Game update method and system based on game development data adjustment, and storage medium

Abstract

The present invention relates to the field of computer technology, and in particular to a game update method and system, and a storage medium based on game development data adjustment. The method obtains fusion data; combines preset key indicators to form a state vector, generates candidate solutions using a deep reinforcement learning model, uses generative adversarial network training to output innovative content, and then forms an integrated update solution through nonlinear weighted fusion; deploys the integrated update solution on a digital twin simulation platform, collects running data through virtual player model simulation testing, generates a verification solution through statistical analysis and multi-objective optimization, and automatically triggers adaptive parameter adjustment based on real-time feedback to form adjustment data; integrates the adjustment data with the verification solution data to generate a full update solution, and implements rolling deployment through an automated deployment platform to achieve closed-loop feedback periodic dynamic iterative updates. The present invention realizes the intelligent, automated, and dynamic optimization of game updates, thereby improving system stability and user experience.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a game update method and system based on game development data adjustment, and a storage medium. Background Art

[0002] As game content becomes increasingly rich and functions continue to expand, game updates have become a key means to enhance user experience and extend product life cycles. However, existing technologies mainly rely on manual or semi-automatic adjustments to the overall game code (Chinese invention patent, publication number: CN119002969B), and achieve updates through code division, manual identification and replacement. Its core means include code segmentation, error log analysis, static code replacement, etc. These technical means have many shortcomings: first, the amount of code is large and complex, resulting in a cumbersome adjustment process and a prolonged development cycle; second, there is a lot of manual intervention, making it difficult to achieve real-time dynamic updates, and when the data volume is large and changes frequently, omissions and errors are prone to occur; third, traditional methods rely on fixed repair strategies for bug handling and lack a complete automated response mechanism, which may result in some bugs not being repaired in a timely manner, thereby affecting the overall stability of the game.

[0003] Overall, existing technologies have obvious defects in update efficiency, automation level and dynamic response capabilities. There is an urgent need for an update method that uses real-time data-driven and intelligent decision-making closed loop to improve game update efficiency and dynamic response capabilities. Summary of the Invention

[0004] To address the numerous issues with the aforementioned prior art, the present invention provides a game update method, system, and storage medium based on game development data adjustments. This method collects and preprocesses multi-source game data to form fused data, then combines it with preset key indicators to form a state vector. Deep reinforcement learning is used to generate candidate solutions, while generative adversarial network training is employed to output innovative content. Nonlinear weighted fusion is used to generate integrated update solutions. This method, tested on a digital twin simulation platform using a virtual player model, automatically triggers adaptive parameter adjustments based on real-time feedback, ultimately achieving dynamic iteration of updates within an intelligent decision-making closed loop, effectively improving game update efficiency and dynamic responsiveness.

[0005] A game update method based on game development data adjustment includes the following steps:

[0006] Collect player behavior data, system performance data, community sentiment data, and bug feedback data from the game environment. After pre-processing all the data, perform feature extraction and fusion processing to form fused data;

[0007] A state vector is formed based on the fused data and preset key indicators, a deep reinforcement learning model is used to generate candidate solutions, and a generative adversarial network is used to train historical scene data and current game style data to generate innovative content. The candidate solutions are then fused with the innovative content through a nonlinear weighted fusion method to form an integrated update solution;

[0008] Deploy the integrated update solution on a digital twin simulation platform, conduct simulation tests on the integrated update solution using a virtual player model, collect operational data, and generate a verification solution using statistical analysis and multi-objective optimization. At the same time, adaptive parameter adjustments are automatically triggered based on real-time feedback to generate adjustment data.

[0009] The adjustment data and the verification scheme data are integrated to generate a full update scheme, the full update scheme is rolled out through an automated deployment platform, and closed-loop feedback is formed using real-time monitoring data to periodically trigger dynamic iterative updates.

[0010] Preferably, the preset key indicators are player retention rate, system response index and economic balance index, and the fusion data and the key indicators are combined with predetermined weights to form a state vector.

[0011] Preferably, the deep reinforcement learning model adopts a multi-task transfer learning strategy, using the state vector as input to generate candidate solutions, wherein the candidate solutions clearly specify level difficulty, currency circulation and character balance parameters.

[0012] Preferably, the generative adversarial network adopts a network structure composed of several convolutional layers, trains historical scene data and current game style data to generate innovative content output, and enhances the innovative content output using feature vectors extracted by variational autoencoders, and fuses the candidate solutions with the enhanced innovative content through a nonlinear weighted fusion method to form integrated updated solution data.

[0013] Preferably, the method of fusing the candidate solutions with the innovative content by a nonlinear weighted fusion method to form an integrated update solution specifically includes:

[0014] By constructing an interaction matrix and applying a nonlinear activation function to fuse the features of candidate solutions and innovative content, an integrated update solution is formed.

[0015] Preferably, the integration and update solution is deployed on a digital twin simulation platform, and simulation testing of the integration and update solution is performed through a virtual player model to collect operation data, specifically including:

[0016] The integrated update solution is implemented and deployed on the digital twin simulation platform using containerization technology, and the integrated update solution is simulated and tested based on the virtual player model of Agent-Based Modeling, so as to collect system load, player behavior, economic balance and BUG generation rate data as operation data.

[0017] Preferably, the generation of a verification plan using statistical analysis and multi-objective optimization specifically includes:

[0018] The running data is processed through standard statistical analysis and parallel multi-objective optimization based on a swarm search strategy to generate a verification plan.

[0019] Preferably, the adjustment data and the verification scheme data are integrated according to a predetermined data format to form a full update scheme, and the full update scheme is implemented in a rolling deployment manner on an automated deployment platform. The rolling deployment method includes first implementing the update on at least 10% of the production nodes, and gradually expanding to all production nodes after confirming that the operation is stable, and real-time monitoring and collection of player behavior data, system performance data, community sentiment data and BUG feedback data to form a closed-loop feedback, and the closed-loop feedback triggers dynamic iterative updates at set time intervals.

[0020] A game update system based on game development data adjustment, for implementing the game update method based on game development data adjustment, the system comprising:

[0021] The collection and preprocessing module is used to collect player behavior data, system performance data, community sentiment data and bug feedback data in the game environment, and perform noise filtering, time correction, format standardization, feature extraction and fusion processing on the data to form fused data;

[0022] A core generation module, which is used to construct a state vector based on the fused data and preset key indicators, use a deep reinforcement learning model to generate candidate solutions, and use a generative adversarial network to train historical scene data and current game style data to generate innovative content, and then use a nonlinear weighted fusion method to fuse the candidate solutions with the innovative content to form an integrated update solution;

[0023] A simulation adjustment module, which is used to deploy the integrated update solution on the digital twin simulation platform, simulate and test the integrated update solution using a virtual player model, collect operational data, generate a verification solution using statistical analysis and multi-objective optimization, and automatically trigger adaptive parameter adjustments based on real-time feedback to generate adjustment data;

[0024] A deployment feedback module is used to integrate the adjustment data and verification plan data according to a predetermined data format to generate a full update plan, implement rolling deployment of the full update plan through an automated deployment platform, and use real-time monitoring data to form a closed-loop feedback to periodically trigger dynamic iterative updates.

[0025] A storage medium stores a computer program, which, when executed by a processor, implements the steps of the game update method adjusted based on game development data.

