Urban parking demand prediction method based on road network traffic carrying capacity

Abstract

The invention discloses an urban parking demand prediction method based on road network traffic bearing capacity, and relates to the urban parking demand prediction method based on the road network traffic bearing capacity. The objective of the invention is to solve the problem that the existing parking demand prediction method does not consider the influence of the road network traffic carrying capacity on the parking demand. The method comprises the following steps: 1, acquiring the total traffic of a road network in a peak time research area; 2, constructing a road network traffic bearing capacity model; 3, enabling the obtained road network traffic total amount of the peak time research area to meet the road network traffic bearing capacity; 4, calculating the parking demand of the research area range in peak hours; 5, obtaining a parking demand prediction value based on the parking demand quantity in the peak-hour research area range and a secondary exponential smoothing method prediction model; and 6, calculating the relative deviation of the predicted value of the parking demand, if the relative deviation is smaller than a threshold value, accepting the predicted value, otherwise, not accepting the predicted value. The method is applied to the field of urban parking demand prediction.

Description

Technical Field

[0001] The invention relates to a method for predicting urban parking demand based on road network traffic carrying capacity. Background Art

[0002] In recent years, the rapid growth of urban economy has led to an increase in the number of cars. Effective parking information services are of great significance to promoting the smooth operation of urban traffic and improving the quality of life of urban residents. Traffic congestion in most cities can be attributed to the difficulty of parking. Research shows that developing accurate and effective prediction technology and providing drivers with real-time parking information is one of the most effective strategies to solve the parking problem.

[0003] However, traditional prediction methods often rely on a single data source, such as population, vehicle or land use data, and ignore the impact of road network traffic carrying capacity on parking demand. In view of this, in order to effectively solve the above problems, it is necessary to propose a parking demand prediction method based on road network traffic carrying capacity. In order to effectively control urban traffic flow, alleviate urban road congestion, reduce the incidence of traffic accidents, and provide reliable data support for parking lot layout design and the formulation of parking-related policies. Summary of the invention

[0004] The purpose of the present invention is to solve the problem that the existing parking demand prediction method does not consider the impact of road network traffic carrying capacity on parking demand, and to propose an urban parking demand prediction method based on road network traffic carrying capacity.

[0005] A method for predicting urban parking demand based on road network traffic carrying capacity. The specific process is as follows:

[0006] Step 1: Obtain the total traffic volume of the road network in the study area during peak hours;

[0007] The peak time is divided into morning peak time and evening peak time;

[0008] Step 2: Construct a road network traffic carrying capacity model;

[0009] Step 3: Make the total traffic volume of the road network in the study area during peak hours obtained in step 1 meet the traffic carrying capacity of the road network;

[0010] Step 4: Calculate the parking demand in the study area during peak hours based on the total traffic volume of the study area road network during peak hours that meets the road network traffic carrying capacity;

[0011] Step 5: Based on the parking demand in the study area during peak hours and the quadratic exponential smoothing prediction model, the parking demand prediction value is obtained;

[0012] Step 6: Calculate the relative deviation of the parking demand prediction value obtained in step 5. If the relative deviation is less than the threshold, the prediction value is accepted, otherwise the prediction value is not accepted.

[0013] The beneficial effects of the present invention are:

[0014] In order to further improve the applicability and accuracy of the urban parking demand prediction method based on the road network traffic carrying capacity, the present invention takes into account the impact of the road network traffic carrying capacity on the parking demand according to the different characteristics of residents' travel modes, and optimizes the method in a targeted manner, so as to determine the number of parking demands under the constraint of the road network traffic carrying capacity.

