A parking resource refined management optimization method based on a two-layer model

Abstract

The present invention provides a parking resource refined management optimization method based on a double-layer model, which belongs to the technical field of traffic management. The refined management optimization method is based on the network balanced allocation theory, characterizes regional parking resources (indoor and outdoor parking areas, roadside parking spaces, etc.) as a traffic network model, and establishes a double-layer planning model that considers the parking time of multiple types of users. The upper-level planning reflects the optimization management decision of the parking operation manager on the regional parking resources, including: block differentiated parking pricing, parking block berth capacity design, parking time limit and other management strategies; the lower-level planning is reflected in the parking behavior decision of parking network travelers considering parking time and user time value under a given system optimization scheme, so as to achieve balanced allocation of network parking traffic, and then generate parking resource zoning planning and differentiated management optimization scheme formulation. The present invention provides a theoretical basis for improving the utilization rate of parking resources and the service level of parking areas.

Description

Technical Field

[0001] The present invention belongs to the technical field of traffic management, and in particular relates to a parking resource refined management optimization method based on a double-layer model. Background Art

[0002] The acceleration of urbanization has led to increasingly prominent urban traffic problems, one of which is the rational management and utilization of parking resources. With the increase in the number of cars, the shortage of parking spaces and the waste of parking resources have gradually become a difficult problem for urban traffic management. Traditional parking areas that adopt a unified and homogenized management method often cannot meet parking needs, and do not consider the diversity of parking users and the differences in parking resources. This is because the parking space resources of large parking areas are different, such as the walking distance to the destination and the driving distance to the entrance and exit. The parking time of travelers in the parking network is not uniform, and the proportion of long-term parking overnight vehicles is relatively large, resulting in a large number of long-term parking occupied by high-quality parking resources, low resource turnover, and high-time value users need to spend a lot of time parking.

[0003] The existing technology has achieved relevant results in optimizing parking resource management, but most of them are aimed at road network-level parking management optimization and there is little analysis of behavioral characteristics and differentiated parking management in parking areas.

[0004] In the traditional road network equilibrium model, the detour behavior of drivers is ignored, that is, a path cannot repeat a section of road that has been traveled. In the parking network, due to the existence of parking and patrolling behavior, vehicles usually pass through the same regular section of road many times to find available parking spaces, that is, detours are inevitable. This is the difference between the regional parking network and the traditional road network allocation. In the parking system, there is not only a game between travelers, but also a game between travelers and parking area planners. Each traveler hopes to obtain the maximum travel benefit, while the parking area planner hopes to maximize the travel benefit of the entire parking system. The parking manager makes relevant decisions to change the parking choice behavior of travelers (such as arrival time, parking area selection, parking space selection, etc.), and the travel choice decision results of travelers will in turn affect the management measures of decision makers. The two play against each other, constantly adjust and feedback.

[0005] In response to the above problems, it is necessary to invent a parking resource refined management optimization method based on a two-layer model, so that the parking system can be optimized in the game, and the zoning and differentiated management optimization within the parking area can be realized, thereby improving the utilization rate of parking resources and the service level of the parking area. Summary of the invention

[0006] The object of the present invention is to provide a parking resource refined management optimization method based on a double-layer model, characterized in that it includes the following steps:

[0007] S1: Collect regional parking resource data and build a regional parking system network based on the data;

[0008] S2: Establish a two-level planning model based on the parking traffic network considering the parking time of multiple types of users;

[0009] S3: Design a genetic algorithm to solve the bi-level programming model;

[0010] S4: Generate a refined management plan for regional parking resources based on the solution results of the two-level programming model.

[0011] Furthermore, S1 is specifically as follows: based on the road network structure of the regional parking system, the regional parking resources are divided into multiple parking blocks according to the distances to the activity destination and the driving distances, and a regional parking system network is formed between the multiple parking blocks and the connected paths.

