A method and system for carpooling at large traffic hubs based on multi-attribute constraints

Abstract

The application discloses a large traffic hub carpooling method and system based on multi-attribute constraints, first, the basic information of passengers who have the intention to participate in the large traffic hub carpooling is acquired, and the information of vehicles waiting for the large traffic hub carpooling is acquired; then, the vehicles meeting the constraint conditions are screened out according to the parameters of the carpooling passenger information and the vehicle information, the multi-attribute values of the carpooling are quantified, a passenger time tolerance coefficient matrix A, a vehicle detour mileage coefficient matrix D and a vehicle seat rate array arrP are created; A and D are reduced from m*n matrices to one-dimensional arrays arrTime and arrMileage containing m elements, m is the number of vehicles participating in this carpooling, and n is the maximum number of seats provided to passengers; then, the vehicle with the optimal result value is acquired to perform order matching; finally, the vehicle information is sent to a new order passenger client, and the carpooling is successful. The application realizes the indexes of good passenger experience, moderate detour and high income in a balanced manner, and achieves optimal passenger matching and line design.

Description

Technical Field

[0001] This invention belongs to the field of urban traffic management and control technology, and relates to a carpooling method and system applicable to large transportation hubs, specifically a carpooling method and system for large transportation hubs based on multi-attribute constraints. Background Technology

[0002] At major transportation hubs (such as airports, train stations, and long-distance bus stations), passengers, due to long journey times and a lot of luggage, prefer to choose high-quality business vehicles or taxis. Because of the characteristics of passenger flow at major transportation hubs, such as concentrated passenger flow at certain times, there is a severe imbalance between the demand for rides and available capacity, and taxi fares are high. Carpooling is a good solution to this problem. However, current "one-to-many" carpooling service models primarily use allocation algorithms based on virtual road network points, map grid division, and proximity area planning. These methods suffer from low carpooling success rates, unreasonable passenger itineraries, high operating costs, and high computational complexity. For example, it often results in long detours, poor value for money, and low service quality for passengers. In terms of operation and management, unreasonable order allocation can occur, where one vehicle may be fully booked but with an inefficient route, while other vehicles only have one order but generate less revenue than private car services, leading to low vehicle utilization and high operating costs. Therefore, under resource constraints, it is of great significance to study scheduling algorithms and systems that can improve the success rate of carpooling systems, shorten the average carpooling completion time for passengers, enhance the passenger travel experience, and reduce operating costs. Summary of the Invention

[0003] To address the aforementioned technical problems, this invention provides a carpooling method and system for large transportation hubs based on multi-attribute constraints.

[0004] The technical solution adopted by the method of the present invention is: a carpooling method at large transportation hubs based on multi-attribute constraints, comprising the following steps:

[0005] Step 1: Obtain basic information on passengers who are willing to participate in group buying at major transportation hubs;

[0006] Step 2: Obtain information on vehicles already waiting for group rides at major transportation hubs;

[0007] Step 3: Based on the passenger and vehicle information of the carpooling, filter out vehicles that meet the constraints, quantify the carpooling multi-attribute values, and create a passenger time tolerance coefficient matrix A, a vehicle detour mileage coefficient matrix D, and a vehicle carpooling rate array arrP.

[0008] Step 4: Reduce the passenger time tolerance coefficient matrix A and the vehicle detour mileage coefficient matrix D from m×n matrices to one-dimensional arrays arrTime and arrMileage containing m elements, where m is the number of vehicles participating in this group purchase and n is the maximum number of seats provided to passengers.

[0009] Step 5: Match the vehicle with the best result value to the order;

[0010] Step 6: The vehicle information is sent to the new passenger's client, and the carpooling is successful.

[0011] Preferably, in step 1, the passenger basic information includes the passenger's name, contact information, desired destination, number of passengers, number and size of luggage, and scheduled departure time.

[0012] Preferably, in step 2, the information of the waiting vehicles for group booking includes the license plate number, driver's name, driver's contact information, the destination required by the matched passengers, the estimated departure time, the number of available seats, the number and size of luggage items that can be placed, the queuing time, and the queuing order.