[0026] Compared with the prior art, the advantages and beneficial effects of the present invention are:

[0027] The present invention realizes the automatic generation of candidate update schemes based on state vectors by introducing deep reinforcement learning technology;

[0028] This invention uses generative adversarial network technology to achieve innovative content training for historical scenarios and current game styles, thereby enriching the creative dimension of the update scheme;

[0029] The present invention uses nonlinear weighted fusion technology to achieve deep feature integration of candidate solutions and innovative content, thereby forming an integrated update solution with high decision-making accuracy and adaptive adjustment capabilities;

[0030] The present invention realizes simulation testing and real-time feedback closed-loop control by utilizing a digital twin simulation platform and a virtual player model, thereby automatically triggering adaptive parameter adjustment and forming a dynamic iterative update effect.

[0031] The comprehensive application of the above technical means has solved the defects of existing technologies such as manual updates, large code adjustments, imperfect BUG handling and insufficient real-time performance. It not only realizes the dynamic iteration of updates in the intelligent decision-making closed loop, but also effectively improves the game update efficiency and dynamic response capabilities, further improving the overall game performance and user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 Schematic diagram of the process of the present invention;

[0033] Figure 2 A flow chart of the integrated update solution is formed in the present invention;

[0034] Figure 3 Schematic diagram of simulation test and adaptive adjustment in the present invention;

[0035] Figure 4 It is a structural block diagram of the system of the present invention. DETAILED DESCRIPTION

[0036] Hereinafter, embodiments of the present disclosure will be described with reference to the accompanying drawings. However, it should be understood that these descriptions are merely exemplary and are not intended to limit the scope of the present disclosure. In the following detailed description, for ease of explanation, many specific details are set forth to provide a comprehensive understanding of the embodiments of the present disclosure. However, it is apparent that one or more embodiments may also be implemented without these specific details. In addition, in the following description, descriptions of well-known structures and technologies are omitted to avoid unnecessary confusion of the concepts of the present disclosure.

[0037] The terms used herein are only for describing specific embodiments and are not intended to limit the present disclosure. The terms "comprise," "include," etc. used herein indicate the presence of the features, steps, operations, and / or components, but do not exclude the presence or addition of one or more other features, steps, operations, or components.

[0038] All terms used herein (including technical and scientific terms) have the meanings commonly understood by those skilled in the art unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning consistent with the context of this specification and should not be interpreted in an idealized or overly rigid manner.

[0039] like Figure 1 As shown, a game update method based on game development data adjustment includes the following steps:

[0040] Collect player behavior data, system performance data, community sentiment data, and bug feedback data from the game environment. After pre-processing all the data, perform feature extraction and fusion processing to form fused data;

[0041] The present invention aims to capture the game running status in real time and provide reliable data support for subsequent intelligent update decisions by implementing a data-driven update mechanism in the game development environment. To this end, the system first collects player behavior data, system performance data, community sentiment data and BUG feedback data in the game environment. Specifically, player behavior data includes the click records, movement trajectories, operation paths and stay time generated by players during the game; system performance data covers server response time, memory usage, CPU utilization and network latency, etc.; community sentiment data mainly comes from game forums, social media and internal feedback, and its content involves player comments, emotion identification and related multimedia information; BUG feedback data captures system anomalies and fault logs through a preset error reporting mechanism.

[0042] All types of collected data are transmitted in a unified format through standardized interfaces, and preliminary processing is performed on player behavior data, system performance data, community sentiment data, and bug feedback data collected from the game environment. Specifically, for player behavior data, a time alignment algorithm is used, combined with a time synchronization protocol, to convert data from different time zones into a unified standard time format to ensure the consistency of data timestamps; for system performance data, an anomaly detection method based on statistics (such as mean and standard deviation) is used to remove outliers and smooth the data, and at the same time, each monitoring indicator is normalized to ensure that the numerical data is compared on a unified scale; for community sentiment data, a pre-trained natural language processing model is used to segment text data, score sentiment, and extract keywords, and normalize the sentiment scores to achieve numerical consistency between different text data; for bug feedback data, the system parses the fixed format of the error log, extracts key attributes (such as error code, error description, and timestamp), and converts them into a standard format compatible with other data.

[0043] After preprocessing, the data is stored in a standard format, and high-dimensional feature representations are extracted using machine learning algorithms (such as convolutional neural networks, recurrent neural networks, or Transformer models). Finally, the features from each source are weighted and fused according to predetermined rules to generate a unified fused data vector. This process effectively unifies the data in terms of time and value, providing a reliable and unified data foundation for subsequent state vector construction and intelligent decision-making.

[0044] After the preprocessing, the data undergoes feature extraction. The system utilizes convolutional neural networks (CNNs), recurrent neural networks (RNNs), and Transformer models to extract high-dimensional feature representations from each data source. After normalization, these extracted feature representations are fused using a Transformer-based multi-head attention mechanism, dynamically adjusting the weights of the features from each data source to form a fused data set that comprehensively describes the game's runtime status. This fused data preserves temporal information while incorporating multimodal features such as text, images, and numerical values, comprehensively reflecting player behavior, system operation, user emotions, and fault conditions. The entire data collection, preprocessing, and fusion process ensures data consistency and real-time performance, seamlessly integrating with subsequent modules through standardized data interfaces (e.g., RESTful APIs), providing a solid foundation for intelligent updates based on game development data. This overall process enables the system to acquire large amounts of multi-dimensional data in a short period of time, and efficiently and accurately preprocesses and fuses features. This lays the foundation for subsequent state assessment and intelligent decision-making, while also ensuring real-time responsiveness and dynamic updates.

[0045] like Figure 2 As shown, a state vector is formed based on the fusion data and preset key indicators, a deep reinforcement learning model is used to generate candidate solutions, and a generative adversarial network is used to train historical scene data and current game style data to generate innovative content. The candidate solutions are fused with the innovative content through a nonlinear weighted fusion method to form an integrated update solution.

[0046] The present invention is based on an update method for adjusting game development data, and aims to achieve intelligent dynamic updates of the game system through data-driven. The overall solution first collects information from multiple data sources in the game environment. These data sources include player behavior data, system performance data, community sentiment data, and BUG feedback data. After the collected data undergoes noise filtering, time correction, format standardization, and feature extraction processing in the pre-processing module, their respective feature representations are generated and integrated into a unified fused data vector through a multi-head attention fusion algorithm. The fused data vector fully reflects the information of each dimension in the game operation status, including the interactive behavior of players in the game, as well as system operation indicators and player feedback emotions.

[0047] On this basis, the system combines the fused data with preset key indicators (defined in the preferred embodiment as player retention rate, system response index, and economic balance index) according to predetermined weights to form a state vector. The state vector serves as the input of the deep reinforcement learning model and is an important data expression describing the overall state of the current game. Based on the constructed state vector, the model generates candidate update plans using deep reinforcement learning technology. The candidate plans mainly provide adjustment suggestions for key parameters in the game, such as changes in level difficulty, regulation of currency circulation, and optimization of character balance parameters.

[0048] At the same time, the system uses a generative adversarial network (GAN) to train historical scene data and current game style data to generate content output with innovative features. Here, the historical scene data provides visual information about previous scenes in the game, while the current game style data reflects the design style of the current game. By jointly training these two types of data, the innovative content output by the generative adversarial network can capture and showcase new trends in the game's visuals and emotions. The generated innovative content and candidate update solutions are then integrated using a nonlinear weighted fusion method. This fusion method deeply fuses the feature information of the candidate solutions with the innovative content by constructing an interaction matrix and applying a nonlinear activation function, thereby forming an integrated update solution that integrates intelligent decision-making and creative output.