[0012] Furthermore, in S2, the two-level planning model for parking duration of multiple types of users includes an upper-level planning model and a lower-level planning model. The upper-level planning of the upper-level planning model reflects: based on the regional parking network constructed in S1, the parking operation manager makes optimal management decisions on regional parking resources, including regional parking system network modeling, regional parking system allocation model, and regional parking system allocation algorithm;

[0013] The lower-level planning embodiment of the lower-level planning model: The lower-level planning embodiment of the lower-level planning model: describes the micro-behavior of each parking network traveler under a given system optimization scheme, that is, the parking behavior decision considering the parking duration and the user's time value;

[0014] The regional parking system management optimization model is solved by the genetic algorithm in S3 to achieve balanced distribution of network parking traffic. The parking resource zoning planning and differentiated management optimization plan are generated by the regional parking system management optimization method, that is, the refined management plan of regional parking resources in S4.

[0015] Furthermore, the regional parking system network modeling is as follows: multiple parking areas realize differentiated management strategies by selecting different pricing and parking time limits. Parking network travelers select parking spaces in different locations at the entrance of the system based on their own time value, parking time, parking area and parking space attributes (parking fee strategy, entry distance and regional parking occupancy rate, etc.);

[0016] The regional parking system allocation model is as follows: Based on the improved Axhausen model, the physical capacity of parking resources and the impact of parking duration are considered to calculate the generalized parking cost of the regional parking system. It is assumed that all parking users rationally choose parking spaces and make optimal parking decisions based on the information they have learned, so that the parking fee is minimized and user equilibrium is formed. The user equilibrium model is established by treating multiple types of parking areas as parallel network links under fixed OD pairs and expressing the road resistance function on the links with generalized parking fees.

[0017] The regional parking system allocation algorithm is specifically as follows: solving the user equilibrium model based on the universal method of successive average (MSA).

[0018] Furthermore, the regional parking system generalized parking fee Parking fee included PC a , driving time cost Walking time cost and search time costs Driving time fee Driving time fee Walking time cost and search time costs By multiplying the time value (VOT) of the corresponding parking user, it is converted into a monetary cost value to construct the difference in socioeconomic value of different travelers (business people or others); for parking users with discrete time value (VOT) variables, the generalized total parking cost of user type ω in parking area a is expressed as:

[0019]

[0020] in, is the generalized parking fee of user ω in parking area a, is the distance that user ω walks to parking area a, g ω is the average walking speed of user ω, v ω is the time value of user ω, is the number of users of type ω who choose parking area a, p is the path selected by user ω, and f a is the parking fee rate, t is the parking time of user ω, S a The time it takes for a vehicle to search for a valid parking space after entering the parking area. is the hourly arrival rate of vehicles in parking area a, k a is the physical capacity of parking area a, A is the parking area (block) selection set, Ω is the multi-user type set, P a The set of paths from the system entrance to parking area a.

[0021] Furthermore, the user equilibrium is expressed as follows under the equilibrium condition for a specific parking user group and the user type ω and the user's parking time are equal:

[0022]

[0023] in, is the minimum total parking fee for user ω.

[0024] Furthermore, the user equilibrium model is expressed as an optimization (minimization) problem:

[0025]

[0026] Among them, f apω is the flow of users of the parking network of type ω on path p from the system entrance to parking area a, q ω Hourly parking demand for type ω users.

[0027] Furthermore, the regional parking system management optimization model is specifically as follows: based on the two-level planning model, the system optimization problem of the parking network is described, and a regional parking system optimization framework is established. The upper-level planning reflects the optimal management decision of the parking operation manager on the regional parking resources, and the lower-level planning reflects the parking user game equilibrium considering the parking decision under a given system optimization scheme; considering the influencing factors of the parking decision behavior of travelers, the total utility of the driving time, cruising time and walking time of travelers in the parking network is taken as the objective function of the upper-level planning.

