Distributed account book asynchronous communication optimization method and device based on Nash equilibrium game and topological sorting

Abstract

The invention provides a distributed account book asynchronous communication optimization method and equipment based on a Nash equilibrium game and topological sorting. The method comprises the following steps: in response to a resource allocation request, solving Nash equilibrium based on a distributed optimal reaction iterative algorithm under a chain so as to obtain equilibrium bid and expected utility of each broker in a specific strategy space; determining a winning agent and locking a target resource according to a four-in-one rule including bid standard reaching, utility maximization, utility difference standard reaching and qualification compliance; constructing a business process directed acyclic graph by using a topological sorting algorithm, and calculating a process priority of each task node based on a critical path method; and calculating a dynamic comprehensive priority in combination with a Nash priority generated by a game and a process priority, and further driving a workflow engine to execute task asynchronous scheduling. According to the scheme, the problem of data consistency in a high-concurrency scene is effectively solved, and the resource utilization rate and the system throughput are remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of digital technology for real estate transaction processes, and more specifically, to a distributed ledger asynchronous communication optimization method and device based on Nash equilibrium game theory and topological sorting. Background Technology

[0002] With the digital transformation of the real estate industry, SaaS (Software as a Service) platforms have become the core hub connecting brokers, customers, developers, and financial institutions, supporting the entire business process from property listing to contract signing and loan processing. However, most existing real estate collaborative systems are built on traditional centralized architectures, which are increasingly revealing their technical limitations when facing asynchronous transaction processes involving multiple roles and long cycles.

[0003] Regarding the synchronization of property listing status, centralized servers are prone to performance bottlenecks under high-concurrency scenarios, resulting in a significant time lag between data updates and distributed clients. This delay not only causes "false existence" of property information but also creates serious resource conflicts at the business level, such as multiple agents simultaneously locking the same property or viewing time slot, which is essentially a "double-spending" problem in a distributed environment.

[0004] Meanwhile, real estate transaction processes are highly asynchronous and involve complex logical dependencies. Existing collaborative models often rely on point-to-point manual triggering or simple sequential flows, lacking modeling of the deep topological relationships between task nodes. When cross-institutional collaboration is involved, such as the parallel processing of bank loan approval and title deed verification, the system often fails to identify the critical path of the business, resulting in a large amount of redundant waiting time in the process.

[0005] A deeper problem lies in the resource allocation mechanism. Currently, most platforms adopt a "first-come, first-served" or simple weighted rules. This rudimentary allocation logic completely ignores the game-theoretic balance between brokers' service capabilities and customer needs, failing to achieve optimal allocation of overall benefits in a dynamic environment of multi-party competition. Therefore, how to ensure high data consistency while improving collaboration efficiency through intelligent game theory and path optimization in a distributed asynchronous communication environment has become a key technological bottleneck that urgently needs to be overcome in the field of real estate digitalization. Summary of the Invention

[0006] The present invention aims to solve at least one of the aforementioned technical problems existing in the prior art.

[0007] To this end, the first aspect of the present invention provides an optimization method for asynchronous communication of distributed ledgers based on Nash equilibrium game theory and topological sorting.

[0008] A second aspect of the present invention provides an electronic device.

[0009] This invention provides an optimization method for asynchronous communication in a distributed ledger based on Nash equilibrium game theory and topological sorting. It is applied to a system comprising a distributed ledger, a workflow engine, and an application service layer, for handling asynchronous collaborative processes such as property viewings, contract signing, and loan processing. The method includes: In response to resource allocation requests for target properties or viewing times, the application service layer automatically selects qualified agents from the qualification service through the API interface to form a participant set, and sets initial game parameters based on property popularity and agent historical performance. Off-chain, Nash equilibrium game calculations are performed based on a distributed optimal response iterative algorithm. The bidding strategies of each participant are updated through multiple rounds of iteration until the change in each participant's bid is less than a preset convergence threshold, thereby obtaining the equilibrium bid combination of each participant and the corresponding expected utility. The indicator function value of each participant is calculated based on the expected utility, and the winning agent is selected by combining the equilibrium bid combination, and the initial allocation plan of the target property or viewing time is generated. For real estate business processes involving cross-role collaboration, the workflow engine decomposes them into multiple task nodes and constructs a directed acyclic graph. It then uses a topological sorting algorithm to determine the task execution sequence and calculate the process priority of each task node. The expected utility of the winning broker is mapped to Nash priority, and a dynamic comprehensive priority is obtained by weighting and combining it with the process priority. This drives the workflow engine to perform asynchronous task scheduling and distribution based on the priority. After the allocation scheme is generated, a smart contract writes status information including resource identifier, winning broker identifier, lock status bit and version number to the distributed ledger. Before executing asynchronous business operations, each distributed node reads and verifies the lock status bit. When a lock status is detected and the ledger version number is higher than the local cache, resource double-spending protection under asynchronous communication is achieved by rejecting the local operation request.

