A carpool matching system and method based on multi-level alliance formation game

Abstract

The application discloses a carpool matching system and method based on a multilevel alliance formation game, collects a trip request of a passenger; initializes an operator alliance structure, and forms an operator alliance by all operators; generates a passenger graph according to a current operator alliance structure, and edges in the passenger graph connect two passengers belonging to the same operator alliance of the selected operator; searches for a passenger alliance on the passenger graph, calculates passenger alliance utility based on an optimal path of the passenger alliance; calculates operator alliance utility based on passenger alliance utility and a mapping relationship, and performs operator utility distribution according to a predefined distribution protocol; performs stability inspection on the current operator alliance structure, adjusts the operator alliance structure and returns to adjust if the stability condition is not met; and if the stability condition is met, calculates passenger utility distribution, and outputs a final order result including a stable operator alliance structure, operator utility distribution, a passenger alliance structure and passenger utility distribution.

Description

Technical Field

[0001] This invention relates to the field of intelligent transportation and travel service optimization technology, specifically to a carpooling matching system and method based on multi-level alliance game theory. Background Technology

[0002] As a shared mobility mode, ride-sharing is widely considered to alleviate traffic congestion, reduce travel costs and carbon emissions, and has significant social and environmental value. However, in reality, ride-sharing adoption remains limited, and system efficiency is difficult to fully realize. A key reason is the highly fragmented market: different operators in the same region operate independently, demand and capacity cannot be integrated across platforms, ride-sharing matching is limited, and the potential for efficiency improvement is weakened.

[0003] To alleviate this problem, some aggregation platforms have attempted to integrate multiple operators. While this has expanded the matching scope to some extent, significant differences in operators' scale, costs, and market position make it difficult to balance order and revenue distribution, leading to unsustainable and unstable partnerships. How to improve efficiency while maintaining stable cooperation has become a key issue for the sustainable development of ride-sharing systems. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention proposes a carpooling matching method based on a multi-level alliance formation game, comprising the following steps: Step S1: Collect passengers' travel requests, which include the passengers' pick-up point, drop-off point, and the operator information selected by the passengers; Step S2: Initialize the operator alliance structure, forming an operator alliance with all operators; Step S3: Generate a passenger graph based on the current operator alliance structure, wherein the edges in the passenger graph connect two passengers belonging to the same operator alliance of the selected operator; Step S4: Search for passenger alliances on the passenger graph and calculate the passenger alliance utility based on the optimal path of the passenger alliance; Step S5: Calculate the operator alliance utility based on passenger alliance utility and mapping relationship, and allocate operator utility according to predefined allocation protocol; Step S6: Perform a stability check on the current operator alliance structure. If the stability condition is not met, adjust the operator alliance structure and return to step S3. If the stability condition is met, calculate the passenger utility allocation and output the final order result, which includes the stable operator alliance structure, operator utility allocation, passenger alliance structure, and passenger utility allocation.

[0005] Preferably, in step S4, the calculation method for the passenger alliance utility is as follows: ; In the formula, It represents the passenger alliance, The operating profit parameter per unit distance. Passenger Alliance The optimal path, Represents the optimal path Operating costs Indicates passenger boarding point and drop-off point The distance between them.

[0006] Preferably, in step S5, the calculation method for the operator alliance utility is as follows: ; in, Indicates the operator alliance, Indicates the operator alliance The passenger alliance served Indicates the operator alliance The total driving distance when each operator operates independently. This indicates the total distance traveled during cooperative operation. This represents the collaboration benefit coefficient resulting from a reduction in unit relative travel distance. Passenger Alliance The passenger alliance effect.

[0007] Preferably, in step S6, the stability test is performed as follows: For any possible subset of operators in the current operator alliance structure Calculate its deviation gain: ; in, For operator subset The effectiveness of alliances when forming new alliances. For operators The current utility allocation value; If a subset of operators exists Make If the stability condition is not met, then it is determined that the stability condition is not satisfied. If for all operator subsets All If the stability condition is met, then the stability condition is satisfied.

