Parking reservation and dynamic allocation method for large-scale high-concurrency users

Abstract

The invention relates to a parking reservation and dynamic allocation method for large-scale high-concurrency users, and the method comprises the following steps: S1, collecting the state information of all parking facilities in a region in real time, and parking request information and user state information from a network-connected automatic driving user vehicle or a manual driving user vehicle; s2, according to the distribution moment information obtained in the S1, determining a feasible parking resource set which accords with user preferences and a total user expected cost; s3, describing a parking reservation and distribution system into a dynamic resource distribution problem by using mixed integer linear programming, and setting an objective function and constraint conditions according to parking demands and supply information; and S4, solving the dynamic resource allocation problem in the S3 through an improved Monte Carlo tree search and heuristic rule combined algorithm, and allocating the most suitable parking resources for each user. According to the invention, while the total parking fee of the system user is minimized, the utilization of parking resources can be optimized, and cruise vehicles are reduced.

Description

Technical Field

[0001] The present invention relates to the field of intelligent transportation systems and Internet of Things technologies, and in particular to a parking reservation and dynamic allocation method for large-scale high-concurrency users. Background Art

[0002] With the rapid development of new-generation information technologies such as the Internet of Things, cloud computing, and artificial intelligence, parking resources in various regions can be connected to the Internet for real-time monitoring and management. This allows the regional management center to obtain real-time parking resource occupancy information, thereby optimizing the allocation of parking resources. At the same time, the popularization of "Internet +" in the transportation field and the rapid advancement of networked autonomous driving technology have enabled both manual and automatic driving parking users to apply for parking resource allocation. This networking model can effectively manage parking demand in the region, but it is necessary to comprehensively consider the different operating characteristics of fully autonomous and manually driven vehicles, and formulate appropriate parking reservation and allocation strategies to meet the diverse needs of users and achieve optimal utilization of parking resources.

[0003] In the past decade, urban traffic management departments have developed and deployed urban parking guidance information systems, which collect parking space occupancy status through ultrasonic, video, and geomagnetic sensing technologies, and publish available parking space quantity and fee information through variable information boards or the Internet. However, this method easily guides a large number of vehicles into the same area during peak hours, resulting in intensified local parking competition and some vehicles having no parking space. In order to solve these problems, personalized parking reservation and resource allocation systems have emerged. Based on smartphone applications and smart car terminals, the service center uniformly allocates appropriate parking spaces based on personal preferences and parking resource utilization.

[0004] In recent years, researchers have viewed the parking reservation process as a resource allocation problem and modeled it using mixed integer linear programming to minimize the comprehensive cost of users and maximize the utilization of parking resources. Related studies have considered factors such as the price, quality, parking space occupancy, user walking distance, parking fees, and traffic congestion of parking resources, and designed smartphone applications and intelligent navigation algorithms to optimize parking allocation. In addition, due to the autonomy and non-cooperation of drivers, some studies have introduced theories such as game theory, Nash equilibrium, and stable marriage problems to analyze and solve parking competition behaviors and ensure that users are guided to the most appropriate parking facilities.

[0005] However, the above method only considers the situation of manual driving. In view of the gradual penetration of autonomous driving technology and the gradual acceptance of the market in the future, there is an urgent need for a parking reservation and dynamic allocation method for large-scale and high-concurrency users that takes into account both fully autonomous driving and manually driven vehicles, ensuring user experience in various driving modes and the efficient utilization of parking resources in the region. Summary of the invention

[0006] The purpose of the present invention is to overcome the defects of the above-mentioned prior art and provide a parking reservation and dynamic allocation method for large-scale high-concurrency users.