[0010] The distributed ledger asynchronous communication optimization method based on Nash equilibrium game and topological sorting according to the above technical solution of the present invention may also have the following additional technical features: In the above technical solution, the execution of Nash equilibrium game calculation includes: The system obtains the quality score of the target property through the data acquisition layer, and reads the weight parameters, bidding cost coefficient and upper and lower limits of the bid from the system configuration service, and the initial bid is input by each participant node; Each participating node exchanges its bid for the current round through an asynchronous communication module and updates its bid for the next round using a gradient ascent method. The update method includes:

[0011] in, This represents broker i's bid in round k; Indicates the learning rate; This represents the sum of the bids from other participants in round k-1; R represents the overall value of agent i, which is obtained by weighted summation based on indicators from different dimensions, including property quality, agent performance, and client matching; R represents the fixed income upon winning. This represents broker i's (k-1)th bid; This represents the bid cost coefficient; Determine if the change in bids of all participants is less than the convergence threshold. If so, terminate the iteration and output the current set of bids as the Nash equilibrium point. Otherwise, repeat the iteration.

[0012] In the above technical solution, the indicator function value of each participant is calculated based on the expected utility, and the winning agent is selected by combining the equilibrium bidding combination. An initial allocation scheme for the target property or viewing time slot is then generated, including: The indicator function value of each participant is determined according to the "four-in-one judgment rule". Specifically, the indicator function value of a participant is determined to be valid when the participant's equilibrium bid is not lower than the platform's preset minimum bid threshold, the bid can maximize its expected utility, the difference between its expected utility and the utility of the second-ranked participant is not lower than the minimum utility difference threshold, and the participant's real-time qualification verification is compliant. Otherwise, the indicator function value of the participant is determined to be invalid.

[0013] In the above technical solution, the method for calculating the expected utility includes:

[0014] in, This represents the expected utility of broker i under equilibrium bidding conditions; This represents the winning probability calculated based on the Tullock competition model, which is equal to the ratio of the participant's bid to the sum of all participants' bids; This represents the equilibrium bid of broker i.

[0015] In the above technical solution, during the calculation of the Nash equilibrium game, a parameter adaptive adjustment mechanism to prevent dead loops is introduced: when the system detects that the change in the bid of the game participants within a consecutive preset number of rounds is less than the convergence threshold but does not meet the uniqueness determination condition, the game is determined to have entered a deadlock state. The system automatically increases the bid cost coefficient or introduces a random disturbance term with limited amplitude based on the current bid, so as to force the game process to converge to a determinable Nash equilibrium solution within a finite number of rounds.

[0016] In the above technical solution, if the same agent wins in the game of multiple housing resources and a time period conflict occurs, the system identifies the expected utility value of the conflicting tasks, retains one task in descending order of expected utility, and releases the remaining conflicting resources; for the released resources, the second-ranked agent is automatically redistributed according to the expected utility ranking of the original participants, without the need to re-trigger the game calculation.

[0017] In the above technical solution, the process priority calculation for each task node includes: Analyze the logical dependencies between nodes in the real estate transaction process and construct a directed acyclic graph; the real estate transaction process nodes include at least one of the following: property listing, customer matching, viewing arrangement, reaching an agreement, contract signing, loan processing, and property transfer; The earliest start time and latest start time of each task node are calculated using the critical path method, and the difference between the two is taken as the total float of that task node. Mapping process priorities based on the total float of task nodes includes:

[0018] in, The process priority of task node n; This represents the total float of task node n; This represents the maximum total float across all task nodes.

[0019] In the above technical solution, the method for calculating the dynamic synthesis priority includes:

[0020] in, Indicates the dynamic synthesis priority; This represents the normalized result of the winning broker's expected utility; This indicates a weight that is dynamically adjusted based on the urgency of the process; when the total float of the critical path tasks approaches 0, the system reduces... To enhance the impact of process priority on task scheduling.

[0021] In the above technical solution, the distributed ledger adopts a hybrid architecture of off-chain computation and on-chain evidence storage. The Nash equilibrium game calculation and topology sorting are both executed off-chain at the application layer. The distributed ledger only confirms consensus on the finalized Nash equilibrium result and changes in resource lock status, while ordinary business scheduling notifications are transmitted non-blockingly through a distributed message queue.

[0022] The present invention provides an electronic device, including a memory and a processor. The memory stores a computer program that can run on the processor. When the processor executes the program, it implements the distributed ledger asynchronous communication optimization method based on Nash equilibrium game and topological sorting as described in any of the above technical solutions.

[0023] In summary, due to the adoption of the above-mentioned technical features, the beneficial effects of the present invention are: This invention significantly improves the operational efficiency of the real estate digital collaborative system in multiple dimensions.

[0024] At the business collaboration level, this invention reduces the average completion time of a real estate transaction by approximately 35% by accurately modeling the topological dependencies of each stage. Thanks to the resource state locking mechanism provided by the distributed ledger, this solution successfully avoids business conflicts caused by redundant resource allocation in an asynchronous communication environment, reducing the resource allocation conflict rate by 80% and effectively preventing "double-spending" conflicts such as "selling one property multiple times" or duplicate use of viewing slots due to message delays. This strong consistency guarantee improves the system's service availability to 99.99%, significantly enhancing the platform's robustness in handling asynchronous collaboration among brokers, clients, and financial institutions.

