Game resource optimization method and system based on artificial intelligence

Abstract

The invention discloses a game resource optimization method and system based on artificial intelligence, and relates to the technical field of resource management, and the method comprises the steps: carrying out the empirical mode decomposition of a time sequence matrix, obtaining an effective IMF and a residual error, carrying out the Hilbert transformation of the effective IMF, constructing a comprehensive load index, and calculating a reference priority trajectory through linear extrapolation. Obtaining a final priority vector; and predicting resource demands in future time, defining a potential function for each resource, calculating a total potential function, predicting a system state by using a state equation, constructing a resource allocation target function, and obtaining an optimized resource allocation result. According to the method, through a signal decomposition method combining EMD and Hilbert transform, the supporting capacity of priority prediction and resource allocation for time sensitivity is enhanced, a comprehensive load index and reference trajectory prediction are introduced, and the performance balance and player experience continuity of a game under multi-task competition are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of resource management, and in particular to a method and system for optimizing game resources based on artificial intelligence. Background Art

[0002] With the rapid development of the electronic game industry, modern games have placed higher demands on the consumption and utilization efficiency of system resources. Especially in large-scale 3D games, multiplayer online competitive games and virtual reality scenarios, the game system's dynamic allocation requirements for CPU, GPU, memory and network bandwidth have become increasingly complex. Traditional resource management strategies are mostly based on static configuration or fixed rules, which fail to fully respond to the rapid changes in player behavior, game status and hardware load. The successful application of artificial intelligence in image recognition, speech processing and adaptive control has prompted researchers to begin to try to introduce AI technology to optimize game resource allocation, in order to achieve energy consumption control and performance maximization while ensuring operational stability and visual effects.

[0003] Existing AI-based game resource optimization methods still have the following shortcomings: resource demand modeling lacks quantitative representation of the dynamic coupling relationship between game behavior and system response, resulting in lag or overfitting problems in resource prediction. Summary of the Invention

[0004] In view of the above existing problems, the present invention is proposed.

[0005] Therefore, the present invention provides an artificial intelligence-based game resource optimization method and system to solve the problem that resource demand modeling lacks quantitative representation of the dynamic coupling relationship between game behavior and system response, resulting in lag or overfitting in resource prediction.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: In a first aspect, the present invention provides a game resource optimization method based on artificial intelligence, which comprises: Collect and preprocess game operation data to generate a time series matrix. Perform empirical mode decomposition on the time series matrix to obtain effective IMF and residuals. Perform Hilbert transform on the effective IMF, calculate the instantaneous amplitude and instantaneous frequency, calculate the mean of the residuals, construct a comprehensive load index, use linear extrapolation to calculate the reference priority trajectory, construct the priority objective function, and constrain the projection of the optimized priority trajectory to obtain the final priority vector. Based on the time series matrix, a sliding window is used to set the time window. A fast Fourier transform is performed on each time window to obtain the spectrum of the time window. The Shannon entropy of the spectrum is calculated. The time window corresponding to the minimum Shannon entropy is selected as the optimal window to predict resource demand in the future. A potential function is defined for each resource, and the total potential function is calculated. The state equation is used to predict the system state, and the resource allocation objective function is constructed to obtain the optimized resource allocation result. Use the PID controller to convert it into control instructions, transmit and execute them, and build a visual interface to display the control instructions.

[0007] As a preferred solution of the artificial intelligence-based game resource optimization method of the present invention, the method of constructing a comprehensive load index, using linear extrapolation to calculate a reference priority trajectory, building a priority objective function, and performing constrained projection on the optimized priority trajectory to obtain a final priority vector includes: The time series matrix is ​​used as the initial signal, and the local maximum and local minimum of the initial signal are identified using the numerical comparison method. The upper and lower envelopes of the local maximum and local minimum are calculated using cubic spline interpolation, and the mean envelope is calculated to extract the candidate IMF. Use the numerical comparison method to verify the candidate IMF, count the number of extreme points and zero crossing points of the candidate IMF, use the statistical distribution method to set a fixed threshold, set the verification conditions, and when the candidate IMF meets the verification conditions, mark the candidate IMF as a valid IMF, otherwise use signal assignment to update the current signal; Use signal subtraction to calculate the update signal, use the standard deviation threshold comparison to set the inspection condition, filter the update signal that meets the inspection condition, and mark it as residual; Perform Hilbert transform on the effective IMF, calculate the instantaneous amplitude and instantaneous frequency, calculate the mean of the residual, construct the comprehensive load index, and perform normalization to generate the initial priority vector; The reference priority trajectory is calculated using linear extrapolation, and the objective function of the priority is constructed using the weighted sum of squares. The optimized priority trajectory is solved using quadratic programming, and the optimized priority trajectory is constrained and projected to obtain the final priority vector.