[0049] The integrated update solution combines the innovative advantages of data-driven intelligent decision-making and visual creativity, so that game updates are not only scientifically adjusted in numerical parameters, but also achieve breakthroughs in game content and experience. The entire process, from data collection, preprocessing, feature fusion, state vector composition, deep reinforcement learning to generate candidate solutions, to the generation of adversarial networks to output innovative content, and then to the nonlinear weighted fusion of candidate solutions and innovative content, forms a complete data-driven closed-loop update link. Through real-time data collection and efficient fusion, the present invention provides a game system with an intelligent update mechanism based on real-time data feedback, which helps to improve the user experience and system stability of the game, and at the same time provides a solid foundation for subsequent automated deployment and dynamic iterative updates. In practical applications, this update method can effectively reduce decision-making deviations caused by data delays and information asymmetry, and realize adaptive optimization of the game system under multi-dimensional indicators, with strong operability and promotion value.

[0050] Preferably, the preset key indicators are player retention rate, system response index and economic balance index, and the fusion data and the key indicators are combined with predetermined weights to form a state vector.

[0051] In the present invention, the preset key indicators are preferably set as player retention rate, system response index, and economic balance index, and these indicators are combined with the fusion data according to predetermined weights to form a state vector. The player retention rate reflects the attractiveness of the game to players and the degree of their continued participation. Its calculation formula is generally expressed as:

[0052]

[0053] Among them, R represents the player retention rate, N active Indicates the number of active players in the set observation period, N init This metric represents the number of players who initially register or participate in the game for the first time. This metric dynamically reflects the player base's mobility through daily or weekly statistics, providing important feedback for game operations.

[0054] The system response index primarily reflects the stability and performance of the server and network when facing high concurrent requests. This metric not only covers response time, but also encompasses multiple parameters such as server CPU utilization, memory usage, and network latency. Response time can be calculated by calculating the time difference between sending a request and receiving a response. The calculation expression is:

[0055] T=t resp -t req

[0056] Where T represents the response time, t resp and t reqThe time it takes to send a request and receive a response, respectively. By monitoring and collecting statistics on these parameters in real time, the system can provide timely warnings when performance fluctuations occur and assist in decision-making updates.

[0057] The economic balance indicator monitors the health of the in-game virtual economy. Its core objective is to assess the balance between virtual currency circulation and player spending. Specifically, the system calculates the economic balance ratio by analyzing the total supply, circulation velocity, and player spending data of the in-game virtual currency. It then sets appropriate thresholds to trigger economic regulation strategies when abnormal values ​​are detected. These key indicators are normalized to ensure consistency on a numerical scale and then weighted and integrated using pre-defined weights.

[0058] After normalizing the fusion data vector F and the key indicator vector K, the system uses the weighted fusion formula:

[0059] S=w f ·F′+w k ·K

[0060] Where F′ and K′ are the normalized fusion data and key indicator vector respectively, w f With w k is a predetermined weight coefficient, satisfying w f +w k =1 condition.

[0061] In practice, the weight w f With w k The initial values ​​of can be set to 0.7 and 0.3, but through historical data backtesting, cross-validation, and Bayesian optimization, the system dynamically adjusts these weights, ensuring that the state vector more accurately reflects the current operational status of the game system. By precisely selecting and calculating pre-set key indicators, the system ensures that the collection, normalization, and weighting of each indicator follow clear technical implementation steps. This fully captures the changes in operational data within the game and provides scientific and accurate input for the subsequent deep reinforcement learning model. The data interfaces and message transmission mechanisms in this process utilize standardized protocols, ensuring seamless data transfer between processing modules and enabling real-time data feedback. Through rigorous quantification and optimized design of key indicators, the system effectively improves the accuracy and reliability of data fusion when constructing the state vector, providing solid data support for the intelligent generation of game update plans and serving as a warning and regulatory mechanism in actual operations, ensuring that the game system achieves its intended goals in terms of user experience, performance stability, and economic health.

[0062] Preferably, the deep reinforcement learning model adopts a multi-task transfer learning strategy, using the state vector as input to generate candidate solutions, wherein the candidate solutions clearly specify level difficulty, currency circulation and character balance parameters.

[0063] In this invention, a deep reinforcement learning model serves as the key technology for generating candidate solutions, and its specific implementation utilizes a multi-task transfer learning strategy. The system first uses the previously constructed state vector S as input. This state vector integrates fused data and key indicator information to comprehensively reflect the game's performance. Based on this, the deep reinforcement learning model generates a series of candidate update solutions through methods such as policy evaluation and value function approximation. These candidate solutions primarily involve adjustments to level difficulty, currency circulation, and character balance parameters within the game.

[0064] In practical implementations, deep reinforcement learning models employ deep Q-networks (DQNs) or actor-critic structures. These networks typically include multiple layers of fully connected networks and convolutional layers to process high-dimensional data in state vectors. The training objective of the model is to minimize a pre-set loss function, L, which can be expressed as:

[0065]

[0066] Where Q(S, a; θ) is the expected reward of the current network taking action a in state S under parameters θ, γ is the discount factor, r represents the immediate reward, S′ is the next state, and θ′ is the target network parameter. This formula, through continuous iterative optimization, enables the model to accurately predict the long-term benefits of various update scenarios in complex state spaces.

[0067] The adoption of a multi-task transfer learning strategy aims to utilize the common information between different game subsystems to achieve cross-task knowledge sharing. For example, in the three tasks of level difficulty, currency circulation, and character balance, although the specific control targets are different, they are all affected by player behavior data and system performance data. By sharing some network layers, the model can transfer existing feature representations between tasks, thereby improving overall learning efficiency and decision-making accuracy. In an embodiment, the model can use a shared convolutional layer to extract common features, and then predict the parameters of each task separately in the task-specific layer. Specifically, during the training process, each task generates a separate loss function, and the system takes the weighted sum of the losses of each task as the overall loss, and jointly updates the shared layer weights through backpropagation to ensure the balanced performance of the model on different tasks.

[0068] The output of candidate solutions clearly specifies recommended values ​​for key game parameters, such as adjusting level difficulty to 1.1 times the original value, increasing or decreasing currency circulation by 15% or 10%, and fine-tuning character balance parameters based on player behavior distribution. These candidate solutions, optimized by a deep reinforcement learning model, provide a quantitative basis for decision-making regarding game updates. The entire process seamlessly connects input and output through standardized data interfaces, ensuring the model's high responsiveness and decision-making accuracy under real-time data updates. Experimental verification has demonstrated that this deep reinforcement learning model demonstrates high robustness and convergence speed in multi-task environments, generating targeted and actionable candidate solutions across diverse operational scenarios. This provides a solid data and strategic foundation for the overall generation of subsequent update plans.

[0069] Preferably, the generative adversarial network adopts a network structure composed of several convolutional layers, trains historical scene data and current game style data to generate innovative content output, and enhances the innovative content output using feature vectors extracted by variational autoencoders, and fuses the candidate solutions with the enhanced innovative content through a nonlinear weighted fusion method to form integrated updated solution data.

[0070] Generative adversarial networks (GANs) are used in the present invention to generate innovative content. Their design uses a network structure composed of multiple convolutional layers. By training historical scene data and current game style data, they output new content that can reflect the visual and emotional creativity of the game. Specifically, the GAN model includes two parts: a generator and a discriminator. The generator consists of several convolutional layers, deconvolutional layers, and normalization layers. Its purpose is to map random noise and conditional vectors to an image space that matches the game scene; the discriminator consists of multiple convolutional layers and fully connected layers, and is used to distinguish between generated images and real images. During the training process, the generator and the discriminator are continuously iterated through adversarial training. The goal of the generator is to make it impossible for the discriminator to distinguish between generated images and real images, thereby achieving the purpose of generating high-quality innovative content.