[0008] Preferably, when the stability condition is not met, the methods for adjusting the operator alliance structure include: Select multiple non-overlapping subsets of operators to form a set of deviation groups; Calculate the total deviation gain for each set of deviation groups, and select the set of deviation groups that maximizes the total deviation gain. Update the operator alliance structure based on the selected set of deviation groups; If the updated operator alliance structure has been explored before, the set of suboptimal deviation groups with positive deviation gain is selected for further updates, and finally the structure with the smallest total deviation gain is selected as the stable structure.

[0009] Preferably, in step S4, the method of searching for passenger alliances on the passenger map includes: The passenger map is broken down into multiple sub-maps, and the operators selected by passengers in each sub-map belong to the same operator alliance. Independently search for the optimal passenger alliance structure within each subgraph; The optimal passenger alliance structures of each subgraph are merged into a unified passenger alliance structure.

[0010] Preferably, the method for searching the optimal passenger alliance structure within each subgraph includes: Initialize each passenger to form a separate passenger alliance, and use the initial passenger alliance structure as the root node of the search tree; For any edge in the subgraph, determine whether shrinking the edge connects two passenger alliances that satisfy the vehicle capacity constraint. ; If vehicle capacity constraints are violated If so, then mark the edge as non-shrinkable; If vehicle capacity constraints are met If the two passenger alliances are merged to form a new passenger alliance structure, the utility of the new passenger alliance structure is calculated. If the utility is higher than the current optimal utility, the current optimal solution is updated. Set a bounding function; if the bounding value of the current passenger alliance structure is less than the current optimal utility, stop exploring the subtree of that branch.

[0011] Preferably, in step S6, the method for calculating passenger utility allocation includes: Within each passenger alliance, based on each passenger's individual travel distance. and detour distance Determine the initial utility allocation; Calculate the maximum surplus difference between passenger pairs within a passenger alliance; If the ε-core stability condition is not met, then the passenger pair with the largest surplus difference is selected, and the utility is transferred along the surplus direction; Repeat the utility transfer until the ε-nucleus stability condition is met.

[0012] Preferably, a time limit is set in step S4. If the search time for the passenger alliance exceeds the time limit Then output the current best approximate solution; after the stability condition is satisfied in step S6, remove the time constraint. And re-search the Passenger Alliance.

[0013] This invention also provides a carpooling matching system based on multi-level alliance formation game theory, applicable to the above method, including: The travel request collection module is used to collect passengers' travel requests, which include the passenger's pick-up point, drop-off point, and the operator information selected by the passenger. The operator alliance formation module is used to initialize the operator alliance structure and form an operator alliance with all operators, as well as to adjust the operator alliance structure during the iteration process; The passenger map generation module is used to generate a passenger map based on the current operator alliance structure, wherein the edges in the passenger map connect two passengers belonging to the same operator alliance of the selected operator; The passenger alliance search module is used to search for passenger alliances on the passenger map and calculate the passenger alliance utility based on the optimal path of the passenger alliance; The operator utility calculation and allocation module is used to calculate operator alliance utility based on passenger alliance utility and mapping relationship, and to allocate operator utility according to a predefined allocation protocol. The stability verification module is used to verify the stability of the current operator alliance structure. If the stability conditions are not met, the operator alliance formation module is triggered to adjust the operator alliance structure. The passenger utility allocation module is used to calculate passenger utility allocation after the stability condition is met. The order output module is used to output the final order results, including a stable operator alliance structure, operator utility allocation, passenger alliance structure, and passenger utility allocation.

[0014] The beneficial effects of the present invention include at least the following: This invention proposes a dynamic alliance formation mechanism that allows operators to dynamically adjust their alliance structure based on real-time travel requests, rather than relying on traditional static competition or purely cooperative models. This approach adapts to ever-changing market demands, ensuring that operators can flexibly and efficiently form cooperative alliances in a highly fragmented shared mobility market, thereby improving the overall stability and resource utilization of the system.

[0015] This invention constructs a multi-level game theory framework to model the decision-making processes of both operators and passengers in shared mobility. Within this framework, operators decide whether and how to form alliances through non-cooperative game theory, while passengers form shared mobility groups through cooperative game theory. This framework fully considers the mutual influence and alliance dynamics between operators and passengers, achieving simultaneous modeling and analysis of the stability of operator-passenger matching.

