Non-complete information game intelligent processing method and system based on secure multi-party computation, terminal and storage medium

Abstract

The application discloses a non-complete information game intelligent processing method and system based on secure multi-party computation, a terminal and a storage medium, relates to the technical field of data processing, and comprises the following steps: acquiring a participant set of a game, an initial game state, a historical action set, a terminal history set, a preset game strategy based on a secure near-end strategy optimization algorithm, and an action set, an information set and a profit function corresponding to each participant; modeling is performed based on an extended game tree modeling framework according to the participant set, the initial game state, the historical action set, the terminal history set, the game strategy, and the action set, the information set and the profit function corresponding to each participant; and a safe extended game model is constructed. According to the preset optimization target and the leaf nodes in the safe extended game model, the game results between the participants are determined and output. In this way, the data security is improved, and the security of the game process is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, and particularly relates to a non-complete information game intelligent processing method, system, terminal and storage medium based on secure multi-party computation. BACKGROUND

[0002] With the development of science and technology, especially the development of artificial intelligence technology, the application of machine game is also more and more extensive. According to whether the game information is completely visible, the computer game can be divided into complete information game and non-complete information game. In the process of non-complete information game, hidden information that is invisible to game participants is contained.

[0003] In the prior art, the non-complete information game can be solved based on a preset solving algorithm, but when data comparison is performed, the data of all participants needs to be directly provided to a third-party platform for processing. The problem of the prior art is that when the game information is processed, the third-party platform is relied on, and the data needs to be directly provided to the third-party platform, which is not conducive to improving the data security, and thus is not conducive to improving the security of the game process.

[0004] Therefore, the related technology still needs to be improved and developed. SUMMARY

[0005] The main purpose of the present application is to provide a non-complete information game intelligent processing method, system, terminal and storage medium based on secure multi-party computation, which aims to solve the technical problem that when the game information is processed in the related technology, the third-party platform is relied on, and the data needs to be directly provided to the third-party platform, which is not conducive to improving the data security, and thus is not conducive to improving the security of the game process.

[0006] In order to achieve the above purpose, the first aspect of the present application provides a non-complete information game intelligent processing method based on secure multi-party computation, wherein the non-complete information game intelligent processing method based on secure multi-party computation comprises:

[0007] obtaining a set of participants of a game, an initial game state, a set of historical actions, a set of terminal histories, a preset game strategy based on a secure near-end strategy optimization algorithm, and a set of actions, a set of information and a profit function corresponding to each participant, wherein the set of historical actions is used to represent a sequence of game actions, the set of terminal histories is used to represent a complete game path, and data comparison in the game strategy is based on a secure multi-party computation protocol;

[0008] According to the participant set, the initial game state, the historical action set, the terminal history set, the game strategy, and the action set, information set, and payoff function corresponding to each participant, a security extended game model is constructed based on an extended game tree modeling framework, wherein a node in the security extended game model represents a game state, and an edge in the security extended game model represents a selected action at the game;

[0009] According to the preset optimization target and the leaf node in the security extended game model, the game result between the participants is determined and output.

[0010] Optionally, the information set includes privacy information of the participants;

[0011] The game strategy is applied to a protected node in the security extended game model;

[0012] The protected node is a node for comparing the privacy information between the participants.

[0013] Optionally, when data comparison is performed based on the game strategy, an encryption tag is obtained according to the privacy information in the information set of the participants, and comparison is performed based on the encryption tag to obtain a comparison result.

[0014] Optionally, the encryption tag is obtained according to the privacy information in the information set of the participants, including:

[0015] The corresponding privacy information of the participants is encoded to obtain an encoding vector;

[0016] The encoding vector is encrypted by a preset scheduling system according to a preset security model to obtain the encryption tag.

[0017] Optionally, the comparison based on the encryption tag to obtain the comparison result includes:

[0018] A Boolean circuit for comparing the encryption tag is generated based on a preconfigured logic circuit by the scheduling system, and a comparison result of the encryption tag between the participants is obtained based on the Boolean circuit.

[0019] Optionally, the participant set includes a first intersection signal controller and a second intersection signal controller;

[0020] The actions in the action set are used to control the vehicle passing actions of the intersection;

[0021] The information set includes data indicating the traffic state of the intersection.

[0022] Optionally, the privacy information in the set of information includes at least one of real-time queue length, average delay time, and local saturation.

[0023] The second aspect of the present application provides a non-complete information game intelligent processing system based on secure multi-party computation, wherein the non-complete information game intelligent processing system based on secure multi-party computation comprises:

[0024] A data acquisition module is configured to acquire a set of participants of a game, an initial game state, a set of historical actions, a set of terminal histories, a preset game strategy based on a secure near-strategy optimization algorithm, and an action set, an information set, and a payoff function corresponding to each participant, wherein the set of historical actions is used to represent a sequence of game actions, the set of terminal histories is used to represent a complete game path, and data comparison in the game strategy is based on a secure multi-party computation protocol;

[0025] A model construction module is configured to construct a secure extended game model based on an extended game tree modeling framework according to the set of participants, the initial game state, the set of historical actions, the set of terminal histories, the game strategy, and the action set, the information set, and the payoff function corresponding to each participant, wherein a node in the secure extended game model is used to represent a game state, and an edge in the secure extended game model is used to represent a selected action during the game;

[0026] A game result acquisition module is configured to determine a game result between the participants and output the game result according to a preset optimization target and a leaf node in the secure extended game model.

[0027] The third aspect of the present application provides a terminal, which comprises a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program, when executed by the processor, implements the steps of any one of the non-complete information game intelligent processing methods based on secure multi-party computation.

[0028] The fourth aspect of the present application provides a computer readable storage medium, wherein the computer readable storage medium stores a computer program, and the computer program, when executed by a processor, implements the steps of any one of the non-complete information game intelligent processing methods based on secure multi-party computation.

[0029] From the above, in the scheme, the participating party set of the game, the initial game state, the historical action set, the terminal history set, the preset game strategy based on the secure near-end strategy optimization algorithm, and the action set, information set and payoff function corresponding to each participating party are obtained. The historical action set is used to represent the game action sequence, the terminal history set is used to represent the complete game path, and the data comparison in the game strategy is based on a secure multi-party computation protocol. Based on the participating party set, the initial game state, the historical action set, the terminal history set, the game strategy, and the action set, information set and payoff function corresponding to each participating party, a modeling framework based on an extended game tree is modeled to construct a secure extended game model. The nodes in the secure extended game model are used to represent game states, and the edges in the secure extended game model are used to represent selected actions in the game. According to the preset optimization target and the leaf nodes in the secure extended game model, the game results between the participating parties are determined and output.