[0071] Specifically, the input of the generator includes a random noise vector z and the current game style condition vector c; the generator generates an image output G(z,c) through multi-layer convolution operations. The discriminator receives the real historical scene image x or the generated image G(z,c) as input and calculates the discrimination probability through multi-layer convolution and activation functions. The training goal is to minimize the generator loss function L G and maximize the discriminator loss function L D , its standard adversarial training objective can be expressed as:

[0072]

[0073] Among them, p data(x) represents the distribution of real image data, p(z) represents the distribution of random noise, and p(c) represents the distribution of conditional vectors. Through this training process, the generator is able to generate new images that match the characteristics of the game scene and the current game style.

[0074] In order to further integrate the innovative content of the generative adversarial network with the candidate solutions generated by deep reinforcement learning, the present invention uses a variational autoencoder (VAE) to extract innovative features. The VAE model compresses the input image into a low-dimensional latent vector z through the encoder, and then reconstructs the image through the decoder, thereby learning the key features of the image. The extracted feature vectors and the innovative content output by GAN are spliced ​​or weighted fused through the feature fusion layer to form a unified description of the innovative content. The fusion process uses simple splicing or weighted averaging, supplemented by nonlinear activation functions (such as ReLU) to ensure that the fusion result has richer semantic expression.

[0075] In practical applications, the innovative content generated by generative adversarial networks not only provides visual novelty but also plays a crucial role in inspiring creativity and optimizing the user experience during game update decisions. In an embodiment, by combining generated images with candidate solution parameters, the system can achieve innovative updates to game scenarios and interaction methods, thereby enhancing the game's appeal and player experience. This entire process seamlessly connects with subsequent update decision modules through standardized data interfaces, and continuously optimizes model parameters using extensive training data to ensure the stability and diversity of generated content across diverse scenarios.

[0076] In summary, the present invention utilizes the combination of generative adversarial networks and variational autoencoders, specifically through a multi-layer convolutional network to train historical scenes and current game style data, generate innovative content, and integrate it with candidate update schemes through feature fusion methods, providing game updates with innovative content output that is both data-driven and creative, effectively improving the intelligence and foresight of game update schemes.

[0077] Preferably, the method of fusing the candidate solutions with the innovative content by a nonlinear weighted fusion method to form an integrated update solution specifically includes:

[0078] By constructing an interaction matrix and applying a nonlinear activation function to fuse the features of candidate solutions and innovative content, an integrated update solution is formed.

[0079] The nonlinear weighted fusion of candidate solutions and innovative content is one of the core technologies of the present invention. Its purpose is to effectively integrate the candidate solutions generated by the deep reinforcement learning model with the innovative content output by the generative adversarial network to form a comprehensive and creative integrated update solution. The fusion process uses the method of constructing an interaction matrix and applying a nonlinear activation function to deeply interact the candidate solutions with the various features of the innovative content. Specifically, let the candidate solution feature vector be C and the innovative content feature vector be I. The system first normalizes C and I to ensure that their numerical ranges are consistent. Subsequently, the interaction matrix M is constructed to capture the interaction relationship between C and I. The calculation formula of the interaction matrix can be expressed as:

[0080] ″T

[0081] M=C′×I

[0082] Among them, C′ and I′ are the normalized candidate solution feature vector and innovative content feature vector, respectively. × represents matrix multiplication and T represents matrix transposition. After constructing the interaction matrix, the system uses a nonlinear activation function (such as ReLU or Swish function) to activate the matrix M. The activated matrix M act Represents the nonlinear interaction results between candidate solutions and innovation content on specific characteristics.

[0083] Then, the system will M act The weighted fusion is performed with the original candidate solution feature vector C and the innovative content feature vector I. The fusion formula can be expressed as:

[0084] U=α·C+β·I+γ·M act

[0085] Here, U represents the final integrated update solution feature vector, and α, β, and γ are predetermined weight coefficients that satisfy the normalization condition of α + β + γ = 1. Through this weighted fusion, the system can integrate candidate solutions and innovative content while retaining their respective strengths while supplementing their complementary information through the interaction matrix, thus forming an integrated update solution that combines data-driven and creative innovation.

[0086] The implementation of this fusion process requires strict control of data format and numerical precision in all aspects of the collection, normalization, matrix construction, and weighted calculation of candidate solution and innovative content features. For example, in specific operations, if the dimension of the candidate solution feature vector C is n and the dimension of the innovative content feature vector I is m, then the size of the interaction matrix M is n×m; a nonlinear activation function is applied to each element to ensure that all values ​​are positive, thereby avoiding the influence of negative values ​​on the fusion effect. This process has been backtested and verified in the laboratory through a large amount of data using a preset algorithm, proving that the integrated update solution after fusion has high accuracy and stability in predicting game performance and optimizing user experience.

[0087] Ultimately, the integrated update solution is output as a high-dimensional feature vector and passed to subsequent deployment modules via a unified data interface, enabling its application in the actual game environment. The advantage of this nonlinear weighted fusion approach lies in its ability to not only simply superimpose candidate solutions and innovative content, but also capture the deeper connections between them through the construction of an interaction matrix, thereby generating more creative and targeted update solutions. This provides a strong technical foundation for the dynamic updating of game systems.

[0088] like Figure 3 As shown, the integration and update scheme is deployed on the digital twin simulation platform, and the integration and update scheme is simulated and tested through a virtual player model, and operation data is collected. Statistical analysis and multi-objective optimization are used to generate a verification scheme, and adaptive parameter adjustment is automatically triggered based on real-time feedback to generate adjustment data.

[0089] The present invention aims at how to use real-time data to drive intelligent decision-making in the process of game updates, and proposes a holistic method for deploying integrated update solutions on a digital twin simulation platform. The present invention first pre-processes and fuses the multi-source data generated in the game environment to form a highly comprehensive fused data vector. Then, the fused data and the preset key indicators together constitute a state vector, which serves as the basic input for the subsequent generation of intelligent update solutions. Next, a deep reinforcement learning model is used to generate candidate update solutions, and a generative adversarial network (GAN) is used to train historical scene data and current game style data to output content with innovative features. The candidate solutions and innovative content are fused through nonlinear weighted fusion to achieve deep interaction between the two at the feature level, thereby generating an integrated update solution that integrates data-driven decision-making and visual creativity.

[0090] Afterwards, the integrated update scheme is deployed in a digital twin simulation platform, which simulates the operating status in a real game environment through a virtual player model and collects operating data including system load, player behavior, economic balance, and bug generation rate. The collected operating data undergoes statistical analysis and parallel multi-objective optimization processing based on a group search strategy to generate a verification scheme for evaluating and verifying the effectiveness of the update scheme; at the same time, the system automatically triggers adaptive parameter adjustment based on real-time feedback data to form the final adjustment data. The overall method constructs a complete closed-loop mechanism through data collection, fusion, intelligent decision-making, simulation testing and feedback adaptive adjustment, which not only realizes the automatic generation and intelligent optimization of game update schemes, but also provides a scientific basis and real-time data support for the dynamic iterative update of the system. The present invention fully utilizes the advantages of multi-source data and uses advanced core technical means such as deep reinforcement learning, generative adversarial networks, nonlinear weighted fusion, and parallel multi-objective optimization to ensure the adaptability and accuracy of the update scheme in different operating environments, thereby effectively improving the overall performance and user experience of the game system. Through the present invention, game developers can capture key indicators of the game's running status in real time, implement dynamic regulation based on data feedback, avoid strategy failures caused by data delays and information asymmetry in traditional updates, and achieve continuous and stable operation of the system under high load and complex environments.