[0007] The purpose of the present invention can be achieved by the following technical solutions:

[0008] A parking reservation and dynamic allocation method for large-scale high-concurrency users, the method comprising the following steps:

[0009] S1. Real-time collection of status information of all parking facilities in the area, as well as parking request information and user status information from connected autonomous driving user vehicles or manually driven user vehicles, including the real-time location of the vehicle, destination and user preferences;

[0010] S2, determining a set of feasible parking resources that meet the user's preferences and the user's expected total cost based on the allocation time information obtained in S1;

[0011] S3. The parking reservation and allocation system is characterized as a dynamic resource allocation problem using mixed integer linear programming, and the objective function and constraints are set according to the parking demand and supply information;

[0012] S4. Solve the dynamic resource allocation problem described in S3 by improving the algorithm combining Monte Carlo tree search with heuristic rules, and allocate the most suitable parking resources to each user.

[0013] Furthermore, the status information and traffic status information of the parking facilities, as well as the parking request information and user status information of the vehicle can be known in real time in the networked environment; the vehicle must be reserved through the center and parked in the parking area allocated by the system. When it arrives at the destination, there is still no parking space allocated, and it is necessary to cruise around the destination waiting for the system allocation plan, and before arriving at the allocated parking resources, the system can dynamically update the parking allocation plan; the allocation time interval depends on the actual traffic system status, and unified parking space reservation and allocation processing is performed at certain time intervals.

[0014] Furthermore, the users arrive independently and randomly and join the corresponding waiting queue with unlimited capacity. After each allocation moment, the available parking resources are allocated to each user and moved to the corresponding reservation queue.

[0015] Furthermore, the S2 specifically includes the following steps:

[0016] S201, dividing the area into a grid model with blocks as units, and determining the allocation time;

[0017] S202, defining a set of feasible parking resources that meet user preferences from two aspects: expected parking fees and parking resource quality;

[0018] S203. Calculate the expected total cost of the user being allocated parking resources. For manually driven users, it is a weighted average of the expected parking cost and the walking distance. For autonomous driving users, it is a weighted average of the expected parking cost and the autonomous idle driving distance.

[0019] Furthermore, all parking resources within the grid are homogeneous.

[0020] Furthermore, the S202 specifically includes the following steps:

[0021] S2021. Determine the maximum parking fee that a user can accept, the maximum walking distance that a manual driving user can tolerate, and the maximum idle driving distance that an autonomous driving user can tolerate;

[0022] S2022, calculating the walking distance between each manual driving user's activity destination and each parking resource and the empty driving distance between each automatic driving user's activity destination and each parking resource;

[0023] S2023, dynamically formulating a unit charging price for parking resources based on the real-time utilization status, i.e., occupancy rate, of parking resources in each grid;

[0024] S2024, calculating the expected parking fee based on the unit charging price of the parking resource, the accumulated time of the user in the corresponding reservation queue, the driving time, and the actual traffic conditions that affect the driving time and the actual parking time at the parking resource;

[0025] S2025, obtaining the parking facility occupancy status and the number of free parking spaces in the grid at the allocation time, and determining a set of parking resources that are free and can be reserved at the allocation time;

[0026] S2026: Determine a feasible parking resource set that meets the user's preference based on the parking fee, walking distance or empty driving distance, and a set of parking resources that can be reserved.

[0027] Furthermore, the weight factor used in the weighted average is determined according to the relative importance of parking fees and parking resources to the user.

[0028] Furthermore, the walking distance and the empty driving distance are both calculated using Manhattan distance.

[0029] Furthermore, the driving time for manual driving users is the remaining driving time, and for automatic driving users is the expected driving time.

[0030] Furthermore, the unit charging price of the parking resource is charged according to the price at the time of the user's first successful allocation after the user has been allocated to the corresponding reservation queue.