[0025] In terms of system performance, by introducing a Nash equilibrium-based resource allocation and topology sorting scheduling mechanism, this technical solution improves system throughput by more than three times compared to traditional architectures, stably supporting the processing needs of over 10,000 concurrent requests per second. Due to the optimization of asynchronous communication logic, the system's average response time is significantly reduced from seconds to milliseconds, while overall resource utilization is increased from 40% to over 85%, greatly maximizing the computing potential of the infrastructure.

[0026] Furthermore, the hybrid architecture of off-chain computation and on-chain evidence storage adopted in this invention cleverly solves the performance bottleneck problem of blockchain consensus algorithms in high-frequency business scenarios. By limiting the PBFT (Practical Byzantine Fault Tolerance) consensus algorithm to only critical state changes such as finalizing the Nash equilibrium result, resource locking, and state version upgrades, the system maintains extremely high non-blocking message transmission efficiency while ensuring high data consistency across multiple nodes. This architectural design, combined with distributed deployment and failover mechanisms, gives the system excellent elasticity and scalability, enabling it to easily handle sudden traffic peaks such as the opening of popular real estate projects. The automated processes significantly reduce manual intervention, achieving a leapfrog improvement in the system's intelligence and reliability.

[0027] Additional aspects and advantages of the invention will become apparent in the following description or may be learned by practice of the invention. Attached Figure Description

[0028] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the description of the embodiments taken in conjunction with the following drawings, in which: Figure 1 This is a flowchart of an embodiment of the distributed ledger asynchronous communication optimization method based on Nash equilibrium game and topological sorting according to one of the present invention; Figure 2 This is an application system architecture diagram of an embodiment of the distributed ledger asynchronous communication optimization method based on Nash equilibrium game and topology sorting according to one of the present invention. Figure 3 This is a flowchart of the Nash equilibrium game model construction in an embodiment of the distributed ledger asynchronous communication optimization method based on Nash equilibrium game and topological sorting according to an embodiment of the present invention. Figure 4 This is a schematic diagram of resource allocation based on game results in a distributed ledger asynchronous communication optimization method based on Nash equilibrium game and topological sorting according to an embodiment of the present invention. Figure 5 This is a flowchart of the topology sorting optimization algorithm in an embodiment of the distributed ledger asynchronous communication optimization method based on Nash equilibrium game and topology sorting according to an embodiment of the present invention. Figure 6 This is a flowchart of distributed asynchronous communication and state synchronization in a distributed ledger asynchronous communication optimization method based on Nash equilibrium game and topological sorting according to an embodiment of the present invention. Figure 7 This is a sequence diagram illustrating the workflow of an asynchronous communication optimization method for distributed ledgers based on Nash equilibrium game theory and topological sorting, according to an embodiment of the present invention. Detailed Implementation

[0029] To better understand the above-mentioned objectives, features, and advantages of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in these embodiments can be combined with each other.

[0030] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and therefore the scope of protection of the invention is not limited to the specific embodiments disclosed below.

[0031] The following reference Figures 1 to 7 This paper describes an asynchronous communication optimization method for distributed ledgers based on Nash equilibrium game theory and topological sorting, according to some embodiments of the present invention.

[0032] Some embodiments of this application provide an optimization method for asynchronous communication of distributed ledgers based on Nash equilibrium game theory and topological sorting.

[0033] like Figure 1 As shown, the first embodiment of this invention proposes an asynchronous communication optimization method for distributed ledgers based on Nash equilibrium game theory and topological sorting. This method is applied to a system comprising a distributed ledger, a workflow engine, and an application service layer, for handling asynchronous collaborative processes such as property viewings, contract signing, and loan processing. (Refer to...) Figure 2 The illustrated embodiment presents a schematic diagram of the overall system architecture. This invention is deployed within a multi-layered distributed system. The underlying infrastructure layer consists of multiple computing nodes and network resources, providing computing power support for the entire platform. Above this, the data support layer includes distributed caching, a Kafka message queue, and a database cluster, responsible for data temporary storage and asynchronous message transmission under high concurrency environments. The application service layer encapsulates core business logic such as property listing services, customer listing services, and transaction services, and interfaces with the unified access layer through an API gateway to respond to requests from external brokers or customers. The core of the system is the engine layer, composed of a Nash equilibrium game engine and a topology sorting workflow engine. These two components work together to achieve optimal resource allocation and dynamic scheduling of business processes.

[0034] Continue reading Figure 1 The method disclosed herein includes the following steps S1 to S7.

[0035] S1. In response to a resource allocation request for a target property or viewing time slot, the application service layer automatically selects brokers who meet the admission criteria from the qualification service through the API interface to form a participant set, and sets the initial game parameters based on the property popularity and the broker's historical performance.

