A trip cost optimization and risk control integrated platform and method

Abstract

The application discloses a trip cost optimization and risk control integrated platform and method. The application obtains trip intention and environment data through a data acquisition module; a topological construction module converts a candidate path sequence into a directed acyclic graph data structure and calculates a time buffer margin; a quantitative evaluation module performs cascade calculation of total risk premium by using a reverse topological sorting sequence; a joint optimization decision module fuses procurement cost and total risk premium to output an optimal path sequence, triggers control flow switching logic based on a dynamic safety threshold, and matches a standby hedging node and establishes an option mapping data structure in memory to lock resources; a state machine control and execution module controls deterministic finite state machine transition based on real-time risk and time-out conditions, and issues a refund and change and a ticket instruction according to a termination state. The application realizes objective quantification of trip interruption risk and low-cost resource pre-hedging.

Description

Technical Field

[0001] This invention relates to the field of computer information processing technology, and more specifically, to an integrated platform and method for travel cost optimization and risk control. Background Technology

[0002] As businesses expand and cross-regional operations become more frequent, travel management systems have become an indispensable information infrastructure for daily operations. The core requirement of corporate travel management is to rationally coordinate employees' travel plans and effectively control the online procurement costs of travel resources such as air tickets, hotels, and train tickets while ensuring the smooth operation of core businesses.

[0003] Existing mainstream travel booking platforms typically rely on static rule engines to perform cost control and itinerary scheduling. When employees submit travel applications, the system mainly filters compliant itinerary options based on the company's preset travel standards and recommends them to users simply according to the ticket purchase price from low to high. For changes in the external environment encountered during the trip, the system usually only relies on independent early warning notices issued by meteorological departments or transportation hubs. When actual flight delays or route interruptions occur, the travel system then transfers the abnormal orders to human customer service or business personnel for temporary intervention to perform passive refunds or rebooking operations.

[0004] However, existing travel management models suffer from significant deficiencies in risk response mechanisms, exhibiting both lag and isolation. Real-world travel itineraries often involve multiple geographical transfers and closely linked business processes. Delays at upstream transportation nodes can directly lead to the failure of subsequent business nodes. Existing technologies can only provide reactive responses to emergencies involving a single mode of transportation or isolated nodes, failing to systematically and objectively quantify the risk of itinerary disruptions caused by multiple environmental factors with time delays and spatial linkages during the booking phase. This scheduling scheme, lacking proactive risk transmission calculations, can easily lead to itineraries with seemingly low unit purchase prices incurring high emergency reset costs due to associated disruptions during actual execution, resulting in hidden loss of control over corporate travel expenditures and severe stagnation of core business operations. Summary of the Invention

[0005] The technical problem to be solved by this invention is that existing technologies can only respond passively to emergencies at single, isolated travel nodes after the fact, and cannot objectively quantify and pre-calculate the risks of travel interruptions with time delays and spatial linkages during the travel booking stage. This can easily lead to high emergency reset costs for enterprises and hidden loss of control over travel expenses. This invention provides an integrated platform and method for travel cost optimization and risk control.

[0006] The first aspect of this invention provides an integrated platform for travel cost optimization and risk control, comprising: a data acquisition module for acquiring travel intention data and multi-source heterogeneous environment data; a topology construction module for converting candidate path sequences into a directed acyclic graph data structure and calculating the time buffer margin between entity event nodes; a quantitative evaluation module for calculating the volatility parameters, basic risk premium, and total risk premium of each entity event node based on the multi-source heterogeneous environment data; a joint optimization decision module for integrating the ticket purchase cost of the entity event nodes and the total risk premium, calculating a comprehensive evaluation cost to output the optimal path sequence, and triggering control flow switching logic based on a dynamic safety threshold; a virtual option pool management module for matching backup hedging nodes and establishing an option mapping data structure in an in-memory database to complete resource locking when the control flow switching logic is triggered; and a state machine control and execution module for controlling the state transitions of a deterministic finite state machine based on real-time risk characteristic values ​​and timeout conditions, and issuing refund / rescheduling instructions and payment / ticketing instructions based on the termination state of the deterministic finite state machine.

[0007] To achieve cascading risk propagation calculations, the topology construction module, when constructing the directed acyclic graph (DAG) data structure, establishes attribute fields for each entity event node, including the geographical location of occurrence, estimated start time, estimated end time, and business importance weight. The business importance weight is obtained by converting the business type label corresponding to the entity event node through a preset business weight mapping relationship. The topology construction module retrieves all path sequences connecting the preceding entity event node to the following entity event node in the DAG, calculates the path time buffer margin for each path sequence, and extracts the path time buffer margin with the smallest value as the cumulative time buffer margin. This cumulative time buffer margin is used to characterize the most stringent risk propagation constraint boundary between the preceding and following entity event nodes.

[0008] When quantifying risk assessment, the quantitative assessment module calculates the total risk premium by summing the basic risk premium of each entity event node with the cascaded risk surcharges fed back by all downstream entity event nodes. The cascaded risk surcharge of a single downstream entity event node is obtained by multiplying the basic risk premium of the corresponding downstream entity event node, the business importance weight of the corresponding downstream entity event node, and the exponential decay term calculated based on the time decay constant and the cumulative time buffer margin. The quantitative assessment module uses a depth-first search algorithm to generate the inverse of the directed acyclic graph. The topological sorting sequence is used, and the total risk premium is calculated by adding the above-mentioned basic risk premium sequentially from back to front according to the reverse topological sorting sequence. The calculation of the basic risk premium specifically includes: obtaining the volatility parameters of the entity event node, calculating the blocking probability through the logistic regression function, wherein the sensitivity adjustment coefficient and risk offset benchmark constant included in the logistic regression function are obtained by performing a regression fitting algorithm on the historical travel blocking sample database; calculating the emergency reset cost by combining the immediate reset resource price, ticket purchase cost and default penalty base, and multiplying the emergency reset cost by the blocking probability to obtain the basic risk premium.

[0009] During joint optimization, the joint optimization decision module extracts candidate path sequences that satisfy the spatiotemporal continuity constraint from the candidate path sequence set, calculates the sum of the node comprehensive costs of each entity event node in the candidate path sequence as the comprehensive evaluation cost, and extracts the candidate path sequence with the smallest comprehensive evaluation cost as the optimal path sequence. The joint optimization decision module calculates the dynamic safety threshold by dividing the enterprise's basic risk tolerance by the business importance weight of the entity event node. When the total risk premium of the entity event node is greater than the dynamic safety threshold, the entity event node is marked as a high-risk master node, and a hedging trigger instruction is generated and distributed to the virtual option pool management module.

[0010] The virtual option pool management module calculates alternative time windows based on the expected start and end times of the high-risk master node and preset travel tolerance parameters, and obtains multiple candidate backup nodes that are within the alternative time windows and meet spatial displacement constraints. After eliminating candidate backup nodes whose total risk premium is greater than or equal to the backup determination threshold, the virtual option pool management module finds the node with the minimum sum of resource locking cost and total risk premium in the remaining set of valid backup nodes and establishes it as the final backup hedging node. The virtual option pool management module sends a standard reservation instruction containing seat reservation requests but not payment settlement request parameters to the global distribution system interface to obtain passenger reservation records and the latest ticketing time limit. The virtual option pool management module constructs an option mapping data structure in the memory database containing master node identifier fields, backup node identifier fields, reservation record fields, ticketing time limit fields, and status enumeration fields, initializes the status enumeration fields to the initial monitoring state values, and establishes a data locking mapping relationship between the high-risk master node and the backup hedging node.