[0008] As a preferred solution of the artificial intelligence-based game resource optimization method of the present invention, wherein: the Shannon entropy of the spectrum is calculated, and the time window corresponding to the minimum Shannon entropy is selected as the optimal window to predict resource requirements in the future time includes: Based on the time series matrix, a sliding window is used to set the time window. Fast Fourier transform is performed on each time window to obtain the spectrum of the time window. The Shannon entropy of the spectrum is calculated using the Shannon entropy formula. The time window corresponding to the minimum Shannon entropy is selected as the optimal window. Fast Fourier transform is performed on the optimal window to obtain the optimal spectrum. The amplitude of the discrete frequency and the spectral phase of the discrete frequency are calculated. Based on the final priority vector, the resource demand in the future time is predicted and normalized in combination with the amplitude of the discrete frequency and the spectral phase of the discrete frequency.

[0009] As a preferred solution of the artificial intelligence-based game resource optimization method of the present invention, the step of constructing a resource allocation objective function to obtain an optimized resource allocation result includes: Define potential function for each resource and calculate the total potential function; The state equation is used to predict the system state. Based on the predicted system state, the resource allocation objective function is constructed and solved using Newton's method to obtain the optimized resource allocation result.

[0010] As a preferred solution of the artificial intelligence-based game resource optimization method of the present invention, the method of converting the control instructions using a PID controller, transmitting and executing the control instructions, includes: Use the difference method to calculate the error between the optimized resource allocation result and the current resource allocation result; A PID controller is used to convert the error into a control instruction, and a network communication protocol is used to transmit and execute the control instruction.

[0011] As a preferred solution of the artificial intelligence-based game resource optimization method of the present invention, the step of constructing a visual interface to display control instructions includes: Use the visualization tool Matplotlib to build a visualization interface to display control instructions and game running data in real time; Users who have passed real-name verification are allowed to view it.

[0012] As a preferred solution of the artificial intelligence-based game resource optimization method of the present invention, the collecting of game operation data and preprocessing to generate a time series matrix includes: Use the APL interface to collect game running data and perform timestamp alignment, denoising, and normalization processing; The game operation data includes CPU occupancy, GPU occupancy, memory usage, frame rate, player input frequency, and network latency data; Sort the game running data in chronological order to generate a time series matrix.

[0013] In a second aspect, the present invention provides a game resource optimization system based on artificial intelligence, comprising: The priority collection module is used to collect and preprocess game operation data, generate a time series matrix, perform empirical mode decomposition on the time series matrix, obtain effective IMF and residuals, perform Hilbert transform on the effective IMF, calculate the instantaneous amplitude and instantaneous frequency, calculate the mean of the residuals, construct a comprehensive load index, use linear extrapolation to calculate the reference priority trajectory, construct the priority objective function, and constrain the projection of the optimized priority trajectory to obtain the final priority vector; The demand allocation module is used to set the time window using a sliding window based on the time series matrix, perform a fast Fourier transform on each time window, obtain the spectrum of the time window, calculate the Shannon entropy of the spectrum, select the time window corresponding to the minimum Shannon entropy as the optimal window, predict resource demand in the future, define a potential function for each resource, calculate the total potential function, use the state equation to predict the system state, construct the resource allocation objective function, and obtain the optimized resource allocation result; The instruction visualization module is used to convert the PID controller into control instructions, transmit and execute them, and build a visual interface to display the control instructions.

[0014] In a third aspect, the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program is executed by the processor, any step of the artificial intelligence-based game resource optimization method as described in the first aspect of the present invention is implemented.

[0015] In a fourth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, any step of the artificial intelligence-based game resource optimization method as described in the first aspect of the present invention is implemented.