[0030] Compared with the prior art, in the non-complete information game intelligent processing method based on secure multi-party computation provided by the application, the game strategy based on the secure near-end strategy optimization algorithm is used in the game process, and the secure multi-party computation protocol is used for data comparison, rather than directly providing all data of the participating parties to the third party platform for data comparison. When data comparison is performed based on the secure multi-party computation protocol, encryption processing of the data can be realized, thereby ensuring the security of the data and avoiding theft, attack or tampering of the original data. In this way, the data security is improved, and the security of the game process is improved. BRIEF DESCRIPTION OF DRAWINGS

[0031] In order to more clearly illustrate the technical solutions in the embodiments of the application, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description are only some embodiments of the application, and other drawings can be obtained by those skilled in the art without creative labor.

[0032] Figure 1 is a flowchart of a non-complete information game intelligent processing method based on secure multi-party computation provided by an embodiment of the application;

[0033] Figure 2 is a game tree structure diagram provided by an embodiment of the application;

[0034] Figure 3 is a component module diagram of a non-complete information game intelligent processing system based on secure multi-party computation provided by an embodiment of the application;

[0035] Figure 4 Figure 1 is a schematic diagram illustrating an internal structure of a terminal according to an embodiment of the present application. DETAILED DESCRIPTION

[0036] In the following description, for purposes of explanation and not limitation, specific details are set forth, such as particular architectures, techniques, etc., in order to provide a thorough understanding of the embodiments of the present application. However, it will be apparent to those skilled in the art that the present application can be practiced in other embodiments that depart from these specific details. In other instances, detailed descriptions of well-known methods, devices, circuits, and

[0037] It is to be understood that the terminology "includes", "has", "holds", "contains" and / or "comprising", when used in this specification, including the accompanying claims, includes the presence of said feature, item, step, operation, element, and / or component but does not exclude the presence or addition of one or more other features, items, steps, operations, elements, components and / or groups thereof.

[0038] It is also to be understood that the terminology used in the present specification is for the purpose of describing specific embodiments only and is not intended to be limiting. As used in this specification and the appended claims, the singular forms "a", "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise.

[0039] It will be further understood that the terms "and / or", as used in this specification, includes a combination of one or more of the associated listed items, as well as all possible combinations of the associated listed items.

[0040] As used in this specification and the appended claims, the term "if" can be construed to mean "when" or "once" or "in response to determining" or "in response to ascertaining", depending on the context. Similarly, the phrase "if it is determined" or "if it is ascertained" can be construed to mean "once it is determined" or "in response to determining" or "once it is ascertained" or "in response to ascertaining", depending on the context.

[0041] The technical solutions in the embodiments of the present application are clearly and completely described below with reference to the drawings of the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative effort fall within the scope of protection of the present application.

[0042] Many specific details are set forth in the following description in order to provide a thorough understanding of the application. However, the application can be practiced according to other embodiments that can not be described in detail herein, and the skilled person can make similar substitutions and modifications without departing from the scope of the application. The application is therefore not limited to the specific embodiments disclosed below.

[0043] At present, the application of machine game is more and more extensive. Game theory has become a powerful theoretical framework for modeling complex intelligent interactions between participants and optimizing their decision-making behavior, and the goal of each participant in these game scenarios is to maximize their own benefits. In recent years, with the rapid development of urbanization and intelligent transportation systems, game theory and machine game methods have also been increasingly applied to traffic signal control, road network coordination optimization and other scenarios to address the increasingly complex traffic flow dynamic scheduling problems.

[0044] According to whether the game information is completely visible, computer game can be divided into complete information game and imperfect information game. Complete information game (CIG, Complete Information Game) refers to the game problem in which the participants master all the information in the game process, including the accurate information of the game environment, the characteristics of all participants, the strategy space and the payoff function, etc. The participants in imperfect information game (IIG, Imperfect Information Game) are different, and the information they master has incompleteness, fuzziness, untrustworthiness, etc. Compared with complete information game, imperfect information game contains hidden information that is not visible to the game participants. This information asymmetry and incompleteness make imperfect information game more truly reflect the complex decision-making process, and the strategy solving research is more difficult and challenging.

[0045] The security problem in imperfect information game cannot be ignored. Due to the difference in the control of information by the participants in the game, there may be security risks such as information leakage, malicious manipulation or attack, which not only destroys the fairness of the game, but also may affect the reliability of the game result. Therefore, how to ensure the security of the game process while maintaining the optimization of the strategy has become another big challenge in imperfect information machine game.

[0046] In the prior art, the imperfect information game can be solved based on a preset solving algorithm, but when comparing data, the data of all participants needs to be directly provided to a third-party platform for processing, which is not conducive to improving data security, and thus is not conducive to improving the security of the game process.

[0047] Specifically, in some application scenarios, a method based on reinforcement learning (RL), a method based on game theory, and a hybrid method combining the two can be used to solve non-complete information games. The method based on reinforcement learning optimizes the strategy to cope with the challenge of incomplete information through repeated interaction between the agent and the environment. Although reinforcement learning shows strong adaptability in strategy optimization, it usually faces the problems of slow convergence speed and huge computational resources required for training. The method based on game theory provides theoretical guidance for the optimal strategy by deriving game equilibrium, and can make effective decisions in dynamic games with incomplete information. However, the game theory method usually assumes that the game participants are completely rational and relies on accurate game modeling, resulting in high computational complexity in some complex games. The combined method of reinforcement learning and game theory combines the advantages of both, uses the exploration ability of reinforcement learning and the equilibrium constraints of game theory, and effectively improves the efficiency and stability of strategy optimization. Through this method, the agent can efficiently explore the strategy space in complex non-complete information games while ensuring the convergence and optimality of the strategy.

[0048] However, when using existing solving algorithms to solve non-complete information games, there are certain security problems, such as dependence on third-party platforms. When these platforms are attacked or return tampered data, the solving algorithm cannot generate the optimal strategy, thereby affecting the overall effect of the game. During the game process, if the processing of information is not fully secured, malicious participants or third parties can exploit vulnerabilities to obtain hidden information and manipulate the game results to seek unfair advantages. Such security risks are particularly critical in environments that require privacy and fairness.