[0011] When performing status monitoring and mutual exclusion control, the virtual option pool management module calculates the timeout condition using the following logic: It determines the time difference between the expected start time of the high-risk master node and the current system time, or the time difference between the latest ticket issuance deadline and the current system time. If either of these values ​​is less than or equal to a preset time tolerance constant, the timeout condition is deemed to be met. When the timeout condition is met, the status enumeration field is updated to the near-expiration trigger state value, and a resource near-expiration determination signal is generated. The state machine control and execution module instantiates a deterministic finite state machine represented by a quintuple. When the input event is the resource near-expiration determination signal,... When the time-sensitive event of the parameter is triggered and the deterministic finite state machine is in a safe or warning state, the deterministic finite state machine transitions to the release state, and the state machine control and execution module sends a resource cancellation instruction for the passenger reservation record; when the input event is the time-sensitive event and the deterministic finite state machine is in the blocking state, the deterministic finite state machine transitions to the final blocking execution state, and the state machine control and execution module starts the distributed transaction control engine based on the two-phase commit protocol, sends a refund and change execution instruction to the master node resource supplier system, and sends the full payment and ticketing instruction of the backup hedging node to the global distribution system interface.

[0012] A second aspect of this invention provides an integrated method for optimizing travel costs and controlling risks, comprising the following steps:

[0013] Step S100: Acquire travel intent data and multi-source heterogeneous environment data through the data acquisition module;

[0014] Step S200: The candidate path sequence is converted into a directed acyclic graph data structure through the topology construction module, and the time buffer margin between entity event nodes is calculated; the volatility parameter, basic risk premium and total risk premium of each entity event node are calculated based on the multi-source heterogeneous environment data through the quantitative evaluation module.

[0015] Step S300: By integrating the ticket purchase cost of the entity event node with the total risk premium through the joint optimization decision module, calculate the comprehensive evaluation cost to output the optimal path sequence, and trigger the control flow switching logic based on the dynamic security threshold.

[0016] Step S400: When the control flow switching logic is triggered by the virtual option pool management module, a backup hedging node is matched and an option mapping data structure is established in the memory database to complete resource locking;

[0017] Step S500: The state machine control and execution module controls the state transition of the deterministic finite state machine based on the real-time risk characteristic value and timeout condition, and issues a refund / rescheduling instruction and a payment / ticketing instruction according to the termination state of the deterministic finite state machine.

[0018] The present invention, by adopting the above technical solution, can bring the following beneficial effects:

[0019] 1. This invention constructs a directed acyclic graph data structure to perform topological mapping on candidate path sequences and calculates the cumulative time buffer margin between entity event nodes. At the same time, it uses a depth-first search algorithm to generate a reverse topological sorting sequence and sequentially performs the calculation of the total risk premium including the exponential decay term from back to front. This allows for the joint calculation of the blocking probability of a single node and the cascading risk added value of downstream nodes, thereby realizing an objective quantitative assessment and pre-blocking prevention of the risk of trip interruption with time delay and spatial linkage.

[0020] 2. This invention calculates a dynamic safety threshold by combining the enterprise's basic risk tolerance and business importance weights to trigger the control flow switching logic. When triggered, the virtual option pool management module extracts multiple candidate backup nodes that meet the spatial displacement constraints for cost and risk screening. Then, a standard reservation instruction containing only the reservation request is sent to the global distribution system interface, and an option mapping data structure is established in the memory database. This enables low-cost pre-locking and dynamic replacement of backup hedging resources for high-risk master nodes without actual capital settlement.

[0021] 3. This invention controls the state transitions of a deterministic finite state machine by using a state machine control and execution module based on real-time risk characteristic values ​​and timeout conditions including the expected start time and the latest ticket issuance deadline. When the transition reaches the final blocking execution state, a distributed transaction control engine based on a two-phase commit protocol is launched to strictly verify the readiness state of the interface between the supplier system and the distribution system. This achieves the control effect of ensuring the atomicity and mutual exclusion of the primary node's refund and change instructions and the backup node's payment and ticket issuance instructions in a concurrent network environment to prevent resource deadlock. Attached Figure Description

[0022] Figure 1 This is the overall flowchart of the travel cost and risk joint control method of the present invention;

[0023] Figure 2 This is a schematic diagram of the directed acyclic graph topology of the travel intent and the risk cascading transmission of the present invention;

[0024] Figure 3 This is the state definition and transition logic diagram of the deterministic finite state machine of this invention;

[0025] Figure 4 This is a schematic diagram of the hardware architecture of the electronic device of the present invention. Detailed Implementation

[0026] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0027] See attached document Figure 1-4 This invention provides an integrated platform for travel cost optimization and risk control, which is a joint control system for travel cost and risk based on directed acyclic graphs and dynamic hedging mechanisms. It may include: a data acquisition module 10, a topology construction module 20, a quantitative evaluation module 30, a joint optimization decision module 40, a virtual option pool management module 50, and a state machine control and execution module 60.

[0028] The data acquisition module 10 connects to external data sources and the travel distribution system network. The data acquisition module 10 receives travel intention data and acquires multi-source heterogeneous environmental data, which includes weather radar echo rate data, air traffic control status data, and local traffic congestion index data.

[0029] The topology construction module 20 is communicatively connected to the data acquisition module 10. The topology construction module 20 converts the travel intention data into a directed acyclic graph data structure. The nodes of the directed acyclic graph represent the entity event nodes in the trip, and the edges of the directed acyclic graph represent the spatiotemporal transfer relationship between entity event nodes. The topology construction module 20 calculates the time buffer margin attribute corresponding to the edge.

[0030] The quantitative evaluation module 30 is communicatively connected to the topology construction module 20. The quantitative evaluation module 30 calculates the volatility parameters of each entity event node based on multi-source heterogeneous environmental data. The quantitative evaluation module 30 then combines the entity event nodes... Blocking probability With emergency replacement costs Calculate the basic risk premium for a single node The calculation formula is:

[0031] ,

[0032] in, For volatility parameters;

[0033] Quantitative evaluation module 30 utilizes the topological structure of a directed acyclic graph and the time buffer margin attribute. Calculate the total risk premium of entity event nodes. The calculation formula is:

[0034] ,

[0035] in, For entity event nodes The set of downstream nodes, For downstream nodes The business importance weight, The time decay constant;

[0036] The joint optimization decision module 40 is communicatively connected to the quantitative evaluation module 30, and the joint optimization decision module 40 obtains candidate path sequences. Each entity event node Ticket procurement cost The joint optimization decision module 40 calculates the ticket purchase cost of each entity event node in the candidate path sequence. With total risk premium Add them together to get the overall evaluation cost The calculation formula is:

[0037] ,

[0038] The joint optimization decision module 40 selects the optimal path sequence based on a comprehensive cost assessment, and determines the relationship between the total risk premium of the core main node in the optimal path sequence and the preset dynamic safety threshold.

[0039] The virtual option pool management module 50 is communicatively connected to the joint optimization decision module 40. When the total risk premium is greater than the preset dynamic safety threshold, the virtual option pool management module 50 generates a backup hedging node. The virtual option pool management module 50 sends a resource reservation instruction to the global distribution system, obtains the passenger reservation records and the latest ticketing time limit attribute of the successfully reserved passengers, and writes the passenger reservation records and the latest ticketing time limit attribute into the memory database to form virtual option pool data, and establishes a mapping and locking relationship between the backup hedging node and the core master node.