[0028] Further, the objective function is expressed as;

[0029]

[0030] Furthermore, the regional parking system management optimization algorithm is specifically as follows: based on the genetic algorithm to solve the two-level programming model, the solution variables are first genetically encoded, and for the generated first-generation population, according to the mechanism of survival of the fittest and survival of the fittest, "excellent" individuals are selected according to individual fitness, and through genetic operators (including crossover, mutation, etc.), better individuals and populations are evolved generation by generation, so as to solve the model, including the following steps:

[0031] S31: Set the maximum evolutionary generation Gmax, crossover probability Pc, mutation probability Pm; population size PopSize; set the evolutionary generation counter g = 0; randomly generate PopSize individuals as the initial population;

[0032] S32: constructing a fitness function and calculating the fitness function value of each chromosome in the population, specifically: using a continuous average method to solve the user equilibrium allocation problem in the bi-level programming, thereby calculating the fitness function;

[0033] S33: According to the fitness function value of the chromosome, a selection operator is executed, and the operator is a roulette selection method. The excellent individuals are inherited to the next generation through the roulette selection method, wherein the fitness of the individual is proportional to the probability of being selected;

[0034] S34: Generate new chromosomes through the crossover operator according to the crossover probability, thereby ensuring the global search capability of the genetic algorithm;

[0035] S35: Select multiple gene positions in the chromosome to change to improve the local search ability of the genetic algorithm;

[0036] S36: When the current genetic generation g reaches the maximum evolutionary generation Gmax, the iteration is terminated. At this time, the individual with the maximum fitness function value is regarded as the optimal solution; otherwise, let g=g+1 and return to S32, repeat S32-S36 until the iteration is terminated.

[0037] For different decision variables, such as parking pricing (continuous variables or discrete variables), parking space capacity design (integer variables), etc., different chromosome encoding methods can be used, such as binary encoding, real number encoding, etc.

[0038] Compared with the prior art, the beneficial effects of the present invention are mainly reflected in: the present invention collects real-time and accurate parking data based on the intelligent parking management system, constructs a parking space selection behavior model that takes into account the differences in parking space resources and the remaining probability of effective parking resources, and realizes the minimization of the travel time, search time and walking distance after parking of travelers in the parking network, as well as the optimization of partial zoning and differentiated management in the parking area, providing a theoretical basis for improving the utilization rate of parking resources and the service level of parking areas. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 The present invention is a flow chart of the refined management optimization method for regional parking resources.

[0040] Figure 2 A framework diagram of the regional parking resource management optimization double-layer programming model of the present invention.

[0041] Figure 3 A flowchart of network allocation of a parking system considering multiple user types provided by an embodiment of the present invention.

[0042] Figure 4 A schematic diagram of a regional parking system provided in an embodiment of the present invention.

[0043] Figure 5 A flowchart of regional parking system management optimization considering multiple user types provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0044] A parking resource refined management optimization method based on a double-layer model of the present invention will be described in more detail below in conjunction with a schematic diagram, wherein a preferred embodiment of the present invention is shown. It should be understood that those skilled in the art can modify the present invention described herein while still achieving the beneficial effects of the present invention. Therefore, the following description should be understood as being widely known to those skilled in the art and not as a limitation to the present invention.

[0045] like Figure 1 As shown, a parking resource refined management optimization method based on a two-layer model includes the following steps:

[0046] Step 1: Collect regional parking resource data and build a regional parking system network based on the data.

[0047] Based on the road network structure of the regional parking system, the regional parking resources are divided into multiple parking blocks according to the distance from the activity destination and the driving distance, and a regional parking system network is formed between the multiple parking blocks and the connected paths.

[0048] Step 2: Based on the parking traffic network, a two-level planning model is established that takes into account the parking time of multiple types of users. Figure 2 shown.

[0049] The two-level planning model is reflected in:

[0050] It includes the upper-level planning model and the lower-level planning model. The upper-level planning of the upper-level planning model is reflected in: the regional parking network constructed in step 1, the network allocation of parking systems considering multiple user types, the optimal management decision of parking operators on regional parking resources, and the reference Figure 3 ,include:

[0051] 1): Regional parking system network modeling;

[0052] Differentiated management strategies are implemented in multiple parking areas by selecting different pricing and parking time limits. Parking network travelers can choose parking spaces in different areas at the entrance of the system based on their own time value, parking time, parking area and parking space attributes (parking fee strategy, entry distance and regional parking occupancy rate, etc.).