[0013] Preferably, in step 3, the vehicle that meets the constraints satisfies the following constraints:

[0014] (5) Each passenger transport shall not exceed its passenger capacity;

[0015] C i ≤L i ;

[0016] Among them, L i C represents the maximum number of seats available for shared rides when vehicle i is empty. i This indicates the number of seats occupied in the vehicle after inserting the new order k;

[0017] (6) The number and size of luggage carried by passengers each time shall not exceed the carrying capacity of the vehicle’s trunk;

[0018]

[0019] Where, N i This indicates the number of pieces of luggage that can be provided in the trunk of carpooling vehicle i when it is empty. V represents the number of luggage items occupied after inserting the new order k; i This indicates the luggage space available in the trunk of carpool vehicle i when it is empty. This indicates the luggage space occupied after inserting the new order k;

[0020] (7) Establish a waiting time of N minutes, that is, the order is completed within N minutes after the first passenger gets on the bus. If the bus departs after N minutes, it will not participate in the subsequent order-sharing service.

[0021] waitTime k＝1 ≤N;

[0022] Among them, waitTime k＝1 This represents the waiting time for the first passenger to board and depart in carpool vehicle i, where N is a preset value.

[0023] (8) For vehicles that have already departed, it is prohibited to add more passengers en route due to temporary empty passenger capacity.

[0024] departed Vehicle i =1;

[0025] Among them, departedVehicle i This is a scalar indicating whether vehicle i in the group-buying program has been dispatched. When it is 1, it means that the vehicle has been dispatched and no new orders will be accepted. When it is 0, it means that the vehicle has not been dispatched.

[0026] Preferably, in step 3, the passenger time tolerance coefficient matrix A = (α ij ) m×n Let A be an m×n matrix, where m is the number of vehicles participating in this group purchase; due to different vehicle models, the number of seats provided by the vehicles in the group purchase may vary, and n is determined by the maximum number of seats provided to passengers, with zeros padded if the number of seats is less than n; α ij This represents the time tolerance coefficient of the j-th passenger in vehicle i in a group-buying vehicle.

[0027] Insert the new order k into all pending vehicle orders waiting for group booking, replan the driving route, and calculate the estimated arrival time for each passenger. After inserting the new order k, the time tolerance coefficient for the first order passenger and available seats is 0, and the time tolerance coefficient for other passengers is:

[0028]

[0029] Where, Δt j This indicates the estimated increase in travel time for the j-th passenger after a new order k is inserted into vehicle i in a group-buying service; t 0,j This represents the optimal time for the j-th passenger to travel directly from the origin to the destination without participating in group bookings; r-1 = 0 represents the origin; t r-1.r This represents the travel time from the destination of passenger r-1 to the destination of passenger r.

[0030] Preferably, in step 3, the vehicle detour mileage coefficient matrix D = (β) ij ) m×n D is an m×n matrix, where n is determined by the maximum number of seats available to passengers, with seats less than n padded with 0s; β ijThis represents the detour distance coefficient for the j-th passenger in vehicle i of a shared ride;

[0031] Insert the new order k into all pending car orders, replan the driving route, and calculate the estimated mileage to the destination for each passenger. After inserting the new order k, the detour mileage coefficient for the first order passenger and the empty seat is 0, and the detour mileage coefficient for other passengers is:

[0032]

[0033] Where, Δs j This indicates the estimated additional mileage for the j-th passenger after vehicle i is inserted into order k in a group-buying service; S j s represents the distance traveled by the j-th passenger under the shared ride route planning; r-1.r s represents the distance traveled from the destination of passenger r-1 to the destination of passenger r; r-2.r |r=k+1 represents the direct travel distance between two adjacent destinations before and after the new order k (i.e., the direct travel distance from the destination of passenger k-1 to the destination of passenger k+1).

[0034] Preferably, in step 3, the vehicle sharing rate array arrP = [p1, p2, ..., p i ,...,p m ], p i This indicates the ride-sharing rate of vehicle i in the group-sharing program. L i C represents the maximum number of seats available for shared rides when vehicle i is empty. i This indicates the actual number of seats occupied by the vehicle after inserting the new order k.