[0040] The state machine control and execution module 60 is communicatively connected to the virtual option pool management module 50. The state machine control and execution module 60 contains deterministic finite state machine logic units. The state machine control and execution module 60 receives real-time environmental data within the time window corresponding to the core master node to trigger state transitions. When in a safe state and the time release condition is met, the state machine control and execution module 60 sends a passenger reservation record cancellation instruction. When in a blocked state, the state machine control and execution module 60 sends a refund and change instruction to the core master node based on the two-phase commit protocol, and simultaneously sends a payment and ticketing instruction to the standby hedging node.

[0041] This invention provides a method for joint control of travel costs and risks, applied to the aforementioned joint control system for travel costs and risks based on directed acyclic graphs and dynamic hedging mechanisms, and may include the following steps:

[0042] In step S100, the data acquisition module 10 acquires travel intention data, which includes departure point, destination, time window, and business event node requirements. The topology construction module 20 generates candidate itinerary sequences based on the travel intention data, converts the candidate itinerary sequences into a directed acyclic graph (DAG), extracts the time buffer margin of the transition edges between entity event nodes in the DAG, and the data acquisition module 10 pulls multi-source heterogeneous environmental data based on the spatiotemporal attributes of the entity event nodes.

[0043] Step S200: The quantitative evaluation module 30 calculates the volatility parameters of each entity event node based on multi-source heterogeneous environmental data. The quantitative evaluation module 30 combines the entity event nodes... Blocking probability With emergency replacement costs Calculate the basic risk premium for entity event nodes The calculation formula includes:

[0044] ,

[0045] in For volatility parameters;

[0046] Quantitative evaluation module 30 obtains downstream nodes in a directed acyclic graph. Business importance weight Quantitative assessment module 30 combines basic risk premium Business importance weight With time buffer margin Calculate the total risk premium for each entity event node. The calculation formula includes:

[0047] ,

[0048] in, For entity event nodes The set of downstream nodes, The time decay constant;

[0049] Step S300: The joint optimization decision module 40 obtains the ticket purchase cost corresponding to each candidate itinerary sequence, and the joint optimization decision module 40 selects the candidate itinerary sequences... Event nodes of each entity in the middle Ticket procurement cost With total risk premium Add them together to get the overall assessment cost The calculation formula includes:

[0050] ,

[0051] The joint optimization decision module 40 selects the sequence with the lowest comprehensive evaluation cost from the candidate route sequences as the optimal path sequence. The joint optimization decision module 40 judges whether the total risk premium of the core main node in the optimal path sequence is greater than the preset dynamic safety threshold. When the total risk premium is greater than the preset dynamic safety threshold, the joint optimization decision module 40 generates a trigger command.

[0052] In step S400, the virtual option pool management module 50 receives the trigger instruction, matches the backup hedging node according to the spatiotemporal attributes of the core master node, sends the resource reservation instruction for the backup hedging node to the global distribution system, obtains the passenger reservation record and the corresponding latest ticketing time limit returned by the global distribution system, and stores the passenger reservation record and the latest ticketing time limit in the memory database to establish a data mapping and locking relationship between the backup hedging node and the core master node.

[0053] In step S500, the state machine control and execution module 60 monitors real-time environmental data and drives the state transition of the internal state machine based on the real-time environmental data. When the state machine is in a safe state and the time release condition is met, the state machine control and execution module 60 sends a passenger reservation record cancellation instruction to the global distribution system. When the state machine transitions to the blocking state, the state machine control and execution module 60 sends a refund and change instruction to the core master node and simultaneously sends a payment and ticketing instruction to the backup offset node.

[0054] This invention provides a method for constructing and defining the topology of travel intentions, executed by the aforementioned topology construction module 20. The method includes: the topology construction module 20 receiving travel intention data and candidate itinerary sequences from the data acquisition module 10; the topology construction module 20 converting the candidate itinerary sequences into a directed acyclic graph (DAG) data structure, where the DAG is represented by a mathematical model as follows: ;

[0055] In a directed acyclic graph, the set This represents the set of entity event nodes in the process. A single entity event node in the data is denoted as Entity event node This refers to the business segments within the actual business trip itinerary, encompassing transportation, accommodation, and business activities.

[0056] Topology building block 20 is for each entity event node Create attribute fields, which include node identifier, geographical location of occurrence, and estimated start time. Expected end time and business importance weight The topology building module 20 is configured with a preset business weight mapping relationship, and the topology building module 20 extracts entity event nodes. The corresponding business type tags are converted into numerical business importance weights through a preset business weight mapping relationship. The preset business weight mapping relationship includes different preset weight values ​​corresponding to different business type tags;

[0057] In a directed acyclic graph, the set Denotes the set of directed edges, set A single directed edge in the middle is denoted as Directed edge Established between two adjacent entity event nodes, it indicates that from the preceding entity event node... To the post-entity event node The spatiotemporal transfer relationship;

[0058] Topology building block 20 is a directed edge Set time buffer margin attribute Time buffer margin attribute The topology building module 20 is used to characterize the time interval between two consecutive entity event nodes, based on the preceding entity event node. With subsequent entity event nodes Time attribute calculation time buffer margin attribute Time buffer margin attribute The calculation formulas include:

[0059] ,

[0060] Time buffer margin attribute In the calculation formula, For post-entity event nodes The expected start time, For the preceding entity event node The expected end time, To start from the preceding entity event node The geographical location of the event is transferred to the subsequent entity event node. The physical travel time required for the geographical location to occur is obtained through an external map application interface;

[0061] Topology building module 20 calculates the cumulative time buffer margin across multiple entity event nodes in a directed acyclic graph, for each entity event node starting from the initial entity event node. Reaching the target entity event node via multiple intermediate nodes The path, the cumulative time buffer margin on the topology building block 20 calculation path. Cumulative time buffer margin The calculation formulas include:

[0062] ,

[0063] Cumulative time buffer margin In the calculation formula, Indicates the event node starting from the initial entity. Transferred sequentially to the target entity event node The set of all directed edges on the path, the topology construction module 20 writes the entity event node attribute field and the directed edge attribute into the memory space for the quantitative evaluation module 30 to read.

[0064] This invention provides a volatility mapping method for multi-source heterogeneous data, executed by a data acquisition module 10 and a quantitative evaluation module 30, and may include:

[0065] Data acquisition module 10 reads entity event nodes from memory space. The attribute fields include the geographical location of occurrence and the estimated start time. and the expected end time The data acquisition module 10 is based on the geographical location of the occurrence and the estimated start time. and the expected end time The application interface request is sent to an external data source, which returns multi-source heterogeneous environmental data containing meteorological radar echo rate data, air traffic control status data, and local traffic congestion index data. The multi-source heterogeneous environmental data has different data formats and dimensions.