[0031] Furthermore, the S3 specifically includes the following steps:

[0032] S301, introduce 0-1 control variables and define a distribution matrix;

[0033] S302, setting an objective function according to the goal of minimizing the total parking fee of the user while optimizing the utilization of parking resources;

[0034] S303, setting capacity constraints, for each autonomous driving user in the waiting queue and each manual driving user in the waiting queue, at most one parking resource can be allocated, or may fail to be successfully allocated, while ensuring that the number of users occupying parking resources does not exceed the number of available parking spaces in the current grid;

[0035] S304, ensuring that each autonomous driving user and each manual driving user in the reservation queue can maintain an allocated parking resource;

[0036] S305, after the allocation is completed at each time, the expected cost of each user in the corresponding scheduled queue is recorded to ensure that the updated reallocation result of the user in the corresponding scheduled queue is not worse than the last allocation result;

[0037] S306, giving higher priority to users in the corresponding reservation queue, because the users who have made reservations in the expected total cost have already paid for the reservation time;

[0038] S307, ensuring that if a user is closer to his feasible parking resource set than a user with the same feasible parking resource set, the user is given a higher allocation priority;

[0039] S308, respectively setting the time for the automatic driving user and the manual driving user to start entering the allocation system;

[0040] S309: Introduce an emergency allocation queue, and move users whose cumulative waiting time in the corresponding waiting queue exceeds their tolerable maximum cumulative waiting time to the emergency allocation queue.

[0041] Furthermore, the objective function is expressed as:

[0042]

[0043] Where k represents the allocation time, X(k)=[x ij (k)] represents the allocation matrix, x ij (k) represents a 0-1 control variable, e ij (k) represents the expected total cost of user i being assigned to parking resource j, Ω i (k) represents the feasible resource set Ω to which the user can be allocatedi (k), aw(k) represents the waiting queue for autonomous driving at time k, hw(k) represents the waiting queue for manual driving at time k, α represents the penalty fee for autonomous driving users, β represents the penalty fee for manual driving users, α, β>0, β is not less than α.

[0044] Furthermore, users in the emergency allocation queue will be allocated preferentially when feasible resources are available.

[0045] Furthermore, the algorithm combining the Monte Carlo tree search with the heuristic rule will continuously improve the optimal resource allocation scheme through the iteration of the tree search, stop iterating before the computing time or resource constraints are reached, and then return the optimal allocation sequence pair.

[0046] Compared with the prior art, the present invention has the following beneficial effects:

[0047] The present invention uses the Internet of Things technology to collect the status information of parking facilities and the parking request information of users in real time, comprehensively considers the diverse needs of fully autonomous driving and manual driving users, and realizes the dynamic allocation of parking resources through an effective coordination mechanism in an environment where autonomous driving and manual driving coexist, ensuring the optimal performance of the system under various driving modes, thereby significantly improving the efficiency of parking resource management and allocation, meeting the parking needs of large-scale users, and providing strong technical support for the development of future intelligent transportation systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 The parking grid management model and parking flow chart of the present invention;

[0049] Figure 2 A flowchart of parking reservation and dynamic allocation for large-scale users of the present invention;

[0050] Figure 3 A schematic diagram of a queuing model in the parking reservation and allocation system of the present invention;

[0051] Figure 4 Schematic diagram of the average user fee distribution of all feasible solutions of the present invention. DETAILED DESCRIPTION

[0052] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. This embodiment is implemented based on the technical solution of the present invention, and provides a detailed implementation method and specific operation process, but the protection scope of the present invention is not limited to the following embodiments.

[0053] Example

[0054] In this embodiment, a 4km×4km area is divided into 40×40 grids, each grid has a side length of 100m, and contains M j Parking spaces.

[0055] Assume that G active destinations are evenly distributed in the region, and the user arrival rate at each destination g is subject to parameter λ g The Poisson distribution of , the proportion of autonomous driving users is κ.

[0056] The location of the newly arrived user in each allocation time interval in the area follows a two-component mixed Gaussian distribution, with a mean and standard deviation of (1.5km, 1.5km) and (1.0km, 1.0km), respectively.

[0057] To simplify the model, we assume that the driving speed is fixed at 15 km / h and the walking speed is 1.2 m / s.