[0016] The method of this invention has significant advantages in improving resource utilization, reducing operating costs, and enhancing passenger travel experience in shared mobility systems. It can adapt to diverse urban traffic environments and dynamically changing market demands. As a solution that can be integrated with multi-platform shared mobility systems, it has broad engineering application value and provides strong technical support for the sustainable development of intelligent transportation and the shared mobility industry. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the interaction mechanism and order generation process of the carpooling plan in an embodiment of the present invention; Figure 2 This is a schematic diagram of the Passenger Alliance Formation Algorithm CRP-CFSS in an embodiment of the present invention. Detailed Implementation

[0018] 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 protection scope of the present invention.

[0019] This invention provides a carpooling matching method based on a multi-level alliance game theory framework, with an aggregation platform as the main body. Its operation can be summarized in three stages. In the first stage, the system collects passenger travel requests and merges them into a travel request pool. In the second stage, the system processes the travel requests in the pool in batches, generating orders. In the third stage, orders are issued, the operator menu is sent to the operators, who then dispatch vehicles to pick up and drop off passengers; the passenger menu is sent to the passengers, who decide whether to participate in the carpooling alliance.

[0020] In this embodiment of the invention, a travel request refers to a single travel plan submitted by a passenger through a carrier's application or directly to the aggregation platform, including information such as departure point and destination. If submitted through a carrier's application, the passenger can explicitly choose whether to participate in the aggregation platform; if submitted directly to the aggregation platform, the passenger can specify a preferred carrier.

[0021] In this embodiment of the invention, the order includes two parts: a carrier menu and a passenger menu. The carrier menu includes a stable carrier alliance structure. And operator utility allocation The former specifies which operators temporarily form a cooperative alliance, while the latter specifies the revenue share each operator receives. The passenger menu includes the optimal passenger alliance structure. and passenger utility allocation The former specifies which passengers are sharing the ride, while the latter determines how passengers will split their payments. Order results guide vehicle dispatching, passenger confirmation, and platform settlement, enabling cross-carrier ride-sharing services.

[0022] The alliance structure definition and the relationship between the operator alliance and the passenger alliance in this embodiment of the invention are as follows.

[0023] Alliance structure is for the complete set That is, a division of all participants in the system, an alliance structure denoted as ,in K To represent the total number of alliances, the following two conditions must be met: .

[0024] set up Represents the set of all operators; a complete operator alliance structure is denoted as . Operator alliance structure This constraint determines the matching space between passengers, forming a passenger graph. ,in This indicates that all passengers have gathered. A carpooling relationship is defined as one between passengers if and only if both passengers have chosen the same operator alliance. Only then does it exist.

[0025] Each operator alliance structure A single optimal passenger alliance structure is determined. There are two mapping relationships between the two: mapping This indicates that each passenger will be part of the alliance. Mapped to the carrier alliance that provides services to it ; mapping This indicates that each operator alliance It is mapped to the set of passenger alliances it serves.

[0026] The passenger alliance utility calculation method includes: for any passenger The boarding and alighting points are respectively denoted as and Defining the transportation network as a geographical map ,in Represents the set of all boarding and alighting points, i.e. , This represents the set of shortest distance paths connecting these points. They are all assigned a weight This indicates the distance between the two points.

[0027] path Defined as a sequence of points: For a given passenger alliance Its effective path set The route must meet three conditions: each point in the route must be a boarding or alighting point for passengers of the alliance; the route must include both the boarding and alighting points for each passenger, with the boarding point preceding the alighting point; and there must be at least two distinct passengers. This caused one of the passengers The boarding point was before another passenger. The drop-off point.

[0028] Passenger Alliance Optimal path For all valid paths The optimal path problem is the one that minimizes operating costs. Since operating costs are directly proportional to travel distance, the optimal path problem can be simplified to the shortest path problem. Based on the optimal path, a passenger alliance is defined. Its effects are as follows: ; ; ; In the formula, The operating profit per unit distance is used as a parameter; therefore, passenger alliance utility... It is actually equivalent to the opposite of the total amount paid by the passenger.

[0029] The methods for calculating the utility of operator alliances include: for operator alliances Its utility calculation formula is defined as: ; ; In the formula, Indicates the operator alliance The passenger alliance served Indicates alliance The total driving distance when each operator operates independently; This indicates the total distance traveled during cooperative operation; This represents the collaboration benefit coefficient resulting from a reduction in the unit relative travel distance.