[0049] Based on this, the present application introduces security technology in the game process and performs data processing based on secure multi-party computation (MPC), to ensure the protection of sensitive information during the decision-making process and guarantee the security of the game process and the fairness of the competition among all parties.

[0050] Specifically, to solve at least one of the above technical problems, in the scheme, a set of participants of a game, an initial game state, a historical action set, a terminal history set, a preset game strategy based on a secure near-end strategy optimization algorithm, and an action set, an information set, and a payoff function corresponding to each participant are obtained, wherein the historical action set is used to represent a game action sequence, the terminal history set is used to represent a complete game path, and data comparison in the game strategy is based on a secure multi-party computation protocol; based on the set of participants, the initial game state, the historical action set, the terminal history set, the game strategy, and the action set, the information set, and the payoff function corresponding to each participant, a modeling framework based on an extended game tree is modeled to construct a secure extended game model, wherein nodes in the secure extended game model are used to represent game states, and edges in the secure extended game model are used to represent selected actions in the game; and based on a preset optimization target and leaf nodes in the secure extended game model, a game result between the participants is determined and output.

[0051] Compared with the prior art, in the non-complete information game intelligent processing method based on secure multi-party computation provided in the application, the game strategy based on the secure near-end strategy optimization algorithm is used in the game process, and the secure multi-party computation protocol is used for data comparison, instead of directly providing all data of the participants to the third-party platform for data comparison. When data comparison is performed based on the secure multi-party computation protocol, encryption processing of the data can be implemented, thereby ensuring the security of the data and avoiding theft, attack, or tampering of the original data. In this way, the data security is improved, and the security of the game process is further improved.

[0052] As shown in Figure 1 , the embodiment of the application provides a non-complete information game intelligent processing method based on secure multi-party computation, and specifically, the method comprises the following steps:

[0053] In step S100, a set of participants of a game, an initial game state, a historical action set, a terminal history set, a preset game strategy based on a secure near-end strategy optimization algorithm, and an action set, an information set, and a payoff function corresponding to each participant are obtained, wherein the historical action set is used to represent a game action sequence, the terminal history set is used to represent a complete game path, and data comparison in the game strategy is based on a secure multi-party computation protocol;

[0054] Step S200, according to the above set of participants, the above initial game state, the above set of historical actions, the above set of terminal histories, the above game strategy, and the action set, information set and payoff function corresponding to each of the above participants, modeling based on the modeling framework of the extended game tree, constructing a secure extended game model, wherein the nodes in the above secure extended game model are used to represent the game state, and the edges in the above secure extended game model are used to represent the selected action at the game time.

[0055] Step S300, according to the preset optimization target, and the leaf nodes in the above secure extended game model, determining the game result between the above participants and outputting.

[0056] Specifically, the set of participants includes at least two participants in the game. In the embodiments of the application, two participants are taken as an example for specific description, but the scheme of the application can also be applied to the scene of multi-party game, which is not specifically limited here.

[0057] Compared with the prior art, in the non-complete information game intelligent processing method based on secure multi-party computation provided by the application, the game strategy based on the secure near-end strategy optimization algorithm is used in the game process, and the secure multi-party computation protocol is used when comparing data, rather than directly providing all data of the participants to the third-party platform for data comparison. When comparing data based on the secure multi-party computation protocol, the encryption processing of the data can be realized, so as to protect the security of the data and avoid the original data being stolen, attacked or tampered with. In this way, it is beneficial to improve the data security, and further beneficial to improve the security of the game process.

[0058] In the embodiments of the application, the information set includes the privacy information of the participants.

[0059] The game strategy is applied to the protected node in the secure extended game model.

[0060] The protected node is a node for comparing the privacy information between the participants.

[0061] It should be noted that the information set is used to store the information corresponding to the participants, and the information set can include privacy information and non-privacy information. In the embodiments of the application, only the privacy information uses the game strategy that can perform secure multi-party computation, so as to reduce the complexity of data processing and improve the efficiency of data processing. In some application scenarios, the game strategy that can perform secure multi-party computation can also be used for the privacy information and non-privacy information in the information set, so as to further improve the security of data processing.

[0062] Specifically, when comparing data based on the game strategy, an encryption label is obtained according to the privacy information in the information set of the participants, and comparison is performed based on the encryption label to obtain a comparison result.

[0063] The encryption label is obtained according to the privacy information in the information set of the participants, and includes:

[0064] The corresponding privacy information of the participants is encoded to obtain an encoding vector.

[0065] The encoding vector is encrypted according to a preset security model through a preset scheduling system to obtain an encryption label.

[0066] The comparison based on the encryption label to obtain a comparison result includes:

[0067] A Boolean circuit for comparing the encryption label is generated based on a preconfigured logic circuit through the scheduling system, and a comparison result of the encryption label between the participants is obtained according to the Boolean circuit.

[0068] In this way, when performing a secure comparison, data comparison is performed based on the processing mode of secure multi-party computation, only necessary comparison results are output, and specific numerical values are not disclosed, which is beneficial to improve the security of data processing.

[0069] It should be noted that the non-complete information game intelligent processing method based on secure multi-party computation can be applied to a traffic scheduling scene to realize city intelligent traffic signal scheduling. Specifically, the participant set includes a first intersection signal controller and a second intersection signal controller, the actions in the action set are used to control the actions of vehicles passing through the intersection, and the information set includes data indicating the traffic state of the intersection. The privacy information in the information set includes at least one of real-time queue length, average delay time, and local saturation. The average delay time can be obtained by statistical calculation of vehicle passing time in a preset scheduling period, the real-time queue length can be obtained according to data collected by a lane sensor or can be estimated according to historical data, and the local saturation can be calculated according to the ratio of the number of vehicles passing through per unit time to the theoretical maximum passing capacity under the signal phase, reflecting the congestion degree and utilization efficiency of the current intersection or passing phase. The value can be estimated in combination with the flow data obtained by the sensor and the signal timing plan.