[0066] The quantitative assessment module 30 receives multi-source heterogeneous environmental data transmitted by the data acquisition module 10. The quantitative assessment module 30 performs normalization processing on the multi-source heterogeneous environmental data according to data categories and outputs dimensionless risk characteristic values. The multi-source heterogeneous environmental data includes... The data category, the quantitative evaluation module 30 will be the first The original collected values ​​corresponding to each data category are denoted as: The quantitative evaluation module 30 extracts the first value from the preset database. Historical maximum value of a data category Compared with historical minimum The quantitative evaluation module 30 is based on historical maximum values. Compared with historical minimum Calculate risk characteristic value Risk eigenvalues The calculation formulas include:

[0067] ,

[0068] Quantitative evaluation module 30 extracts entity event nodes The business type tags, the quantitative evaluation module 30, in the preset weight matrix database, based on entity event nodes... The feature weight vector corresponding to the business type tag query. Feature weight vector Include Specific weight coefficients corresponding to each data category Specific weight coefficients satisfy Specific weight coefficients Used to adjust the impact of different data categories on entity event nodes. The extent of the impact;

[0069] Quantitative assessment module 30 will use risk characteristic values With specific weighting coefficients Multiply and sum to calculate entity event nodes. volatility parameter Volatility parameter The calculation formulas include:

[0070] ,

[0071] In volatility parameter In the calculation formula, volatility parameter The numerical range is located in Internal volatility parameter The larger the value, the stronger the representation of the entity event node. The higher the probability of facing external environmental disruptions, the more the quantitative assessment module 30 will adjust the volatility parameter. Write to entity event node The attribute field, the quantitative evaluation module 30 will include the volatility parameter. As a computational entity event node Input parameters for the base risk premium.

[0072] This invention provides a quantitative calculation method for a single-node basic risk premium, executed by the aforementioned quantitative assessment module 30, and may include:

[0073] Quantitative evaluation module 30 reads entity event nodes volatility parameter The quantitative assessment module 30 is based on the volatility parameter. Calculate entity event nodes Blocking probability Blocking probability Used to characterize entity event nodes The probability that the plan cannot be executed as scheduled under the current external environmental influences; the probability of disruption. The calculation formulas include:

[0074] ,

[0075] In blocking probability In the calculation formula, The preset sensitivity adjustment coefficient, As a preset risk offset benchmark constant, the quantitative assessment module 30 is equipped with a historical travel disruption sample database. This database contains volatility parameter records for historical entity event nodes and corresponding binary classification results of actual disruptions, along with a sensitivity adjustment coefficient. Risk offset benchmark constant The quantitative assessment module 30 obtains the determined value by performing a logistic regression fitting algorithm on the historical travel disruption sample database;

[0076] Quantitative evaluation module 30 reads entity event nodes Ticket procurement cost The quantitative evaluation module 30 sends a real-time inquiry instruction to the external resource supplier system. The real-time inquiry instruction includes entity event nodes. Based on the geographical location of the occurrence and the expected start time, the quantitative assessment module 30 receives the real-time reset resource price returned by the external resource supplier system. Reset resource prices instantly Indicates at entity event node In the event of a disruption, the system temporarily purchases the equivalent alternative resources at the immediate market price.

[0077] Quantitative evaluation module 30 based on entity event nodes The business type tag is matched with the corresponding penalty base in the preset travel policy database. Base amount for penalties for breach of contract Representing entity event nodes The cancellation and rescheduling fees directly incurred due to the inability to execute the ticket;

[0078] Quantitative assessment module 30 combined with real-time resource price reset Ticket procurement cost Base amount of penalty for breach of contract Calculate emergency replacement costs Emergency replacement cost The calculation formulas include:

[0079] ,

[0080] Quantitative assessment module 30 calculates and resets resource prices instantly. With ticket procurement cost When the quantitative assessment module 30 determines that the difference is less than zero, it assigns the difference a value of zero. The quantitative assessment module 30 then compares the assigned difference with the default penalty base. Perform an addition operation and output the emergency reset cost. ;

[0081] Quantitative assessment module 30 combined with emergency replacement cost With blocking probability Calculate entity event nodes Basic risk premium Basic risk premium The calculation formulas include:

[0082] ,

[0083] The quantitative assessment module 30 will calculate the underlying risk premium. Write to entity event node The attribute fields of the quantitative assessment module 30 will include the basic risk premium. As a computational entity event node The basic input variables for total risk premium.

[0084] This invention provides a method for calculating the cascaded risk premium of a directed acyclic graph based on backpropagation, executed by the aforementioned quantitative assessment module 30, and may include:

[0085] The quantitative evaluation module 30 reads the topology data of the directed acyclic graph in the memory space. The topology data includes a set of entity event nodes and a set of directed edges. The quantitative evaluation module 30 reads the basic risk premium, business importance weight, and cumulative time buffer margin of each entity event node.

[0086] For any entity event node in a directed acyclic graph The quantitative evaluation module 30 traverses the set of directed edges in the directed acyclic graph and determines the entity event nodes based on the set of directed edges. Construct a downstream node set from all downstream nodes. Downstream node set A single node in the event is denoted as an entity event node. ;

[0087] The quantitative evaluation module 30 retrieves entity event nodes from the directed acyclic graph. Connect to downstream entity event nodes For all path sequences, the quantization evaluation module 30 calculates the path time buffer margin corresponding to each path sequence, and extracts the path time buffer margin with the smallest value as the cumulative time buffer margin. The path time buffer margin with the smallest value represents the entity event node. downstream entity event nodes The most stringent risk transmission constraint boundary between them;

[0088] The quantitative evaluation module 30 is equipped with a time decay constant. Time decay constant It is a preset value greater than zero. The quantitative evaluation module 30 uses an exponential decay function. Calculate the time buffer attenuation term, which characterizes the numerical attenuation effect of the time interval on risk transmission;

[0089] Quantitative evaluation module 30 extracts downstream entity event nodes Basic risk premium Weighting based on business importance The quantitative evaluation module 30 will analyze the downstream entity event nodes. Basic risk premium Business importance weight Multiplying by the time buffer decay term, the downstream entity event node is calculated. Reverse propagation to entity event nodes The transmission risk value;

[0090] The quantitative evaluation module has 30 pairs of downstream node sets. The transmission risk values ​​of all entity event nodes are summed to obtain the cascaded risk added value. The quantitative assessment module 30 then calculates the risk value of the entity event nodes. Basic risk premium The entity event node is calculated by summing the cascaded risk added value. Total risk premium Total risk premium The calculation formulas include:

[0091] ,

[0092] In total risk premium In the calculation formula, For entity event nodes The basic risk premium, downstream entity event nodes The basic risk premium, downstream entity event nodes The business importance weight, The time decay constant, For entity event nodes Downstream entity event nodes Cumulative time buffer margin For entity event nodes The set of downstream nodes;

[0093] The quantitative evaluation module 30 uses a depth-first search algorithm to generate a reverse topological sorting sequence of the directed acyclic graph. Following this reverse topological sorting sequence, the quantitative evaluation module 30 sequentially applies the total risk premium. The calculation steps ensure that the entity event node is calculated. Total risk premium At that time, the downstream node set The parameters of all nodes have been calculated;

[0094] The quantitative assessment module 30 will calculate the total risk premium for each entity event node. The total risk premium is written into memory space and the quantitative assessment module 30 writes it into memory space. The total risk premium is transmitted to the joint optimization decision-making module 40. This serves as the input value for calculating the overall assessment cost.

[0095] This invention provides a comprehensive evaluation method for algebraic fusion of costs and path optimization, executed by the aforementioned joint optimization decision module 40, and may include:

[0096] The joint optimization decision module 40 reads the candidate path sequence set in memory space. The candidate path sequence set contains multiple candidate path sequences that satisfy the spatiotemporal continuity constraint. Spatiotemporal continuity constraints include: candidate path sequence Two adjacent entity event nodes and satisfy Arrival location and Their departure locations are the same, and Expected start time Later Expected end time ;

[0097] Joint optimization decision module 40 reads candidate path sequence Event nodes of each entity in the middle Ticket procurement cost Ticket procurement cost To execute entity event nodes The direct financial amount required to be paid.