[0015] The urban parking demand prediction method based on the road network traffic carrying capacity described in the present invention is crucial to the judgment of the scale of urban parking lots. When determining the scale of a parking lot, in addition to meeting the parking needs of residents in the service area, the traffic load capacity of the existing road network must also be considered to prevent traffic congestion in local areas of the road network caused by incoming traffic. The method of the present invention can better optimize the allocation of parking resources and improve the utilization rate of parking spaces. It has significant guiding significance for the planning, design and implementation of parking guidance systems, and has effectively promoted the evolution of parking facilities to an efficient and modern model. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 A flowchart of the method for predicting urban parking demand based on road network traffic carrying capacity according to the present invention;

[0017] Figure 2 It is a schematic diagram of the traffic flow in different directions on the main roads in the central urban area of ​​Beijing during the morning rush hour (8:00-9:00) in the urban parking demand prediction based on the road network traffic carrying capacity of the present invention;

[0018] Figure 3 It is a schematic diagram of the traffic flow in different directions on the main roads in the central urban area of ​​Beijing during the evening rush hour (18:00-19:00) in the urban parking demand prediction based on the road network traffic carrying capacity of the present invention;

[0019] Figure 4 A schematic diagram of the full-day travel purpose composition of residents in the central urban area of ​​Beijing in the urban parking demand prediction based on the road network traffic carrying capacity of the present invention;

[0020] Figure 5 It is a schematic diagram of the number of parking demands in the central urban area of ​​Beijing obtained by the quadratic exponential smoothing method in the urban parking demand prediction based on the road network traffic carrying capacity described in the present invention. DETAILED DESCRIPTION

[0021] Specific implementation method 1: This implementation method is a method for predicting urban parking demand based on road network traffic carrying capacity. The specific process is as follows:

[0022] The technical problem to be solved by the present invention is to overcome the problem that the existing parking demand prediction model does not consider the impact of road network traffic carrying capacity on parking demand, and to provide a method for predicting urban parking demand by using relevant technologies of urban parking facility system planning. In the formulation of the method for predicting parking demand based on road network traffic carrying capacity: first, the parking supply and demand relationship model is selected as the framework for predicting parking demand, effectively analyzing the supply and demand relationship, and better understanding the operation mechanism of the parking market; second, the regional road network traffic composition analysis is studied, and the automobile travel paths are divided according to the diversity of automobile travel starting and ending points, and the parking supply and demand model is optimized and processed; then the road network traffic carrying capacity is introduced into the parking supply and demand relationship model, and the parking demand under the road network capacity constraint in the research area is calculated, and finally the exponential smoothing method is used to predict the parking demand in the future.

[0023] Establish a relationship model between parking facility supply and demand;

[0024] Taking into account the diversity of various travel purposes and modes, combined with the specific parameters of urban parking policy management, the relationship model between parking facility supply and demand is expressed as follows:

[0025]

[0026] Where: D skc is the standard supply of parking facilities for the kth mode of transportation (bicycle, bus, subway, etc.) and the cth type of travel purpose (going to work or school, etc.); t kc is the average parking time coefficient of vehicles with the kth mode of transportation and cth travel purpose; σ is the parking policy management parameter; β is the average utilization rate of parking spaces during peak hours; D dkc is the standard demand for parking facilities for the kth mode of transportation and cth travel purpose; r is the total number of types of transportation modes; q is the total number of types of travel purpose;

[0027] Different categories of parking demand determine different characteristics between parking space supply and vehicle use. When discussing parking demand, the following distinctions can be made:

[0028] For parking for work or school purposes on weekdays, this type of demand is usually manifested as long-term occupation of parking spaces during the day, and vehicles will continue to park during working hours, so their turnover rate is relatively low. This means that each car will occupy a fixed parking space for most of the day. In this case, a relatively stable correspondence is formed between parking spaces and vehicles, which is almost a one-to-one match. Since this type of parking demand is relatively fixed and continuous, the number of such parking spaces is not dependent on the number of repeated uses of parking spaces per unit time, but directly estimates the total number of parking spaces required based on factors such as the number of office workers or students and the assumed vehicle ownership rate.

[0029] Parking at home mostly refers to nighttime parking in residential areas, which is also highly stable. Vehicles will be parked in their own communities or nearby areas for a long time at night and on weekends. Therefore, the supply of parking spaces here is also designed based on the residents' vehicle ownership to ensure that each household has at least one basic parking space.