[0070] In the embodiments of the present application, the above-mentioned intelligent processing method for non-complete information game based on secure multi-party computation is also specifically described based on a specific application scenario. The above-mentioned intelligent processing method for non-complete information game based on secure multi-party computation provided in the embodiments of the present application can effectively prevent potential data tampering, information leakage and other attacks in the scheduling of urban intelligent traffic signal lights, and ensure the privacy and integrity of the key traffic flow data of each intersection or regional control unit in the whole process of collection, transmission and decision calculation. In this scenario, each control unit (for example, the first intersection signal controller and the second intersection signal controller) usually only masters the real-time traffic information of the region, such as the queue length, the average speed, the signal state of the adjacent intersection, etc., and these information collectively constitutes a non-complete information game environment. Each traffic control unit needs to consider the balance and stability of the overall road network while optimizing the intersection signal timing to maximize its own traffic efficiency under the condition of lacking global transparent data. With the help of the secure expansion game model and the secure proximal policy optimization (SPPO) method proposed in the present application, the collaborative decision among multiple nodes can be realized without exposing the private traffic and control preferences of each node, and the signal light control scheme close to the global optimum can be jointly solved. This not only helps to improve the overall operation efficiency of the urban road network and the robustness of adapting to sudden traffic changes, but also significantly enhances the security and credibility of the traffic signal system when facing potential network attacks or data pollution risks, thereby providing a practical technical path for non-complete information game and secure optimization in complex traffic networks.

[0071] Non-complete information game has the following technical problems in the processing process due to factors such as information hiding and strategy complexity: In the game process, the private information of the participants is easy to be maliciously stolen, which destroys the fairness of decision-making; the dependence on third-party platforms easily leads to manipulation risks, and attackers or compromised servers may tamper with strategies or results, thereby affecting the correctness and credibility of the game; the existing methods are difficult to meet the demand of large-scale application in terms of calculation and communication efficiency in complex game scenarios, especially when high efficiency and security are pursued simultaneously, the balance between the two is difficult to achieve. Therefore, it is necessary to efficiently solve the non-complete information game while ensuring the security of private information.

[0072] In the embodiments of the present application, firstly, a security extended game model is constructed, which includes a set of game participants, a strategy space of each participant and a corresponding payoff function, and can depict the strategy behavior and information interaction process of the participants, while fully considering the factors of incomplete information and interaction security; based on the constructed game model, a secure proximal policy optimization method is used to iteratively optimize the strategy space. This method combines the efficiency of the proximal policy optimization (PPO, Proximal Policy Optimization) algorithm in the strategy updating process, and introduces a secure multi-party computation protocol to realize the encryption of sensitive information in the strategy evaluation and updating, and prevent strategy leakage. Further, the TinyGarble2 framework can be combined to improve the communication and calculation efficiency in the security calculation process, so as to meet the performance requirements in a large-scale game environment.

[0073] In the strategy execution process, the information stealing and manipulation risks can be effectively resisted, and the safe solution of the non-complete information game is realized. Through the application of reinforcement learning and game theory, the efficiency and stability of the strategy solution are improved, and the correctness and fairness of the game result are ensured. Through verification in a typical non-complete information game environment, the proposed SPPO method significantly improves the average utility compared with the traditional method, and shows excellent strategy performance and system robustness. The method can be widely applied to complex game scenarios such as trusted resource scheduling involving privacy protection and security requirements.

[0074] In the embodiments of the present application, the comparison and calculation steps in the game process can be performed through secure multi-party computation, instead of relying on a third-party server. Even if a third-party server is involved, it cannot access the specific data of the participants, and it can only return the final encrypted result, thereby ensuring the security and fairness of the game. Specifically, a complete solution is provided in the present application for safely solving the strategy in the non-complete information game. The core concept is to protect the sensitive data of the participants in each comparison process, while allowing them to calculate the accurate game result.

[0075] In one application scenario, for game participants and , each has secret inputs and . These inputs are encrypted into and through secure multi-party computation, and further mapped into final sequences and , so as to ensure that the participants only know the correct output result without leaking the input information of the other party.

[0076] Specifically, when modeling the security extended game, the corresponding input information includes: a set of participants where, denotes the number of participants; the action set, information set and payoff function of each participant ; the initial game state, the set of historical actions and the set of terminal histories .

[0077] The incomplete information game is modeled as an extended game tree, where each node represents a decision point (for representing the game state), the edge represents the executable action (i.e. the action selected at the game), and the leaf node represents the terminal. By information partitioning the observable state of the participants is demarcated, and the information hiding modeling is achieved.

[0078] In the game process, the nodes that need to compare the privacy information (i.e. private information such as strategy strength, traffic flow indicators, etc.) of the participants are identified and marked as "protected computing points" so that the secure computing mechanism can be introduced subsequently.

[0079] In the intelligent traffic signal optimization scenario, each intersection or regional controller can be regarded as an independent participant in the game. They respectively master private traffic information such as queue length, vehicle speed, adjacent intersection signal state, etc. Through this modeling method, the privacy data of each node interaction and control logic can be effectively encapsulated while the interaction and control logic of each node are completely represented.

[0080] In the embodiments of the present application, the game strategy is realized by using the game strategy based on the secure proximal policy optimization algorithm to safely solve the incomplete information game strategy. Specifically, at the protected computing point, each participant encodes its private information (such as strategy score, traffic scheduling indicator) into an equal-length bit string and inputs it into the secure computing module. Then, the MPC comparison is embedded, the comparison logic is circuitized using the TinyGarble2 framework, the corresponding Boolean circuit is generated and the obfuscation coding is completed; each participant cooperatively participates in the secure multi-party computation protocol based on the obfuscated circuit, only exchanges encrypted labels, and does not leak the original input; finally, the comparison result , indicates whether it is greater than , the output result is transparent but the input is still in the ciphertext state, wherein, denotes the secure comparison function, denotes the private input of the participant , and denotes the private input of the participant .

[0081] It should be noted that the strategy used in the game process in the embodiments of the present application can be pre-trained based on the actual scene to realize the strategy update based on the actual scene and improve the performance of the strategy in the corresponding scene. In one application scenario, the strategy update process is as follows:

[0082] State observation: Participants observe the available game state based on their respective information sets

[0083] Action sampling: Action sampling is performed using the policy function in Proximal Policy Optimization (PPO).

[0084] Reward evaluation and gradient update: After completing a round of the game (reaching the terminal history ), the payoff is calculated based on the value of each participant at that state . The clip loss function in PPO and the Generalized Advantage Estimation (GAE) algorithm are applied to update the parameters of the policy network .

[0085] Multiple iterations of updates: The action sampling, reward evaluation, and gradient update processes are repeated continuously until the preset iteration termination condition is reached (e.g., reaching the preset number of iterations), allowing the agent to converge to a better strategy under the premise of ensuring information security, thereby improving its overall performance in the game of incomplete information.