[0098] The joint optimization decision module 40 reads the event nodes of each entity. Total risk premium Total risk premium The quantitative evaluation module 30 calculates and generates the data, which is then written into the memory space.

[0099] The joint optimization decision-making module has 40 pairs of entity event nodes. Ticket procurement cost With total risk premium Algebraic addition is performed to obtain the overall cost of each node, and the joint optimization decision module selects 40 candidate path sequences. All entity event nodes The node comprehensive costs are summed to calculate the candidate path sequence. Comprehensive assessment cost Comprehensive cost assessment The calculation formulas include:

[0100] ,

[0101] In comprehensive cost assessment In the calculation formula, For the first Candidate path sequence, Candidate path sequence Entity event nodes in the data, For entity event nodes The cost of purchasing tickets, For entity event nodes Total risk premium;

[0102] The joint optimization decision module iterates through all candidate path sequences in the candidate path sequence set 40 times. The joint optimization decision module 40 calculates each candidate path sequence. Corresponding comprehensive assessment cost ;

[0103] The joint optimization decision module has 40 pairs of all candidate path sequences. Comprehensive assessment cost The numerical values ​​are sorted and compared, and the joint optimization decision module 40 extracts the comprehensive evaluation cost from the candidate path sequence set. The candidate path sequence with the smallest numerical value is selected as the optimal path sequence. Optimal path sequence The solution formulas include:

[0104] ,

[0105] In the optimal path sequence In the solution formula, This represents the set of variable objects that minimize the objective function.

[0106] The joint optimization decision module 40 will generate the optimal path sequence The entity event node sequence data contained therein is written into the memory space, and the joint optimization decision module 40 writes the optimal path sequence. This serves as the benchmark data for determining whether the virtual option pool management module 50 will subsequently lock resources.

[0107] This invention provides a dynamic security threshold triggering and control flow distribution method, executed by the aforementioned joint optimization decision module 40, and may include:

[0108] The joint optimization decision module 40 reads the optimal path sequence from the memory space. Optimal path sequence The optimal path sequence is generated by the joint optimization decision module 40 during the comprehensive cost evaluation optimization step. Included entity event nodes ;

[0109] The joint optimization decision-making module 40 reads the enterprise's basic risk tolerance from the preset database. Enterprise's basic risk tolerance The fixed financial constraint benchmark values ​​pre-configured for the travel system administrator are used by the joint optimization decision module 40 to read entity event nodes. Business importance weight ;

[0110] Joint Optimization Decision Module 40 combines the enterprise's basic risk tolerance Weighting based on business importance Calculate entity event nodes Dynamic security threshold Dynamic security threshold Used to define the system's entity event nodes. The maximum risk premium that can be tolerated, and the dynamic safety threshold. The calculation formulas include:

[0111] ,

[0112] Dynamic security threshold In the calculation formula, For entity event nodes Dynamic security threshold, As the company's basic risk tolerance, For entity event nodes Business importance weight, business importance weight The larger the value, the higher the dynamic safety threshold. The smaller the value, the lower the system's risk tolerance for core business nodes;

[0113] Joint Optimization Decision Module 40 reads entity event nodes Total risk premium The joint optimization decision-making module 40 performs a numerical comparison operation, and then calculates the total risk premium. The numerical value and dynamic security threshold The values ​​are used to determine their magnitude;

[0114] When the joint optimization decision module 40 determines the total risk premium Less than or equal to the dynamic security threshold At that time, the joint optimization decision module 40 determines the entity event node. Once the safety execution conditions are met, the joint optimization decision-making module 40 generates a regular booking instruction. The joint optimization decision-making module 40 then distributes the regular booking instruction to the external supplier system interface, completing the entity event node. Procurement of normal business travel resources;

[0115] When the joint optimization decision module 40 determines the total risk premium Greater than the dynamic security threshold At that time, the joint optimization decision module 40 triggers the control flow switching logic, and the joint optimization decision module 40 will adjust the total risk premium. Greater than the dynamic security threshold Entity event nodes Marked as a high-risk master node ;

[0116] The joint optimization decision module 40 generates hedging trigger instructions, which include high-risk master nodes. Based on the identification information, geographical location, estimated start time, and estimated end time, the joint optimization decision module 40 distributes the hedging trigger instruction to the virtual option pool management module 50. The virtual option pool management module 50 receives the hedging trigger instruction and initiates a hedging action targeting the high-risk master node. The backup hedging node matching and resource locking procedure.

[0117] This invention provides a method for matching hedging candidate nodes, executed by the aforementioned virtual option pool management module 50, and may include:

[0118] The virtual option pool management module 50 receives a hedging trigger instruction from the joint optimization decision module 40. The hedging trigger instruction includes a high-risk master node. Identification information, geographical location of occurrence, and estimated start time and the expected end time ;

[0119] The virtual option pool management module 50 parses hedging trigger instructions and extracts high-risk master nodes based on their geographical location. The virtual option pool management module 50 reads the preset travel tolerance parameters from the start and end coordinates. ;

[0120] Virtual Options Pool Management Module 50 based on expected start time Expected end time With travel tolerance parameter Calculate alternative time windows Alternative time windows The calculation formulas include:

[0121] ,

[0122] In the alternative time window In the calculation formula, The set of time intervals that allow the execution of alternative routes. High-risk master node The expected start time, High-risk master node The expected end time, This refers to the travel tolerance parameter;

[0123] The virtual option pool management module 50 sends a resource retrieval request to the external resource provider system. The resource retrieval request includes the starting coordinates, the ending coordinates, and alternative time windows. The virtual option pool management module 50 receives a set of candidate hedging resources returned by an external resource supplier system. The set of candidate hedging resources includes those within the available time window. Multiple candidate backup nodes that satisfy spatial displacement constraints ;

[0124] The virtual option pool management module 50 sends the candidate hedging resource set to the quantitative evaluation module 30. The quantitative evaluation module 30 calculates each candidate backup node based on the aforementioned backpropagation-based cascaded calculation method. Total risk premium The quantitative assessment module 30 will calculate the total risk premium. The value is returned to the virtual option pool management module 50;

[0125] The virtual option pool management module 50 is configured with a backup security threshold coefficient. Backup safety threshold coefficient The range of values ​​lies within the interval Within the virtual options pool management module, 50 reads high-risk master nodes. Dynamic security threshold ;

[0126] Virtual option pool management module 50 combined with dynamic security threshold With backup safety threshold coefficient Calculate the backup decision threshold Backup decision threshold The calculation formulas include:

[0127] ,

[0128] In the backup decision threshold In the calculation formula, As a backup decision threshold, As a backup safety threshold coefficient, High-risk master node Dynamic security threshold;

[0129] The virtual option pool management module 50 executes numerical comparison logic, and removes the total risk premium from the candidate hedging resource set. Greater than or equal to the backup decision threshold Candidate backup nodes The virtual option pool management module 50 will manage the remaining candidate backup nodes. Combined into a set of effective standby nodes Effective standby node set The nodes in the table represent security resources that meet the substitution requirements in terms of risk.