[0016] The beneficial effects of the present invention are as follows: the present invention enhances the support capability of priority prediction and resource allocation for time sensitivity through a signal decomposition method combining EMD and Hilbert transform, introduces comprehensive load indicators and reference trajectory prediction, and improves the performance balance and player experience continuity of games under multi-task competition. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0018] Figure 1This is a flowchart of the game resource optimization method based on artificial intelligence in Example 1.

[0019] Figure 2 Schematic diagram of the game resource optimization system based on artificial intelligence in Example 1. DETAILED DESCRIPTION

[0020] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0021] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0022] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive of other embodiments.

[0023] Example 1, reference Figure 1 and Figure 2 , which is the first embodiment of the present invention, provides a game resource optimization method based on artificial intelligence, comprising the following steps: S1. Collect game operation data and preprocess it to generate a time series matrix. Perform empirical mode decomposition on the time series matrix to obtain effective IMF and residuals. Perform Hilbert transform on the effective IMF, calculate the instantaneous amplitude and instantaneous frequency, calculate the mean of the residuals, construct a comprehensive load index, use linear extrapolation to calculate the reference priority trajectory, construct the objective function of the priority, and constrain the projection of the optimized priority trajectory to obtain the final priority vector. Specifically, collect game running data and preprocess it to generate a time series matrix, including: Use the APL interface to collect game running data and perform timestamp alignment, denoising, and normalization processing; The game operation data includes CPU occupancy, GPU occupancy, memory usage, frame rate, player input frequency, and network latency data; Sort the game running data in chronological order to generate a time series matrix.

[0024] Multi-source data fusion avoids the bias of single resource indicators in prediction and improves the accuracy of overall evaluation. Normalization processing helps data of different dimensions to be comparable in subsequent modeling, thereby improving model stability and convergence speed.

[0025] Furthermore, a comprehensive load index is constructed, a reference priority trajectory is calculated using linear extrapolation, an objective function of the priority is constructed, and the optimized priority trajectory is constrained and projected to obtain the final priority vector, including: The time series matrix is ​​used as the initial signal, and the local maximum and local minimum of the initial signal are identified using the numerical comparison method. The upper and lower envelopes of the local maximum and local minimum are calculated using cubic spline interpolation, and the mean envelope is calculated to extract the candidate IMF. The formula is: , , in is the mean envelope of the kth iteration at time t, which represents the average value of the upper envelope and the lower envelope, and are the upper and lower envelopes of the kth iteration at time t, respectively, is the current signal at time t, kth iteration, is the candidate IMF at time t, kth iteration; The numerical comparison method is used to verify the candidate IMF, and the number of extreme points and zero crossing points of the candidate IMF is counted. The statistical distribution method is used to set a fixed threshold and set the verification conditions. The formula is: , , in is the number of extreme points of the candidate IMF, is the number of zero-crossing points of the candidate IMF, is a fixed threshold, is the standard deviation of the current signal, is the absolute mean of the mean envelope; When the candidate IMF meets the verification conditions, the candidate IMF is marked as a valid IMF, otherwise the signal value is used to update the current signal (the candidate IMF is assigned to the current signal); The update signal is calculated using signal subtraction, as follows: , in is the update signal of the k+1th iteration at time t, is the effective IMF of the kth iteration of resource i at time t; Use the standard deviation threshold comparison to set the inspection condition. The formula is: , in is the standard deviation, is the input time series of resource i; Filter the update signals that meet the inspection conditions and mark them as residuals; Perform Hilbert transform on the effective IMF to calculate the instantaneous amplitude and instantaneous frequency. The formula is: , , in is the Hilbert transform of the effective IMF, is the instantaneous amplitude at time t, is the instantaneous frequency at time t, and d is the mathematical differential operator; Calculate the mean of the residuals, construct the comprehensive load index, perform normalization, and generate the initial priority vector. The formula is: , in is the comprehensive load index of resource i at time t, K is the number of valid IMFs, which is the same as the total number of iterations k, is the residual weighting coefficient, which is set through experimental verification. is the residual mean of resource i; The reference priority trajectory is calculated using linear extrapolation as follows: , in Forecast time The reference priority trajectory, is the initial priority vector at time t, is the predicted time step of the priority trajectory, is the time step; The objective function of priority is constructed using the weighted sum of squares, and solved using quadratic programming to obtain the optimized priority trajectory. The formula is: , Where J is the priority objective function, balancing the priority tracking error and adjustment smoothness, For the prediction time , the initial priority vector of resource i, For the prediction time , the reference priority trajectory of resource i, is the control penalty coefficient of the priority, which is set using grid search, and T is the prediction time span of the priority; The optimized priority trajectory is constrained and projected to obtain the final priority vector, which is: , , , in is the final priority vector of resource i, and n is the number of game running data types.