[0091] Preferably, the integration and update solution is deployed on a digital twin simulation platform, and simulation testing of the integration and update solution is performed through a virtual player model to collect operation data, specifically including:

[0092] The integrated update solution is implemented and deployed on the digital twin simulation platform using containerization technology, and the integrated update solution is simulated and tested based on the virtual player model of Agent-Based Modeling, so as to collect system load, player behavior, economic balance and BUG generation rate data as operation data.

[0093] In this invention, the integrated update solution, deployed on the digital twin simulation platform, employs containerization technology and a virtual player model based on Agent-Based Modeling to ensure consistency and high fidelity between the simulation environment and the actual production environment. Specifically, the digital twin simulation platform utilizes containerization technologies such as Docker to deploy various services and automatically allocates and balances resources through the Kubernetes scheduling mechanism, ensuring stable system operation under high concurrency and resource constraints. A virtual player model is built within the platform, which simulates player behavior based on the principles of Agent-Based Modeling (ABM). The virtual player model simulates the player's behavioral decision-making process, path selection, interaction patterns, and emotional fluctuations in the game, enabling simulation testing to fully capture the system's response in real-world game scenarios. During the simulation test, the system collects data including, but not limited to, server system load, CPU utilization, memory usage, network latency, as well as player behavior patterns, dwell time, task completion rate, and interaction frequency in the game. It also records the virtual currency circulation and bug generation rate in the economic system. This data is stored in a structured format and uploaded to the data processing center in real time.

[0094] To ensure the accuracy and real-time nature of data collection, the digital twin simulation platform utilizes high-frequency data collection technology and time-series database storage. Each virtual player model collects key data at preset intervals (e.g., every 30 seconds or every minute). Data transmission utilizes standard RESTful APIs and Kafka message queues for low-latency transmission. Containerization technology enables the platform to be scalable and fault-tolerant. If a node fails, its services are automatically taken over by other nodes, ensuring uninterrupted data collection. Furthermore, by configuring a load balancer, the system can dynamically adjust task allocation across nodes to ensure a balanced distribution of collected data. The implementation of containerization not only improves deployment efficiency but also reduces system maintenance costs, providing a solid foundation for subsequent simulation testing and data analysis. In this way, the digital twin simulation platform accurately simulates the game environment and generates high-fidelity data for subsequent verification of solutions, providing data support for the scientific nature and reliability of the overall update plan. The platform's implementation details include the image construction of each container, network configuration, log collection and monitoring mechanisms, as well as the behavioral parameter settings and simulation rule definitions of the virtual player model. These have undergone multiple verifications and debugging to ensure that they can truly reflect the game's operating status in actual applications, thereby providing accurate and timely data feedback for game update decisions.

[0095] Preferably, the generation of a verification plan using statistical analysis and multi-objective optimization specifically includes:

[0096] The running data is processed through standard statistical analysis and parallel multi-objective optimization based on a swarm search strategy to generate a verification plan.

[0097] In the present invention, the collected operating data is subjected to standard statistical analysis and parallel multi-objective optimization processing based on a group search strategy, and is used to generate a verification scheme, and further automatically triggers adaptive parameter adjustment based on real-time feedback to form adjustment data. Specifically, the operating data collected by the system from the digital twin platform after the simulation test includes server system load, player behavior data, economic balance indicators and BUG generation rate data. These data enter the data analysis module after preprocessing. The data analysis module first uses conventional statistical methods to calculate statistical quantities such as mean, variance, and standard deviation for each data, thereby obtaining the basic distribution characteristics of the data. In order to further extract deep-level information from the data, the system uses a group search strategy (such as particle swarm optimization or genetic algorithm) to perform data processing under a parallel multi-objective optimization framework. The parallel multi-objective optimization algorithm takes multiple indicators as targets and constructs an objective function. The objective function is in the form of:

[0098]

[0099] Among them, N represents the number of monitoring indicators, Error i represents the deviation of the i-th indicator, α i The system uses a swarm search strategy to iteratively screen a large number of candidate solutions, ultimately selecting the verification solution that minimizes the objective function. This verification solution reflects the performance of the current integrated update solution in the simulation environment, including predicted indicators such as system stability, player satisfaction, and economic balance.

[0100] After the verification scheme is generated, the system further makes adaptive adjustments to the verification scheme based on real-time feedback data. Real-time feedback data is continuously collected by the monitoring system during the deployment phase and transmitted to the adaptive adjustment module through the data interface. This module uses preset rules to automatically trigger parameter adjustment when it detects that a certain indicator exceeds the preset threshold. The parameter adjustment algorithm constructs a new objective function based on the feedback data and re-optimizes the update scheme parameters through the gradient descent method or evolutionary algorithm. For example, when the system monitors that the server response time T exceeds the preset upper limit T max When the adjustment module immediately calculates the response time deviation ΔT=TT max Based on this deviation, system resource parameters are reallocated, such as adjusting task scheduling priorities or reconfiguring load balancing weights. Throughout this process, the system records parameter changes before and after the adjustment and the corresponding feedback effects. Through a closed-loop feedback mechanism, the adjustment results are incorporated into the overall data analysis, achieving dynamic iterative updates.

[0101] In addition, during the data processing process, there are strict regulations on the algorithm parameters of statistical analysis and multi-objective optimization to ensure that each step of the operation has a clear technical implementation plan. All calculations are performed through standardized data interfaces, and the transmission and processing of data in each link adopt a unified data format (such as JSON format) to ensure data consistency and reliability. After the above process is processed, the system can generate a verification plan that reflects the current operating status and prediction effect. At the same time, it automatically optimizes the update plan based on real-time data feedback, and finally forms adjustment data to provide an accurate basis for the generation of a full update plan. This process not only improves the scientific nature of the update plan, but also verifies the efficiency and adaptability of the data-driven decision-making model in actual applications, laying a solid foundation for the intelligent and dynamic update of the game system.

[0102] After generating the integration and update plan, the present invention conducts simulation testing on the present invention through the digital twin simulation platform, and automatically triggers adaptive parameter adjustment based on real-time feedback data, thereby generating the final adjustment data. The deployment of the integration and update plan on the digital twin platform adopts a standardized operating process to ensure that the operation of the plan in the simulation environment is highly consistent with the real production environment. During the deployment process, the system uses containerization technology to package the integration and update plan into a standard container image, and realizes automatic scheduling and elastic expansion through Kubernetes. After the deployment is completed, the platform simulates the actual player behavior in the game through a virtual player model based on Agent-Based Modeling, covering various interactive scenarios, task execution and economic transaction processes. The parameter settings of the virtual player model are calibrated based on historical player behavior data and real-time collected behavioral statistics to ensure that its simulation effect is representative.

[0103] During the simulation test, the system collects various operating data in real time, including but not limited to server system load (such as CPU utilization, memory occupancy and network latency), player behavior (such as click-through rate, stay time and task completion rate), economic balance data (such as virtual currency circulation and consumption behavior) and BUG generation rate. The collected data is uniformly stored in the data storage module and transmitted to the data analysis module through the real-time data interface. The data analysis module uses standard statistical methods to calculate statistical quantities such as mean, variance, and percentile of the operating data, and uses a parallel multi-objective optimization algorithm based on a group search strategy to jointly optimize various indicators. By constructing a multi-objective optimization problem, the system sets the objective function to minimize the deviation of each indicator to ensure that the verification scheme can fully reflect the overall performance of the integrated update scheme in the simulation environment.