[0058] Manhattan distance can be used to calculate the distance between any two points, and thus the driving time and walking time between any two points can be obtained.

[0059] The allocation time interval is fixed τ(T)=τ, T=1,2,... The parameter settings are shown in the table:

[0060]

[0061] like Figures 1-2 As shown, a parking reservation and dynamic allocation method for large-scale high-concurrency users, the method comprising the following steps:

[0062] S1. Real-time collection of status information of all parking facilities in the area, as well as parking request information and user status information from connected autonomous driving user vehicles or manually driven user vehicles, including the real-time location of the vehicle, destination and user preferences;

[0063] S2, determining a set of feasible parking resources that meet the user's preferences and the user's expected total cost based on the allocation time information obtained in S1;

[0064] S201, dividing the area into a model with N grids in units of blocks, and determining the allocation time as k;

[0065] S202: Define a feasible parking resource set Ω that meets the user's preferences from two aspects: expected parking fees and parking resource quality. i (k);

[0066] S2021. Determine the maximum parking fee M that user i can accept i , the maximum walking distance W that can be tolerated by manual driving user i i , the maximum idle distance D that the autonomous driving user i can toleratei , Indicates the type of user, autonomous driving user and manual driving user, such as:

[0067]

[0068] S2022. Calculate the walking distance between each manual driving user's activity destination and each parking resource, and the empty driving distance between each automatic driving user's activity destination and each parking resource.

[0069] For manual driving users, we define yj to represent the spatial location of parking resource j. At the allocation time k, the remaining driving distance between user i and parking resource j is d ij (k), remaining travel time t ij (k) is t ij (k) = f(d ij (k), ε(k)), ||·|| represents the Manhattan distance, and ε(k) represents the random traffic state vector at time k.

[0070] We define d i is the actual activity destination of the user. If parking resource j is assigned to user i, the walking distance between user i’s activity destination and the assigned parking resource j is W ij =||d i -y j ||.

[0071] For autonomous driving users, the first step is to find the destination of the user's activity. i Drop off passengers nearby, then automatically drive to the assigned parking resource j, located at y j Therefore, at time k, the remaining driving distance between user i and his activity destination is od i (k) and expected travel time ot i (k) is ot i (k) = f(od i (k),ε(k)), the empty driving distance D from its activity destination to the parking resource location ij D ij =||d i -y j ||;

[0072] S2023. Dynamically formulate the unit charging price of parking resources based on the real-time utilization status, i.e., occupancy rate, of parking resources in each grid.

[0073] Use p j (k) represents the unit parking fee of parking resource j at a certain moment. The parking fee is dynamically determined based on the real-time utilization status of parking resources in each grid, i.e., the occupancy rate, such as:

[0074]

[0075] In the formula, γ j (k) represents the parking price update factor at time k, represents the initial unit parking charge price of parking resource j;

[0076] S2024, based on the unit charging price of parking resources p j (k), and the cumulative time r of the user in the corresponding reservation queue i (k), driving time t ij (k) or autonomous driving driving time ot i (k) and the actual traffic conditions will affect the travel time and the actual parking time T at the parking resource. i , calculate the expected parking cost, such as:

[0077]

[0078] Among them, the cumulative duration and two types of driving time are directly obtained.

[0079] S2025. Obtain the parking facility occupancy status at the allocation time and the number of free parking spaces in the grid.

[0080] Define G(k)={g1(k),...,g N (k)}, represents the parking resource state set of the system, where g j (k) represents the number of free parking spaces in grid j, j=1,...,N, i.e., parking resource j; determine the set of parking resources Γ(k)={j:g j (k)>0,j=1,...N};

[0081] S2026: Determine a feasible parking resource set Ω that meets the user's preference based on the parking fee, walking distance or empty driving distance, and a set of parking resources that can be reserved. i (k), that is, assigning any parking resource j∈Ω to user i i (k):

[0082]

[0083] S203: Calculate the expected total cost e of user i being assigned to parking resource j ij (k).