[0030] On the contrary, for temporary and short-term parking needs such as official travel, shopping, or picking up children, the parking time of vehicles is shorter and the turnover rate of parking spaces is significantly improved. Parking spaces in these scenarios need to have a higher utilization efficiency, that is, they can accommodate more vehicles in a short period of time. Therefore, when planning such parking lots, not only the total number of vehicles is considered, but also the turnover number and utilization rate of parking spaces per unit time are paid more attention to ensure that the parking needs during peak hours can be met, while also trying to avoid the waste of resources caused by vacant parking spaces.

[0031] Step 1: Obtain the total traffic volume of the road network in the study area during peak hours;

[0032] The peak time is divided into morning peak time and evening peak time;

[0033] The morning peak period is 8:00-9:00 (8:00 and 9:00 inclusive);

[0034] The evening peak period is from 18:00 to 19:00 (inclusive);

[0035] Step 2: Construct a road network traffic carrying capacity model;

[0036] Step 3: Make the total traffic volume of the road network in the study area during peak hours obtained in step 1 meet the traffic carrying capacity of the road network;

[0037] Step 4: Calculate the parking demand in the study area during peak hours based on the total traffic volume of the study area road network during peak hours that meets the road network traffic carrying capacity;

[0038] Step 5: Based on the parking demand in the study area during peak hours and the quadratic exponential smoothing prediction model, the parking demand prediction value is obtained;

[0039] Step 6: Calculate the relative deviation of the parking demand prediction value obtained in step 5. If the relative deviation is less than the threshold, the prediction value is accepted, otherwise the prediction value is not accepted.

[0040] Specific implementation method 2: This implementation method is different from specific implementation method 1 in that the total amount of road network traffic in the study area during peak hours is obtained in step 1;

[0041] The peak time is divided into morning peak time and evening peak time;

[0042] The specific process is:

[0043] Vehicles traveling on the road network in the study area are divided into four categories: intra-regional traffic in the study area, transit traffic in the study area, attracted traffic in the study area, and generated traffic in the study area;

[0044] Traffic volume within the study area refers to the volume of motor vehicles that start and end within the study area;

[0045] The transit traffic volume in the study area refers to the motor vehicle flow whose origin and destination are not within the study area;

[0046] The traffic volume attracted by the study area refers to the motor vehicle flow whose starting point is not in the study area but whose end point is in the study area;

[0047] Traffic volume generated in the study area refers to the motor vehicle flow that starts in the study area but ends outside the study area;

[0048] The total traffic volume of the road network in the study area during peak hours is expressed as:

[0049] Q=Q 产 +Q 吸 +Q 区内 +Q 过境 (1)

[0050] Where: Q is the total traffic volume in the study area;

[0051] Q 产 Generate traffic volume for the study area, Q 吸 To attract traffic to the study area, Q 区内 is the traffic volume in the study area, Q 过境 The transit traffic volume in the study area.

[0052] The other steps and parameters are the same as those in the first embodiment.

[0053] Specific implementation method three: This implementation method is different from specific implementation methods one or two in that the road network traffic carrying capacity model is constructed in step two;

[0054] The specific process is:

[0055] There are two ways to quantify the traffic carrying capacity of a road network. The first is to calculate the sum of the traffic carrying capacities of the basic units using the average travel speed of the road network. The second is to calculate the traffic carrying capacity of each basic unit using its own average travel speed, and then add up the carrying capacities of each basic unit.

[0056] The calculation formula of the road network traffic carrying capacity model is expressed as:

[0057]

[0058] The constraints are:

[0059]

[0060] Where: A is the traffic carrying capacity of the road network; A u is the traffic carrying capacity of the u-th basic unit, u=1,2,…,B; is the equivalent lane traffic carrying capacity in the u-th basic unit; is the traffic carrying capacity of the non-equivalent lane in the u-th basic unit; is the traffic carrying capacity of the stacked release zone in the u-th basic unit; is the traffic carrying capacity of the free travel area in the u-th basic unit;

[0061] · is the multiplication sign; B represents the set of basic traffic units in the road network;

[0062] is the traffic proportion of non-equivalent lanes in the u-th basic unit; is the traffic proportion of the equivalent lane in the u-th basic unit;

[0063] p w is the phase sequence of the wth signalized intersection, w∈W; W is the set of signalized intersections in the road network; p u represents the total number of phases contained in the u-th basic unit;

[0064] L u is the length of the basic unit u;