[0053] 2): Regional parking system allocation model;

[0054] Based on the improved Axhausen model, the generalized parking cost of the regional parking system is calculated by considering the physical capacity of parking resources and the impact of parking duration. It is assumed that all parking users rationally choose parking spaces and make optimal parking decisions based on the information they have learned to minimize parking fees and form a user equilibrium. By treating multiple types of parking areas as parallel network links under fixed OD pairs, the road resistance function on the link is represented by generalized parking fees, and a user equilibrium model is established.

[0055] Regional parking system generalized parking fee Parking fee included PC a , driving time cost Walking time cost and search time costs Driving time fee Driving time fee Walking time cost and search time costs By multiplying the time value (VOT) of the corresponding parking user, it is converted into a monetary cost value to construct the difference in socioeconomic value of different travelers (business people or others); for parking users with discrete time value (VOT) variables, the generalized total parking cost of user type ω in parking area a is expressed as:

[0056]

[0057] in, is the generalized parking fee of user ω in parking area a, is the distance that user ω walks to parking area a, g ω is the average walking speed of user ω, v ω is the time value of user ω, is the number of users of type ω who choose parking area a, p is the path selected by user ω, and f a is the parking fee rate, t is the parking time of user ω, S a The time it takes for a vehicle to search for a valid parking space after entering the parking area. is the hourly arrival rate of vehicles in parking area a, k a is the physical capacity of parking area a, A is the parking area (block) selection set, Ω is the multi-user type set, P a The set of paths from the system entrance to parking area a.

[0058] 3): Regional parking system allocation algorithm

[0059] The user equilibrium model is solved based on the universal method of successive average (MSA).

[0060] The user equilibrium under the equilibrium condition for a specific parking user group with equal user type ω and user parking time is expressed as:

[0061]

[0062] in, is the minimum total parking fee of user ω. The user equilibrium model is expressed as an optimization (minimization) problem:

[0063]

[0064] Among them, f apω is the flow of users of the parking network of type ω on path p from the system entrance to parking area a, q ω Hourly parking demand for type ω users.

[0065] The lower-level planning of the lower-level planning model is reflected in: considering the regional parking system management optimization of multiple user types, considering the parking behavior decision of parking network travelers considering parking duration and user time value under a given system optimization scheme, solving through genetic algorithm, achieving balanced distribution of network parking flow, and then generating parking resource zoning planning and differentiated management optimization scheme formulation, reference Figure 5 ,include:

[0066] 4): Regional parking system management optimization model

[0067] Based on the two-level planning model, the system optimization problem of the parking network is described, and a regional parking system optimization framework is established. The upper-level planning reflects the optimal management decision of the parking operation manager on the regional parking resources, and the lower-level planning reflects the parking user game equilibrium considering the parking time under a given system optimization scheme; considering the influencing factors of the travelers' parking decision-making behavior, the total utility of the travel time, cruising time and walking time of the travelers involved in the parking network is taken as the objective function of the upper-level planning.

[0068] The objective function is expressed as;

[0069] 5): Regional parking system management optimization method, i.e. step 3.

[0070] Step 3: Design a genetic algorithm to solve the bi-level programming model

[0071] To solve the bi-level programming model based on the genetic algorithm, the solution variables are first genetically encoded. For the generated first-generation population, the "excellent" individuals are selected according to the mechanism of survival of the fittest and survival of the fittest, and according to the individual fitness, and genetic operators (including crossover, mutation, etc.) are used to evolve better individuals and populations from generation to generation, so as to solve the model. The steps include:

[0072] (1): Initialization. Parameter setting: Set the maximum evolutionary generation Gmax, crossover probability Pc, mutation probability Pm; population size PopSize; set the evolutionary generation counter: g = 0; randomly generate PopSize individuals as the initial population.