[0084] For manual driving users, it is the expected parking fee M ij (k) and walking distance W ij For autonomous driving users, it is the expected parking fee Mij (k) and automatic empty distance D ij The weighted average of

[0085]

[0086] In the formula, is the corresponding weight factor, indicating the relative importance of parking fees and parking resources to user i.

[0087] a indicates autonomous driving;

[0088] h means manual driving;

[0089] S3. The parking reservation and allocation system is characterized as a dynamic resource allocation problem using mixed integer linear programming, and the objective function and constraints are set according to the parking demand and supply information;

[0090] S301, introduce 0-1 control variables And define the distribution matrix X(k) = [x ij (k)];

[0091] S302, according to the goal of minimizing the total parking fee of the user while optimizing the utilization of parking resources, the objective function is set as:

[0092]

[0093] Where k represents the allocation time, X(k)=[x ij (k)] represents the allocation matrix, x ij (k) represents a 0-1 control variable, e ij (k) represents the expected total cost of user i being assigned to parking resource j, Ω i (k) represents the feasible resource set Ω to which the user can be allocated i (k), aw(k) represents the waiting queue for autonomous driving at time k, hw(k) represents the waiting queue for manual driving at time k, α represents the penalty fee for autonomous driving users, β represents the penalty fee for manual driving users, α,β>0, β is not less than α, Figure 3 It is a schematic diagram of the queuing model in the parking reservation and allocation system;

[0094] S303: Set capacity constraints.

[0095] For each autonomous driving user in the waiting queue and each manual driving user in the waiting queue, at most one parking resource can be allocated, or it may fail to be allocated successfully, while ensuring that the number of users occupying parking resources does not exceed the number of available parking spaces in the current grid, such as:

[0096]

[0097] S304: Ensure that each autonomous driving user and each manual driving user in the reservation queue can maintain an allocated parking resource, such as:

[0098]

[0099] Among them, aR(k) represents the scheduled queue for automatic driving at time k, and hR(k) represents the scheduled queue for manual driving at time k;

[0100] S305. After the allocation is completed at each time, the expected cost of each user in the corresponding scheduled queue is recorded to ensure that the updated redistribution result of the user in the corresponding scheduled queue is not worse than the last allocation result, such as:

[0101]

[0102] S306, giving higher priority to users in the corresponding reservation queue, because the users who have made reservations in the expected total cost have already paid for the reservation time;

[0103] S307: Ensure that if user i is away from his feasible parking resource set Ω i (k) is closer than user m with the same set of feasible parking resources, i.e., t ij (k)<t mj (k), the user i will be given a higher allocation priority, that is, if x ij (k)=0, then x mj (k) must be 0, such as:

[0104]

[0105] t mj (k)>t ij (k)

[0106] S308: Set the time for the automatic driving user and the manual driving user to enter the allocation system respectively. The expected driving time for the manual driving user to reach his destination is The expected driving time for the autonomous driving user to reach his destination is When entering the distribution system, such as:

[0107]

[0108] S309: Introduce an emergency allocation queue, and move users whose cumulative waiting time in the corresponding waiting queue exceeds their tolerable maximum cumulative waiting time to the emergency allocation queue.

[0109] If the cumulative waiting time c of user i in the corresponding waiting queuei (k) Exceeds the maximum cumulative waiting time that can be tolerated The user will be moved to the urgent allocation queue I(k):

[0110]

[0111] If the emergency allocation queue is not empty, when a feasible resource j is free, the system will give priority to allocating it to user i in the emergency allocation queue I(k). If user i satisfies

[0112] S4. Solve the dynamic resource allocation problem described in S3 by improving the algorithm combining Monte Carlo tree search with heuristic rules, and allocate the most suitable parking resources to each user.