[0086] In this application, a secure extended game model for games of incomplete information is constructed during the game process, which combines extended game theory and security comparison techniques. Through this model, traffic participating units (such as intersection signal controllers) can find solutions under the condition of incomplete information while ensuring the privacy and security of sensitive traffic data and control preferences involved in the scheduling process.

[0087] This model is based on an extended game structure and represents the traffic game process as a decision tree. Each node in the tree corresponds to a traffic control unit's decision point at a specific time, and the branches represent the available signal phase adjustments or timing actions.

[0088] Figure 2 is a game tree structure diagram provided by an embodiment of the present application, as shown in Figure 2 The secure extended game model for games of incomplete information described above includes the following key components:

[0089] A finite set of participants representing various intersection signal controllers in the city road network, each participant having its own traffic detection and control system. It should be noted that Figure 2 includes two participants as an example.

[0090] A finite sequence set ​​(History Action Set), recording possible traffic states and control decision sequences, where the empty sequence belongs to , and The prefix of each sequence in also belongs to Figure 2 Take , Figure 2 for example, represents the initial state, with no action taken; represents the left-turn traffic control action; represents the right-turn traffic control action; represents the combined action of right-turn followed by left-turn; represents the consecutive two right-turn actions. Set represents the terminal history (i.e., the history that is not the prefix of any other sequence), i.e., the converged state reached at a certain scheduling period, e.g., the traffic flow reaches stability after a number of periods. For a non-terminal history , i.e., the history sequence that has not reached the termination condition of the game or scheduling process, the available action set is defined as , e.g., the possible operation of green light extension, shortening, or keeping unchanged at a certain state.

[0091] A player function assigns an element in to each non-terminal history node, indicating which intersection controller makes the decision, where represents the natural player, i.e., the action selector determined by environmental factors rather than human control (e.g., random events, traffic fluctuations, etc.). For a non-terminal history , represents the player who performs the action after the history. If , it represents that the action after the history is determined by a random event (e.g., natural factors). In the above game tree, the nodes marked with represent decision points where the intersection controller makes the selection, and the nodes marked with represent decision points where the intersection controller makes the selection, with both making decisions alternately until the game ends.

[0092] A safety comparison function is used to compare the private indicators (e.g., real-time queue length, average delay time, local saturation, etc.) of different intersections without revealing their respective traffic data, and returns the comparison result for subsequent decision-making.

[0093] For each player , defines a payoff function , maps the set of terminal histories to the set of real numbers , which is used to measure the traffic performance under a certain scheduling result, such as the average vehicle delay time, the queue length or the intersection capacity. For a certain intersection controller (party) , its payoff function is , , which represents the utility value of the party under the terminal history (the larger the value, the better the scheduling effect, because the smaller the delay time and the queue length, the better, so the negative sign is taken in front). represents the weight parameter, which is used to adjust the influence degree of the average delay and the queue length on the payoff function, and needs to be set according to the actual application scenario. is the average vehicle delay time (unit: second) of the intersection in the whole scheduling period, which can be calculated by the vehicle passing time statistics. is the average queue length (unit: meter) of the intersection in the whole scheduling period, which can be estimated by the lane sensor or historical data.

[0094] It should be noted that Figure 2 in and represent the corresponding payoff values of the two parties (or controllers) at the termination node. For the leaf node, represents the payoff of party 1, represents the payoff of party 2. Figure 2 in represents the left-turn action, represents the right-turn action, represents the left-turn action, represents the right-turn action, wherein and both represent left turns, and different letters are used to distinguish the identity of the parties and emphasize the attribution of the action, and for the same reason.

[0095] To ensure the security of the sensitive traffic data (such as the accurate queue length of a certain road section or the control preference) involved in the traffic solution scheduling process in the interaction, in this application, a secure multi-party computation (MPC) mechanism is introduced at the nodes where private traffic indicators need to be compared. At these decision points, each intersection control unit cooperates to complete the comparison of key traffic indicators in the ciphertext state through a secure comparison protocol without revealing their original data.

[0096] This secure comparison can be represented as:

[0097] ;

[0098] where, represents the private input of the participant , and represents the private input of the participant . The comparison is performed based on a secure protocol, ensuring that only the comparison result is revealed, without exposing the original numerical values.

[0099] In actual scheduling processes, when two intersections need to perform signal timing comparisons based on their respective private traffic situations, traditional methods are prone to expose the original inputs. The present application scheme uses secure multi-party computation, enabling each party to jointly calculate the comparison function under ciphertext, and only output the necessary results, thereby ensuring information security. The specific process is as follows:

[0100] Private traffic information input: Each intersection controller encodes the key indicators such as queue length, traffic pressure, or green light demand observed locally to obtain an encoded vector, and converts it into an encrypted representation suitable for subsequent comparison through a secure protocol.

[0101] Key generation and interaction: The system generates an encrypted label for each encrypted input according to a pre-agreed security model (such as honest but curious or stronger malicious model), and securely transmits it between different controllers through a secure protocol.

[0102] Constructing comparison logic: The scheduling system generates a Boolean circuit for comparing different intersection encrypted inputs based on a pre-configured logic circuit, to determine the solution conditions such as "whether to prioritize increasing green light time for a certain intersection".

[0103] Encrypted execution and output: Circuit evaluation is completed under ciphertext, and only the necessary comparison results (such as which intersection has a more serious queue) are output, without revealing the specific numerical values.

[0104] Thus, the security game model of the application retains the concept of standard game theory solution while adding security. A privacy protection mechanism is further introduced to enhance security. The model effectively safeguards the privacy of private data of the participants and the fairness of game interaction in the entire game process by encrypting sensitive information in the decision-making process. Specifically, the system introduces a secure comparison mechanism based on secure multi-party computation (MPC) at key nodes of the game, uses a garbled circuit to compare the private inputs of the participants, and only reveals the necessary comparison results without revealing the original data. In addition, with the aid of technologies such as TinyGarble2, the model can achieve both security and computational efficiency in large-scale games, thereby preventing any party from illegally obtaining information of others by observing or manipulating the game process, and ensuring the credibility and robustness of the entire game environment. The fusion of the extended game structure and the encryption mechanism provides a unified and scalable secure computing framework for modeling and solving non-complete information games.

[0105] The security extended game model proposed in the embodiments of the application realizes accurate modeling of non-complete information games and identification of sensitive interaction points by introducing information sets, decision node division, a secure comparison mechanism, and a payoff function definition, thereby providing a complete and rigorous state representation space for subsequent policy optimization algorithms.