[0104] After the verification plan is generated, the system implements adaptive parameter adjustments based on real-time feedback data. The adaptive module automatically calculates the parameter adjustment amount by comparing the real-time data with the verification plan, and immediately triggers the update. For example, when real-time monitoring data shows that the player's task completion rate decreases and the bug generation rate increases, the adaptive module will analyze the system performance data, calculate the various deviation values, and re-optimize the parameter settings of the candidate update plan through the preset adjustment strategy. The adaptive parameter adjustment algorithm can use the gradient descent method to solve the objective function. Its specific calculation process includes:

[0105]

[0106] Where θ represents the parameter vector in the integrated update scheme, η is the learning rate, is the objective function gradient. In this way, the system can automatically correct inappropriate parameters in the update scheme during operation to generate the final adjustment data.

[0107] The entire simulation testing and adaptive adjustment process utilizes a unified data interface and message queue to achieve seamless data transmission and real-time processing, ensuring data consistency and reliability across all processing modules. The system not only pre-verifies the effectiveness of integrated update solutions in a simulated environment, but also implements closed-loop control through real-time monitoring feedback, enabling continuous optimization and iteration of update solutions, providing the gaming system with dynamic adjustment and intelligent optimization capabilities. Through this rigorous technical process, the resulting adjustment data fully reflects the real-time changes in the game's operating status, providing a scientific basis and reliable support for the final deployment of the full update solution.

[0108] The adjustment data and the verification scheme data are integrated according to a predetermined data format to generate a full update scheme, the full update scheme is rolled out through an automated deployment platform, and closed-loop feedback is formed using real-time monitoring data to periodically trigger dynamic iterative updates.

[0109] This invention proposes an innovative approach for leveraging multi-source data to build an intelligent decision-making closed loop during game updates. This approach, driven by data, integrates multidimensional data, which has undergone preliminary preprocessing, feature extraction, and fusion, with further processed adjustment data and verification solution data according to a predetermined data format to generate a full update solution. This full update solution is then rolled out via an automated deployment platform. A closed-loop feedback mechanism is established using real-time monitoring data to periodically trigger dynamic iterative updates, thereby enabling intelligent adaptive control and continuous optimization of the game system. This integrated approach implements closed-loop control throughout the entire process, from data acquisition, preprocessing, feature fusion, to state assessment, candidate solution generation, simulation verification, automatic adjustment, and full deployment. First, during the data acquisition phase, the system acquires real-time data on player behavior, system performance, community sentiment, and bug feedback from the game environment, ensuring that each data source reflects the game's operational status in both time and space. Next, after noise filtering, time correction, and format standardization are applied to this data by the preprocessing module, the feature extraction module extracts key information from each data source. Multi-head attention fusion technology is then used to integrate these data sources into a unified, fused data set. This fused data serves as the basis for the subsequent construction of state vectors and the generation of intelligent update plans. Its characteristics include user behavior patterns, system load conditions, emotional tendencies, and failure frequencies.

[0110] On this basis, the system combines the fused data with pre-set key indicators according to predetermined weights to form a state vector. This state vector reflects the overall operational status of the game and provides high-quality input for the deep reinforcement learning model to generate candidate update solutions. During the candidate solution generation process, the system utilizes a deep reinforcement learning model and a generative adversarial network to generate candidate solutions and innovative content, respectively. It then integrates these two using a nonlinear weighted fusion method to form an integrated update solution that combines numerical control and creative expression. This integrated update solution is then deployed on a digital twin simulation platform. This platform utilizes containerization technology and virtual player models to simulate the real-world game environment and collects operational data, including system load, player behavior, economic balance, and bug generation rates. After statistical analysis and parallel multi-objective optimization, this operational data is used to generate a verification solution to scientifically evaluate the effectiveness of the update solution. Simultaneously, the system collects feedback data through a real-time monitoring module. When monitoring indicators exceed preset thresholds, adaptive parameter adjustments are automatically triggered, resulting in the final adjustment data. Finally, this adjustment data and the verification solution data are integrated using a predetermined data format to generate a full update solution. This is then rolled out via an automated deployment platform. A closed-loop feedback mechanism, leveraging real-time monitoring data, is established to enable periodic dynamic iterative updates. The entire process, from data collection to full updates, forms a closed-loop control system, ensuring that update solutions can adapt to the ever-changing gaming environment, enabling dynamic regulation and continuous optimization of the system, thereby improving player experience and system stability. In practical applications, this invention demonstrates excellent operability and scalability, accurately identifying various abnormal conditions that may arise during operation and enabling timely adjustments to update solutions to ensure the gaming system is always operating optimally.

[0111] Preferably, the adjustment data and the verification scheme data are integrated according to a predetermined data format to form a full update scheme, and the full update scheme is implemented in a rolling deployment manner on an automated deployment platform. The rolling deployment method includes first implementing the update on at least 10% of the production nodes, and gradually expanding to all production nodes after confirming that the operation is stable, and real-time monitoring and collection of player behavior data, system performance data, community sentiment data and BUG feedback data to form a closed-loop feedback, and the closed-loop feedback triggers dynamic iterative updates at set time intervals.

[0112] In the present invention, the adjustment data and verification scheme data are integrated in JSON data format according to a predetermined data format to generate a full update plan. To ensure the standardization and compatibility of data integration, the system uses JSON as the data exchange format. The reason is that the JSON format has the characteristics of clear structure, easy parsing, and strong cross-platform compatibility. During the specific operation, the system encodes the adjustment data generated from the adaptive adjustment module and the verification scheme data generated by statistical analysis and parallel multi-objective optimization processing in standard JSON format. Each data item is expressed in the form of a key-value pair to ensure that the field name and data type remain consistent between different modules. For example, key parameters in the adjustment data such as "difficulty_adjustment", "currency_flow_adjustment" and "balance_parameter" are all recorded in numerical form, while monitoring indicators in the verification scheme data such as "cpu_usage", "memory_usage", "latency", "player_engagement" and "bug_rate" are all stored in floating point or integer form. When integrating, the system merges the two JSON data objects into a unified JSON object through a predetermined data interface call. The integration process can be expressed by the following pseudo code:

[0113] JSON_Adjustment={"difficulty_adjustment":X,"currency_flow_adjustment":Y,"balance_parameter":Z,...}

[0114] JSON_Verification={"cpu_usage":A,"memory_usage":B,"latency":C,"player_engagement":D,"bug_rate":E,...}

[0115] Full_Update_Scheme=merge(JSON_Adjustment,JSON_Verification)

[0116] The merge function retains or renames identical fields according to predefined rules, ensuring the integrity and uniform format of the merged JSON object data. The merged full update plan data not only includes adaptively adjusted real-time parameters but also includes various monitoring metrics used to verify the update's effectiveness, ensuring data input compatibility with subsequent processing modules. To ensure stable and secure data transmission, the system utilizes standard HTTPS protocols and RESTful APIs for data exchange. Furthermore, to ensure data consistency, the system implements data validation during the integration process, checking the completeness of required fields in each JSON object and verifying the data format. After these steps, the generated full update plan data is stored in the data warehouse in JSON format and serves as input for the subsequent automated deployment module, further enabling fully automated updates. The advantage of this data integration approach lies in the standardized data format, which facilitates cross-module data calls and efficiently transmits complex, multi-dimensional data. This ensures high data consistency and traceability during the implementation of the full update plan, providing accurate and real-time data support for game update decision-making. In this way, developers can quickly locate the adjustment needs of various parameters in the system and achieve fine-grained control at the data level to ensure the efficient operation and scientific nature of the entire update process.

