Game state synchronization optimization system based on edge computing
Patent Information
- Application Number
- CN202611339865.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-09-01
- Publication Date
- 2026-10-09
AI Technical Summary
[0003]针对上述情况，为克服现有技术的缺陷，本发明提供了基于边缘计算的游戏状态同步优化系统，针对常规游戏状态同步优化系统存在难以适应高动态游戏环境下状态变化的不确定性、无法覆盖不同游戏运行模式，降低同步优化对复杂场景的适应能力的问题，本方案基于历史时间窗口采集的游戏状态数据，构建状态变化关联空间，提高边缘节点对动态游戏状态变化规律的感知能力；通过状态邻近聚集度筛选代表性参考状态，同时设置参考状态间隔约束，避免多个参考状态集中于相似区域，使最终参考状态覆盖更广泛的游戏状态变化空间，提高同步优化鲁棒性；通过游戏影响因子实现边缘服务器对不同游戏状态的动态可信度评价，降低异常状态传播，提高游戏状态同步稳定性；针对常规游戏状态同步优化系统存在容易受到异常节点影响、容易产生局部同步突变以及低可靠节点影响最终同步结果的问题，本方案引入状态参数同步损失，并构建参数同步因子，实现了由所有状态参数统一同步向重要状态参数优先同步的转变，提高关键玩家行为状态同步精度；针对不同边缘服务器计算能力和网络状态差异，引入节点同步损失计算节点融合比例，避免单一异常边缘节点影响全局同步结果；设计状态连续置信度构造同步约束损失，提高游戏状态演化过程的连续性
（1）针对常规游戏状态同步优化系统存在难以适应高动态游戏环境下状态变化的不确定性、无法覆盖不同游戏运行模式，降低同步优化对复杂场景的适应能力的问题，本方案基于历史时间窗口采集的游戏状态数据，构建状态变化关联空间，提高边缘节点对动态游戏状态变化规律的感知能力；通过状态邻近聚集度筛选代表性参考状态，同时设置参考状态间隔约束，避免多个参考状态集中于相似区域，使最终参考状态覆盖更广泛的游戏状态变化空间，提高同步优化鲁棒性；通过游戏影响因子实现边缘服务器对不同游戏状态的动态可信度评价，降低异常状态传播，提高游戏状态同步稳定性。
Smart Images
Figure CN122891584A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of data processing, specifically to a game state synchronization optimization system based on edge computing. Background Technology
[0002] A game state synchronization optimization system optimizes the collection, transmission, and updating of game state data, enabling multiple player clients to obtain a consistent game state from the server in a timely manner, thereby reducing latency and desynchronization. However, conventional game state synchronization optimization systems struggle to adapt to the uncertainty of state changes in highly dynamic game environments, cannot cover different game operation modes, and have reduced adaptability to complex scenarios. Furthermore, conventional game state synchronization optimization systems are susceptible to abnormal nodes, prone to local synchronization mutations, and the impact of unreliable nodes on the final synchronization result. Summary of the Invention
[0003] To address the above issues and overcome the shortcomings of existing technologies, this invention provides a game state synchronization optimization system based on edge computing. Addressing the problems of conventional game state synchronization optimization systems, such as difficulty adapting to the uncertainty of state changes in highly dynamic game environments, inability to cover different game operation modes, and reduced adaptability of synchronization optimization to complex scenarios, this solution constructs a state change correlation space based on game state data collected through historical time windows, improving the edge nodes' ability to perceive the dynamic game state change patterns. It selects representative reference states through state proximity clustering and sets reference state interval constraints to avoid multiple reference states concentrating in similar areas, ensuring that the final reference states cover a wider range of game state change spaces and improving the robustness of synchronization optimization. Furthermore, it implements game influence factors... The current edge server dynamically evaluates the reliability of different game states, reducing the propagation of abnormal states and improving the stability of game state synchronization. Addressing the issues of conventional game state synchronization optimization systems being susceptible to abnormal nodes, prone to local synchronization mutations, and having low-reliability nodes affecting the final synchronization result, this solution introduces state parameter synchronization loss and constructs a parameter synchronization factor. This achieves a shift from unified synchronization of all state parameters to priority synchronization of important state parameters, improving the synchronization accuracy of key player behavior states. To address the differences in computing power and network conditions among different edge servers, a node synchronization loss is introduced to calculate the node fusion ratio, preventing a single abnormal edge node from affecting the global synchronization result. A state continuity confidence score is designed to construct a synchronization constraint loss, improving the continuity of the game state evolution process.