[0106] Based on the above security extended game model, the embodiments of the application further propose a secure proximal policy optimization method as a policy solving module. This method integrates a secure multi-party computation mechanism in the game policy optimization process, allowing game participants to safely perform policy comparison and updating without revealing their private information. By combining the structural expression capability of the extended model with the policy optimization advantages of the PPO algorithm, SPPO constructs a learning framework that realizes the continuous evolution of agents under the premise of privacy protection. The specific algorithm process is as follows:

[0107] To select the optimal strategy in the game, the system first encrypts the candidate strategies of the participants. The strategy of each party is converted into ciphertext form through an encryption protocol (such as TinyGarble2), ensuring that the content of each strategy remains private during the comparison process. Subsequently, the system uses a garbled circuit mechanism (SecureCompare) to perform secure comparison on the ciphertext, obtains the comparison result, and calculates the score corresponding to the strategy in the current environment. Next, the system traverses the strategy space, simulates and evaluates each possible strategy, and uses the CLIP loss function in PPO to evaluate its performance. If a strategy performs better than the optimal score recorded in the evaluation, the optimal strategy record is updated. Finally, the strategy with the highest score is returned as the optimal strategy selection in the current game round.

[0108] In the embodiments of the present application, each intersection controller acts as a game participant and has traffic state information that is invisible to each other, such as queue length, average delay time, etc. These information have important influence on the signal timing strategy of each self, but if directly disclosed, it may lead to privacy leakage or strategy manipulation. Therefore, in the process of strategy comparison and selection, a secure comparison module SecureCompare() is introduced to realize the information protection type comparison operation between participants. For this purpose, the present application adopts TinyGarble2 (TG2) framework as the basis for realizing secure comparison. TG2 is a high-efficiency garbled circuits (Garbled Circuits) implementation platform with good flexibility and security, which can be applied to privacy-sensitive computing tasks. It has the following advantages: supporting arbitrary combination logic function, can construct the corresponding comparison logic according to the traffic index. Introducing sequential logic processing mechanism, suitable for processing continuous period traffic data, meeting the demand of dynamic game. Low computing and communication overhead, can maintain efficient operation in city-level multi-intersection game tasks.

[0109] In actual execution process, the SecureCompare() module includes the following key steps:

[0110] Local input preparation: each intersection controller extracts an index value for comparison according to the traffic state (such as real-time queue length, green light waiting vehicle number, historical average delay, etc.) observed by itself. The value is encoded locally and converted into a bit string format as the subsequent input.

[0111] Garbled circuit generation (provided by TG2): the system will generate a Boolean comparison circuit according to the comparison logic (such as "whether the queue is more serious than the opponent"), and perform garbled processing in TG2. The generation process of garbled circuit is invisible to the outside and does not contain the original input information.

[0112] Label distribution and protocol execution: each participant obtains a set of ciphertext labels through the TG2 protocol, which correspond to the encrypted representation of its private input. Then the participants jointly run the garbled circuit protocol of TG2, only exchange encrypted labels without disclosing any original data.

[0113] Output decryption and use: after the protocol execution is completed, each participant can obtain a clear comparison result, such as "whether the queue length of participant 1 is greater than participant 2", but they still cannot know the specific value of the other party. This result is used for subsequent strategy judgment and game path selection.

[0114] The algorithm PPO.LCLIP() is partially implemented using a proximal policy optimization (PPO) algorithm. In the previously constructed non-complete information game environment model, the agent makes decisions based on interactions with the environment (Env). In each task, the agent analyzes the current environment state and selects an action from its policy space The policy space refers to a set of possible actions that the agent can choose in the current game state. These actions are selected by the policy network based on the probability distribution generated by the environment state. The policy space is represented by a neural network model, and its parameters depend on the state of the environment. From any state , the agent performs an action and enters the next environment state.

[0115] Specifically, the game policy network needs to be pre-trained to obtain the game policy to achieve game solving based on the game policy. In the embodiments of the present application, the PPO algorithm is used as the core training idea. Based on the parameterized policy idea, the main goal of the PPO algorithm adopted in the present application is to analyze various possible action spaces and evaluate which action space branch can maximize the advantage evaluation function score, and then make the probabilities of these action spaces larger, so that the policy is more likely to choose these actions. In this way, the optimal solution is constantly optimized and found. Therefore, in the process of policy training, the defined maximum objective function is as follows:

[0116] ;

[0117] wherein, represents a set of parameters of the policy network, represents a policy function determined by the parameters , represents a complete trajectory for representing “state-action-reward”, represents the total return value (cumulative income) of the trajectory , represents the probability of generating the trajectory under the policy parameters , represents the mathematical expectation operation on the trajectory reward under the trajectory distribution defined by the current policy.

[0118] In the traffic game framework constructed in the present application, the agent (Agent) represents each intelligent traffic signal control unit (such as an intersection controller) in the road network, and the environment (Env) represents the external traffic dynamics such as the traffic flow state, adjacent intersection feedback, and information set evolution under the safety game mechanism. In each round of signal decision process, the Agent selects the timing scheme of the next period according to the current observable local traffic information (such as queue length, vehicle speed, and adjacent intersection signal state), and the Env feeds back the updated road network state and the benefits brought by the delay or congestion relief, which constitutes the basis of the closed loop of strategy optimization. Among them, represents a signal timing trajectory generated by the continuous interaction of the intelligent traffic control unit and the environment. By summing all possible trajectories , the traffic evolution that may be caused by the interaction of Agent and Env can be systematically explored. The core of PPO is to gradually increase the probability of the occurrence of the timing trajectory that can obtain higher traffic efficiency reward through the dynamic update of the strategy network, so as to drive the intelligent signal control strategy to converge to a better scheduling solution under the premise of ensuring local and global coordination.

[0119] The specific strategy updating steps of the PPO algorithm are as follows:

[0120] Data collection: in the intelligent traffic signal scheduling scenario, the Agent (i.e. the signal control unit of each intersection) interacts with the traffic environment by executing the current signal timing strategy, and collects data such as state (such as queue length, vehicle speed, and adjacent intersection timing state), action (selected green period and phase scheme), obtained traffic benefits, and next state in real time.

[0121] Advantage estimation: in the non-complete information traffic game model established in the present application, each signal period completion can be regarded as an action execution, and then the advantage function value under the current state can be calculated according to the feedback information of the traffic environment, which is used to evaluate the advantage degree of the adopted timing strategy compared with the benchmark.