[0117] In this invention, the full update plan is deployed using a rolling deployment method implemented by an automated deployment platform. The goal is to gradually roll out the update plan to all production nodes while ensuring stable system operation, thereby reducing update risks and achieving a smooth transition. The rolling deployment method initially implements the update on at least 10% of production nodes. During this phase, the system pushes the full update plan data to a subset of nodes via a standardized interface for a small-scale trial run. Specifically, the automated deployment platform divides the production environment into several node groups based on pre-set policies, with each node group independently performing the update task. After the update begins, the system monitors the operating status of each node group, including key metrics such as server load, response time, error rate, and player feedback. This monitoring data is collected by a real-time monitoring module and fed back to the central data processing unit in JSON format. If, during the trial run, the operating metrics of each node group show no abnormal fluctuations, the system gradually expands the update scope, for example, from an initial 10% to 30%, and then to 60%, ultimately completing the full node update.

[0118] During rolling deployments, the automated deployment platform uses containerization technologies (such as Docker and Kubernetes) to achieve deployment flexibility and fault tolerance. Each node undergoes a health check before updating to ensure that node resources (such as memory, CPU, and network bandwidth) meet the update requirements. During the update process, the platform adopts a blue-green deployment or canary release strategy. This strategy involves running the old and new versions in parallel and using real-time monitoring data to determine whether the new version outperforms the old version. If the new version demonstrates the expected stability and performance improvements on the pilot nodes, the platform automatically expands the update to additional nodes. Otherwise, the system quickly rolls back to the old version to ensure that the entire production environment is not affected.

[0119] To ensure efficient and reliable data transmission and task scheduling during the rolling deployment process, the system adopts a distributed task scheduling mechanism. Each node receives update commands through a standard RESTful API and implements asynchronous task processing through a message queue. The rolling deployment strategy not only greatly reduces the risk of system failures caused by a single full deployment, but also enables developers to monitor and adjust deployment strategies in real time through phased implementation, further optimizing the update process. During this process, the update status, operating indicators, and user feedback of each node are saved in a standardized log record format for subsequent data analysis and decision support. Through this rolling deployment method, the system has achieved a gradual update from a small-scale pilot to a full-scale promotion, effectively ensuring the smoothness of the update process and the continuity of the system. In this example, if the average response time T of a certain node group is less than the set threshold T th And the error rate E is lower than the preset standard E th , the node group is considered stable and the system expands the update scope. During the rolling deployment process, detailed data analysis and monitoring reports at each stage provide data support for decision-making in the next stage, ensuring that the update plan is fully verified before full rollout, making the entire update process efficient and secure.

[0120] The present invention uses real-time monitoring data to form a closed-loop feedback mechanism to periodically trigger dynamic iterative updates to achieve continuous optimization of the full update plan. After the system is fully deployed, the automated monitoring module collects player behavior data, system performance data, and BUG feedback data in real time. These data not only reflect the current operating status of the full update plan in the actual production environment, but also provide a basis for the system to detect potential problems. The monitoring data is uploaded to the data center in JSON format through a standardized interface. The data includes server load, response time, error rate, player interaction frequency, and BUG generation rate. The data center performs statistical analysis on these real-time data and compares them with the preset normal operation threshold.

[0121] When the system detects that one or more indicators exceed the preset range, the closed-loop feedback mechanism will automatically trigger dynamic iterative updates. In specific operations, the feedback module calculates the deviation of each indicator by constructing an objective function. The objective function is in the form of:

[0122]

[0123] Among them, x i Indicates the current value of the i-th monitoring indicator, x i,th is the preset threshold of the indicator, α i are the corresponding weights, and N is the total number of monitored indicators. The value of the objective function L reflects the deviation between the system's operating state and the ideal state. When L exceeds a set critical value, the adaptive adjustment module immediately intervenes and automatically adjusts key parameters in the update plan, such as level difficulty, resource allocation, and economic balance parameters, through parameter tuning methods based on gradient descent or evolutionary algorithms. The adjusted parameters are then integrated again through the data interface to generate a new update plan, which is then rolled out to the production environment through the automated deployment platform.

[0124] The design of the closed-loop feedback mechanism ensures that the system can achieve adaptive iterative updates driven by real-time monitoring data. Feedback data collection uses high-frequency data transmission technology to ensure that the system status is automatically checked at preset time intervals (such as 24 hours or less) and parameters are adjusted dynamically. All data transmission in the system is carried out using the standard HTTPS protocol and RESTfulAPI to ensure data security and no delay. During the feedback process, the monitoring module not only records the current data status, but also generates historical data reports for subsequent trend analysis. For example, when real-time monitoring data shows that the average player response time T exceeds the set upper limit T max When the system automatically calculates the response time deviation ΔT=TT max , and conduct a comprehensive evaluation in combination with other relevant indicators, and then adjust the load balancing strategy or modify the task scheduling parameters to reduce the system response time to within the preset standard. Similarly, when the BUG generation rate B is higher than the preset standard B th When a fault occurs, the system will trigger a fault warning, automatically adjust resource allocation or trigger a patch update to ensure stable system operation.

[0125] Through this closed-loop feedback mechanism, the system can not only capture and respond to abnormal conditions in the production environment in a timely manner, but also continuously optimize and update plans during long-term operation, forming an intelligent closed loop of dynamic iterative updates. The entire feedback process consists of data collection, data transmission, statistical analysis, objective function calculation and adaptive parameter adjustment. Standardized processes are implemented in each link to ensure data consistency and processing efficiency. The present invention has proved in many experiments that its adaptive iterative update can significantly improve the stability of the game system and user experience, while reducing operational risks caused by system anomalies. The closed-loop feedback mechanism ultimately realizes the full-process closed loop of the system from data monitoring to dynamic optimization, providing a scientific, accurate and real-time decision support platform for game updates, thereby promoting the game system to maintain continuous competitiveness in a changing operating environment.

[0126] like Figure 4 As shown, a game update system based on game development data adjustment is used to implement the game update method based on game development data adjustment, and the system includes:

[0127] The acquisition preprocessing module is used to collect player behavior data, system performance data, community sentiment data, and bug feedback data from the gaming environment. It then performs noise filtering, time correction, format standardization, feature extraction, and fusion processing on this data to form fused data. This module typically consists of dedicated sensors, an embedded processor, and a data acquisition card. Sensors are used to capture various data from the gaming environment in real time, such as player input signals (e.g., keystrokes, mouse clicks), system status monitoring (CPU, memory, network load sensors), and network interfaces to collect community feedback and bug reporting data. The data acquisition card converts these analog or digital signals into digital data, after which the embedded processor performs preliminary noise filtering, time correction, and format standardization. This processor can use a high-performance ARM or DSP chip, supporting real-time data preprocessing. The feature-extracted data is then transmitted to subsequent processing units via a bus (e.g., PCIe or high-speed Ethernet) to form fused data.

[0128] The core generation module is used to form a state vector based on the fusion data and preset key indicators, generate candidate solutions using a deep reinforcement learning model, and use a generative adversarial network to train historical scene data and current game style data to generate innovative content, and then fuse the candidate solutions with the innovative content through a nonlinear weighted fusion method to form an integrated update solution; in hardware implementation, the core generation module relies on a high-performance computing platform, such as a server equipped with a GPU or a dedicated accelerator (such as a TPU or FPGA). The platform receives the fusion data transmitted by the preprocessing module and uses preset key indicators (stored in preset hardware registers or dedicated memory) to form a state vector. Both the deep reinforcement learning model and the generative adversarial network can run on a GPU accelerator card, and parallel computing is used to realize candidate solution generation and innovative content training. In hardware implementation, a dedicated deep learning server is often used, which is equipped with high-bandwidth memory and high-speed interconnection. The overall integrated update solution is finally formed through nonlinear weighted fusion circuits or parallel computing at the software level.