[0130] Virtual Options Pool Management Module 50: Obtaining a Set of Valid Backup Nodes Candidate backup nodes Resource lock-in costs Resource lock-in costs The amount of reserve funds or interface call fees required to be paid for the system to call the global distribution system's time-limited retention instruction;

[0131] The virtual options pool management module has a set of 50 valid standby nodes. Each candidate backup node Perform hedging assessment calculations to obtain the hedging assessment value. hedging assessment value The calculation formulas include:

[0132] ,

[0133] The virtual option pool management module 50 has a set of valid standby nodes. Finding hedging valuation values The smallest node, the virtual option pool management module 50, will hedge the assessed value. The smallest node is determined as the final backup hedging node. Backup hedging node The solution formulas include:

[0134] ,

[0135] The virtual options pool management module 50 will have backup hedging nodes. Attribute data and backup hedging nodes With high-risk master nodes The associated identifier is written into the memory space for subsequent calls to the underlying global distribution system interface.

[0136] This invention provides a virtual option pool data structure and lifecycle management method, executed by the aforementioned virtual option pool management module 50, and may include:

[0137] The virtual option pool management module 50 reads high-risk master nodes from memory space. With backup hedging nodes The attribute data of the virtual option pool management module 50 is based on the standby hedging node. The attribute data generates a standard reservation instruction. The virtual option pool management module 50 sends the standard reservation instruction to the global distribution system interface. The standard reservation instruction contains a reservation request but does not contain payment settlement request parameters.

[0138] The virtual options pool management module 50 receives a confirmation message from the global distribution system interface, which includes passenger reservation records. With the latest ticket issuance deadline Passenger reservation records Reserved identification code assigned to the global distribution system, latest ticketing deadline The deadline for maintaining the validity of the reserved identification code in the global distribution system;

[0139] The virtual option pool management module 50 constructs an option mapping data structure in the in-memory database. This data structure includes a master node identifier field, a backup node identifier field, a reservation record field, a ticket issuance time limit field, and a status enumeration field. The virtual option pool management module 50 assigns high-risk master nodes... The identifier is written into the master node identifier field, and the virtual option pool management module 50 will assign the backup hedging node. The identifier is written into the backup node identifier field, and the virtual option pool management module 50 records the passenger reservation records. Write the reservation record field, and the virtual option pool management module 50 will set the latest ticket issuance time. Write the ticket issuance time limit field, and the virtual option pool management module 50 initializes the status enumeration field to the initial monitoring status value;

[0140] The virtual option pool management module 50 completes the writing of the option mapping data structure and establishes a high-risk master node in the in-memory database. With backup hedging nodes Data locking mapping relationship;

[0141] The virtual option pool management module 50 starts a lifecycle timer, which reads the current system time from an external time server at a preset frequency. The virtual options pool management module reads high-risk master nodes. Expected start time The virtual options pool management module calculates the estimated start time. With the current system time The time difference value is used by the virtual option pool management module to calculate the latest ticket issuance time. With the current system time The time difference value;

[0142] The virtual option pool management module 50 is configured with a time tolerance constant. Time tolerance constant Characterizes the reserved buffer time required for the system to execute network interface calls;

[0143] The virtual option pool management module 50 executes the judgment logic and calculates the timeout condition. Timeout conditions The determination formulas include:

[0144] ,

[0145] Timeout condition In the determination formula, This represents a logical OR operation, with a timeout condition. When the determination result is true, the virtual option pool management module 50 updates the state enumeration field in the option mapping data structure to the near-expiration trigger state value;

[0146] The virtual option pool management module 50 synchronously generates a resource expiration determination signal. The virtual option pool management module 50 transmits the resource expiration determination signal and the updated option mapping data structure to the state machine control and execution module 60. The resource expiration determination signal is used as an input parameter to drive the state machine control and execution module 60 to execute instruction distribution.

[0147] This invention provides a method for defining and transitioning states in a deterministic finite state machine, executed by the aforementioned state machine control and execution module 60, and may include:

[0148] The state machine control and execution module 60 reads the option mapping data structure in the memory database whose state is the initial monitoring state value, and extracts the high-risk master node associated with the option mapping data structure. ;

[0149] The state machine control and execution module 60 instantiates a deterministic finite state machine in memory. Deterministic finite state machine Deterministic finite state machine represented by a quintuple. The formula for representing a quintuple includes:

[0150] ,

[0151] In deterministic finite state machines In the formula, For a finite set of states, For the set of input events, This is the state transition function. This is the initial state. For the set of terminating states;

[0152] Finite State Set Includes security status Warning status and blocking status Set of terminating states Includes release status With final blocking execution state The state machine control and execution module 60 will initialize the state. Assigned to a safe state ;

[0153] Input event collection Including environmental safety incidents Environmental early warning events Environmental disruption events and time-triggered events ;

[0154] The state machine control and execution module 60 calls the data acquisition module 10 according to a preset time period, and the state machine control and execution module 60 acquires high-risk master nodes. Real-time risk characteristics of the geographical location The state machine control and execution module 60 reads the early warning threshold constant. With blocking threshold constant ;

[0155] The state machine control and execution module 60 executes condition judgment logic to generate input events for the current cycle. When the real-time risk characteristic value At that time, the state machine control and execution module 60 generates an environmental safety event. ,when At that time, the state machine control and execution module 60 generates an environmental warning event. ,when At that time, the state machine control and execution module 60 generates an environmental blocking event. ;

[0156] The state machine control and execution module 60 is configured with a state transition matrix, which defines the state transition function. Mapping rules, state transition function Mapping a finite set of states With input event set Cartesian product to finite set of states With the set of termination states State transition function The basic mapping formula is expressed as:

[0157] ,

[0158] in This is the current state. The next state after the jump;

[0159] The state machine control and execution module 60 performs the following state transition operations based on the state transition matrix: when the deterministic finite state machine In a safe state And input events Environmental early warning events When, execute When deterministic finite state machines In a safe state or warning status And input events For environmental disruption events When, execute or When deterministic finite state machines Under alert status And input events For environmental safety incidents When, execute ;

[0160] The state machine control and execution module 60 receives the resource expiration determination signal sent by the virtual option pool management module 50, and uses the resource expiration determination signal as the input event set. Time-triggered events in ;

[0161] When input event Time-triggered events And deterministic finite state machine In a safe state or warning status At that time, the state machine control and execution module 60 executes... or ;

[0162] State machine control and execution module 60 locks and releases states And terminate receiving input events generated by external environmental data sources;

[0163] When input event Time-triggered events And deterministic finite state machine In a blocked state At that time, the state machine control and execution module 60 executes... ;

[0164] The state machine control and execution module 60 locks the final blocking execution state. And terminate receiving input events generated by external environmental data sources;

[0165] The state machine control and execution module 60 writes the final determined termination state into the memory space, and the state machine control and execution module 60 triggers subsequent concurrent control flow execution instructions based on the termination state in the memory space.

[0166] This invention provides a method for operating a mutual exclusion execution engine within a time window, executed by the aforementioned state machine control and execution module 60, and may include:

[0167] The state machine control and execution module 60 reads the termination state of the deterministic finite state machine from the memory space. The termination state includes the release state. With final blocking execution state ;

[0168] When the termination state is the release state At that time, the state machine control and execution module 60 reads the reservation record field from the corresponding option mapping data structure and extracts the passenger reservation record. The state machine control and execution module 60 generates a resource cancellation instruction, which includes passenger reservation records. The state machine control and execution module 60 sends a resource cancellation instruction to the global distribution system interface after receiving the reserved identification code. The state machine control and execution module 60 then receives a cancellation confirmation message from the global distribution system interface, deletes the option mapping data structure from the memory database, and releases the standby hedging node. Resource locking, state machine control and execution module 60 maintains high-risk master node The routine business execution control flow;

[0169] When the termination status is the final blocked execution status At that time, the state machine control and execution module 60 starts the distributed transaction control engine based on the two-phase commit protocol, and the state machine control and execution module 60 will control the high-risk master node. Refund and rescheduling operations and backup hedging nodes The payment and ticketing operation is encapsulated as a global transaction. The state machine control and execution module 60 executes the preparation phase of the two-phase commit protocol. The state machine control and execution module 60 then sends a request to the high-risk master node. The corresponding supplier system sends a refund / change status verification request, and the state machine control and execution module 60 sends a request for passenger reservation records to the global distribution system interface. The payment channel connectivity and timeliness verification request is received by the state machine control and execution module 60 from the high-risk master node. The state machine control and execution module 60 extracts the ready state flag from the verification response returned by the corresponding supplier system and global distribution system interface.