[0026] EMD has a good ability to retain local features, so that subtle fluctuations in resource utilization over time can be restored. Compared with frequency domain methods such as FFT, EMD is more adaptable when processing non-stationary signals. By decoupling complex system behaviors into multiple IMF components, it is convenient to extract features separately and perform targeted optimization control. The verification process controls the physical reliability of each component through explicit formulas to ensure that the retained IMF is representative in terms of frequency and amplitude, effectively eliminating calculation errors and artifacts. The residual part retains system trends and low-frequency disturbance information, which is helpful for the overall evaluation of the system's "underlying load pressure". Compared with statistical indicators such as average frame rate, instantaneous frequency can reflect performance fluctuations such as sudden freezes and operation response delays. The instantaneous amplitude reveals the peak response capability of the system under high load. This feature Extraction can achieve more sensitive real-time state identification, provide a forward-looking basis for dynamic scheduling, and avoid the problem of a single resource bottleneck masking the overall performance through multi-source load fusion. The comprehensive load indicator is computable in real time, avoiding dependence on prediction models and improving response speed. The introduction of residuals ensures that trend changes are not ignored during optimization and enhances the system's global judgment ability. The linear extrapolation model has low computational complexity and is suitable for real-time deployment in high-frequency game running scenarios. The generation of reference trajectories avoids "blind optimization" and gives subsequent control strategies a clear direction, which helps to achieve stable scheduling rather than frequent mutations. The quadratic programming solution has stability and global convergence, and is suitable for deployment in real-time embedded environments. The allocation boundaries are explicitly controlled through linear constraints to avoid crashes or deadlocks due to over-allocation of resources in actual scheduling.

[0027] S2. Based on the time series matrix, use a sliding window to set the time window, perform a fast Fourier transform on each time window, obtain the spectrum of the time window, calculate the Shannon entropy of the spectrum, select the time window corresponding to the minimum Shannon entropy as the optimal window, predict resource demand in the future, define a potential function for each resource, calculate the total potential function, use the state equation to predict the system state, construct the resource allocation objective function, and obtain the optimized resource allocation result; Specifically, the Shannon entropy of the spectrum is calculated, and the time window corresponding to the minimum Shannon entropy is selected as the optimal window to predict resource requirements in the future, including: Based on the time series matrix, a sliding window is used to set the time window, and a fast Fourier transform is performed on each time window to obtain the spectrum of the time window. The formula is: , in For the i-th resource in the time window Next, discrete frequency The spectrum, is a discrete frequency index corresponding to the frequency point in the frequency domain. Shannon entropy needs to traverse the normalized energy of all frequency points to calculate the distribution concentration of the spectrum. is the length of the jth time window, l is the index of the time window, and o is the virtual unit; Use the Shannon entropy formula to calculate the Shannon entropy of the spectrum, and select the time window corresponding to the minimum Shannon entropy as the optimal window. The formula is: , , in For the i-th resource in the time window Shannon entropy under is the total energy; Perform fast Fourier transform on the optimal window to obtain the optimal spectrum, calculate the amplitude of the discrete frequency and the spectrum phase of the discrete frequency, the formula is: , , in is a discrete frequency The amplitude, is a discrete frequency The phase, For the i-th resource in the time window Next, discrete frequency The optimal spectrum of Based on the final priority vector, combined with the amplitude of the discrete frequency and the spectral phase of the discrete frequency, the resource demand in the future time is predicted and normalized. The formula is: , in For the i-th resource in the future time Forecast demand, is the final priority vector of resource i at time t, is the predicted time step of resource demand, is the number of time window points.