[0113] In order to verify the computational efficiency and effect of the Monte Carlo tree search algorithm, the user costs of all feasible allocation results are calculated and drawn into a distribution histogram, such as Figure 4 As shown in Figure 1. We enumerate all the nearly 100 million feasible solutions and find that the result calculated based on the Monte Carlo tree search algorithm ranks 428th.

[0114] The preferred specific embodiments of the present invention are described in detail above. It should be understood that a person skilled in the art can make many modifications and changes based on the concept of the present invention without creative work. Therefore, any technical solution that can be obtained by a person skilled in the art through logical analysis, reasoning or limited experiments based on the concept of the present invention on the basis of the prior art should be within the scope of protection determined by the claims.

Claims

1. A parking reservation and dynamic allocation method for large-scale high-concurrency users, characterized in that: The method comprises the following steps: S1. Real-time collection of status information of all parking facilities in the area, as well as parking request information and user status information from connected autonomous driving user vehicles or manually driven user vehicles, including the real-time location, destination and user preferences of the vehicles; S2, determining a set of feasible parking resources that meet the user's preferences and the user's expected total cost based on the allocation time information obtained in S1; S3. The parking reservation and allocation system is characterized as a dynamic resource allocation problem using mixed integer linear programming, and the objective function and constraints are set according to the parking demand and supply information; S4. Solve the dynamic resource allocation problem described in S3 by improving the algorithm combining Monte Carlo tree search with heuristic rules, and allocate the most suitable parking resources to each user.

2. A parking reservation and dynamic allocation method for large-scale high-concurrency users according to claim 1, characterized in that: The parking facility status information and traffic status information as well as the vehicle parking request information and user status information are available in real time in the network environment; The vehicle must be booked through the center and parked in the parking area assigned by the system. If there is still no parking space assigned when arriving at the destination, it must cruise around the destination waiting for the system allocation plan. Before arriving at the allocated parking resource, the system can dynamically update the parking allocation plan; The allocation time interval depends on the actual traffic system status, and unified parking space reservation and allocation processing is performed at certain time intervals.

3. The parking reservation and dynamic allocation method for large-scale high-concurrency users according to claim 1 is characterized in that: The users described in step S1 arrive independently and randomly and join the corresponding waiting queue with unlimited capacity. After each allocation moment, the available parking resources are allocated to each user and moved to the corresponding reservation queue.

4. The parking reservation and dynamic allocation method for large-scale high-concurrency users according to claim 1 is characterized in that: The S2 specifically comprises the following steps: S201, dividing the area into a grid model with blocks as units, and determining the allocation time; S202, defining a set of feasible parking resources that meet user preferences from two aspects: expected parking fees and parking resource quality; S203. Calculate the expected total cost of the user being allocated parking resources. For manually driven users, it is a weighted average of the expected parking cost and the walking distance. For autonomous driving users, it is a weighted average of the expected parking cost and the autonomous idle driving distance.

5. A parking reservation and dynamic allocation method for large-scale high-concurrency users according to claim 4, characterized in that: All parking resources within the grid of the grid model are homogeneous.

6. A parking reservation and dynamic allocation method for large-scale high-concurrency users according to claim 4, characterized in that: The S202 specifically includes the following steps: S2021. Determine the maximum parking fee that a user can accept, the maximum walking distance that a manual driving user can tolerate, and the maximum idle driving distance that an autonomous driving user can tolerate; S2022, calculating the walking distance between each manual driving user's activity destination and each parking resource and the empty driving distance between each automatic driving user's activity destination and each parking resource; S2023, dynamically formulating a unit charging price for parking resources based on the real-time utilization status, i.e., occupancy rate, of parking resources in each grid; S2024, calculating the expected parking fee based on the unit charging price of the parking resource, the accumulated time of the user in the corresponding reservation queue, the driving time, and the actual traffic conditions that affect the driving time and the actual parking time at the parking resource; S2025, obtaining the parking facility occupancy status and the number of free parking spaces in the grid at the allocation time, and determining a set of parking resources that are free and can be reserved at the allocation time; S2026: Determine a feasible parking resource set that meets the user's preference based on the parking fee, walking distance or empty driving distance, and a set of parking resources that can be reserved.