[0035] Preferably, in step 4, the row vector group in the passenger time tolerance coefficient matrix A is taken, and the time tolerance coefficient value of the vehicles participating in the group buying is calculated using the overall standard deviation algorithm, forming an array arrTime = (t1, t2, ..., t m );

[0036] Take the row vector α of matrix A i =(α i1 ,a i2 ,...,α in ), and their average values Then the time tolerance coefficient value of vehicle i in the group buying

[0037] t i A higher value indicates that passengers on ride-sharing vehicle i will have to endure a longer journey time due to the insertion of new orders compared to a direct trip from the origin to the destination, resulting in a worse passenger experience.

[0038] Preferably, in step 4, the row vector group in the vehicle detour mileage coefficient moment D is taken, and the detour mileage coefficient value of the vehicles participating in the group purchase is calculated using the overall standard deviation algorithm, forming an array arrMileage=(d1,d2,...,d m );

[0039] Take the row vector β of matrix D i =(β) i1 ,β i2 ,...,β in ), and their average values Then the detour mileage coefficient value of carpooling vehicle i

[0040] d i A higher value indicates that the shared vehicle i takes a longer detour due to the insertion of new orders, resulting in higher vehicle order costs.

[0041] As a preferred option, in step 5, a comprehensive analytical mathematical model with better service quality and moderate cost is built using arrTime, arrMileage, arrP, and carpooling constraints, and the vehicle with the optimal result value f is selected to publish the order task.

[0042] The comprehensive analytical mathematical model is as follows:

[0043]

[0044] Where a, b, and c are weight values, and d is other fixed constant factors that need to be considered in the operation and management of shared vehicles. If there are no other factors, it can be set to 0.

[0045] The technical solution adopted by the system of this invention is: a carpooling system for large transportation hubs based on multi-attribute constraints, comprising:

[0046] One or more processors;

[0047] A storage device for storing one or more programs that, when executed by one or more processors, enable the one or more processors to implement the carpooling method for large transportation hubs based on multi-attribute constraints.

[0048] Compared with existing technical solutions, the beneficial effects of the technical solution of the present invention include:

[0049] 1. Compared to matching and scheduling based on virtual road site locations, map grid division, and neighboring area planning, this invention globally considers the carpooling attribute values ​​of vehicles waiting for carpooling, quantifies carpooling rate, passenger time tolerance, and detour mileage into specific coefficient values, creates a multi-attribute comprehensive analysis mathematical model for carpooling, selects the vehicle with the optimal result value to match orders, effectively improves the service quality of passenger experience, reduces detour time and mileage, and reduces operating costs while balancing vehicle carpooling rates.

[0050] 2. This invention uses an m×n matrix to create a multi-attribute matrix for carpooling, and uses the overall standard deviation algorithm to reduce the m×n matrix to a one-dimensional array with m elements, so that it can be used in the mathematical model for comprehensive analysis of carpooling multi-attribute to calculate the optimal result value. Attached Figure Description

[0051] The technical solutions described herein are further illustrated below using examples and specific implementation methods. Additionally, accompanying drawings are used in the description of the technical solutions. Those skilled in the art can, without any creative effort, obtain other drawings and the intent of the present invention based on these drawings.

[0052] Figure 1 This is a flowchart illustrating the method implemented in this invention;

[0053] Figure 2 This is a flowchart illustrating the process of reducing a multi-attribute matrix to a one-dimensional array in the method implemented in this invention. Detailed Implementation

[0054] To facilitate understanding and implementation of the present invention by those skilled in the art, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0055] Please see Figure 1 and Figure 2 This embodiment provides a carpooling method for large transportation hubs based on multi-attribute constraints, including the following steps:

[0056] Step 1: Obtain basic information on passengers who are willing to participate in group buying at major transportation hubs;

[0057] In one implementation, the basic information includes passenger name, contact information, desired destination, number of passengers, number and size of luggage, and scheduled departure time.

[0058] Step 2: Obtain information on vehicles already waiting for group rides at major transportation hubs;

[0059] In one implementation, the information of the waiting shared ride vehicle includes the license plate number, driver's name, driver's contact information, the destination required by the matched passengers, estimated departure time, number of available seats, number and size of luggage items that can be placed, queuing time, and queuing order. The destination required by the matched passengers refers to the expected destination of the matched passengers and the planned route. The estimated departure time refers to the departure within 20 minutes after the first passenger boards the vehicle.