[0004] The technical solution adopted by the present invention is as follows: The game state synchronization optimization system based on edge computing provided by the present invention includes a game state data acquisition module, a reference state construction module, a game influence factor calculation module, a synchronization factor calculation module, an edge node fusion ratio calculation module, a state credibility continuous constraint module, and a final synchronization state generation module. The game status data acquisition module uses an edge server to collect the game running status and construct a game status sample set. The reference state construction module filters representative stable state patterns based on game state change data within a historical time window to obtain a reference state set. The game impact factor calculation module calculates the similarity between the game state sample and the reference state based on the matching relationship between the current game state and the reference state, and generates the corresponding impact factor. The synchronization factor calculation module calculates the contribution of each state parameter to the synchronization loss based on the influence factor and the deviation relationship between different state parameters and the reference state, and dynamically adjusts the synchronization factor of each state parameter. The edge node fusion ratio calculation module calculates the node synchronization loss based on the deviation between the game state and the reference state, combined with the influencing factor, and dynamically generates the fusion ratio of each edge node based on the node synchronization loss. The state confidence continuity constraint module constructs a game state association range based on the similarity relationship of game states, establishes a state continuity constraint relationship by analyzing the continuous state changes within the association range, generates a state continuity confidence, and constructs a comprehensive optimization objective. The final synchronization state generation module uses the game state predicted by each edge node for the next moment, and combines it with the influence factor, the state parameter synchronization factor and the edge node fusion ratio for weighted fusion to generate the final game synchronization state.
[0005] Furthermore, the game state data acquisition module uses an edge server to collect game running status, preprocesses the collected data, performs max-min normalization on continuous data, and performs one-hot encoding on discrete data to obtain a game state sample set.
[0006] Furthermore, the reference state construction module constructs a state change correlation space based on game state samples within each edge server, and filters reference states by analyzing the similarity between different game states; specifically including: State similarity distance calculation involves performing nonlinear similarity calculations on game states collected within different time windows within edge nodes. The state similarity distance is expressed as: ; ; It is the similarity distance between the i-th and j-th game state samples; It is the influence factor of the m-th state parameter; M is the total number of state parameters; and These are the m-th state parameters of the i-th and j-th game state samples, respectively; It is the variance of the change of the m-th state parameter within the historical synchronization window; It is a smoothing term; The state proximity clustering degree is calculated by counting the number of states with similar change patterns around each game state based on the state similarity distance calculation results, thus obtaining the state proximity clustering degree. ; Introducing minimum difference distance ; The reference state screening value is obtained by fusing the state proximity clustering degree and the minimum difference distance, as follows: ; This is a reference status filter value; A candidate reference state sequence is formed by sorting the reference state screening values of all states in descending order. Reference state spacing constraint, setting the reference state interval threshold. , It is the state dispersion adjustment coefficient; for the candidate reference state sequence, it retains The state as a reference state set ; This is the reference state obtained through filtering.
[0007] Furthermore, the game influence factor calculation module calculates the influence factor of each state sample based on the similarity between the current game state and the reference state, and determines the influence factor of the game state sample through the distance relationship between the state and the reference state, expressed as: ; It is the influence factor of the i-th game state sample; It is the m-th parameter in the k-th reference state; K is the number of reference states.
[0008] Furthermore, the synchronization factor calculation module, based on the influence factor, dynamically calculates the contribution of different state parameters to the synchronization loss according to the state reliability and the degree of deviation of the current state from the reference state; specifically including: Based on the influence factor and the deviation between the current state and the reference state, the contribution of each state parameter to the synchronization loss is calculated and expressed as follows: ; It is the contribution of the m-th game state parameter to the synchronization loss; It is the confidence level of state continuity. ; It is the Euclidean distance between the current state and the k-th reference state; It is the m-th parameter of the nearest reference state to the i-th game state sample; Update the parameter synchronization factor based on the parameter error contribution: ; It is the synchronization factor of the m-th state parameter; It is the contribution value of the s-th state parameter; q is the factor smoothing factor; It is a smoothing term.
