units can be integrated into one module / unit. The integrated modules / units described above can be implemented in hardware or as software functional modules / units.

[0186] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in this application, and these modifications or substitutions should all be covered within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for optimizing device power consumption, characterized in that, include: A hardware feature library is constructed based on the device's real-time hardware electrical parameters, a software feature library is constructed based on the device's real-time software operating parameters, and an interaction feature library is constructed based on the device's real-time data interaction volume. The hardware feature library includes hardware features of multiple components; the software feature library includes task scheduling features among multiple components; and the interaction feature library includes the total data transmission volume of each component and the data transmission volume between each component and other components. Calculate the correlation degree of hardware parameters between every two components in the hardware feature library; Calculate the task scheduling cycle overlap degree between every two components in the software feature library, and obtain the software parameter correlation degree between the two components based on the task scheduling cycle overlap degree; for every two components in the interaction feature library, calculate the interaction correlation degree between the two components based on the total data transmission volume of each of the two components and the data transmission volume between the two components; construct a real-time component coupling strength matrix based on the hardware parameter correlation degree, the software parameter correlation degree, and the interaction correlation degree between every two components; Obtain the historical component coupling strength matrix of the device, and perform a weighted summation based on the real-time component coupling strength matrix and the historical component coupling strength matrix to obtain the dynamic coupling weight matrix; The historical component coupling strength matrix is ​​determined based on the device's historical hardware electrical parameters, historical software operating parameters, and historical data interaction volume. The historical component coupling strength matrix includes the historical average component coupling vector among multiple components in the device. The dynamic coupling weight matrix includes the target component coupling vector among multiple components in the device. The target component coupling vector is used to characterize the interaction influence strength of the multiple components in the current operating scenario of the device. The component coupling constraints are determined based on the dynamic coupling weight matrix, and the power consumption parameter boundary conditions are determined based on the device's scenario parameters. The device's scenario parameters include the device's ambient temperature, task type, and operating mode. The component coupling constraints refer to the rules that restrict the interaction relationship of the power consumption parameters of each component, which are used to ensure that the adjustment of component parameters conforms to the current interaction influence intensity. An initial particle swarm is randomly generated based on the initial power consumption parameters of all components and the hard constraint range of the parameters of each component. The initial particle swarm includes multiple initial particles, and each initial particle includes the power consumption parameters corresponding to each component. The initial particle swarm is updated based on the component coupling constraint and the power consumption parameter boundary condition to obtain a first particle swarm that satisfies the component coupling constraint and the power consumption parameter boundary condition. The first particle swarm is iteratively updated until the iterative convergence condition is met. The power consumption parameters of all components corresponding to the particle with the lowest fitness value during the iterative update process are used as the component power consumption optimization parameters of the device. A higher fitness value indicates higher power consumption; The step of determining component coupling constraints based on the dynamic coupling weight matrix includes: extracting the target coupling vector values ​​of core component pairs from the dynamic coupling weight matrix; requiring strict coordination for highly coupled component pairs and easing constraints for low-coupled component pairs; and generating a reference table of component pair-parameter adjustment constraints as the basis for subsequent initial particle swarm updates of component coupling constraints.

2. The device power consumption optimization method as described in claim 1, characterized in that, The amount of data transmission between the two components includes the first amount of data transmission from the first component to the second component and the second amount of data transmission from the second component to the first component. The calculation of the interaction correlation degree between the two components based on the total data transmission volume of each component and the data transmission volume between the two components includes: Based on the total data transmission volume and the first data transmission amount of the first component, the first dependency coefficient of the first component on the second component is calculated using the first dependency coefficient calculation formula. The formula for calculating the first dependency coefficient is: ； in, The first dependency coefficient, This is the first data transmission volume. For correction items, This represents the total data transmission volume of the first component. Based on the total data transmission volume and the second data transmission amount of the second component, the second dependency coefficient of the second component to the first component is calculated using the second dependency coefficient calculation formula; The formula for calculating the second dependency coefficient is: ； in, The second dependency coefficient, This is the second data transmission volume. This represents the total data transmission volume of the second component. The interaction correlation degree between the two components is calculated based on the first dependency coefficient and the second dependency coefficient.

3. The device power consumption optimization method as described in claim 2, characterized in that, The calculation of the interaction correlation degree between the two components based on the first dependency coefficient and the second dependency coefficient includes: Based on the first dependency coefficient and the second dependency coefficient, the interaction degree between the two components is calculated using the interaction degree calculation formula. The formula for calculating the degree of interaction correlation is: ； Where C represents the degree of interaction correlation. and All are preset weighting coefficients. This represents the total amount of data transferred across all components.

4. The device power consumption optimization method as described in claim 1, characterized in that, The construction of a real-time component coupling strength matrix based on the hardware parameter correlation, software parameter correlation, and interaction correlation between every two components includes: The synergistic influence coefficient between two components is calculated based on the correlation degree of hardware parameters, the correlation degree of software parameters, and the correlation degree of interaction between each pair of components; a real-time component coupling vector between the two components is constructed based on the correlation degree of hardware parameters, the correlation degree of software parameters, the correlation degree of interaction, and the synergistic influence coefficient. The real-time component coupling strength matrix is ​​constructed by using the real-time component coupling vector between each pair of components as the matrix elements corresponding to those two components.