[0028] FFT can identify periodic or oscillatory characteristics in resource usage and reveal important patterns hidden in time domain noise. Through frequency domain analysis, it can effectively weaken the interference of occasional fluctuations on resource trend modeling and improve prediction stability. By selecting the window with the minimum Shannon entropy, it ensures that the information contained in the window is highly structured, centralized, and representative. A lower entropy value indicates a clearer spectral structure, which facilitates the subsequent extraction of stable phase and amplitude features and improves prediction accuracy. Compared with using only amplitude, adding phase information can better restore the trend and turning points of the original signal. The phase reflects the offset of the fluctuation and is suitable for modeling resource surge scenarios in a short period of time in the game. The normalized prediction value can be applied to system architectures and resource pools of different specifications, enhancing the portability of the algorithm.

[0029] Furthermore, a resource allocation objective function is constructed to obtain the optimized resource allocation results, including: Define the potential function for each resource, the formula is: , in is the potential function of the i-th resource, is the allocation of the i-th resource, obtained based on the game running data, is the hardware upper limit of the i-th resource, collected from manufacturer specifications, is the normalized resource requirement; Calculate the total potential function, the formula is: , in is the total potential function of the i-th resource; Use the state equation to predict the system state, the formula is: , in For time The system status, is the system status at time t, set based on game running data, is the resource allocation at time t, Y and B are the state transition matrices, which are set using empirical rules; Based on the predicted system status, the resource allocation objective function is constructed, and the formula is: , in is the resource allocation objective function, The forecast time horizon for resource allocation, Forecast time The system status, is the target frame rate, which is set based on the average frame rate. The control penalty coefficient for resource allocation is set using Bayesian optimization. Forecast time The total potential function of Use Newton's method to solve and obtain optimized resource allocation results, including CPU resources, memory resources, and rendering resources; The CPU resources indicate the number of CPU cores allocated to the game process, the memory resources indicate the memory capacity allocated to the game process, and the rendering resources indicate the resolution of the game rendering.

[0030] The potential function establishes a nonlinear trade-off model for resource usage, effectively reflecting the "pressure" of resource over-allocation and under-allocation systems. By optimizing the total potential function, a coordination relationship is automatically established between multiple resources to prevent local resource bottlenecks. Hardware specifications are introduced as parameters in the potential function design to make the optimization algorithm hardware adaptable. State equation modeling can simulate the system response process after resource adjustment, capture the "inertia" characteristics of the system, and incorporate the frame rate response into the system state, which is conducive to joint optimization with perception goals. The frame rate deviation indicator is introduced into the objective function to make the optimization process oriented towards user experience goals, rather than just minimizing computing load. The prediction results drive the optimization process, and the optimization feedback feeds back to the prediction state to form a closed-loop feedback system. The Newton method is used to iteratively solve the problem to ensure the global optimality of resource allocation rather than local greedy selection.

[0031] S3, use the PID controller to convert it into control instructions, transmit and execute them, and build a visual interface to display the control instructions; Specifically, the PID controller is used to convert the control instructions, transmit and execute them, including: Use the difference method to calculate the error between the optimized resource allocation result and the current resource allocation result; A PID controller is used to convert the error into a control instruction, and a network communication protocol is used to transmit and execute the control instruction.

[0032] The control instructions generated by the PID controller need to be transmitted to the system resource manager through the network communication protocol.

[0033] The differential method is used to quickly capture the smallest granularity of resource offsets, improving the real-time and accuracy of controller response. Resource usage fluctuates frequently during game operation. The differential method has the characteristics of high sensitivity and good stability and can adapt to this highly dynamic environment. The proportional term responds to the current error, the integral term suppresses long-term deviations, and the differential term predicts trend changes. The three work together to improve resource control accuracy and stability. In high-frequency interactive games, PID responds sensitively. In slow-heating resource change games, the integral term effectively corrects long-term imbalances. The use of lightweight protocols (such as UDP+control confirmation mechanism) can significantly reduce data interaction delays and improve resource adjustment responsiveness.

[0034] Furthermore, a visual interface is constructed to display control instructions, including: Use the visualization tool Matplotlib to build a visualization interface to display control instructions and game running data in real time; Users who have passed real-name verification are allowed to view it.

[0035] Observe control behaviors and effects through real-time charts, providing a basis for algorithm tuning and system diagnosis. Only users who have passed the real-name verification mechanism are allowed to access the graphical interface, enhancing system data security.