[0009] Furthermore, the edge node fusion ratio calculation module dynamically calculates the node contribution factor based on the historical synchronization loss and state matching degree of the edge nodes; specifically including: Edge node synchronization loss calculation: For the h-th edge node, based on the deviation between the game state collected by the node and the reference state, combined with the influencing factor, the node synchronization loss is calculated, expressed as: ; It is the synchronization loss of the h-th edge node, where h is the edge node number; It is the similarity distance between the i-th game state sample and the k-th reference state; The edge node fusion ratio is calculated adaptively based on the synchronization loss of each edge node, and is expressed as follows: ;constraint ; It is the fusion ratio of the h-th edge node; is the synchronization loss of the l-th edge node; H is the total number of edge nodes; It is the factor smoothing parameter. ; and These are the maximum factor smoothing parameter and the initial factor smoothing parameter, respectively.
[0010] Furthermore, the state credibility continuous constraint module constructs the game state association range relationship and adds continuous constraints to game states with similar state change patterns; specifically including: The player association range is constructed by determining the game state association range based on the game state distance calculation results, and is represented as follows: ; ; R is the set of game state association ranges for the i-th game state; R is the threshold value for the state association range. It controls the scope and scale of related activities; State synchronization and continuity constraints, introducing state continuity evaluation factors ; It is the continuous confidence of the i-th game state; It is the state distance scaling factor; It is the similarity distance between the player's current state and the k-th reference state; Construct a synchronization constraint loss based on the continuous state relationship between players within the associated range: ; It is a continuous loss of the associated range of states; It is the correlation range constraint strength coefficient; It is the average number of associated range states; It is the continuous confidence of the j-th associated range game state; It is the set of associated ranges for the i-th game state; Construct a comprehensive synchronous optimization objective that simultaneously considers: state prediction error, edge node reliability, parameter importance, and player local continuity; comprehensive optimization objective ; It is the nearest reference state to the i-th game state sample; q is the smoothing factor; It is the continuity constraint coefficient of the associated range; It is the edge collaboration constraint coefficient. t is the number of iterations; It represents the maximum collaborative constraint strength; T represents the maximum number of iterations.
[0011] Furthermore, the final synchronization state generation module fuses the predicted states of each node according to the edge node fusion ratio to generate a unified client synchronization state; specifically including: Edge node state prediction: Each edge node extracts state change features based on its own collected game state data through a state feature encoding network, and inputs them into a state prediction network to obtain the game state prediction result for the next moment. Multi-edge node state fusion utilizes the edge node fusion ratio to weight and fuse the predicted states of different nodes to obtain the global predicted state.
[0012] The beneficial effects achieved by the present invention using the above solution are as follows: (1) To address the problems of conventional game state synchronization optimization systems being unable to adapt to the uncertainty of state changes in highly dynamic game environments and being unable to cover different game operation modes, thus reducing the adaptability of synchronization optimization to complex scenarios, this solution constructs a state change association space based on game state data collected in historical time windows, thereby improving the edge nodes' ability to perceive the dynamic game state change patterns; representative reference states are selected by state proximity clustering, and reference state interval constraints are set to avoid multiple reference states being concentrated in similar areas, so that the final reference states cover a wider range of game state change spaces and improve the robustness of synchronization optimization; the game influence factor is used to realize the dynamic credibility evaluation of different game states by the edge server, reduce the propagation of abnormal states, and improve the stability of game state synchronization.
[0013] (2) To address the problems of conventional game state synchronization optimization systems being susceptible to abnormal nodes, prone to local synchronization mutations, and having low-reliability nodes affecting the final synchronization result, this solution introduces state parameter synchronization loss and constructs parameter synchronization factors, realizing the transformation from unified synchronization of all state parameters to priority synchronization of important state parameters, thereby improving the synchronization accuracy of key player behavior states; to address the differences in computing power and network status of different edge servers, node synchronization loss is introduced to calculate the node fusion ratio, avoiding the impact of a single abnormal edge node on the global synchronization result; and a state continuity confidence degree is designed to construct synchronization constraint loss, thereby improving the continuity of the game state evolution process. Attached Figure Description
[0014] Figure 1 A flowchart illustrating the game state synchronization optimization system based on edge computing provided by this invention.