[0073] (2): Fitness evaluation. Construct a fitness function and calculate the fitness function value of each chromosome in the population. In order to calculate the fitness function, the continuous average method is used to solve the lower model (user equilibrium allocation problem) in the bi-level programming.

[0074] (3): Selection. The greater the individual fitness, the higher the probability of being selected. According to the fitness function value of the chromosome, a selection operator is executed, such as the roulette wheel selection method, to pass excellent individuals to the next generation.

[0075] (4): Crossover. Crossover is to exchange parts of the structures of two parent chromosomes according to the crossover probability to generate new chromosomes. Crossover is the core genetic operation of genetic algorithms, which is mainly completed through the crossover operator.

[0076] (5): Mutation. Mutation is to select certain gene positions in the chromosome for change with a certain mutation probability. Crossover is to ensure the global search capability of the genetic algorithm, while mutation is to improve the local search capability of the genetic algorithm.

[0077] (6): Termination test. If the current genetic generation g has reached the maximum evolutionary generation Gmax, the iteration is terminated, and the individual with the maximum fitness function value is considered the optimal solution; otherwise, let g = g + 1, return to step (2), and repeat steps (2) to (6) until the iteration is terminated.

[0078] For different decision variables, such as parking pricing (continuous variables or discrete variables), parking space capacity design (integer variables), etc., different chromosome encoding methods can be used, such as binary encoding, real number encoding, etc.

[0079] Step 4: Generate a refined management plan for regional parking resources based on the solution of the two-level programming model

[0080] Based on the solution of the two-level programming model, the differentiated management strategy of parking resource partitions is determined, flow distribution and parking guidance are performed on parking resource partitions, and a differentiated management optimization plan for parking resource partitions is generated to achieve refined management and optimization of parking resources.

[0081] The present invention is further described below by means of specific embodiments.

[0082] Example

[0083] Regional Parking System Network Modeling:

[0084] Figure 4 A schematic diagram of a regional parking system provided for an embodiment of the present invention. Taking a parking system near an activity center (such as a public transportation hub or a large shopping mall office building) as an example, the two entrances and exits of the parking system are regarded as nodes, denoted as node 0 and node 7, and the parking area is divided into three blocks, namely A = {1, 2, 3}, corresponding to the network virtual nodes {4, 5, 6}. One of the parking areas is shorter in walking distance from the activity center than the other two parking areas, and the driving time is relatively short, but due to land restrictions, the parking space capacity is limited; the other parking area is moderately walking distance from the activity center, the driving time is moderate, and the parking space capacity is relatively large; the last parking area is the longest in walking distance from the activity center, the driving time is also longer, and the parking space capacity is relatively abundant.

[0085] Parking users are generally divided into two types based on the value of time (VOT), high time value and low time value, representing business affairs and leisure and entertainment purposes respectively; parking duration is also mainly divided into two types, short-term parking and long-term parking.

[0086] Regional parking system allocation model:

[0087] Calculation of generalized parking costs in regional parking systems based on the improved Axhausen model The parking data parameter settings used to establish the user equilibrium model based on the road resistance function of the network link are shown in Table 1-Table 3:

[0088]

[0089] Table 1 Parking area attribute parameters

[0090]

[0091] Table 2 Parking user attribute parameters

[0092]

[0093]

[0094] Table 3 Parking driving section attribute parameters

[0095] Regional parking system allocation algorithm:

[0096] The user equilibrium model is solved based on the universal method of successive average (MSA). The algorithm based on Matlab can be written as follows:

[0097] Initialize the algorithm. Input the initial parameters (q, k, f, s, v).

[0098] Select the initial solution,

[0099] make

[0100] For i=1:3000.

[0101] Utility calculation. Calculate the generalized parking cost of the regional parking system and get

[0102] Augmentation-based objective function Perform an all-or-nothing assignment.