[0170] State machine control and execution module 60 establishes transaction commit judgment conditions Transaction commit criteria Logical operation formulas include:

[0171] ,

[0172] In the transaction commit determination conditions In logical operation formulas, High-risk master node The boolean value indicating whether the cancellation / rescheduling is ready. As a backup hedging node The boolean value indicating ticket issuance readiness. This represents a logical AND operation, used as a condition for transaction commit. When the calculation result is true, the state machine control and execution module 60 enters the submission phase of the two-phase commit protocol, and the state machine control and execution module 60 calls the underlying multi-threaded concurrent scheduling interface.

[0173] State machine control and execution module 60 directs to high-risk master node The corresponding supplier system sends high-risk master nodes. The state machine control and execution module 60 concurrently sends backup hedging nodes to the global distribution system interface upon receiving the refund / change order. Full payment and ticketing instructions;

[0174] The state machine control and execution module 60 receives the ticket issuance success certificate returned by the global distribution system interface, and writes the ticket issuance success certificate and business change parameters into the memory space.

[0175] When the transaction commit condition is met If the calculation result is false, or if a communication timeout occurs during the preparation phase of the two-phase commit protocol, the state machine control and execution module 60 executes a transaction rollback instruction and terminates the global transaction.

[0176] The state machine control and execution module 60 sends passenger reservation records to the global distribution system interface. The resource cancellation command is used to release the standby hedging node. The underlying system resources occupied, the state machine control and execution module 60 will be a high-risk master node. The status data is reset to the manual pending status, the underlying exception log is recorded, and a manual intervention signal is output.

[0177] The present invention provides an electronic device, which may include: a processor 71, a memory 72, a communication interface 73 and a bus 74. The processor 71, the memory 72 and the communication interface 73 are electrically connected and communicate with each other through the bus 74. The bus 74 includes an address bus, a data bus and a control bus.

[0178] Communication interface 73 is configured to establish a data transmission link between the electronic device and the external network environment. Communication interface 73 interacts with external data sources and the global distribution system interface through network protocols.

[0179] Communication interface 73 receives weather radar echo rate data, air traffic control status data and local traffic congestion index data transmitted from external data sources. Communication interface 73 sends resource reservation instructions, resource cancellation instructions and payment ticketing instructions to the global distribution system interface.

[0180] The memory 72 is configured to store computer programs, business operation data and system configuration parameters. The memory 72 includes high-speed random access memory and non-volatile memory.

[0181] The memory 72 stores a computer program that implements the aforementioned method for joint control of travel costs and risks. The computer program is logically divided into multiple program code segments.

[0182] The computer program stored in memory 72 includes program code corresponding to data acquisition module 10, topology construction module 20, quantitative evaluation module 30, joint optimization decision module 40, virtual option pool management module 50, and state machine control and execution module 60.

[0183] Processor 71 includes a central processing unit, an application-specific integrated circuit, or a field-programmable gate array;

[0184] The processor 71 is configured to read computer programs from memory 72 and control electronic devices to perform specific calculation and data scheduling operations by executing instructions in the computer programs;

[0185] The processor 71 acquires travel intention data by calling the program code of the data acquisition module 10, and the processor 71 constructs a directed acyclic graph data structure and calculates the time buffer margin by calling the program code of the topology construction module 20.

[0186] The processor 71 calculates the volatility parameters, basic risk premium, and total risk premium for each entity event node by calling the program code of the quantitative evaluation module 30;

[0187] The processor 71 integrates the ticket purchase cost and the total risk premium by calling the program code of the joint optimization decision module 40, calculates the comprehensive evaluation cost, and outputs the optimal path sequence;

[0188] The processor 71 extracts candidate standby nodes by calling the program code of the virtual option pool management module 50, performs the matching of standby hedging nodes, and establishes the option mapping data structure in the memory database;

[0189] The processor 71 initializes the deterministic finite state machine by calling the program code of the state machine control and execution module 60, and executes the two-phase commit protocol and issues the refund and change instruction and the payment ticketing instruction when the time timeout condition is met or an environmental blocking event is received;

[0190] This invention provides a computer-readable storage medium, which may include:

[0191] A computer-readable storage medium stores a computer program, which consists of a series of computer-executable instructions.

[0192] When the processor of an electronic device executes computer-executable instructions, the processor implements the aforementioned method for joint control of travel costs and risks;

[0193] The processor executes computer-executable instructions and calls the data acquisition module 10 and the topology construction module 20. The processor acquires travel intent data by executing the program code of the data acquisition module 10 and the topology construction module 20, constructs a directed acyclic graph data structure, and determines the time buffer margin between entity event nodes.

[0194] The processor executes computer-executable instructions and calls the quantitative evaluation module 30. The processor performs mapping processing on multi-source heterogeneous environment data by executing the program code of the quantitative evaluation module 30, and calculates the basic risk premium and total risk premium of entity event nodes.

[0195] The processor executes computer-executable instructions and calls the joint optimization decision module 40. The processor calculates the comprehensive evaluation cost of each candidate path sequence by executing the program code of the joint optimization decision module 40, performs numerical sorting and comparison, and outputs the optimal path sequence.

[0196] The processor executes computer-executable instructions and calls the virtual option pool management module 50. The processor locks the backup hedging node resources by executing the program code of the virtual option pool management module 50 and establishes an option mapping data structure in the memory database.

[0197] The processor executes computer-executable instructions and calls the state machine control and execution module 60. The processor controls the state transitions of the deterministic finite state machine by executing the program code of the state machine control and execution module 60, and issues refund / rescheduling instructions and payment / ticket issuance instructions based on the termination state of the deterministic finite state machine.

[0198] Computer-readable storage media are non-transient computer-readable storage media. Non-transient computer-readable storage media are physical physical media with data storage function, and do not contain electromagnetic signals and optical signals that propagate instantaneously. Physical physical media include non-volatile storage media and volatile storage media.

[0199] Non-volatile storage media include read-only memory, flash memory, disk storage, optical disk storage, and magnetic random access memory. Non-volatile storage media are used for persistent storage of computer programs and system operation logs.

[0200] Volatile storage media include dynamic random access memory and static random access memory. Volatile storage media are used to provide memory space for instruction fetching and data interaction when the processor performs computational tasks.

[0201] Computer-readable storage media exist as independent physical modules or are integrated into the internal motherboard architecture of electronic devices. Computer-readable storage media establish a physical link with the processor for data access and instruction transmission via the system bus.