7. The parking reservation and dynamic allocation method for large-scale high-concurrency users according to claim 4 is characterized in that: The weight factor used in the weighted average in step S203 is determined according to the relative importance of parking fees and parking resources to the user.

8. The parking reservation and dynamic allocation method for large-scale high-concurrency users according to claim 6 is characterized in that: The walking distance and the empty driving distance in step S2021 are both calculated using Manhattan distance.

9. The parking reservation and dynamic allocation method for large-scale high-concurrency users according to claim 6 is characterized in that: The driving time in step S2024 is the remaining driving time for manual driving users and the expected driving time for automatic driving users.

10. The parking reservation and dynamic allocation method for large-scale high-concurrency users according to claim 6, characterized in that: The method for formulating the initial unit parking charge price in the unit charge price of the parking resources in step S2023 is: when the user has been allocated to the corresponding reservation queue, the price at the time of the user's first successful allocation is used as the initial unit parking charge price.

11. The parking reservation and dynamic allocation method for large-scale high-concurrency users according to claim 1 is characterized in that: The S3 specifically comprises the following steps: S301, introducing control variables and defining a distribution matrix; wherein the control variables include 0 and 1; S302, setting an objective function according to the goal of minimizing the total parking fee of the user while optimizing the utilization of parking resources; S303, setting capacity constraints, for each autonomous driving user in the waiting queue and each manual driving user in the waiting queue, at most one parking resource can be allocated, or may fail to be successfully allocated, while ensuring that the number of users occupying parking resources does not exceed the number of available parking spaces in the current grid; S304, ensuring that each autonomous driving user and each manual driving user in the reservation queue can maintain an allocated parking resource; S305, after the allocation is completed at each time, the expected cost of each user in the corresponding scheduled queue is recorded to ensure that the updated reallocation result of the user in the corresponding scheduled queue is not worse than the last allocation result; S306, giving higher priority to users in the corresponding reservation queue, because the users who have made reservations in the expected total cost have already paid for the reservation time; S307, ensuring that if a user is closer to his feasible parking resource set than a user with the same feasible parking resource set, the user is given a higher allocation priority; S308, respectively setting the time for the automatic driving user and the manual driving user to start entering the allocation system; S309: Introduce an emergency allocation queue, and move users whose cumulative waiting time in the corresponding waiting queue exceeds their tolerable maximum cumulative waiting time to the emergency allocation queue.

12. A parking reservation and dynamic allocation method for large-scale high-concurrency users according to claim 11, characterized in that: The objective function is expressed as: Where k represents the allocation time, X(k)=[x ij (k)] represents the allocation matrix, x ij (k) represents a 0-1 control variable, e ij (k) represents the expected total cost of user i being assigned to parking resource j, Ω i (k) represents the feasible resource set Ω to which the user can be allocated i (k), aw(k) represents the waiting queue for autonomous driving at time k, hw(k) represents the waiting queue for manual driving at time k, α represents the penalty fee for autonomous driving users, β represents the penalty fee for manual driving users, α, β>0, β is not less than α.

13. The parking reservation and dynamic allocation method for large-scale high-concurrency users according to claim 11, characterized in that: Users in the emergency allocation queue will be allocated with priority when feasible resources are available.

14. The parking reservation and dynamic allocation method for large-scale high-concurrency users according to claim 1, characterized in that: The algorithm combining the Monte Carlo tree search with the heuristic rule will continuously improve the optimal resource allocation scheme through the iteration of the tree search, stop iterating before the computing time or resource constraints are reached, and then return the optimal allocation sequence pair.

Images