[0065] is the expected average travel speed of the road network; is the expected average travel speed of the u-th basic unit;

[0066] Indicates that in the u-th basic unit, g pw The saturated headway of the traffic flow on the road section ij within the time period;

[0067] g pwis the green light time of the pth phase of the wth signalized intersection;

[0068] E[] is the mathematical expectation;

[0069] In the u-th basic unit, g is pw The original number of lanes allocated;

[0070] λ pw is the green-to-signal ratio of the p-th phase at the w-th signalized intersection;

[0071] C w is the period of the wth signalized intersection;

[0072] is the traffic proportion of the τth direction in the uth basic unit, τ = {right, through, left}, right is right turn, through is straight, left is left turn;

[0073] Traffic ratio refers to the proportion of traffic in each direction to the total traffic, with a total value of 1. The specific proportions of straight, left turn and right turn directions are determined based on actual traffic conditions, such as left turn accounting for 0.3, straight accounting for 0.4, and right turn accounting for 0.3;

[0074] The basic traffic unit is the road between the upstream and downstream intersection stop lines;

[0075] The equivalent lane is a lane that converts the intermittent traffic flow controlled by the signal on the import lane into a continuous release traffic flow;

[0076] The non-equivalent lane is a lane on the import lane that continuously releases traffic flow without signal control;

[0077] Free travel means that vehicles travel at free flow speed on the roads within sections and intersections, and the corresponding area is called the free travel area;

[0078] Cascading release means that the traffic flow is released in a cascading manner before the stop line, and the corresponding area is called the cascading release driving area;

[0079] According to the operation mode of the traffic flow in the stacked release area after the traffic flow is abstracted, the vehicles stand in the set area in a stacked form and are released one by one from the top; one is stacked on top of another and released one by one.

[0080] The basic traffic unit is defined as the smallest repeating structure of the road network, that is, the road between the upstream and downstream intersection stop lines, specifically including the upstream section driving area and the downstream intersection driving area;

[0081] An equivalent lane is a new type of import lane. The key is to convert the intermittent traffic flow controlled by signals on it into a continuous release traffic flow while keeping the number of vehicles unchanged. The lane that can continuously release traffic flow after the equivalent is the equivalent lane.

[0082] Lanes that are not controlled by signals and continuously release traffic, such as right-turn lanes at signalized intersections and lanes at non-signalized intersections, are called non-equivalent lanes.

[0083] The other steps and parameters are the same as those in the first or second embodiment.

[0084] Specific implementation method 4: This implementation method is different from any one of specific implementation methods 1 to 3 in that in step 3, the total amount of road network traffic in the study area during peak hours obtained in step 1 satisfies the road network traffic carrying capacity;

[0085] The traffic volume of transit traffic does not generate parking demand in the study area. Therefore, when calculating the total parking demand in the area during peak hours, the impact of this part of the traffic volume is excluded and it does not need to be taken into account. On the other hand, there is a special case where vehicles depart from the starting point in the study area to the destination outside the area, especially during peak hours. After such vehicles leave the study area, the parking spaces they previously occupied in the area will be released, which is actually equivalent to increasing the effective parking supply in the area during that time period. Therefore, the traffic volume generated by such vehicles that start in the area, end outside the area, and leave during peak hours is also not taken into account when calculating parking demand, because they not only do not increase parking demand, but help alleviate parking tension in the area.

[0086] Therefore, when analyzing the parking demand in the study area, only Q 吸 With Q 区内 Two parts of traffic volume; when actually collecting these two parts of data, they will be determined based on relevant statistical data from the traffic management department.

[0087] The expression is:

[0088] Q≤A (5).

[0089] The other steps and parameters are the same as those in Specific Embodiments 1 to 3.