[0060] Step 3: Based on the passenger and vehicle information of the carpooling, filter out vehicles that meet the constraints, quantify the carpooling multi-attribute values, and create a passenger time tolerance coefficient matrix A, a vehicle detour mileage coefficient matrix D, and a vehicle carpooling rate array arrP.

[0061] In one implementation, the vehicles selected that meet the constraints must satisfy the following constraints:

[0062] (1) Each passenger shall not exceed its passenger capacity.

[0063] C i ≤L i

[0064] L i C represents the maximum number of seats available for shared rides when vehicle i is empty. i This indicates the number of seats occupied by the vehicle after inserting the new order k.

[0065] (2) The number and size of luggage items carried by passengers each time shall not exceed the carrying capacity of the vehicle's trunk.

[0066]

[0067] N i This indicates the number of pieces of luggage that can be provided in the trunk of carpooling vehicle i when it is empty. V represents the number of luggage items occupied after inserting the new order k; i V represents the luggage space available in the trunk of the shared vehicle i when it is empty. i k This indicates the luggage space occupied after inserting a new order k.

[0068] (3) Establish a 20-minute waiting time cycle, that is, the order is completed within 20 minutes after the first passenger boards the vehicle. If the vehicle departs after 20 minutes, it will not participate in the subsequent order-sharing service.

[0069] waitTime k＝1 ≤20

[0070] waitTime k＝1This represents the waiting time for the first passenger to board the shared vehicle i. (This 20-minute timeframe is used for illustrative purposes only and is not intended to limit the scope of the invention.)

[0071] (4) For vehicles that have already departed, it is prohibited to add more passengers en route due to temporary empty passenger capacity.

[0072] departed Vehicle i =1

[0073] departed Vehicle i This is a control scalar indicating whether the vehicle i in the group-buying program has been dispatched. When it is 1, it means that the vehicle has been dispatched and no new orders will be accepted; when it is 0, it means that the vehicle has not been dispatched.

[0074] In one implementation, the passenger time tolerance coefficient matrix A = (α ij ) m×n Let A be an m×n matrix, where m is the number of vehicles participating in this group purchase (i.e., the number of vehicles waiting to depart in the waiting area); due to different vehicle models, the number of seats provided by the group purchase vehicles may vary, and n is determined by the maximum number of seats provided to passengers, with zeros padded if the number of seats is less than n; α ij This represents the time tolerance coefficient for the j-th passenger on ride-sharing vehicle i. The passengers are ordered according to the planned route based on the expected destination of all orders on that vehicle, regardless of the order time. The time lost when a passenger gets off the vehicle is negligible.

[0075] Insert the new order k into all pending ride orders, replan the routes, and calculate the estimated arrival time for each passenger. After inserting the new order k, the time tolerance coefficient for the first order's passengers and empty seats (empty seats or seats created by unfilled vehicles) is 0, and the time tolerance coefficients for other passengers are:

[0076]

[0077] Where, Δt j This indicates the estimated increase in travel time for the j-th passenger after a new order k is inserted into vehicle i in a group-buying service; t 0,j This represents the optimal time for the j-th passenger to travel directly from the origin to the destination without participating in group bookings; r-1 = 0 represents the origin; t r-1.r This represents the travel time from the destination of passenger r-1 to the destination of passenger r.

[0078] α ij The larger the value, the longer it takes for the j-th passenger to travel by carpool compared to traveling directly from the origin to the destination.

[0079] Travel time is affected by factors such as road conditions and weather at the time, and the optimal estimated arrival time recommended by commonly used maps is used in the calculation.

[0080] In one implementation, the vehicle detour mileage coefficient matrix D = (β) ij ) m×n D is an m×n matrix, where n is determined by the maximum number of seats available to passengers, with seats less than n padded with 0s; β ij β represents the detour distance coefficient for the j-th passenger in vehicle i of a shared ride; i and α i β represents the coefficients of the same vehicle i in different dimensions. ij and α ij These correspond to coefficients of different dimensions for the same vehicle i and the same passenger j. The passenger ranking is determined by the planned route based on the expected destination of all orders for that vehicle, and is independent of the order time.