Claims

1. An integrated platform for travel cost optimization and risk control, characterized in that, The platform includes: The data acquisition module (10) is used to acquire travel intention data and multi-source heterogeneous environment data; The topology building module (20) is used to convert the candidate path sequence into a directed acyclic graph data structure and calculate the time buffer margin between entity event nodes; The quantitative assessment module (30) is used to calculate the volatility parameters, basic risk premium and total risk premium of each entity event node based on the multi-source heterogeneous environment data. The joint optimization decision module (40) is used to integrate the ticket purchase cost of the entity event node with the total risk premium, calculate the comprehensive evaluation cost to output the optimal path sequence, and trigger the control flow switching logic based on the dynamic security threshold. The virtual option pool management module (50) is used to match the backup hedging node and establish an option mapping data structure in the memory database to complete resource locking when the control flow switching logic is triggered; The state machine control and execution module (60) is used to control the state transition of the deterministic finite state machine based on the real-time risk characteristic value and timeout condition, and to issue refund and rescheduling instructions and payment and ticketing instructions according to the termination state of the deterministic finite state machine.

2. The integrated platform for travel cost optimization and risk control according to claim 1, characterized in that, When constructing the directed acyclic graph data structure, the topology construction module (20) performs the following operations: it establishes attribute fields for each entity event node, including the geographical location of occurrence, the expected start time, the expected end time, and the business importance weight; The business importance weight is obtained by converting the business type label corresponding to the entity event node through a preset business weight mapping relationship. All path sequences connecting the preceding entity event node to the following entity event node are retrieved in the directed acyclic graph. The path time buffer margin corresponding to each path sequence is calculated. The path time buffer margin with the smallest value is extracted as the cumulative time buffer margin. The path time buffer margin with the smallest value represents the most stringent risk transmission constraint boundary between the preceding entity event node and the following entity event node.

3. The integrated platform for travel cost optimization and risk control according to claim 2, characterized in that, When calculating the total risk premium of the entity event node, the quantitative assessment module (30) specifically executes the following formula: ； In the formula for calculating the total risk premium, For entity event nodes Total risk premium For entity event nodes The basic risk premium, downstream entity event nodes The basic risk premium, downstream entity event nodes The business importance weight, The time decay constant, For entity event nodes Downstream entity event nodes Cumulative time buffer margin For entity event nodes The set of downstream nodes; The quantitative evaluation module (30) uses a depth-first search algorithm to generate a reverse topological sorting sequence of the directed acyclic graph, and executes the calculation steps of the total risk premium in sequence according to the reverse topological sorting sequence.

4. The integrated platform for travel cost optimization and risk control according to claim 3, characterized in that, The underlying risk premium is calculated as follows: the volatility parameters of the entity event node are obtained, the blocking probability is calculated using a logistic regression function, and the sensitivity adjustment coefficient and risk offset benchmark constant in the logistic regression function are obtained by performing a regression fitting algorithm on a historical travel blocking sample database; the emergency reset cost is calculated by combining the immediate reset resource price, ticket purchase cost and default penalty base, and the underlying risk premium is obtained by multiplying the emergency reset cost by the blocking probability.

5. The integrated platform for travel cost optimization and risk control according to claim 4, characterized in that, When the joint optimization decision module (40) outputs the optimal path sequence and triggers the control flow switching logic, it performs the following operations: extracts candidate path sequences that satisfy the spatiotemporal continuity constraints from the candidate path sequence set; calculates the sum of the node comprehensive costs of each entity event node in the candidate path sequence as the comprehensive evaluation cost, and extracts the candidate path sequence with the smallest comprehensive evaluation cost as the optimal path sequence; and calculates the dynamic security threshold by combining the enterprise's basic risk tolerance with the business importance weight of the entity event node. When the total risk premium of the entity event node is greater than the dynamic security threshold, the corresponding entity event node is marked as a high-risk master node, a hedging trigger instruction is generated and distributed to the virtual option pool management module (50).

6. The integrated platform for travel cost optimization and risk control according to claim 5, characterized in that, When the virtual option pool management module (50) matches standby hedging nodes, it performs the following operations: calculates alternative time windows based on the expected start time, expected end time and preset travel tolerance parameters of the high-risk master node; obtains multiple candidate standby nodes that are within the alternative time windows and meet the spatial displacement constraints; and removes candidate standby nodes whose total risk premium is greater than or equal to the standby determination threshold to combine them into a set of effective standby nodes. The node with the minimum sum of resource lock-in cost and total risk premium in the set of effective standby nodes is selected and established as the final standby hedging node.

7. The integrated platform for travel cost optimization and risk control according to claim 6, characterized in that, When the virtual option pool management module (50) establishes the option mapping data structure, it performs the following operations: sends a standard reservation instruction containing a seat reservation request but not a payment settlement request parameter to the global distribution system interface to obtain passenger reservation records and the latest ticketing time limit; constructs the option mapping data structure in the memory database containing a master node identifier field, a backup node identifier field, a reservation record field, a ticketing time limit field, and a status enumeration field; initializes the status enumeration field to the initial monitoring state value, and establishes a data locking mapping relationship between the high-risk master node and the backup hedging node.

8. The integrated platform for travel cost optimization and risk control according to claim 7, characterized in that, When the virtual option pool management module (50) performs lifecycle management, it calculates the timeout conditions. Its logical operation formula include: ； In the logical operation formula, The estimated start time for high-risk master nodes. The current system time. This is the latest time limit for ticket issuance. The time tolerance constant, The result of the timeout condition is true. The virtual option pool management module (50) updates the state enumeration field to the near-expiration trigger state value and generates a resource near-expiration determination signal to be transmitted to the state machine control and execution module (60).

9. The integrated platform for travel cost optimization and risk control according to claim 8, characterized in that, The state machine control and execution module (60) performs the following operations: instantiates the deterministic finite state machine represented by a quintuple, the finite state set including a safe state, a warning state, and a blocking state, the termination state set including a release state and a final blocking execution state, the input event set including an environmental safety event, an environmental warning event, an environmental blocking event, and a time-sensitive triggering event with the resource nearing expiration determination signal as an input parameter; when the input event is the time-sensitive triggering event and the deterministic finite state machine is in a safe state or a warning state, the state transitions to the release state; the state machine control and execution module (60) sends a resource cancellation instruction for passenger reservation records to the global distribution system interface; when the input event is the time-sensitive triggering event and the deterministic finite state machine is in a blocking state, the state transitions to the final blocking execution state; The state machine control and execution module (60) starts the distributed transaction control engine based on the two-phase commit protocol, sends the refund and resale execution instruction to the master node resource supplier system, and sends the full payment and ticketing instruction of the backup hedging node to the global distribution system interface.

10. An integrated method for travel cost optimization and risk control, characterized in that, Includes the following steps: Step S100: Acquire travel intention data and multi-source heterogeneous environment data through the data acquisition module (10); Step S200: Convert the candidate path sequence into a directed acyclic graph data structure through the topology building module (20), and calculate the time buffer margin between entity event nodes; The quantitative assessment module (30) calculates the volatility parameters, basic risk premium, and total risk premium of each entity event node based on the multi-source heterogeneous environment data. Step S300: By integrating the ticket purchase cost of the entity event node with the total risk premium through the joint optimization decision module (40), the comprehensive evaluation cost is calculated to output the optimal path sequence, and the control flow switching logic is triggered based on the dynamic security threshold. Step S400: When the control flow switching logic is triggered by the virtual option pool management module (50), a backup hedging node is matched and an option mapping data structure is established in the memory database to complete the resource locking; Step S500: The state machine control and execution module (60) controls the state transition of the deterministic finite state machine based on the real-time risk characteristic value and timeout condition, and issues a refund / rescheduling instruction and a payment / ticketing instruction according to the termination state of the deterministic finite state machine.

Images