[0036] Reference Figure 4 The resource allocation diagram shown is as follows: Figure 7 The workflow sequence diagram shows that when a client submits a viewing appointment or resource request through the API gateway, the system first enters the access verification phase. The application service layer automatically filters out a set of qualified participants from the agent qualification service, such as requiring agents to have a service rating of no less than 80 points in the past 30 days and no violation records. At this time, the system reads game environment parameters such as the property quality rating Q, initial weight parameters, and upper and lower limits of bids from the configuration service.

[0037] More specifically, the final participant set is usually kept to a minimum of 20 people to balance the intensity of competition and the efficiency of iteration.

[0038] The strategy space in this scheme defines the behavioral boundaries of each broker. Each broker's bid B_i is a non-negative real number, expressed in platform points, and must strictly fall within the interval [Bmin, Bmax]. The minimum bid Bmin is dynamically set by the platform based on the property type; for example, 200 points for popular properties and 100 points for ordinary properties. The maximum bid Bmax is determined by the real-time available points limit returned by the broker's points account service. This strategy space design ensures that the game process conforms to market competition principles while being effectively constrained by the broker's actual asset capabilities.

[0039] In some embodiments, in addition to the original time and resource constraints, detailed business rules constraints are added. Regarding business rules, the system strictly limits the interval between bids by the same agent for the same property to be greater than or equal to 10 minutes, and the number of times a single agent participates in property bidding per day cannot exceed 5. Furthermore, the bid amount cannot exceed the expected commission conversion points calculated by the property valuation service (which can be calculated by the "property valuation service," e.g., an expected commission of 10,000 yuan corresponds to a maximum of 500 points).

[0040] S2. Off-chain, Nash equilibrium game calculation is performed based on the distributed optimal response iterative algorithm. The bidding strategies of each participant are updated through multiple rounds of iteration until the change in each participant's bid is less than the preset convergence threshold, thereby obtaining the equilibrium bid combination of each participant and the corresponding expected utility.

[0041] Figure 3 This illustrates the specific process of the Nash equilibrium game engine performing the solution off-chain. The system first calculates the combined value of each broker:

[0042] Among them, the property quality rating Broker performance rating And customer matching degree All are provided in real time by the corresponding microservices, while the quality of the property listings is weighted. Agent performance weighting And customer matching weight Then, adjustments will be made dynamically based on the actual situation.

[0043] Specifically, the weighting of property quality is dynamically allocated based on property popularity, with more popular properties receiving higher weighting. This weighting can be dynamically adjusted based on the platform's "Property Popularity Index" (daily page views, inquiry volume), typically ranging from 0.3 to 0.5; Property Quality Score The weighting is based on factors such as apartment type (20%), location (30%), price (30%), and amenities (20%), typically ranging from 1 to 10, and can be calculated in real-time using the "Property Information Service". For agent performance, higher performance results in higher weighting, incentivizing top agents. The weighting is determined by factors such as the agent's 30-day closing rate (40%), customer satisfaction (30%), and viewing conversion rate (30%), typically ranging from 0.2 to 0.4. Agent performance scores are standardized platform scores reflecting service capabilities, updated daily by the "Agent Performance Service", typically ranging from 1 to 10. For customer matching weighting, higher matching results in higher weighting, increasing the probability of a sale. The weighting is based on the degree of matching between customer needs and property attributes, typically ranging from 0.2 to 0.3. Customer matching is calculated based on historical communication records and demand tags, and can be generated in real-time by the "Customer Demand Matching Service".

[0044] It should be noted that the specific algorithm for the comprehensive value of broker i described above is merely illustrative and can be flexibly adjusted by those skilled in the art according to needs and actual circumstances. Its main purpose is to quantify the competitiveness of brokers, that is, the comprehensive value of broker i. The higher the value, the stronger its competitiveness.

[0045] In some embodiments, this disclosure employs a multi-objective weighted approach, comprehensively considering property quality, agent performance, client matching degree (the overall value of agent i), and bidding costs, and determines the optimal bidding strategy by maximizing expected utility. Specifically, the method for calculating expected utility includes:

[0046] in, This represents the expected utility of broker i under equilibrium bidding conditions; This represents the winning probability calculated based on the Tullock competition model, which is equal to the ratio of the participant's bid to the sum of all participants' bids; R represents the overall value of agent i, which is obtained by weighted summation based on indicators from different dimensions, including property quality, agent performance, and client matching; R represents the fixed reward for winning (set by the platform based on the property value, such as 5-15 points). This represents the equilibrium bid of broker i; This represents the bid combination of all participants except broker i; This represents the bidding cost coefficient, used to balance bidding incentives and costs, avoiding blindly high prices, especially when resources are scarce. Increase and curb blindly high-price competition; when resources are abundant, Lower the price to encourage reasonable bidding.

[0047] Based on the above winning probability model, it can be guaranteed that the higher the bid, the greater the winning probability. The sum of all participants' bids determines the intensity of competition. Even if the bid is low, there is still a possibility of winning (avoiding the extreme case of "all or nothing").