[0081] Insert the new order k into all pending car orders, replan the routes, and calculate the estimated mileage to the destination for each passenger. After inserting the new order k, the detour mileage coefficient for the first order's passengers and empty seats (empty seats in empty cars or cars that are not full) is 0, and the detour mileage coefficient for other passengers is:

[0082]

[0083] Where, Δs j S represents the estimated additional mileage for the j-th passenger after vehicle i is inserted into order k in a group-buying service; j s represents the distance traveled by the j-th passenger under the shared travel route planning; r-1.r s represents the distance traveled from the destination of passenger r-1 to the destination of passenger r; r-2.r |r=k+1 represents the direct travel distance between two adjacent destinations before and after the new order k (i.e., the direct travel distance from the destination of passenger k-1 to the destination of passenger k+1).

[0084] β ij A larger value indicates that the j-th passenger takes a longer route.

[0085] In one implementation, the new order k is inserted into all waiting carpooling orders, and the carpooling rate for each car is calculated. The carpooling rate array arrP = [p1, p2, ..., p i ,...,p m ], p i This indicates the ride-sharing rate of vehicle i in the group-sharing program. L i C represents the maximum number of seats that carpooling vehicle i can provide when it is empty. iThis indicates the actual number of seats occupied by the vehicle after inserting a new order k.

[0086] Due to differences in vehicle type, the number of seats offered by shared vehicles may vary. i A higher value indicates that the vehicle transports more passengers, which to some extent reflects higher revenue and lower costs.

[0087] Step 4: Using the overall standard deviation algorithm, reduce the passenger time tolerance coefficient matrix A and the vehicle detour mileage coefficient matrix D from m×n matrices to one-dimensional arrays arrTime and arrMileage containing m elements, where m is the number of vehicles participating in this group purchase and n is the maximum number of seats provided to passengers.

[0088] In one implementation, the time tolerance coefficient array arrTime for shared vehicles is derived by taking the row vector group from the passenger time tolerance coefficient matrix A, and using the overall standard deviation algorithm to calculate the time tolerance coefficient values ​​of the vehicles participating in the shared ride, thus forming the array arrTime = (t1, t2, ..., t m For example: take the row vector α of matrix A. i =(α i1 ,a i2 ,...,α in ), and their average values Calculate the time tolerance coefficient t for vehicle i in the group-buying service. i :

[0089]

[0090] t i A higher value indicates that passengers on ride-sharing vehicle i will have to endure a longer journey time due to the insertion of new orders, resulting in a worse passenger experience.

[0091] In one implementation, the detour mileage coefficient array arrMileage for shared vehicles is obtained by taking the row vector group from the vehicle detour mileage coefficient moment D, and using the overall standard deviation algorithm to calculate the detour mileage coefficient values ​​of the vehicles participating in the shared ride, forming the array arrMileage = (d1, d2, ..., d m For example: take the row vector β of matrix D. i =(β) i1 ,β i2 ,...,β in ), and their average values Calculate the detour mileage coefficient value d for the i-th vehicle. i :

[0092]

[0093] d iA higher value indicates that the shared vehicle i takes a longer detour due to the insertion of new orders, resulting in higher vehicle operating costs.

[0094] Step 5: Use a multi-attribute comprehensive analysis mathematical model for carpooling to obtain the vehicle with the best result value for order matching;

[0095] In one implementation, a comprehensive analytical mathematical model with superior service quality and moderate cost is built using arrTime, arrMileage, arrP, and carpooling constraints, and the vehicle with the optimal result value f is selected to issue order tasks.

[0096] The comprehensive analytical mathematical model is as follows:

[0097]

[0098] Where a, b, and c are weight values, and d is other fixed constant factors that need to be considered in the operation and management of shared vehicles. If there are no other factors, it can be set to 0.

[0099] If the weight values ​​are set to a=0.5, b=0.35, c=0.15, d=0, it means that while taking into account the balance of vehicle sharing rate, passenger experience is valued and detour costs are reduced. The values ​​can be adjusted appropriately according to the actual situation.