[0090] Specific implementation method 5: This implementation method is different from any one of specific implementation methods 1 to 4 in that, in step 4, the total amount of road network traffic in the study area during peak hours that meets the road network traffic carrying capacity is used to calculate the parking demand in the study area during peak hours;

[0091] The specific process is:

[0092] Step 4.1. Parking demand in the study area during peak hours is calculated by the traffic volume Q attracted by the study area. 吸 and the traffic volume Q in the study area 区内 Composition, its expression is:

[0093]

[0094] Where: D d is the parking demand in the research scope during peak hours; γ is the average parking generation rate of various travel modes and travel purposes during peak hours; k is the traffic mode classification; c is the travel purpose classification; S t-1 The regional parking demand generated by private car trips for category C travel purposes in the study area during peak hours;

[0095] For the peak time zone traffic volume Q 区内 For peak time, the traffic volume Q 吸 For the parking demand in the study area, its size is directly affected by the road network supply conditions in the study area. When this part of the traffic flow increases to the threshold of the traffic carrying capacity that the road network can bear, its increment will reach saturation and stop growing. Therefore, the parking demand in the study area during peak hours is actually limited by the traffic carrying capacity of the road network. At this time, the parking demand model in the study area during peak hours is as follows:

[0096]

[0097] Buses and taxis usually have their own dedicated parking lots, which are usually located near the city's transportation hubs or public transportation stations. Since the main task of these vehicles is to provide convenient travel services for passengers rather than parking for a long time, there is almost no parking demand for these vehicles during peak hours. Therefore, when calculating the peak hour parking demand, only private cars are considered;

[0098] Step 42: Calculate the parking demand for private cars in the study area during peak hours; the expression is:

[0099] D′ d =(Q′ 区内 +Q′ 吸 )×γ′ (7)

[0100] in:

[0101]

[0102] Where: D′ d Q′ is the parking demand for private cars in the study area during peak hours; 吸 The traffic volume attracted by private car travel in the study area; Q′ 区内is the traffic volume in the private car travel area of ​​the study area; γ′ is the average parking generation rate of private cars for various travel purposes; μ c The proportion of private cars for category C trips in the total number of trips; γ c The rate of private car parking generation for type c travel purposes;

[0103] Step 4.3: Study the parking demand D′ of private cars in the area based on peak hours d , calculate the regional parking demand S generated by private car travel for type c travel purpose in the study area t-1 ;

[0104] The calculation formula is:

[0105]

[0106] The other steps and parameters are the same as those in Specific Embodiments 1 to 4.

[0107] Specific implementation method 6: This implementation method is different from any one of specific implementation methods 1 to 5 in that, in step 5, a parking demand forecast value is obtained based on the parking demand in the study area during peak hours and a quadratic exponential smoothing method forecast model;

[0108] The specific process is:

[0109] The quadratic exponential smoothing method not only makes up for the limitation that the first exponential smoothing method can only predict the data of the future period, but also is applicable to the market phenomenon with obvious trend changes. Therefore, the quadratic exponential smoothing method is selected as the prediction method of parking demand.

[0110] The quadratic exponential smoothing forecasting model is as follows:

[0111] Y t+T =a t +b t T (10)

[0112] Where: Y t+T is the parking demand forecast value for the t+T period; a t 、b t is the coefficient of the quadratic exponential smoothing method; T is the time period, which is a positive integer; t is the time.

[0113] The other steps and parameters are the same as those in Specific Implementation Methods 1 to 5.

[0114] Specific implementation method 7: This implementation method is different from any one of the specific implementation methods 1 to 6 in that the quadratic exponential smoothing coefficient a t 、b t The expression is:

[0115]

[0116] Where: for The first exponential smoothing value of ;

[0117] for The double exponential smoothing value of ;

[0118] a is the smoothing coefficient, 0≤a≤1.

[0119] The other steps and parameters are the same as those in Specific Embodiments 1 to 6.

[0120] Specific implementation eight: This implementation differs from any one of the specific implementations one to seven in that the first exponential smoothing method value The expression is:

[0121]

[0122] Where: y t is the actual value in period t;

[0123] The regional parking demand S generated by private car travel for type C travel purpose in the study area t-1 An exponential smoothing of

[0124] The quadratic exponential smoothing value The expression is:

[0125]

[0126] Where: The regional parking demand S generated by private car travel for type C travel purpose in the study area t-1 The second exponential smoothing value of .

[0127] The other steps and parameters are the same as those in Specific Embodiments 1 to 7.