[0030] The following describes the specific steps of the carpooling matching method based on multi-level alliance formation game provided in this invention. Figure 1 The steps shown are as follows.

[0031] Step S1: Collect passenger travel requests The system collects passengers' travel requests, including their pick-up and drop-off points, as well as the operator information they have selected. This information forms the basis for subsequent alliance formation and matching optimization.

[0032] Step S2: Initialize the operator alliance structure, forming an operator alliance with all operators.

[0033] System setting iteration counter and initialize the operator set. In the initial state, the system forms an operator alliance with all operators, and the initial alliance structure is defined as follows: The design intent of this initialization strategy is to start with the maximum level of cooperation and gradually adjust the alliance structure through subsequent iterations to achieve a stable state.

[0034] Step S3: Generate a passenger graph based on the current operator alliance structure. The edges in the passenger graph connect two passengers belonging to the same operator alliance of the selected operator.

[0035] Specifically, a passenger graph is generated based on the current operator alliance structure, where edges connect two passengers belonging to the same operator alliance as the selected operator. In each iteration, the aggregation platform generates a new operator alliance structure based on the results of the previous round. ,when When using the initial structure, when Then the structure is adjusted based on the results of the previous stability test.

[0036] After determining the operator alliance structure, the system generates a constrained set of valid passenger alliances: ; in: It represents a passenger alliance (i.e., a group of passengers traveling together). This step represents vehicle capacity and results in generating a complete set of feasible passenger alliances for each operator alliance.

[0037] Step S4: Search for passenger alliances on the passenger graph and calculate the passenger alliance utility based on the optimal path of the passenger alliance.

[0038] The system within a limited time Within the effective passenger alliance set, the optimal passenger carpooling plan is searched. If the search time exceeds the time limit. Then output the current best approximate solution. This step involves processing each passenger alliance. Solve the vehicle routing problem to determine its utility. .

[0039] The calculation method of passenger alliance utility reflects the economic value of carpooling services: by optimizing routes to reduce total travel distance, operating costs are reduced, while ensuring a fair sharing of passenger payments.

[0040] Step S5: Calculate the operator alliance utility based on passenger alliance utility and mapping relationship, and allocate operator utility according to predefined allocation protocol.

[0041] Based on the utility of each passenger alliance and its mapping relationship, calculate the utility of each operator alliance. The utility The calculation of operator alliance utility takes into account both the total amount paid by passengers and the collaborative benefits brought about by cooperation, the latter being quantified by comparing the difference in travel distance between independent and cooperative operations.

[0042] After calculating the utility of the operator alliance, the system allocates the utility according to a predefined allocation protocol, thus obtaining the operator utility allocation. ,in Indicates the operator The share of utility gained. Common distribution agreements include equal distribution, investment weight distribution, and performance-based distribution, which can be flexibly selected according to the platform's objectives.

[0043] Step S6: Stability check and iterative adjustment A stability test is performed on the current operator alliance structure. The core idea of ​​the stability test is to determine whether there is any subset of operators that has an incentive to deviate from the current alliance structure. For any possible subset of operators... Calculate its deviation gain: ; in For operator subset The effectiveness of alliances when forming new alliances. For operators The current utility allocation value. Deviation from gain. This indicates the additional utility that could be gained if these operators deviate from the current structure and form a new alliance.

[0044] If a subset of operators exists Make If the stability condition is not met, it means that the subset has a dynamic tendency to deviate from the current structure. At this point, the system sets... Adjust the operator alliance structure and return to step S3 to continue iteration. If for all operator subsets... If the stability condition is met, then the stable structure is denoted as [structure name missing]. .

[0045] After the stability condition is met, the system removes the time constraint. Then, a new search for passenger alliances is conducted to obtain the precise optimal passenger alliance structure: ; Then, the passenger utility allocation stable to the ε-nuclear was calculated. This ensures the stability of the final passenger alliance structure. The final output includes a stable operator alliance structure. Distribution of operator utility Passenger alliance structure and passenger utility allocation The final order result.

[0046] One embodiment of the present invention proposes an iterative algorithm, CRP-CFOS, for efficiently forming stable operator alliance structures. This algorithm avoids traversing all possible alliance structures through iterative optimization and can efficiently converge to a stable alliance.