[0103] Generate auxiliary allocation solution

[0104] Mobile update. Update the current path flow value

[0105]

[0106] Terminate the test. If the relative difference meets the predetermined accuracy requirement, terminate the iterative calculation, otherwise set i=i+1 and return to continue the iteration.

[0107]

[0108] Here, ε is the predetermined accuracy requirement (eg, 0.001).

[0109] The optimization steps of the regional parking system considering multiple user types include: regional parking system optimization model and regional parking system optimization algorithm. The specific steps are as follows:

[0110] S4: Regional parking system optimization model:

[0111] Based on the bi-level programming model, the system optimization problem of the parking network is described, and the optimization framework of the parking system in the parking area is established. The upper model is the optimal social total benefit, and the lower model is the balanced distribution. The example optimizes the parking fee of each parking area to optimize the social benefit.

[0112] S5: Regional parking system optimization algorithm:

[0113] The genetic algorithm is used to solve the bi-level programming model, and the following results are obtained. When the parking fee of parking area 1 (VIP parking area) increases from 12 yuan / hour to 18 yuan / hour, the parking fee of parking area 2 (ordinary short-term parking area) remains unchanged at 9 yuan / hour, and the parking fee of parking area 3 (long-term parking area) increases from 6 yuan / hour to 8 yuan / hour, the social benefit of the system is the highest, that is, the total cost of parking time, cruising time and walking time is the smallest. The flow distribution results of each parking block and the corresponding path are shown in Table 4:

[0114]

[0115] Table 4 Parking block flow distribution results

[0116] The above is only a preferred embodiment of the present invention and does not limit the present invention in any way. Any technician in the relevant technical field, without departing from the scope of the technical solution of the present invention, makes any form of equivalent replacement or modification to the technical solution and technical content disclosed in the present invention, which does not depart from the content of the technical solution of the present invention and still falls within the protection scope of the present invention.

Claims

1. A parking resource refined management optimization method based on a double-layer model, characterized in that: The steps include: S1: Collect regional parking resource data and build a regional parking system network based on the data; S2: Establish a two-level planning model based on the parking traffic network considering the parking time of multiple types of users; S3: Design a genetic algorithm to solve the bi-level programming model; S4: Generate a refined management plan for regional parking resources based on the solution results of the two-level programming model; In S2, the two-layer planning model of parking duration of multiple types of users includes an upper-layer planning model and a lower-layer planning model, and the upper-layer planning of the upper-layer planning model reflects: based on the regional parking network constructed in S1, the parking operation manager makes an optimal management decision on regional parking resources, including regional parking system network modeling, regional parking system allocation model and regional parking system allocation algorithm; The regional parking system allocation model is specifically as follows: based on the improved Axhausen model, the generalized parking fee of the regional parking system is calculated by considering the physical capacity of parking resources and the impact of parking duration. Assume that all parking users choose parking spaces rationally and make optimal parking decisions based on the information they have learned, so that parking costs are minimized and user equilibrium is achieved; By treating multiple types of parking areas as parallel network links under fixed OD pairs, and expressing the road resistance function on the links with generalized parking fees, a user equilibrium model is established. Generalized parking fees for the regional parking system Parking fee included PC a , driving time cost Walking time cost and search time costs Driving time fee The travel time fee Walking time cost and search time costs By multiplying the time value of the corresponding parking user, it is converted into a monetary cost value to construct the difference in the socioeconomic value of different travelers; for parking users with discrete time value variables, the generalized total parking cost of user type ω in parking area a is expressed as: in, is the generalized parking fee of user ω in parking area a, is the distance that user ω walks to parking area a, g ω is the average walking speed of user ω, v ω is the time value of user ω, is the number of users of type ω who choose parking area a, p is the path selected by user ω, and f a is the parking fee rate, t is the parking time of user ω, S a The time it takes for a vehicle to search for a valid parking space after entering the parking area. is the hourly arrival rate of vehicles in parking area a, k a is the physical capacity of parking area a, A is the parking area selection set, Ω is the multi-user type set, P a The set of paths from the system entrance to parking area a.