[0015] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used together with the embodiments of the invention to explain the invention and do not constitute a limitation thereof. Detailed Implementation
[0016] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.
[0017] In the description of this invention, it should be understood that the terms "upper", "lower", "front", "rear", "left", "right", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.
[0018] Example 1, see Figure 1 The game state synchronization optimization system based on edge computing provided by the present invention includes a game state data acquisition module, a reference state construction module, a game influence factor calculation module, a synchronization factor calculation module, an edge node fusion ratio calculation module, a state credibility continuous constraint module, and a final synchronization state generation module. The game status data acquisition module uses an edge server to collect the game running status and construct a game status sample set. The reference state construction module filters representative stable state patterns based on game state change data within a historical time window to obtain a reference state set. The game impact factor calculation module calculates the similarity between the game state sample and the reference state based on the matching relationship between the current game state and the reference state, and generates the corresponding impact factor. The synchronization factor calculation module calculates the contribution of each state parameter to the synchronization loss based on the influence factor and the deviation relationship between different state parameters and the reference state, and dynamically adjusts the synchronization factor of each state parameter. The edge node fusion ratio calculation module calculates the node synchronization loss based on the deviation between the game state and the reference state, combined with the influencing factor, and dynamically generates the fusion ratio of each edge node based on the node synchronization loss. The state confidence continuity constraint module constructs a game state association range based on the similarity relationship of game states, establishes a state continuity constraint relationship by analyzing the continuous state changes within the association range, generates a state continuity confidence, and constructs a comprehensive optimization objective. The final synchronization state generation module uses the game state predicted by each edge node for the next moment, and combines it with the influence factor, the state parameter synchronization factor and the edge node fusion ratio for weighted fusion to generate the final game synchronization state.
[0019] Example 2, see Figure 1 This embodiment is based on the above embodiment. The game state data acquisition module uses an edge server to collect the game's running state. Instead of uploading all state data to the central cloud, it completes state acquisition, feature extraction, and initial data processing at edge nodes close to the player, thereby reducing synchronization latency caused by cross-network transmission. The collected data includes: player behavior state, network state, game environment state, and edge node state. The behavior state includes location coordinates, movement speed, skill release state, and operation events. The network state includes RTT latency, packet loss rate, and bandwidth usage. The game environment state includes the scene area, the number of surrounding players, and the number of interactive objects. The edge node state includes CPU utilization, memory usage, and the number of current connections. The collected data is preprocessed; continuous data is processed using max-min normalization, and discrete data is processed using one-hot encoding to obtain a game state sample set. ; N represents the game state samples, and N is the number of state samples collected by the current edge node.
[0020] Example 3, see Figure 1This embodiment is based on the above embodiment. To avoid being unable to adapt to high-concurrency battles, rapid player movement, and network fluctuations, the reference state construction module constructs a state change correlation space based on game state samples within each edge server. By analyzing the similarity between different game states, it filters reference states that can represent the main state change patterns of the current edge node; specifically including: State similarity distance calculation involves performing nonlinear similarity calculations on game states collected within different time windows within edge nodes. The state similarity distance is expressed as: ; ; It is the similarity distance between the i-th and j-th game state samples; It is the influence factor of the m-th state parameter; M is the total number of state parameters; and These are the m-th state parameters (already normalized) of the i-th and j-th game state samples, respectively. It is the variance of the change of the m-th state parameter within the historical synchronization window; It is a smoothing term; the more stable the changes in state parameters, the smaller the historical fluctuation variance, and the corresponding... The larger the value, the greater its contribution to difference calculation; for highly fluctuating parameters such as network latency and player speed, its influence is automatically reduced to avoid incorrect reference state selection due to instantaneous abnormal changes; The state proximity clustering degree is calculated by counting the number of states with similar change patterns around each game state based on the state similarity distance calculation results. This is to determine whether the corresponding state can represent the main state change pattern in the current edge node. The state proximity clustering degree is expressed as: ; ; It is the number of locally similar states for the i-th game state sample; It is the indicator function; R is the radius of the state-related range; K is the number of reference states, used to control the number of reference states; N is the number of historical state samples of the current edge node. and These are the maximum and minimum state distances in the current state set, respectively. To avoid selecting states with a large number of similar states but indistinguishable from other change patterns, a minimum difference distance is introduced: ; It is the minimum difference distance between the i-th state and the higher representative state; It is the neighborhood clustering degree of the j-th game state sample; The reference state screening value is obtained by fusing the state proximity clustering degree and the minimum difference distance, as follows: ; This is a reference status filter value; A candidate reference state sequence is formed by sorting the reference state screening values of all states in descending order. Reference state spacing constraints are used to prevent multiple reference states from clustering in the same area; a reference state interval threshold is set. , This is the state dispersion adjustment coefficient, ranging from 1 to 5; for candidate reference state sequences, it is retained. The state as a reference state set ; This is the reference state obtained through filtering.