[0048] In some embodiments, the core form of the Nash equilibrium is the "optimal combination of bids from all brokers," meaning that no broker can increase their expected utility by adjusting their bids individually. This is solved using a distributed optimal response iterative algorithm. Specifically, the Nash equilibrium game calculation includes: The system obtains the quality score of the target property through the data acquisition layer, and reads the weight parameters, bidding cost coefficient, and upper and lower limits of the bid from the system configuration service. Each participating node inputs its initial bid. Each participating node exchanges its bid for the current round through the asynchronous communication module, and updates the bid for the next round using the gradient ascent method. The update method includes:

[0049] in, This represents broker i's bid in round k; This represents the learning rate, which can be between 0.1 and 0.3. This represents the sum of the bids from other participants in round k-1; This represents broker i's (k-1)th bid; The gradient is used; when the calculated bid in the kth round exceeds the boundary range, the boundary value should be taken.

[0050] Determine if the change in bids of all participants is less than the convergence threshold (which can be 0). If so, terminate the iteration and output the current set of bids as the Nash equilibrium point (theoretically, the optimal bid is reached when the gradient is zero). Otherwise, repeat the iteration. Understandably, the iteration should also terminate when the maximum number of iterations (e.g., 50) has been reached and the result has not converged.

[0051] It should be noted that during the iteration process, each distributed node exchanges the bids of all participants in the current round through the "asynchronous communication module" (based on the message queue Kafka).

[0052] In the distributed optimal response iteration process, if the system repeatedly performs game calculations under the same parameter configuration, a deadlock state may occur in which the strategies of multiple participants remain unchanged for a long time and the game results cannot be further distinguished.

[0053] Therefore, in some embodiments, during the calculation of the Nash equilibrium game, a parameter adaptive adjustment mechanism to prevent dead loops is introduced: when the system detects that the change in the bids of the game participants within a consecutive preset number of rounds is less than the convergence threshold but does not meet the uniqueness determination condition, the game is determined to have entered a deadlock state, and the system automatically increases the bid cost coefficient (increases the marginal cost of high bid behavior) or introduces a random disturbance term with limited amplitude based on the current bid, so as to force the game process to converge to a determinable Nash equilibrium solution within a finite number of rounds.

[0054] Understandably, the criteria for determining a deadlock in a game can also include: the ranking of participants' utility remaining unchanged across multiple iterations and no substantial change in resource allocation. In addition to introducing parameters to prevent loops, the deadlock can also be broken by adjusting the learning rate parameter and changing the policy update step size.

[0055] Based on the equilibrium bids of each broker at the determined Nash equilibrium point, the expected utility is calculated using the method described in the above embodiment.

[0056] S3. Calculate the indicator function value of each participant based on the expected utility, combine the equilibrium bid combination to select the winning agent, and generate the initial allocation plan for the target property or viewing time.

[0057] In some embodiments, determining the winning broker and generating an initial allocation scheme based on the expected utility includes: The indicator function value of each participant is determined according to the "four-in-one judgment rule". Specifically, the indicator function value of a participant is determined to be valid when the participant's equilibrium bid is not lower than the platform's preset minimum bid threshold, the bid can maximize its expected utility, the difference between its expected utility and the utility of the second-ranked participant is not lower than the minimum utility difference threshold, and the participant's real-time qualification verification is compliant. Otherwise, the indicator function value of the participant is determined to be invalid.

[0058] Specifically, such as Figure 4 As shown, after iterative convergence, the system determines the indicator function through a four-in-one judgment rule. For an agent to win, they must simultaneously meet four conditions: the equilibrium bid is not lower than the minimum threshold, expected utility is maximized, the difference in utility with the second-place agent is not lower than the minimum utility difference threshold (e.g., 0.5, to avoid allocation disputes caused by "tied first place"), and real-time qualification compliance (agent i currently has no abnormal states such as "viewing time conflict," "expired qualification," or "not cleared violation records"). The indicator function value is 1 (winning). If any condition is not met, the indicator function value is 0 (not winning).

[0059] For brokers with an indicator function value of 1, trigger the resource allocation service and lock the target resource; for brokers with an indicator function value of 0, trigger the points refund service.

[0060] If the same agent wins multiple property listings simultaneously, resulting in time slot conflicts (such as simultaneously obtaining two viewing resources in the same time slot), the system will identify the expected utility value of each task through the task conflict detection service, retain the best resource in descending order, and release the remaining conflict items. The released resources will be directly allocated to the agent ranked second according to their original utility, effectively avoiding the overhead of duplicate calculations.

[0061] The following illustrates a specific example for calculating expected utility and indicator function: Parameter values: Property quality Q = 9.2 (popular school district property); weights α, β, γ are 0.4, 0.3, 0.3 respectively; cost coefficient δ = 0.06 (scarce resources); bid limits Bmin = 200, Bmax = 500; fixed income R = 12; learning rate λ = 0.2.