2. The parking resource refined management optimization method based on the double-layer model according to claim 1 is characterized in that: Specifically, S1 is as follows: based on the road network structure of the regional parking system, the regional parking resources are divided into a plurality of parking blocks according to the distances to the activity destination and the driving distances, and a regional parking system network is formed between the plurality of parking blocks and the interconnected paths.

3. The parking resource refined management optimization method based on the double-layer model according to claim 1 is characterized in that: The lower-level planning of the lower-level planning model is as follows: describing the micro-behavior of each parking network traveler under a given system optimization scheme, that is, the parking behavior decision considering the parking duration and the user's time value; The regional parking system management optimization model is solved by the genetic algorithm in S3 to achieve balanced distribution of network parking traffic. The parking resource zoning planning and differentiated management optimization plan are generated by the regional parking system management optimization method, that is, the refined management plan of regional parking resources in S4.

4. The parking resource refined management optimization method based on the double-layer model according to claim 3 is characterized in that: The regional parking system network modeling is specifically as follows: multiple parking areas realize differentiated management strategies by selecting different pricing and parking time limits, and parking network travelers select parking spaces in different locations at the entrance of the system according to their own time value, parking time, parking area and parking space attributes; The regional parking system allocation algorithm is specifically: solving the user equilibrium model based on a universal continuous average method.

5. The parking resource refined management optimization method based on the double-layer model according to claim 4 is characterized in that: The user equilibrium is expressed as follows under the equilibrium condition for a specific parking user group and the user type ω and the user's parking time are equal: in, is the minimum total parking fee for user ω.

6. The parking resource refined management optimization method based on the double-layer model according to claim 5 is characterized in that: The user equilibrium model is expressed as an optimization problem: Subject to: Among them, f apω is the flow of users of the parking network of type ω on path p from the system entrance to parking area a, q ω Hourly parking demand for type ω users.

7. The parking resource refined management optimization method based on the double-layer model according to claim 6 is characterized in that: The regional parking system management optimization model is specifically as follows: based on a two-level planning model to describe the system optimization problem of the parking network, a regional parking system optimization framework is established, the upper-level planning reflects the optimal management decision of the parking operation manager on the regional parking resources, and the lower-level planning reflects the parking user game equilibrium considering the parking decision under a given system optimization scheme; considering the influencing factors of the parking decision behavior of travelers, the total utility of the travel time, cruising time and walking time of travelers in the parking network is used as the objective function of the upper-level planning.

8. The parking resource refined management optimization method based on the double-layer model according to claim 7 is characterized in that: The objective function is expressed as; 9. The parking resource refined management optimization method based on the double-layer model according to claim 8 is characterized in that: The regional parking system management optimization algorithm is specifically: solving a two-level programming model based on a genetic algorithm, including the following steps: S31: Set the maximum evolutionary generation Gmax, crossover probability Pc, mutation probability Pm; population size PopSize; set the evolutionary generation counter g = 0; randomly generate PopSize individuals as the initial population; S32: constructing a fitness function and calculating the fitness function value of each chromosome in the population, specifically: using a continuous average method to solve the user equilibrium allocation problem in the bi-level programming, thereby calculating the fitness function; S33: executing a selection operator according to the fitness function value of the chromosome, wherein the operator is a roulette wheel selection method, and the excellent individuals are inherited to the next generation through the roulette wheel selection method, wherein the fitness of an individual is proportional to the probability of its being selected; S34: Generate new chromosomes through the crossover operator according to the crossover probability, thereby ensuring the global search capability of the genetic algorithm; S35: Select multiple gene positions in the chromosome to change to improve the local search ability of the genetic algorithm; S36: When the current genetic generation g reaches the maximum evolutionary generation Gmax, the iteration is terminated. At this time, the individual with the maximum fitness function value is regarded as the optimal solution; otherwise, let g=g+1 and return to S32, repeat S32-S36 until the iteration is terminated.

Images