[0021] Example 4, see Figure 1 This embodiment is based on the above embodiment. The game influence factor calculation module addresses the issue of different state samples having varying degrees of influence on the synchronization result during game operation. It calculates the influence factor for each state sample based on the similarity between the current game state and the reference state. Since stable game states are usually located near historical state change patterns, while abnormal states often deviate from existing state evolution patterns, the influence factor of the game state sample is determined by the distance relationship between the state and the reference state, expressed as: ; It is the influence factor of the i-th game state sample; It is the m-th parameter in the k-th reference state; When the current game state is close to multiple historical stable reference states, the exponential function output increases, indicating that the state conforms to existing change patterns and increases the influence factor. When the state is affected by network jitter, abnormal operations, or prediction offsets, and the distance from the reference state increases, the exponential function output decreases, thereby reducing the interference of abnormal states on the synchronization results.
[0022] By performing the above operations, this solution addresses the problems of conventional game state synchronization optimization systems, such as difficulty in adapting to the uncertainty of state changes in highly dynamic game environments, inability to cover different game operation modes, and reduced adaptability of synchronization optimization to complex scenarios. It constructs a state change correlation space based on game state data collected through historical time windows, improving the edge nodes' ability to perceive the dynamic game state change patterns. Representative reference states are selected through state proximity clustering, while reference state interval constraints are set to avoid multiple reference states concentrating in similar areas, ensuring that the final reference states cover a wider range of game state change spaces and improving the robustness of synchronization optimization. Finally, the game influence factor enables edge servers to dynamically evaluate the credibility of different game states, reducing the propagation of abnormal states and improving the stability of game state synchronization.
[0023] Example 5, see Figure 1This embodiment is based on the above embodiment. The synchronization factor calculation module, based on the influence factor, dynamically calculates the contribution of different state parameters to the synchronization loss according to the state reliability and the degree of deviation of the current state from the reference state; specifically including: Based on the influence factor and the deviation between the current state and the reference state, the contribution of each state parameter to the synchronization loss is calculated and expressed as follows: ; It is the contribution of the m-th game state parameter to the synchronization loss; It is the confidence level of state continuity. ; It is the Euclidean distance between the current state and the k-th reference state; It is the m-th parameter of the nearest reference state to the i-th game state sample; Update the parameter synchronization factor based on the parameter error contribution: ; It is the synchronization factor of the m-th state parameter; It is a smoothing term; is the contribution value of the s-th state parameter; q is the factor smoothing factor, which takes the value 1.5~3; when the state parameter causes a large synchronization loss, the contribution value of the corresponding state parameter increases, and a higher synchronization factor is obtained; when the parameter changes little over a long period of time and has a limited impact on the synchronization result, the synchronization factor automatically decreases. Parameter filtering reduces the transmission of invalid state parameters: .
[0024] Example 6, see Figure 1 This embodiment is based on the above embodiment. Due to the varying capabilities of different edge servers, the edge node fusion ratio calculation module dynamically calculates the node contribution factor based on the historical synchronization loss and state matching degree of the edge nodes; specifically, it includes: Edge node synchronization loss calculation: For the h-th edge node, based on the deviation between the game state collected by the node and the reference state, combined with the influencing factor, the node synchronization loss is calculated, expressed as: ; It is the synchronization loss of the h-th edge node, where h is the edge node number; It is the similarity distance between the i-th game state sample and the k-th reference state; by selecting the nearest reference state, the state deviation is calculated so that the node synchronization loss can reflect the degree of matching between the current node state change pattern and the historical stable pattern; The edge node fusion ratio is calculated adaptively based on the synchronization loss of each edge node, ensuring that edge nodes with lower synchronization losses occupy a higher proportion in the final state generation process. This is expressed as: ;constraint ; It is the fusion ratio of the h-th edge node; is the synchronization loss of the l-th edge node; H is the total number of edge nodes; It is the factor smoothing parameter. ; and These are the maximum factor smoothing parameter and the initial factor smoothing parameter, respectively. When the synchronization loss of edge nodes is low, the corresponding fusion ratio automatically increases; when the synchronization loss of nodes increases due to network fluctuations or state offsets, the fusion ratio automatically decreases, thereby reducing the impact of abnormal nodes on the final synchronization state.