[0100] Step 6: The vehicle information is sent to the new passenger's client, and the carpooling is successful.

[0101] This embodiment also provides a carpooling system for large transportation hubs based on multi-attribute constraints, including:

[0102] One or more processors;

[0103] A storage device for storing one or more programs that, when executed by one or more processors, enable the one or more processors to implement the carpooling method for large transportation hubs based on multi-attribute constraints.

[0104] This invention is suitable for scheduling and matching passengers and vehicles waiting for ride-sharing at large transportation hubs. By comprehensively analyzing factors such as matched vehicle orders, passenger service quality, travel costs, and ride-sharing constraints, it optimizes vehicles waiting for ride-sharing at large transportation hubs. A mathematical model is built to achieve a balanced balance of indicators such as good passenger experience, moderate detours, and high profitability, thereby achieving optimal passenger matching and route design.

[0105] It should be understood that the above description of the preferred embodiments is quite detailed, but it should not be considered as a limitation on the scope of protection of this invention. Those skilled in the art, under the guidance of this invention, can make substitutions or modifications without departing from the scope of protection of the claims of this invention, and all such substitutions or modifications fall within the scope of protection of this invention. The scope of protection of this invention should be determined by the appended claims.

Claims

1. A carpooling method for large transportation hubs based on multi-attribute constraints, characterized in that, Includes the following steps: Step 1: Obtain basic information on passengers who are willing to participate in group buying at major transportation hubs; Step 2: Obtain information on vehicles already waiting for group rides at major transportation hubs; Step 3: Based on the passenger and vehicle information of the carpooling, filter out vehicles that meet the constraints, quantify the carpooling multi-attribute values, and create a passenger time tolerance coefficient matrix A, a vehicle detour mileage coefficient matrix D, and a vehicle carpooling rate array arrP. The passenger time tolerance coefficient matrix Let A be an m×n matrix, where m is the number of vehicles participating in this group purchase; n is determined by the maximum number of seats provided to passengers, with zeros padded if the number of seats is less than n. Indicates shared vehicles i Upper j Passenger time tolerance coefficient; New orders k Insert the order into the waiting list of all vehicles that have not yet departed, replan the driving route, and calculate the estimated arrival time for each passenger; New orders k After insertion, the time tolerance coefficient for the first order's passengers and available seats is 0, while the time tolerance coefficients for other passengers are: ； in, Indicates shared vehicles i Insert new order k The later j The estimated additional travel time for each passenger; Indicates the first j The optimal time for passengers to travel directly from their origin to their destination without participating in group bookings; r -1=0 indicates the place of origin; Indicates from the first r -1 passenger's destination to the r The travel time to the destination for each passenger; The vehicle detour mileage coefficient matrix D is m × n matrix, n Determined by the maximum number of seats available to passengers, if the number of seats is less than... n Use 0 to fill in the blanks; Indicates the first i vehicle number j Detour distance coefficient for each passenger; New orders k Insert the new order into the list of all pending rides and replan the route, calculating the estimated distance to the destination for each passenger; k After insertion, the detour mileage coefficient for the first order's passenger and empty seats is 0, and the detour mileage coefficient for other passengers is: ； in, Indicates shared vehicles i Insert order k The later j The estimated additional mileage for each passenger; Indicates the first j The mileage traveled by each passenger under the shared ride route planning; Indicates the first r -1 passenger's destination to the r The distance traveled to the passenger's destination; Indicates new orders k The direct distance between two adjacent destinations; Step 4: Combine the passenger time tolerance coefficient matrix A and the vehicle detour mileage coefficient matrix D from... m × n Matrix dimensionality reduction to include m One-dimensional arrays arrTime and arrMileage containing 1 element each. m The number of vehicles participating in this group purchase, n To provide the maximum number of seats for passengers; Step 5: Match the vehicle with the best result value to the order; Step 6: The vehicle information is sent to the new passenger's client, and the carpooling is successful.