[0062] Participants: Agent 1: , , ; Agent 2: , , ; Iterative process: Round 0 (Initial): , ; Round 1: Agent 1: Gradient = 280 × (6.896 + 12) / (300 + 280)^2 - 0.06 ≈ 0.0158; ; Agent 2: Gradient = 300 × (6.575 + 12) / (300 + 280)^2 - 0.06 ≈ 0.0166; ; ...(After approximately 20 iterations)... Round 20 (convergence): , The change in bid ΔB < the convergence threshold ε; The Nash equilibrium point for the two agents is determined to be (305.2, 285.1). Expected utility calculation: Agent 1: ; Agent 2: ; Indicator function determined: Agent 1: Meets all 4 conditions → Indicator function value is 1 (wins), obtains a property in the allocation results.

[0063] Agent 2: Utility difference = 1.51 - 0.86 = 0.65 > 0.5 → Indicator function value is 0 (no win), failed to obtain a property, points are refunded.

[0064] S4. For real estate business processes involving cross-role collaboration, the workflow engine decomposes them into multiple task nodes and constructs a directed acyclic graph. The topological sorting algorithm is used to determine the task execution sequence and calculate the process priority of each task node.

[0065] In some embodiments, calculating the process priority of each task node includes: Analyze the logical dependencies between nodes in the real estate transaction process and construct a directed acyclic graph; the real estate transaction process nodes include at least one of the following: property listing, customer matching, viewing arrangement, reaching an agreement, contract signing, loan processing, and property transfer; The earliest start time and latest start time of each task node are calculated using the critical path method, and the difference between the two is taken as the total float of that task node. Mapping process priorities based on the total float of task nodes includes:

[0066] in, The process priority of task node n; This represents the total float of task node n; This represents the maximum total float across all task nodes.

[0067] S5. Map the expected utility of the winning broker to a Nash priority, and combine it with the process priority to obtain a dynamic comprehensive priority, thereby driving the workflow engine to perform asynchronous task scheduling and distribution according to the priority.

[0068] In some embodiments, the method for calculating the dynamic synthesis priority includes:

[0069] in, Indicates the dynamic synthesis priority; This represents the normalized result of the winning broker's expected utility. Specifically, linear normalization can be used to map the winning broker's expected utility value at the equilibrium point to the [0,1] interval. When the number of participants is large or the difference in utility is small, softmax normalization can be used to enhance the distinguishability. This indicates a weight that is dynamically adjusted based on the urgency of the process; when the total float of the critical path tasks approaches 0, the system reduces... To enhance the impact of process priority on task scheduling.

[0070] This disclosure presents a process modeling and optimization system based on the principle of directed acyclic graphs (DAGs). First, the real estate transaction process is decomposed into multiple task nodes, clarifying the dependencies between nodes to form a DAG. Sequence constraints and parallel execution conditions are defined for each node (e.g., "contract signing" depends on "agreement reached," and "loan processing" and "property title search" can be executed in parallel). A topological sorting algorithm (Kahn's algorithm) is used to determine the task execution sequence, critical path method is used to identify process bottlenecks, and a dynamic priority adjustment mechanism optimizes the overall process execution efficiency. An asynchronous communication mechanism is established, task scheduling is implemented based on message queues, and an event-driven architecture is used to achieve loosely coupled communication, supporting breakpoint resume and state recovery functions.

[0071] Design a resource allocation algorithm to perform optimal resource allocation based on Nash equilibrium calculations and handle conflicts by combining utility value ranking. Develop a process optimization algorithm to optimize the critical path based on topology ranking results, and dynamically adjust execution priorities through a dynamic comprehensive priority scheduling algorithm (combining Nash equilibrium allocation priorities and critical path process priorities). Establish a performance optimization mechanism to dynamically adjust algorithm parameters by monitoring the system's operating status in real time, and possess the ability to re-trigger the Nash equilibrium game for rescheduling in the event of resource conflicts, ensuring the system is always in an optimal operating state.

[0072] Specifically, such as Figure 5 As shown, the topology sorting workflow engine is responsible for constructing a directed acyclic graph (DAG) for the transaction process involving property listings, viewing arrangements, and contract signing. The system uses the Kahn algorithm to generate the optimal execution sequence and combines it with the Critical Path Method (CPM) to identify process bottlenecks. The total float is calculated by determining the earliest / latest start time of each task. It should be noted that the Critical Path Method (CPM) is well-known to those skilled in the art and can be calculated using either forward or backward methods, so it will not be elaborated here. Tasks with a float of 0 constitute the critical path that determines the overall efficiency. For nodes such as loan processing, the system also has a timeout handling strategy, such as triggering compensation measures or manual intervention if the timeout is 3 days; compensation measures can also be set (such as rolling back to the "intent reached" node for reconfirmation if "contract signing fails") to ensure the robustness of the process.

[0073] Reference Figure 6 The system adopts an event-driven architecture, records every resource allocation and scheduling decision event through an event sourcing pattern, and uses a distributed message queue to achieve loosely coupled asynchronous interaction between components, ensuring that state changes can be broadcast to all relevant nodes efficiently and reliably.