[0025] Example 7, see Figure 1 This embodiment is based on the above embodiment. To avoid synchronization anomalies caused by sudden changes in local player positions and discontinuous action states due to instantaneous network jitter, data loss, and prediction errors, the state reliability continuity constraint module constructs a game state association range relationship and adds continuous constraints to game states with similar state change patterns; specifically including: The player association range is constructed by determining the game state association range based on the game state distance calculation results. This ensures that the association range can adapt to changes in player interaction patterns under different game scenarios. The set of game state association ranges is represented as follows: ; ; R is the set of game state association ranges for the i-th game state; R is the threshold value for the state association range. This is the size of the association range control, ranging from 5 to 30. When two game states are highly similar in position, speed, interaction behavior, and network state, and the state distance is small, they are classified into the same association range. When the game state change patterns are significantly different, no association range relationship is established. To measure the consistency of continuous confidence among associated game states, a state continuity evaluation factor is introduced as part of the state synchronization continuity constraint. ; It is the continuous confidence of the i-th game state; It is the state distance scaling factor, with a value ranging from 0.1 to 2; It is the similarity distance between the player's current state and the k-th reference state; when the game state approaches the historical stable state pattern, Increase; when the game state is affected by abnormal network latency or prediction offset. reduce; Construct a synchronization constraint loss based on the continuous state relationship between players within the associated range: ; It is a continuous loss of the associated range of states; It is the constraint strength coefficient of the associated range, with a value ranging from 0.01 to 10; It is the average number of associated range states; It is the continuous confidence of the j-th associated range game state; It is the set of associated ranges for the i-th game state; Construct a comprehensive synchronous optimization objective that simultaneously considers: state prediction error, edge node reliability, parameter importance, and player local continuity; comprehensive optimization objective ; It is the nearest reference state to the i-th game state sample; q is the smoothing factor, with a value of 1.5 to 3; It is the continuous constraint coefficient of the association range, with a value ranging from 0.01 to 10; It is the edge collaboration constraint coefficient. t is the number of iterations; It represents the maximum collaborative constraint strength; T represents the maximum number of iterations. First item The state reference matching error measures the deviation between the current game state and the reference state; the second item For the scope of association to be continuous; the third item This is the edge node collaborative loss, used to control the error during the fusion process of prediction results from different edge nodes.
[0026] Example 8, see Figure 1 This embodiment is based on the above embodiment. The final synchronization state generation module fuses the predicted states of each node according to the edge node fusion ratio to generate a unified client synchronization state; specifically, it includes: Edge node state prediction: Each edge node, based on its own collected game state data, extracts state change features through a state feature encoding network and inputs them into a state prediction network to obtain the game state prediction result for the next time step. The predicted state of the h-th edge node is represented as: ; It is the game state predicted by the h-th edge node at the next moment; It is the historical state sequence of the h-th edge node within a continuous time window; It is a state prediction network; These are the trainable parameters of the network; Multi-edge node state fusion utilizes the edge node fusion ratio to weight and fuse the predicted states of different nodes, making the synchronization results more dependent on edge nodes with smaller prediction errors and higher state continuity. The global synchronization state is represented as follows: ; It is the global predicted state after fusion; by using node factor fusion, it avoids synchronization errors caused by network fluctuations or local abnormal states of a single edge node, and improves state continuity in a multi-edge computing environment. The state feature encoding network uses a multilayer perceptron (MLP). structure: Input layer: M game state parameters; Hidden layers: 3 layers: 128→64→32; Activation: ReLU; Output: State representation vector; The state prediction network uses a reverse mapping structure to recover the game state parameters for the next time step based on the state feature representation. The structure is: 32→64→M. Output: The state at the next time step; Training uses synchronous loss ; These are the predicted corresponding state parameters; During online operation, each edge node collects the game state in real time and uses the trained prediction network to generate the state for the next moment. Then, based on the real-time calculated state reliability, parameter contribution factor, and node fusion ratio, the fusion ratio is dynamically adjusted to output the final synchronized state.