2. The carpooling method for large transportation hubs based on multi-attribute constraints according to claim 1, characterized in that: In step 1, the passenger's basic information includes the passenger's name, contact information, desired destination, number of passengers, number and size of luggage, and scheduled departure time; In step 2, the information of the waiting vehicles includes the license plate number, driver's name, driver's contact information, the destination required by the matched passengers, the estimated departure time, the number of available seats, the number and size of luggage items that can be placed, the queuing time, and the queuing order.

3. The carpooling method for large transportation hubs based on multi-attribute constraints according to claim 1, characterized in that: In step 3, the vehicles that meet the constraints satisfy the following constraints: (1) Each passenger transport shall not exceed its passenger capacity; ； in, Indicates shared vehicles i The maximum number of seats available for group booking when there are no empty buses. This indicates the number of seats occupied in the vehicle after inserting the new order k; (2) The number and size of luggage carried by passengers each time shall not exceed the carrying capacity of the vehicle's trunk; ； in, Indicates shared vehicles i The number of pieces of luggage that can be provided in the trunk when the car is empty. Indicates inserting a new order k The number of pieces of luggage subsequently occupied; Indicates shared vehicles i The amount of luggage space available in the trunk when the car is empty. Indicates inserting a new order k The luggage space that will be occupied later; (3) Establish a waiting time of N minutes, that is, the order is completed within N minutes after the first passenger gets on the bus. If the bus departs after N minutes, it will not participate in the subsequent order-sharing service. ； in, Indicates shared vehicles i The waiting time for the first passenger to board the bus, where N is a preset value; (4) For vehicles that have already departed, it is prohibited to add more passengers en route due to temporary empty passenger capacity. ； in, Indicates shared vehicles i The value controls whether a vehicle has been dispatched. When it is 1, it means that the vehicle has been dispatched and no new orders will be accepted. When it is 0, it means that the vehicle has not been dispatched.

4. The carpooling method for large transportation hubs based on multi-attribute constraints according to claim 1, characterized in that: In step 3, the vehicle sharing rate array , Indicates shared vehicles i Seating rate , Indicates shared vehicles i The maximum number of seats available for group booking when there are no empty buses. Indicates inserting a new order k Afterwards, the number of seats in the vehicle was already occupied.

5. The carpooling method for large transportation hubs based on multi-attribute constraints according to claim 1, characterized in that: In step 4, the row vector group in passenger time tolerance coefficient matrix A is taken, and the time tolerance coefficient value of the vehicles participating in the group buying is calculated using the overall standard deviation algorithm, forming an array. ; Take the row vectors of matrix A and their average values Then the shared vehicle i Time tolerance coefficient value ; The larger the value, the more likely it is to be a shared vehicle. i Passengers on the platform will have a worse experience due to the insertion of new orders, which means they will have to endure longer travel times compared to a direct journey from their origin to their destination.

6. The carpooling method for large transportation hubs based on multi-attribute constraints according to claim 1, characterized in that: In step 4, the row vector group in the vehicle detour mileage coefficient moment D is taken, and the detour mileage coefficient value of the vehicles participating in the group purchase is calculated using the overall standard deviation algorithm, forming an array. ; Take the row vectors of matrix D and their average values Then the first Detour distance coefficient of vehicle ; The larger the value, the more likely it is to be a shared vehicle. i The insertion of new orders has resulted in longer detours and higher vehicle order costs.

7. The carpooling method for large transportation hubs based on multi-attribute constraints according to any one of claims 1-5, characterized in that: In step 5, a comprehensive analytical mathematical model is built using arrTime, arrMileage, arrP, and carpooling constraints to achieve optimal service quality and moderate cost, and the optimal result value is selected. Vehicle order placement tasks; The comprehensive analytical mathematical model is as follows: ； Where a, b, and c are weight values, and d is other fixed constant factors in the operation and management of shared vehicles. If there are no other fixed constant factors, it is set to 0. The other fixed constant factors are constant factors other than the passenger time tolerance coefficient matrix A, the vehicle detour mileage coefficient matrix D, and the vehicle sharing rate array arrP.

8. A carpooling system for large transportation hubs based on multi-attribute constraints, characterized in that, include: One or more processors; A storage device for storing one or more programs that, when executed by one or more processors, cause the one or more processors to implement the carpooling method for large transportation hubs based on multi-attribute constraints as described in any one of claims 1 to 7.

Images