[0074] S6. After the allocation scheme is generated, a smart contract writes status information including resource identifier, winning broker identifier, lock status bit and version number to the distributed ledger. Before executing asynchronous business operations, each distributed node reads and verifies the lock status bit. When a lock status is detected and the ledger version number is higher than the local cache, resource double-spending protection under asynchronous communication is achieved by rejecting the local operation request.

[0075] The distributed ledger adopts a hybrid architecture of off-chain computation and on-chain evidence storage. The Nash equilibrium game calculation and topology sorting are both executed off-chain at the application layer. The distributed ledger only confirms consensus on the finalized Nash equilibrium result and changes in resource lock status, while ordinary business scheduling notifications are transmitted non-blockingly through a distributed message queue.

[0076] Specifically, to avoid resource conflicts in an asynchronous communication environment, the system adopts a hybrid architecture of off-chain computation and on-chain notarization. Nash equilibrium and topology sorting are performed off-chain, with PBFT consensus confirmation and writing to the ledger only triggered for the finalized result and changes in resource lock status. The resource status bits recorded in the ledger include ID, winner, version number, and timestamp. Each node must verify the ledger before operation. If it detects that a resource is locked and the version number is higher than the local cache, it rejects the local request, thus achieving consistency assurance similar to double-spending protection. This strategy avoids complex algorithms from entering smart contracts, significantly improving the system's throughput in high-frequency scenarios.

[0077] Regarding security and access control, this disclosure employs a multi-factor authentication method for identity verification, combining digital certificates (client certificates), biometric recognition (such as facial recognition), and dynamic verification codes to ensure the authenticity of participants' identities. Brokers must complete dual verification of "account password + facial recognition" when logging in, while financial institutions (such as banks) must pass certificate authentication when accessing the system.

[0078] Regarding access control policies, this disclosure adopts a role-based access control (RBAC) model for fine-grained permission management. It defines data access permissions (e.g., agents can only view the property data they are responsible for, while administrators can view all data) and operation permissions (e.g., agents can initiate viewing appointments, and title specialists can sign contracts), and supports dynamic permission adjustments (e.g., agents automatically gain "team property allocation" permission after being promoted).

[0079] For data security, this disclosure employs end-to-end encryption technology (such as TLS 1.3) to protect data transmission security and uses the AES-256 encryption algorithm to protect static data (such as customer information and contract documents in the database). A data backup and disaster recovery mechanism is established, adopting a "local backup + off-site disaster recovery" model to ensure that data can be recovered within one hour after loss.

[0080] For auditing and compliance, all operation logs and status change history are recorded (e.g., "Agent 1 modified the bid for property A to 330.5 points at 10:00 on 2024-06-10"). The log information is immutable and synchronized to the distributed ledger. Operation traceability and auditing are supported, and compliance reports that comply with the Personal Information Protection Law can be generated.

[0081] Other embodiments of the present invention provide an electronic device, including a memory and a processor, wherein the memory stores a computer program executable on the processor, and the processor executes the program to implement the distributed ledger asynchronous communication optimization method based on Nash equilibrium game and topological sorting as described in any of the above technical solutions.

[0082] In this specification, the illustrative expressions of the terms used do not necessarily refer to the same embodiments or examples. Moreover, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0083] Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this invention shall be included within the scope of protection of this invention.

Claims

1. A distributed ledger asynchronous communication optimization method based on Nash equilibrium game and topological sorting, characterized in that, This method is applied to systems containing distributed ledgers, workflow engines, and application service layers to handle asynchronous collaborative processes such as property viewings, contract signing, and loan processing; the methods include: In response to resource allocation requests for target properties or viewing times, the application service layer automatically selects qualified agents from the qualification service through the API interface to form a participant set, and sets initial game parameters based on property popularity and agent historical performance. Off-chain, Nash equilibrium game calculations are performed based on a distributed optimal response iterative algorithm. The bidding strategies of each participant are updated through multiple rounds of iteration until the change in each participant's bid is less than a preset convergence threshold, thereby obtaining the equilibrium bid combination of each participant and the corresponding expected utility. The indicator function value of each participant is calculated based on the expected utility, and the winning agent is selected by combining the equilibrium bid combination, and the initial allocation plan of the target property or viewing time is generated. For real estate business processes involving cross-role collaboration, the workflow engine decomposes them into multiple task nodes and constructs a directed acyclic graph. It then uses a topological sorting algorithm to determine the task execution sequence and calculate the process priority of each task node. The expected utility of the winning broker is mapped to Nash priority, and a dynamic comprehensive priority is obtained by weighting and combining it with the process priority. This drives the workflow engine to perform asynchronous task scheduling and distribution based on the priority. After the allocation scheme is generated, a smart contract writes status information including resource identifier, winning broker identifier, lock status bit and version number to the distributed ledger. Before executing asynchronous business operations, each distributed node reads and verifies the lock status bit. When a lock status is detected and the ledger version number is higher than the local cache, resource double-spending protection under asynchronous communication is achieved by rejecting the local operation request.