[0036] This embodiment also provides a game resource optimization system based on artificial intelligence, including: The priority collection module is used to collect and preprocess game operation data, generate a time series matrix, perform empirical mode decomposition on the time series matrix, obtain effective IMF and residuals, perform Hilbert transform on the effective IMF, calculate the instantaneous amplitude and instantaneous frequency, calculate the mean of the residuals, construct a comprehensive load index, use linear extrapolation to calculate the reference priority trajectory, construct the priority objective function, and constrain the projection of the optimized priority trajectory to obtain the final priority vector; The demand allocation module is used to set the time window using a sliding window based on the time series matrix, perform a fast Fourier transform on each time window, obtain the spectrum of the time window, calculate the Shannon entropy of the spectrum, select the time window corresponding to the minimum Shannon entropy as the optimal window, predict resource demand in the future, define a potential function for each resource, calculate the total potential function, use the state equation to predict the system state, construct the resource allocation objective function, and obtain the optimized resource allocation result; The instruction visualization module is used to convert the PID controller into control instructions, transmit and execute them, and build a visual interface to display the control instructions.

[0037] This embodiment also provides a computer device suitable for the case of an artificial intelligence-based game resource optimization method, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute computer-executable instructions to implement the artificial intelligence-based game resource optimization method proposed in the above embodiment.

[0038] The computer device may be a terminal, comprising a processor, memory, a communication interface, a display, and an input device connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores an operating system and computer programs. The internal memory provides an environment for the operating system and computer programs stored in the non-volatile storage media. The communication interface of the computer device is used to communicate with external terminals via wired or wireless communication. Wireless communication may be achieved via Wi-Fi, a carrier network, NFC (near-field communication), or other technologies. The display of the computer device may be a liquid crystal display or an electronic ink display. The input device may be a touchscreen overlay on the display, buttons, a trackball, or a touchpad on the computer device housing, or an external keyboard, touchpad, or mouse.

[0039] This embodiment also provides a storage medium having a computer program stored thereon, which, when executed by a processor, implements the game resource optimization method based on artificial intelligence proposed in the above embodiment; the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

[0040] In summary, the present invention enhances the support capabilities of priority prediction and resource allocation for time sensitivity through a signal decomposition method that combines EMD and Hilbert transform, introduces comprehensive load indicators and reference trajectory prediction, and improves the performance balance and player experience continuity of games under multi-task competition.

[0041] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method for optimizing game resources based on artificial intelligence, characterized by: include, Collect and preprocess game operation data to generate a time series matrix. Perform empirical mode decomposition on the time series matrix to obtain effective IMF and residuals. Perform Hilbert transform on the effective IMF, calculate the instantaneous amplitude and instantaneous frequency, calculate the mean of the residuals, construct a comprehensive load index, use linear extrapolation to calculate the reference priority trajectory, construct the priority objective function, and constrain the projection of the optimized priority trajectory to obtain the final priority vector. Based on the time series matrix, a sliding window is used to set the time window. A fast Fourier transform is performed on each time window to obtain the spectrum of the time window. The Shannon entropy of the spectrum is calculated. The time window corresponding to the minimum Shannon entropy is selected as the optimal window to predict resource demand in the future. A potential function is defined for each resource, and the total potential function is calculated. The state equation is used to predict the system state, and the resource allocation objective function is constructed to obtain the optimized resource allocation result. Use the PID controller to convert it into control instructions, transmit and execute them, and build a visual interface to display the control instructions.

2. The game resource optimization method based on artificial intelligence according to claim 1, characterized in that: The method of constructing a comprehensive load index, using linear extrapolation to calculate a reference priority trajectory, building a priority objective function, and performing constraint projection on the optimized priority trajectory to obtain a final priority vector includes: The time series matrix is ​​used as the initial signal, and the local maximum and local minimum of the initial signal are identified using the numerical comparison method. The upper and lower envelopes of the local maximum and local minimum are calculated using cubic spline interpolation, and the mean envelope is calculated to extract the candidate IMF. Use the numerical comparison method to verify the candidate IMF, count the number of extreme points and zero crossing points of the candidate IMF, use the statistical distribution method to set a fixed threshold, set the verification conditions, and when the candidate IMF meets the verification conditions, mark the candidate IMF as a valid IMF, otherwise use signal assignment to update the current signal; Use signal subtraction to calculate the update signal, use the standard deviation threshold comparison to set the inspection condition, filter the update signal that meets the inspection condition, and mark it as residual; Perform Hilbert transform on the effective IMF, calculate the instantaneous amplitude and instantaneous frequency, calculate the mean of the residual, construct the comprehensive load index, and perform normalization to generate the initial priority vector; The reference priority trajectory is calculated using linear extrapolation, and the objective function of the priority is constructed using the weighted sum of squares. The optimized priority trajectory is solved using quadratic programming, and the optimized priority trajectory is constrained and projected to obtain the final priority vector.