5. The device power consumption optimization method as described in claim 1, characterized in that, The iterative update of the first particle swarm until the iterative convergence condition is met includes: The first fitness value of each particle in the first particle swarm is calculated based on the fitness function. The fitness function is: ； ； ； ； ； in, Let n be the set of power consumption parameters for all components. Let be the parameters of component i, s be the scene parameters of the device, and W be the dynamic coupling weight matrix. Let i be the coupling strength between component i and component j. >0, All are weighting coefficients. , This is the total power consumption item. For the power consumption model of component i, This is a coupling penalty term. Used to quantify the additional power consumption caused by coupling between components. A scene adaptation penalty term is used to characterize the degree to which parameters deviate from the ideal range of the scene. Let i be the ideal parameter range for component i in scenario S. This is a robustness penalty term used to simulate disturbances to power consumption parameters during actual operation. The fluctuation range is , To preset the fluctuation threshold, This represents the fluctuation value of the power consumption parameter of component i during actual operation. To introduce the power consumption model corresponding to component i after the disturbance; The first particle swarm is iteratively updated based on the first fitness value of all particles in the first particle swarm until the iteration stopping condition is met. The iteration stopping condition is that the current iteration number reaches the preset maximum iteration number, the change in the fitness value of the globally optimal particle in K consecutive iterations is not greater than a preset precision threshold, or the standard deviation of the fitness values ​​of all particles in the particle swarm is not greater than a preset diversity threshold. The globally optimal particle refers to the particle with the lowest fitness value.

6. A device for optimizing device power consumption, characterized in that, include: The coupling vector acquisition module is used to construct a hardware feature library based on the device's real-time hardware electrical parameters, a software feature library based on the device's real-time software operating parameters, and an interaction feature library based on the device's real-time data interaction volume. The hardware feature library includes hardware features of multiple components; the software feature library includes task scheduling features among multiple components; and the interaction feature library includes the total data transmission volume of each component and the data transmission volume between each component and other components. Calculate the correlation degree of hardware parameters between every two components in the hardware feature library; Calculate the task scheduling cycle overlap degree between every two components in the software feature library, and obtain the software parameter correlation degree between the two components based on the task scheduling cycle overlap degree; for every two components in the interaction feature library, calculate the interaction correlation degree between the two components based on the total data transmission volume of each of the two components and the data transmission volume between the two components; construct a real-time component coupling strength matrix based on the hardware parameter correlation degree, the software parameter correlation degree, and the interaction correlation degree between every two components; The dynamic coupling analysis module is used to obtain the historical component coupling strength matrix of the device, and to obtain the dynamic coupling weight matrix by weighted summation based on the real-time component coupling strength matrix and the historical component coupling strength matrix. The historical component coupling strength matrix is ​​determined based on the device's historical hardware electrical parameters, historical software operating parameters, and historical data interaction volume. The historical component coupling strength matrix includes the historical average component coupling vector among multiple components in the device. The dynamic coupling weight matrix includes the target component coupling vector among multiple components in the device. The target component coupling vector is used to characterize the interaction influence strength of the multiple components in the current operating scenario of the device. The constraint analysis module is used to determine component coupling constraints based on the dynamic coupling weight matrix and to determine power consumption parameter boundary conditions based on the device's scenario parameters. The device's scenario parameters include the device's ambient temperature, task type, and operating mode. The component coupling constraints refer to the rules that restrict the interaction relationship of power consumption parameters of each component, and are used to ensure that the adjustment of component parameters conforms to the current interaction influence intensity. The constraint analysis module is specifically used to extract the target coupling vector values ​​of core component pairs from the dynamic coupling weight matrix. Highly coupled component pairs need strict coordination, while lowly coupled component pairs have relaxed constraints. It generates a reference table of component pair parameter adjustment constraints, which serves as the basis for component coupling constraints in subsequent initial particle swarm updates. An initialization module is used to randomly generate an initial particle swarm based on the initial power consumption parameters of all components and the hard constraint range of the parameters of each component. The initial particle swarm includes multiple initial particles, and each initial particle includes the power consumption parameters corresponding to each component. The condition constraint module is used to update the initial particle swarm based on the component coupling constraint condition and the power consumption parameter boundary condition to obtain a first particle swarm that satisfies the component coupling constraint condition and the power consumption parameter boundary condition. The power consumption optimization module is used to iteratively update the first particle swarm until the iterative convergence condition is met, and to use the power consumption parameters of all components corresponding to the particle with the lowest fitness value during the iterative update process as the component power consumption optimization parameters of the device. A higher fitness value indicates higher power consumption.

7. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method as described in any one of claims 1 to 5.

8. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method as described in any one of claims 1 to 5.

Images