[0027] By performing the above operations, this solution addresses the problems of conventional game state synchronization optimization systems being susceptible to abnormal nodes, prone to local synchronization mutations, and having low-reliability nodes affecting the final synchronization result. It introduces a state parameter synchronization loss and constructs a parameter synchronization factor, achieving a shift from unified synchronization of all state parameters to priority synchronization of important state parameters, thus improving the synchronization accuracy of key player behavior states. To address the differences in computing power and network conditions among different edge servers, a node synchronization loss is introduced to calculate the node fusion ratio, preventing a single abnormal edge node from affecting the global synchronization result. Finally, a state continuity confidence level is designed to construct a synchronization constraint loss, improving the continuity of the game state evolution process.
[0028] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention.
[0029] The present invention and its embodiments have been described above. This description is not restrictive, and the accompanying drawings are only one embodiment of the present invention; the actual structure is not limited thereto. In conclusion, if those skilled in the art are inspired by this description and design similar structures and embodiments without departing from the spirit of the invention, such designs should fall within the protection scope of the present invention.
Claims
1. A game state synchronization optimization system based on edge computing, characterized in that: The system includes a game state data acquisition module, a reference state construction module, a game influence factor calculation module, a synchronization factor calculation module, an edge node fusion ratio calculation module, a state credibility continuous constraint module, and a final synchronization state generation module. The game status data acquisition module uses an edge server to collect the game running status and construct a game status sample set. The reference state construction module filters representative stable state patterns based on game state change data within a historical time window to obtain a reference state set. The game impact factor calculation module calculates the similarity between the game state sample and the reference state based on the matching relationship between the current game state and the reference state, and generates the corresponding impact factor. The synchronization factor calculation module calculates the contribution of each state parameter to the synchronization loss based on the influence factor and the deviation relationship between different state parameters and the reference state, and dynamically adjusts the synchronization factor of each state parameter. The edge node fusion ratio calculation module calculates the node synchronization loss based on the deviation between the game state and the reference state, combined with the influencing factor, and dynamically generates the fusion ratio of each edge node based on the node synchronization loss. The state confidence continuity constraint module constructs a game state association range based on the similarity relationship of game states, establishes a state continuity constraint relationship by analyzing the continuous state changes within the association range, generates a state continuity confidence, and constructs a comprehensive optimization objective. The final synchronization state generation module uses the game state predicted by each edge node for the next moment, and combines it with the influence factor, the state parameter synchronization factor and the edge node fusion ratio for weighted fusion to generate the final game synchronization state.
2. The energy efficiency monitoring and analysis system based on big data according to claim 1, characterized in that: The reference state construction module constructs a state change correlation space based on game state samples within each edge server, and filters reference states by analyzing the similarity between different game states; specifically including: State similarity distance calculation involves performing nonlinear similarity calculations on game states collected within different time windows within edge nodes. The state similarity distance is expressed as: ; ; It is the similarity distance between the i-th and j-th game state samples; It is the influence factor of the m-th state parameter; M is the total number of state parameters; and These are the m-th state parameters of the i-th and j-th game state samples, respectively; It is the variance of the change of the m-th state parameter within the historical synchronization window; It is a smoothing term; The state proximity clustering degree is calculated by counting the number of states with similar change patterns around each game state based on the state similarity distance calculation results, thus obtaining the state proximity clustering degree. ; Introducing minimum difference distance ; The reference state screening value is obtained by fusing the state proximity clustering degree and the minimum difference distance, as follows: ; This is a reference status filter value; A candidate reference state sequence is formed by sorting the reference state screening values of all states in descending order. Reference state spacing constraint, setting the reference state interval threshold. , It is the state dispersion adjustment coefficient; for the candidate reference state sequence, it retains The state as a reference state set ; This is the reference state obtained through filtering.