[0129] A simulation adjustment module is used to deploy the integrated update scheme on the digital twin simulation platform, and use a virtual player model to simulate and test the integrated update scheme, collect operation data, generate a verification scheme using statistical analysis and multi-objective optimization, and automatically trigger adaptive parameter adjustment based on real-time feedback to form adjustment data; at the hardware level, this module usually relies on a dedicated simulation system of the digital twin platform, which can be a simulation server that integrates a multi-core CPU and GPU and is equipped with virtualization technology or a containerized platform to simulate the actual game environment. In this system, a virtual player model is pre-loaded (real-time simulation can be achieved through a dedicated FPGA or GPU) to simulate player behavior and system operation. During the simulation process, hardware sensors and monitoring chips continuously collect operation data (such as system load sensors, temperature sensors, etc.). After the data is internally stored and quickly cached, a multi-objective optimization algorithm is executed through a built-in statistical analysis chip (or using computing resources in the server) to generate a verification scheme. Real-time feedback is transmitted through a high-speed data interface, and the adaptive adjustment controller (embedded microprocessor) automatically triggers parameter adjustment based on the set algorithm to form the final adjustment data.

[0130] The deployment feedback module integrates the adjustment data and verification plan data according to a predetermined data format to generate a full update plan. This full update plan is then rolled out via an automated deployment platform, using real-time monitoring data to form a closed-loop feedback loop to periodically trigger dynamic iterative updates. The hardware implementation of the deployment feedback module typically relies on an automated deployment platform that implements rolling deployment based on distributed servers and cluster management technology. This hardware implementation can consist of a standard server cluster, a load balancer, and network monitoring equipment. The adjustment data and verification plan data are formatted uniformly (e.g., in JSON) and transmitted via a high-speed network to a deployment manager, which is responsible for rolling out the update plan by node, initially testing it on a subset of nodes (e.g., 10%) before gradually expanding. During the deployment process, network monitoring equipment (e.g., SNMP monitors) and dedicated hardware monitoring cards collect real-time data on player behavior, system performance, and bug feedback. This data is processed in a closed-loop by data center servers and fed back to the deployment manager at preset intervals, triggering the next dynamic iterative update. The entire system ensures low latency and high reliability of data transmission through standardized hardware interfaces and high-speed buses, enabling automated updates and dynamic optimization throughout the entire process.

[0131] A storage medium stores a computer program, which, when executed by a processor, implements the steps of the game update method adjusted based on game development data.

[0132] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware.

[0133] The above are merely embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various modifications and variations. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application should be included within the scope of the claims of the present application.

Claims

1. A game update method based on game development data adjustment, characterized in that: The following steps are involved: Collect player behavior data, system performance data, community sentiment data, and bug feedback data from the game environment. After pre-processing all the data, perform feature extraction and fusion processing to form fused data; A state vector is formed based on the fused data and preset key indicators, a deep reinforcement learning model is used to generate candidate solutions, and a generative adversarial network is used to train historical scene data and current game style data to generate innovative content. The candidate solutions are then fused with the innovative content through a nonlinear weighted fusion method to form an integrated update solution; Deploy the integration and update solution on a digital twin simulation platform, conduct simulation tests on the solution using a virtual player model based on Agent-Based Modeling, collect operational data, and generate a verification solution using statistical analysis and multi-objective optimization. At the same time, adaptive parameter adjustments are automatically triggered based on real-time feedback to generate adjustment data. The adjustment data and the verification scheme data are integrated to generate a full update scheme, the full update scheme is rolled out through an automated deployment platform, and closed-loop feedback is formed using real-time monitoring data to periodically trigger dynamic iterative updates.

2. The method according to claim 1, characterized in that The preset key indicators are player retention rate, system response index and economic balance index, and the fusion data and the preset key indicators are combined with predetermined weights to form a state vector.

3. The method according to claim 1, characterized in that The deep reinforcement learning model adopts a multi-task transfer learning strategy, using the state vector as input to generate candidate solutions, wherein the candidate solutions clearly specify level difficulty, currency circulation and character balance parameters.

4. The method according to claim 1, wherein The generative adversarial network adopts a network structure composed of several convolutional layers to train historical scene data and current game style data to generate innovative content output, and enhances the innovative content output using feature vectors extracted by variational autoencoders. The candidate solutions and the enhanced innovative content are fused through a nonlinear weighted fusion method to form integrated updated solution data.

5. The method according to claim 1, wherein The method of fusing the candidate solutions with the innovative content through a nonlinear weighted fusion method to form an integrated update solution specifically includes: By constructing an interaction matrix and applying a nonlinear activation function to fuse the features of candidate solutions and innovative content, an integrated update solution is formed.

6. The method according to claim 1, characterized in that The integration and update solution is deployed on the digital twin simulation platform, and simulation testing of the integration and update solution is performed through a virtual player model to collect operation data, specifically including: The integrated update solution is implemented and deployed on the digital twin simulation platform using containerization technology, and the integrated update solution is simulated and tested based on the virtual player model of Agent-Based Modeling, so as to collect system load, player behavior, economic balance and BUG generation rate data as operation data.

7. The method according to claim 1, characterized in that The statistical analysis and multi-objective optimization are used to generate the verification plan, specifically including: The running data is processed through standard statistical analysis and parallel multi-objective optimization based on a swarm search strategy to generate a verification plan.

8. The method according to claim 1, characterized in that The adjustment data and the verification scheme data are integrated according to a predetermined data format to form a full update scheme, and the full update scheme is implemented in a rolling deployment manner on an automated deployment platform. The rolling deployment method includes first implementing the update on at least 10% of the production nodes, and gradually expanding to all production nodes after confirming that the operation is stable, and real-time monitoring and collection of player behavior data, system performance data, community sentiment data and BUG feedback data to form a closed-loop feedback, and the closed-loop feedback triggers dynamic iterative updates at set time intervals.

9. A game update system based on game development data adjustment, used to implement the game update method based on game development data adjustment according to any one of claims 1 to 8, characterized in that: The system includes: The collection and preprocessing module is used to collect player behavior data, system performance data, community sentiment data, and bug feedback data in the game environment. After preprocessing all the data, feature extraction and fusion processing are performed to form fused data; A core generation module, which is used to construct a state vector based on the fused data and preset key indicators, use a deep reinforcement learning model to generate candidate solutions, and use a generative adversarial network to train historical scene data and current game style data to generate innovative content. Then, a nonlinear weighted fusion method is used to fuse the candidate solutions with the innovative content to form an integrated update solution; A simulation adjustment module, which is used to deploy the integrated update solution on the digital twin simulation platform, simulate and test the integrated update solution using a virtual player model based on Agent-Based Modeling, collect operational data, generate a verification solution using statistical analysis and multi-objective optimization, and automatically trigger adaptive parameter adjustments based on real-time feedback to generate adjustment data; A deployment feedback module is used to integrate the adjustment data and verification plan data according to a predetermined data format to generate a full update plan, implement rolling deployment of the full update plan through an automated deployment platform, and use real-time monitoring data to form a closed-loop feedback to periodically trigger dynamic iterative updates.

10. A storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the game update method based on game development data adjustment according to any one of claims 1 to 8 are implemented.

Images