[0047] During the initial phase of the alliance structure, all operators form an alliance: The utility calculation formula for each operator alliance is used to calculate the utility of each alliance. The utility And calculate the allocation of each operator within the alliance according to the predefined allocation protocol. The utility .

[0048] During the deviation evaluation phase, for any possible subset of operators Its deviation gain is defined as: ; like This indicates that the subset of operators has an incentive to deviate from the current structure. CRP-CFOS simultaneously selects multiple non-overlapping deviation groups for structure updates, and the set of deviation groups... Find the set that maximizes the total deviation gain: ; The set of deviation groups The constraint condition that the deviation groups do not overlap must be met.

[0049] During the alliance structure adjustment phase, if the total deviation from gain... If the result is positive, it indicates that the current alliance structure has reached a stable state. Otherwise, the system updates the alliance structure based on the selected optimal set of deviation groups. ; symbol The definition is as follows: , As a set family, For a set, For a family of sets Each subset Remove the ones that are related to The intersection of the results, and only the non-empty result set is retained. .

[0050] The core of the update operation is to separate the deviation group from the original alliance to form a new independent alliance, while maintaining the integrity of other alliance members.

[0051] To avoid the algorithm looping in the structure space, if a new alliance structure is formed... If the problem has already been explored in previous iterations, the algorithm then considers the next suboptimal set of deviation groups with a positive deviation gain. This process is repeated until no new alliance structure can be generated. Finally, the total deviation gain is selected. The smallest structure is considered an approximately stable structure.

[0052] In one embodiment of the present invention, to determine the optimal passenger alliance structure, the present invention proposes an improved search algorithm CRP-CFSS, such as... Figure 2 As shown in the figure, this algorithm optimizes the graph constraints and computational process, which can significantly improve computational efficiency while ensuring solution quality.

[0053] Passenger map It can be decomposed into multiple subgraphs: Each subgraph satisfy: ; ; To avoid repeatedly validating operator constraints during the search process, the search is restricted to each subgraph. Internally conducted, optimal passenger alliance structure It can be represented as the union of the optimal alliance structures of each subgraph: .

[0054] During the structure initialization phase, each subgraph All edges are initially marked as solid lines, indicating that these passengers may form a carpooling alliance. The initial passenger alliance structure is defined as follows: This means that each passenger initially forms a separate alliance. This structure serves as the root node of the search tree. Its utility is defined as: And use it as the initial optimal value.

[0055] During the branching and contraction phase, for any edge Edge contraction refers to the contraction of the two passenger alliances connected by the edge. and Merge into a new alliance. From the current node. Structure Starting from a solid line edge, the algorithm selects a solid line edge. Then determine whether shrinking the edge satisfies the capacity constraint. If the capacity constraint is violated, the edge is marked as a dashed line, indicating that the edge cannot be shrunk. If the capacity constraint is satisfied, the edge is shrunk and the edge is marked as a dashed line to avoid repeated exploration.

[0056] The new structure after contraction is: Then calculate the utility of the new structure. If the value is higher than the current optimal utility, then update the current optimal solution.

[0057] To avoid redundant calculations, a boundary function is designed. If the current structure's bounding value If the utility is less than the current optimal utility, then prune that branch to prevent further exploration of its subtree. Let the structure be... And sorted by the effectiveness of each alliance from highest to lowest: Then the boundary function is defined as: , Indicates vehicle capacity constraints The minimum possible number of alliances.

[0058] In practical applications, operator utility allocation can be achieved by selecting different allocation protocols based on platform objectives. Common allocation protocols include the following three: (1) Equal distribution: ; (2) Investment weight allocation: ,in Indicates the operator's investment weight; (3) Performance allocation: Utility is allocated based on actual operational performance, so that operators with higher request volume and longer operating distance can obtain a larger share of utility.

[0059] In one embodiment of the present invention, an improved PK algorithm is used to determine the utility allocation for each passenger to ensure the stability of the passenger alliance. During the allocation initialization phase, in each passenger alliance... Within this framework, the initial utility allocation is determined based on each passenger's individual travel distance and detour distance: ; in: Indicates passenger Individual travel distance; Indicates the passenger's carpooling route The actual driving distance in the data, of which and They represent passengers respectively. The index positions of the boarding and alighting points in this optimal route; Indicates the detour distance.