3. The game state synchronization optimization system based on edge computing according to claim 2, characterized in that: The game influence factor calculation module calculates the influence factor of each state sample based on the similarity between the current game state and the reference state. The influence factor of the game state sample is determined by the distance relationship between the current state and the reference state, and is expressed as follows: ; It is the influence factor of the i-th game state sample; It is the m-th parameter in the k-th reference state; K is the number of reference states.
4. The game state synchronization optimization system based on edge computing according to claim 3, characterized in that: The synchronization factor calculation module dynamically calculates the contribution of different state parameters to the synchronization loss based on the influence factor, the state reliability, and the degree of deviation of the current state from the reference state. Specifically, it includes: Based on the influence factor and the deviation between the current state and the reference state, the contribution of each state parameter to the synchronization loss is calculated and expressed as follows: ; It is the contribution of the m-th game state parameter to the synchronization loss; It is the confidence level of state continuity. ; It is the Euclidean distance between the current state and the k-th reference state; It is the m-th parameter of the nearest reference state to the i-th game state sample; Update the parameter synchronization factor based on the parameter error contribution: ; It is the synchronization factor of the m-th state parameter; It is the contribution value of the s-th state parameter; q is the smoothing factor; It is a smoothing term.
5. The game state synchronization optimization system based on edge computing according to claim 4, characterized in that: The edge node fusion ratio calculation module dynamically calculates the node contribution factor based on the historical synchronization loss and state matching degree of the edge nodes; specifically including: Edge node synchronization loss calculation: For the h-th edge node, based on the deviation between the game state collected by the node and the reference state, combined with the influencing factor, the node synchronization loss is calculated, expressed as: ; It is the synchronization loss of the h-th edge node, where h is the edge node number; It is the similarity distance between the i-th game state sample and the k-th reference state; The edge node fusion ratio is calculated adaptively based on the synchronization loss of each edge node, and is expressed as follows: ;constraint ; It is the fusion ratio of the h-th edge node; is the synchronization loss of the l-th edge node; H is the total number of edge nodes; It is the factor smoothing parameter. ; and These are the maximum factor smoothing parameter and the initial factor smoothing parameter, respectively.
6. The game state synchronization optimization system based on edge computing according to claim 5, characterized in that: The state credibility continuous constraint module constructs the game state association range relationship and adds continuous constraints to game states with similar state change patterns; specifically including: The player association range is constructed by determining the game state association range based on the game state distance calculation results, and is represented as follows: ; ; R is the set of game state association ranges for the i-th game state; R is the threshold value for the state association range. It controls the scope and scale of related activities; State synchronization and continuity constraints, introducing state continuity evaluation factors ; It is the continuous confidence of the i-th game state; It is the state distance scaling factor; It is the similarity distance between the player's current state and the k-th reference state; Construct a synchronization constraint loss based on the continuous state relationship between players within the associated range: ; It is a continuous loss of the associated range of states; It is the strength coefficient of the associated range constraint; It is the average number of associated range states; It is the continuous confidence of the j-th associated range game state; It is the set of associated ranges for the i-th game state; Construct a comprehensive synchronous optimization objective that simultaneously considers: state prediction error, edge node reliability, parameter importance, and player local continuity; comprehensive optimization objective ; It is the nearest reference state to the i-th game state sample; q is the smoothing factor; It is the continuity constraint coefficient of the associated range; It is the edge collaboration constraint coefficient. t is the number of iterations; It represents the maximum collaborative constraint strength; T represents the maximum number of iterations.
7. The game state synchronization optimization system based on edge computing according to claim 1, characterized in that: The game state data acquisition module uses an edge server to collect game running status data, preprocesses the collected data, performs max-min normalization on continuous data, and performs one-hot encoding on discrete data to obtain a game state sample set.
8. The game state synchronization optimization system based on edge computing according to claim 1, characterized in that: The final synchronization state generation module merges the predicted states of each node according to the edge node fusion ratio to generate a unified client synchronization state; specifically including: Edge node state prediction: Each edge node extracts state change features based on its own collected game state data through a state feature encoding network, and inputs them into a state prediction network to obtain the game state prediction result for the next moment. Multi-edge node state fusion utilizes the edge node fusion ratio to weight and fuse the predicted states of different nodes to obtain the global predicted state.