3. The game resource optimization method based on artificial intelligence according to claim 2, characterized in that: The step of calculating the Shannon entropy of the spectrum, selecting a time window corresponding to the minimum Shannon entropy as the optimal window, and predicting resource requirements in the future includes: Based on the time series matrix, a sliding window is used to set the time window. Fast Fourier transform is performed on each time window to obtain the spectrum of the time window. The Shannon entropy of the spectrum is calculated using the Shannon entropy formula. The time window corresponding to the minimum Shannon entropy is selected as the optimal window. Fast Fourier transform is performed on the optimal window to obtain the optimal spectrum. The amplitude of the discrete frequency and the spectral phase of the discrete frequency are calculated. Based on the final priority vector, the resource demand in the future time is predicted and normalized in combination with the amplitude of the discrete frequency and the spectral phase of the discrete frequency.

4. The method for optimizing game resources based on artificial intelligence according to claim 3, wherein: The constructing of the resource allocation objective function to obtain the optimized resource allocation result includes: Define potential function for each resource and calculate the total potential function; The state equation is used to predict the system state. Based on the predicted system state, the resource allocation objective function is constructed and solved using Newton's method to obtain the optimized resource allocation result.

5. The game resource optimization method based on artificial intelligence according to claim 4, characterized in that: The method of converting the control instructions into control instructions using a PID controller, and transmitting and executing the control instructions includes: Use the difference method to calculate the error between the optimized resource allocation result and the current resource allocation result; A PID controller is used to convert the error into a control instruction, and a network communication protocol is used to transmit and execute the control instruction.

6. The game resource optimization method based on artificial intelligence according to claim 5, characterized in that: The construction of a visual interface to display control instructions includes: Use the visualization tool Matplotlib to build a visualization interface to display control instructions and game running data in real time; Users who have passed real-name verification are allowed to view it.

7. The game resource optimization method based on artificial intelligence according to claim 6, characterized in that: The collecting of game operation data and preprocessing to generate a time series matrix includes: Use the APL interface to collect game running data and perform timestamp alignment, denoising, and normalization processing; The game operation data includes CPU occupancy, GPU occupancy, memory usage, frame rate, player input frequency, and network latency data; Sort the game running data in chronological order to generate a time series matrix.

8. An artificial intelligence-based game resource optimization system, based on the artificial intelligence-based game resource optimization method according to any one of claims 1 to 7, characterized in that: include, The priority collection module is used to collect and preprocess game operation data, generate a time series matrix, perform empirical mode decomposition on the time series matrix, obtain effective IMF and residuals, perform Hilbert transform on the effective IMF, calculate the instantaneous amplitude and instantaneous frequency, calculate the mean of the residuals, construct a comprehensive load index, use linear extrapolation to calculate the reference priority trajectory, construct the priority objective function, and constrain the projection of the optimized priority trajectory to obtain the final priority vector; The demand allocation module is used to set the time window using a sliding window based on the time series matrix, perform a fast Fourier transform on each time window, obtain the spectrum of the time window, calculate the Shannon entropy of the spectrum, select the time window corresponding to the minimum Shannon entropy as the optimal window, predict resource demand in the future, define a potential function for each resource, calculate the total potential function, use the state equation to predict the system state, construct the resource allocation objective function, and obtain the optimized resource allocation result; The instruction visualization module is used to convert the PID controller into control instructions, transmit and execute them, and build a visual interface to display the control instructions.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the artificial intelligence-based game resource optimization method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the artificial intelligence-based game resource optimization method according to any one of claims 1 to 7 are implemented.

Images