[0060] In the utility transfer phase, starting from the initial allocation, the system has the largest surplus difference. For the iterative transfer of utility between passenger pairs, in order to keep the alliance stable in an approximate sense, the allocation must satisfy: ; The maximum surplus difference is defined as: , For passengers right The surplus.

[0061] in If the current allocation does not meet this condition, then in the alliance The internal utility transfer process selects the passenger pair with the largest surplus difference. and along the direction of surplus from Towards The transfer of a portion of utility, the amount of utility transferred is defined as: ; This ensures that passengers only experience a transfer of utility. Its utility is still no less than its utility for traveling alone. Utility transfer only occurs when there are multiple passengers. Having the same maximum surplus difference To avoid oscillations during the algorithm's iteration process, a unique passenger pair is selected. Perform utility transfer operations to ensure the uniqueness and convergence of each iteration step.

[0062] This invention also provides a carpooling matching system based on multi-level alliance formation game theory, applicable to the above method, including: The travel request collection module is used to collect passengers' travel requests, which include the passenger's pick-up point, drop-off point, and the operator information selected by the passenger. The operator alliance formation module is used to initialize the operator alliance structure and form an operator alliance with all operators, as well as to adjust the operator alliance structure during the iteration process; The passenger map generation module is used to generate a passenger map based on the current operator alliance structure, wherein the edges in the passenger map connect two passengers belonging to the same operator alliance of the selected operator; The passenger alliance search module is used to search for passenger alliances on the passenger map and calculate the passenger alliance utility based on the optimal path of the passenger alliance; The operator utility calculation and allocation module is used to calculate operator alliance utility based on passenger alliance utility and mapping relationship, and to allocate operator utility according to a predefined allocation protocol. The stability verification module is used to verify the stability of the current operator alliance structure. If the stability conditions are not met, the operator alliance formation module is triggered to adjust the operator alliance structure. The passenger utility allocation module is used to calculate passenger utility allocation after the stability condition is met. The order output module is used to output the final order results, including a stable operator alliance structure, operator utility allocation, passenger alliance structure, and passenger utility allocation.

[0063] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described; only preferred embodiments of the present invention are illustrated. The descriptions are relatively specific and detailed, but they should not be construed as limiting the scope of the present invention. As long as the combination of these technical features does not contradict each other, it should be considered within the scope of this specification.

[0064] It should be noted that those skilled in the art can make various modifications and improvements without departing from the inventive concept, and these all fall within the scope of protection of this invention. Therefore, the scope of protection of this invention should be determined by the appended claims.

Claims

1. A carpooling matching method based on multi-level alliance formation game theory, characterized in that, Includes the following steps: Step S1: Collect passengers' travel requests, which include the passengers' pick-up point, drop-off point, and the operator information selected by the passengers; Step S2: Initialize the operator alliance structure, forming an operator alliance with all operators; Step S3: Generate a passenger graph based on the current operator alliance structure, wherein the edges in the passenger graph connect two passengers belonging to the same operator alliance of the selected operator; Step S4: Search for passenger alliances on the passenger graph and calculate the passenger alliance utility based on the optimal path of the passenger alliance; Step S5: Calculate the operator alliance utility based on passenger alliance utility and mapping relationship, and allocate operator utility according to predefined allocation protocol; Step S6: Perform a stability check on the current operator alliance structure. If the stability condition is not met, adjust the operator alliance structure and return to step S3. If the stability condition is met, calculate the passenger utility allocation and output the final order result, which includes the stable operator alliance structure, operator utility allocation, passenger alliance structure, and passenger utility allocation.

2. The method according to claim 1, characterized in that: In step S4, the calculation method for the passenger alliance utility is as follows: ； In the formula, It represents the passenger alliance, The operating profit parameter per unit distance. Passenger Alliance The optimal path, Represents the optimal path Operating costs Indicates passenger boarding point and drop-off point The distance between them.

3. The method according to claim 1, characterized in that: In step S5, the calculation method for the operator alliance utility is as follows: ； in, Indicates the operator alliance, Indicates the operator alliance The passenger alliance served Indicates the operator alliance The total driving distance when each operator operates independently. This indicates the total distance traveled during cooperative operation. This represents the collaboration benefit coefficient resulting from a reduction in unit relative travel distance. Passenger Alliance The passenger alliance effect.