2. The distributed ledger asynchronous communication optimization method based on Nash equilibrium game and topological sorting according to claim 1, characterized in that, The execution of the Nash equilibrium game calculation includes: The system obtains the quality score of the target property through the data acquisition layer, and reads the weight parameters, bidding cost coefficient and upper and lower limits of the bid from the system configuration service, and the initial bid is input by each participant node; Each participating node exchanges its bid for the current round through an asynchronous communication module and updates its bid for the next round using a gradient ascent method. The update method includes: in, This represents broker i's bid in round k; Indicates the learning rate; This represents the sum of the bids from other participants in round k-1; R represents the overall value of agent i, which is obtained by weighted summation based on indicators from different dimensions, including property quality, agent performance, and client matching; R represents the fixed income upon winning. This represents broker i's (k-1)th bid; This represents the bid cost coefficient; Determine if the change in bids of all participants is less than the convergence threshold. If so, terminate the iteration and output the current set of bids as the Nash equilibrium point. Otherwise, repeat the iteration.

3. The distributed ledger asynchronous communication optimization method based on Nash equilibrium game and topological sorting according to claim 1, characterized in that, Calculate the indicator function value for each participant based on the expected utility, select the winning agent by combining the equilibrium bid combination, and generate an initial allocation plan for the target property or viewing time slot, including: The indicator function value of each participant is determined according to the "four-in-one judgment rule". Specifically, the indicator function value of a participant is determined to be valid when the participant's equilibrium bid is not lower than the platform's preset minimum bid threshold, the bid can maximize its expected utility, the difference between its expected utility and the utility of the second-ranked participant is not lower than the minimum utility difference threshold, and the participant's real-time qualification verification is compliant. Otherwise, the indicator function value of the participant is determined to be invalid.

4. The distributed ledger asynchronous communication optimization method based on Nash equilibrium game and topological sorting according to claim 2, characterized in that, The method for calculating the expected utility includes: in, This represents the expected utility of broker i under equilibrium bidding conditions; This represents the winning probability calculated based on the Tullock competition model, which is equal to the ratio of the participant's bid to the sum of all participants' bids; This represents the equilibrium bid of broker i.

5. The distributed ledger asynchronous communication optimization method based on Nash equilibrium game and topological sorting according to claim 2, characterized in that, During the Nash equilibrium game calculation, an adaptive parameter adjustment mechanism to prevent dead loops is introduced: when the system detects that the change in bids of the game participants within a preset number of consecutive rounds is less than the convergence threshold but does not meet the uniqueness determination condition, the game is determined to be in a deadlock state. The system automatically increases the bid cost coefficient or introduces a random disturbance term with limited amplitude based on the current bid, so as to force the game process to converge to a determinable Nash equilibrium solution within a finite number of rounds.

6. The distributed ledger asynchronous communication optimization method based on Nash equilibrium game and topological sorting according to claim 1, characterized in that, If the same agent wins in the game of multiple housing resources and a time period conflict occurs, the system identifies the expected utility value of the conflicting tasks, retains one task in descending order of expected utility, and releases the remaining conflicting resources; for the released resources, the second-ranked agent is automatically redistributed according to the expected utility ranking of the original participants, without the need to re-trigger the game calculation.

7. The distributed ledger asynchronous communication optimization method based on Nash equilibrium game and topological sorting according to claim 1, characterized in that, The process priority for calculating each task node includes: Analyze the logical dependencies between nodes in the real estate transaction process and construct a directed acyclic graph; the real estate transaction process nodes include at least one of the following: property listing, customer matching, viewing arrangement, reaching an agreement, contract signing, loan processing, and property transfer; The earliest start time and latest start time of each task node are calculated using the critical path method, and the difference between the two is taken as the total float of that task node. Mapping process priorities based on the total float of task nodes includes: in, The process priority of task node n; This represents the total float of task node n; This represents the maximum total float across all task nodes.

8. The distributed ledger asynchronous communication optimization method based on Nash equilibrium game and topological sorting according to claim 7, characterized in that, The method for calculating the dynamic synthesis priority includes: in, Indicates dynamic synthesis priority; This represents the normalized result of the winning broker's expected utility; This indicates a weight that is dynamically adjusted based on the urgency of the process; when the total float of the critical path tasks approaches 0, the system reduces... To enhance the impact of process priority on task scheduling.

9. The distributed ledger asynchronous communication optimization method based on Nash equilibrium game and topological sorting according to claim 1, characterized in that, The distributed ledger adopts a hybrid architecture of off-chain computation and on-chain evidence storage. The Nash equilibrium game calculation and topology sorting are both executed off-chain at the application layer. The distributed ledger only confirms consensus on the finalized Nash equilibrium result and changes in resource lock status, while ordinary business scheduling notifications are transmitted non-blockingly through a distributed message queue.

10. An electronic device comprising a memory and a processor, characterized in that, The memory stores a computer program that can run on the processor, and when the processor executes the program, it implements the distributed ledger asynchronous communication optimization method based on Nash equilibrium game and topological sorting as described in any one of claims 1 to 9.

Images