[0122] Optimization objective function: by using , the proportion of different strategies can be distinguished according to the size of the advantage function to select the next step.

[0123] Strategy update: the gradient ascent method is used to update the strategy parameters . Among them, represents the learning rate, i.e. the step length of each parameter update, represents the gradient of the loss function with respect to the parameter , which reflects the direction of the current strategy improvement, represents the loss function.

[0124] Repeat step: constantly sample the traffic state trajectory in the iteration process, calculate the advantage and update the strategy until the advantage value reaches a dynamic balance or the number of iterations reaches a preset threshold, so as to obtain a stable traffic signal control strategy.

[0125] Further, the intelligent processing of game information can be realized based on the stable traffic signal control strategy obtained by training (i.e., the game strategy based on the safe proximal policy optimization algorithm).

[0126] Therefore, in the embodiments of the present application, a safe extended game model is proposed, which systematically formalizes the behaviors and information flow of the participants in the incomplete information game, and provides a safe theoretical framework for solving the game strategy. By introducing a multi-party secure computing mechanism, the strategies and decision-making processes of the participants are embedded into the secure computing framework, so that the participants can ensure the security of their private data even in complex game interactions. By limiting and restricting the information flow of the participants, the model can effectively avoid unfair competition caused by privacy leakage. Not only does it improve the security and efficiency of game solving, but it also provides a more universal and robust solution for incomplete information games.

[0127] The intelligent processing strategy solving method for incomplete information game based on secure multi-party computation proposed in the present application is the core technology to solve the problem of incomplete information game strategy solving. This method innovatively combines the efficiency of proximal policy optimization (PPO) algorithm and the privacy protection ability of secure multi-party computation (MPC), and embeds an encryption computing mechanism in the process of strategy solving to ensure the data security in the process of strategy solving.

[0128] By combining secure multi-party computation technology with PPO algorithm, the present application can effectively prevent any party from obtaining the strategy data of another party during the strategy optimization process. Especially in a complex game environment, the present application optimizes the generation and transmission process of the encryption circuit using the TinyGarble2 technical framework, greatly reducing the computational complexity and communication overhead, so that even in a high-dimensional complex strategy space, the strategy solving can be completed in an efficient and secure manner. Through the integration of these technologies, not only excellent performance is exhibited in experiments, but also a scalable, efficient and safe universal solution for incomplete information game solving is provided.

[0129] It should be noted that in the embodiments of the present application, the multi-party secure computing mechanism is used to ensure the data security during the game process. In actual application, other secure computing protocols can also be used to improve the calculation efficiency and reduce the communication overhead, which are not limited here.

[0130] It needs to be further explained that in the embodiments of the present application, the PPO algorithm and the MPC technology are combined. In actual application process, other game optimization algorithms can also be used to adapt to different types of game scenarios. For example, for specific game environment, other policy updating methods in reinforcement learning can be combined, such as A3C (Asynchronous Advantage Actor-Critic) algorithm, which is not specifically limited herein.

[0131] As shown in Figure 3 Corresponding to the above-mentioned intelligent processing method for non-complete information game based on secure multi-party computation, the embodiments of the present application also provide an intelligent processing system for non-complete information game based on secure multi-party computation, which comprises:

[0132] The data acquisition module 310 is configured to acquire a set of game participants, an initial game state, a set of historical actions, a set of terminal histories, a preset game strategy based on secure proximal policy optimization algorithm, and an action set, an information set and a payoff function corresponding to each participant, wherein the set of historical actions is used to represent a game action sequence, the set of terminal histories is used to represent a complete game path, and data comparison in the game strategy is based on a secure multi-party computation protocol;

[0133] The model construction module 320 is configured to model based on an extended game tree modeling framework according to the set of participants, the initial game state, the set of historical actions, the set of terminal histories, the game strategy, and the action set, the information set and the payoff function corresponding to each participant, and construct a secure extended game model, wherein the nodes in the secure extended game model are used to represent game states, and the edges in the secure extended game model are used to represent selected actions in the game;

[0134] The game result acquisition module 330 is configured to determine the game results between the participants according to a preset optimization target and the leaf nodes in the secure extended game model and output.

[0135] In this way, the game strategy based on the secure proximal policy optimization algorithm is used in the game process, and the secure multi-party computation protocol is used for data comparison, rather than directly providing all the data of the participants to the third-party platform for data comparison. When data comparison is performed based on the secure multi-party computation protocol, encryption processing of the data can be realized, thereby ensuring the security of the data and avoiding the original data from being stolen, attacked or tampered with. In this way, it is beneficial to improve the data security, and further beneficial to improve the security of the game process.

[0136] It should be noted that the specific structure and implementation of the above-mentioned non-complete information game intelligent processing system based on secure multi-party computation and each module or unit thereof can refer to the corresponding description in the above method embodiment, which will not be repeated here.

[0137] It should be noted that the division of each module of the above-mentioned non-complete information game intelligent processing system based on secure multi-party computation is not unique and is not specifically limited here.

[0138] Based on the above embodiments, the present application also provides a terminal, the principle block diagram of which can be as shown in Figure 4 The terminal includes a processor, a memory, a network interface and a display screen connected through a system bus. The processor of the terminal is used to provide computing and control capabilities. The memory of the terminal includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operating system and the computer program in the non-volatile storage medium to run. The network interface of the terminal is used to communicate with external terminals through network connection. The computer program is executed by the processor to implement the steps of any one of the above-mentioned non-complete information game intelligent processing methods based on secure multi-party computation. The display screen of the terminal can be a liquid crystal display screen or an electronic ink display screen.

[0139] Those skilled in the art can understand, Figure 4 The principle block diagram shown in the above-mentioned figure is only a block diagram of part of the structure related to the scheme of the present application, and does not constitute a limitation on the terminal to which the scheme of the present application is applied. The specific terminal can include more or fewer components than those shown in the figure, or combine certain components, or have a different component arrangement.

[0140] In one embodiment, a terminal is provided, which includes a memory, a processor, and a computer program stored on the memory and executable on the processor. When the computer program is executed by the processor, the steps of any one of the non-complete information game intelligent processing methods based on secure multi-party computation provided by the embodiments of the present application are implemented.

[0141] The embodiments of the present application also provide a computer readable storage medium, which stores a computer program. When the computer program is executed by a processor, the steps of any one of the non-complete information game intelligent processing methods based on secure multi-party computation provided by the embodiments of the present application are implemented.