4. The method according to claim 1, characterized in that: In step S6, the stability test is performed as follows: For any possible subset of operators in the current operator alliance structure Calculate its deviation gain: ； in, For operator subset The effectiveness of alliances when forming new alliances. For operators The current utility allocation value; If a subset of operators exists Make If the stability condition is not met, then it is determined that the stability condition is not satisfied. If for all operator subsets All If the stability condition is met, then the stability condition is satisfied.

5. The method according to claim 4, characterized in that: When stability conditions are not met, the methods for adjusting the operator alliance structure include: Select multiple non-overlapping subsets of operators to form a set of deviation groups; Calculate the total deviation gain for each set of deviation groups, and select the set of deviation groups that maximizes the total deviation gain. Update the operator alliance structure based on the selected set of deviation groups; If the updated operator alliance structure has been explored before, the set of suboptimal deviation groups with positive deviation gain is selected for further updates, and finally the structure with the smallest total deviation gain is selected as the stable structure.

6. The method according to claim 1, characterized in that: In step S4, the method of searching for passenger associations on the passenger map includes: The passenger map is broken down into multiple sub-maps, and the operators selected by passengers in each sub-map belong to the same operator alliance. Independently search for the optimal passenger alliance structure within each subgraph; The optimal passenger alliance structures of each subgraph are merged into a unified passenger alliance structure.

7. The method according to claim 6, characterized in that: The methods for searching the optimal passenger alliance structure within each subgraph include: Initialize each passenger to form a separate passenger alliance, and use the initial passenger alliance structure as the root node of the search tree; For any edge in the subgraph, determine whether shrinking the edge connects two passenger alliances that satisfy the vehicle capacity constraint. ; If vehicle capacity constraints are violated If so, then mark the edge as non-shrinkable; If vehicle capacity constraints are met If the two passenger alliances are merged to form a new passenger alliance structure, the utility of the new passenger alliance structure is calculated. If the utility is higher than the current optimal utility, the current optimal solution is updated. Set a bounding function; if the bounding value of the current passenger alliance structure is less than the current optimal utility, stop exploring the subtree of that branch.

8. The method according to claim 1, characterized in that: In step S6, the methods for calculating passenger utility allocation include: Within each passenger alliance, based on each passenger's individual travel distance. and detour distance Determine the initial utility allocation; Calculate the maximum surplus difference between passenger pairs within a passenger alliance; If the ε-core stability condition is not met, then the passenger pair with the largest surplus difference is selected, and the utility is transferred along the surplus direction; Repeat the utility transfer until the ε-nucleus stability condition is met.

9. The method according to claim 1, characterized in that: Set the time limit in step S4 If the search time for the passenger alliance exceeds the time limit Then output the current best approximate solution; after the stability condition is satisfied in step S6, remove the time constraint. And re-search the Passenger Alliance.

10. A carpooling matching system based on multi-level alliance formation game theory, applicable to the method described in any one of claims 1 to 9, characterized in that, include: The travel request collection module is used to collect passengers' travel requests, which include the passenger's pick-up point, drop-off point, and the operator information selected by the passenger. The operator alliance formation module is used to initialize the operator alliance structure and form an operator alliance with all operators, as well as to adjust the operator alliance structure during the iteration process; The passenger map generation module is used to generate a passenger map based on the current operator alliance structure, wherein the edges in the passenger map connect two passengers belonging to the same operator alliance of the selected operator; The passenger alliance search module is used to search for passenger alliances on the passenger map and calculate the passenger alliance utility based on the optimal path of the passenger alliance; The operator utility calculation and allocation module is used to calculate operator alliance utility based on passenger alliance utility and mapping relationship, and to allocate operator utility according to a predefined allocation protocol. The stability verification module is used to verify the stability of the current operator alliance structure. If the stability conditions are not met, the operator alliance formation module is triggered to adjust the operator alliance structure. The passenger utility allocation module is used to calculate passenger utility allocation after the stability condition is met. The order output module is used to output the final order results, including a stable operator alliance structure, operator utility allocation, passenger alliance structure, and passenger utility allocation.

Images