[0142] It should be understood that the serial numbers of the steps in the above-mentioned embodiments do not mean the order of execution. The execution order of each process should be determined according to its function and inherent logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0143] Those skilled in the art can clearly understand that, for the convenience and brevity of description, only the division of the above functional units and modules is taken as an example for description, and in actual application, the above functions can be completed by different functional units and modules according to needs, that is, the internal structure of the apparatus is divided into different functional units or modules to complete all or part of the functions described above. Each functional unit and module in the embodiment can be integrated in one processing unit, or each unit can exist physically, or two or more units can be integrated in one unit, and the integrated unit can be realized in the form of hardware or in the form of software functional unit. In addition, the specific names of each functional unit and module are only for convenient distinction, and do not limit the protection scope of the application. The specific working process of the units and modules in the apparatus can refer to the corresponding process in the foregoing method embodiments, which will not be described here.

[0144] In the above embodiments, the description of each embodiment has its own emphasis, and the parts not described or recorded in a certain embodiment can be referred to the related description of other embodiments.

[0145] Those of ordinary skill in the art can realize that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be realized in electronic hardware or in combination of computer software and electronic hardware. Whether the functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different ways to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0146] In the embodiments provided in the present application, it should be understood that the disclosed system / terminal device and method can be implemented in other ways. For example, the above-described system / terminal device embodiments are only schematic, for example, the division of the above modules or units is only a logical function division, and actual implementation can be in another division manner, for example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed.

[0147] The above integrated modules / units, if realized in the form of software function units and sold or used as independent products, can be stored in a computer readable storage medium. Based on such understanding, all or part of the processes in the above embodiment methods can also be completed by a computer program instructing related hardware, and the above computer program can be stored in a computer readable storage medium. When the processor executes the computer program, the steps of the above various method embodiments can be implemented. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or some intermediate forms, etc. The computer readable medium can include any entity or device capable of carrying the above computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, and software distribution medium, etc. It should be noted that the content contained in the above computer readable storage medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction.

[0148] The above embodiments are only used to illustrate the technical solutions of the present application, rather than limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand; it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for part of the technical features; and these modifications or replacements do not deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application.

Claims

1. A method for intelligent processing of non-complete information game based on secure multi-party computation, characterized in that, The method comprises: obtaining a set of participants of a game, an initial game state, a set of historical actions, a set of terminal histories, a preset game strategy based on a secure proximal policy optimization algorithm, and an action set, an information set and a payoff function corresponding to each participant, wherein the set of historical actions is used to represent a game action sequence, the set of terminal histories is used to represent a complete game path, and data comparison in the game strategy is based on a secure multi-party computation protocol; based on the set of participants, the initial game state, the set of historical actions, the set of terminal histories, the game strategy, and the action set, the information set and the payoff function corresponding to each participant, modeling is performed based on an extended game tree modeling framework to construct a secure extended game model, wherein nodes in the secure extended game model are used to represent game states, and edges in the secure extended game model are used to represent selected actions during the game; determining the game results between the participants and outputting according to a preset optimization target and leaf nodes in the secure extended game model; wherein, when data comparison is performed based on the game strategy, an encryption label is obtained according to privacy information in the information set of the participant, and comparison is performed based on the encryption label to obtain a comparison result; the set of participants includes a first intersection signal controller and a second intersection signal controller; the actions in the action set are used to control vehicle traffic actions corresponding to the intersection; the information set includes data indicating the traffic state of the intersection.

2. The non-complete information game intelligent processing method based on secure multi-party computation according to claim 1, characterized in that, The information set includes privacy information of the participants; the game strategy is applied to a protected node in the secure extended game model; the protected node is a node used to compare privacy information between the participants. 3.The security multi-party computation based non-complete information game intelligent processing method according to claim 1, characterized in that, The encryption label is obtained according to the privacy information in the information set of the participant, comprising: the participant encodes the corresponding privacy information to obtain an encoding vector; through a preset scheduling system, the encoding vector is encrypted according to a preset security model to obtain an encryption label.

4. The non-complete information game intelligent processing method based on secure multi-party computation according to claim 3, characterized in that, The comparison result is obtained by comparing the encryption labels, comprising: a Boolean circuit for comparing the encryption labels is generated based on a preconfigured logic circuit through the scheduling system, and a comparison result of the encryption labels between the participants is obtained based on the Boolean circuit.

5. The non-complete information game intelligent processing method based on secure multi-party computation according to claim 3, characterized in that, The privacy information in the information set includes at least one of real-time queue length, average delay time and local saturation.

6. A secure multi-party computation based intelligent processing system for non-complete information game, characterized in that, The system comprises: a data acquisition module for acquiring a set of participants of a game, an initial game state, a set of historical actions, a set of terminal histories, a preset game strategy based on a secure proximal policy optimization algorithm, and an action set, an information set and a payoff function corresponding to each participant, wherein the set of historical actions is used to represent a game action sequence, the set of terminal histories is used to represent a complete game path, and data comparison in the game strategy is based on a secure multi-party computation protocol; The model construction module is configured to construct a secure extended game model based on an extended game tree modeling framework according to the set of participants, the initial game state, the set of historical actions, the set of terminal histories, the game strategy, and the action set, information set, and payoff function corresponding to each participant, wherein a node in the secure extended game model represents a game state, and an edge in the secure extended game model represents a selected action at a game time; The game result acquisition module is configured to determine and output a game result between the participants according to a preset optimization target and a leaf node in the secure extended game model. When data comparison is performed based on the game strategy, an encryption label is obtained according to privacy information in the information set of the participant, and comparison is performed based on the encryption label to obtain a comparison result. The set of participants includes a first intersection signal controller and a second intersection signal controller. The actions in the action set are used to control corresponding vehicle traffic actions at an intersection. The information set includes data used to indicate a traffic state at an intersection.

7. A terminal, characterized by comprising: The terminal includes a memory, a processor, and a computer program stored on the memory and executable on the processor, and the computer program implements the steps of the secure multi-party computation based non-complete information game intelligent processing method according to any one of claims 1 to 5 when executed by the processor.

8. A computer-readable storage medium, characterized in that, The computer program is stored on the computer readable storage medium, and the computer program implements the steps of the secure multi-party computation based non-complete information game intelligent processing method according to any one of claims 1 to 5 when executed by the processor.

Images