[0128] Specific implementation method 9: This implementation method is different from any one of the specific implementation methods 1 to 8 in that the value of the smoothing coefficient a is determined by:

[0129] Substituting a=0.1 into formula (13) and formula (14), we can get the new exponential smoothing value and the new quadratic exponential smoothing value The new exponential smoothing value and the new quadratic exponential smoothing value Substitute into formula (11) and formula (12) respectively to obtain the new quadratic exponential smoothing coefficient a t 、b t ; Set the new quadratic exponential smoothing coefficient at 、b t Substitute into formula (10) to obtain the new parking demand forecast value;

[0130] Substituting a=0.2 into formula (13) and formula (14), we can get the new exponential smoothing value and the new quadratic exponential smoothing value The new exponential smoothing value and the new quadratic exponential smoothing value Substitute into formula (11) and formula (12) respectively to obtain the new quadratic exponential smoothing coefficient a t 、b t ; Set the new quadratic exponential smoothing coefficient a t 、b t Substitute into formula (12) to obtain the new parking demand forecast value;

[0131] Substituting a=0.3 into formula (13) and formula (14), we can get the new exponential smoothing value and the new quadratic exponential smoothing value The new exponential smoothing value and the new quadratic exponential smoothing value Substitute into formula (11) and formula (12) respectively to obtain the new quadratic exponential smoothing coefficient a t 、b t ; Set the new quadratic exponential smoothing coefficient a t 、b t Substitute into formula (10) to obtain the new parking demand forecast value;

[0132] Substituting a=0.4 into formula (13) and formula (14), we can obtain the new exponential smoothing value and the new quadratic exponential smoothing value The new exponential smoothing value and the new quadratic exponential smoothing value Substitute into formula (11) and formula (12) respectively to obtain the new quadratic exponential smoothing coefficient a t 、b t ; Set the new quadratic exponential smoothing coefficient a t 、b t Substitute into formula (10) to obtain the new parking demand forecast value;

[0133] Substituting a=0.5 into formula (13) and formula (14), we can get the new exponential smoothing value and the new quadratic exponential smoothing value The new exponential smoothing value and the new quadratic exponential smoothing value Substitute into formula (11) and formula (12) respectively to obtain the new quadratic exponential smoothing coefficient a t 、b t ; Set the new quadratic exponential smoothing coefficient a t 、b t Substitute into formula (10) to obtain the new parking demand forecast value;

[0134] Substituting a=0.6 into formula (13) and formula (14), we can get the new exponential smoothing value and the new quadratic exponential smoothing value The new exponential smoothing value and the new quadratic exponential smoothing value Substitute into formula (11) and formula (12) respectively to obtain the new quadratic exponential smoothing coefficient a t 、b t ; Set the new quadratic exponential smoothing coefficient a t 、b t Substitute into formula (10) to obtain the new parking demand forecast value;

[0135] Substituting a=0.7 into formula (13) and formula (14), we can get the new exponential smoothing value and the new quadratic exponential smoothing value The new exponential smoothing value and the new quadratic exponential smoothing value Substitute into formula (11) and formula (12) respectively to obtain the new quadratic exponential smoothing coefficient a t 、b t ; Set the new quadratic exponential smoothing coefficient a t 、b t Substitute into formula (10) to obtain the new parking demand forecast value;

[0136] Substituting a=0.8 into formula (13) and formula (14), we can obtain the new exponential smoothing value and the new quadratic exponential smoothing value The new exponential smoothing value and the new quadratic exponential smoothing value Substitute into formula (11) and formula (12) respectively to obtain the new quadratic exponential smoothing coefficient a t 、b t ; Set the new quadratic exponential smoothing coefficient a t 、b t Substitute into formula (10) to obtain the new parking demand forecast value;

[0137] Substituting a=0.9 into formula (13) and formula (14), we can get the new exponential smoothing value and the new quadratic exponential smoothing value The new exponential smoothing value and the new quadratic exponential smoothing value Substitute into formula (11) and formula (12) respectively to obtain the new quadratic exponential smoothing coefficient a t 、b t ; Set the new quadratic exponential smoothing coefficient a t 、b t Substitute into formula (10) to obtain the new parking demand forecast value;

[0138] The mean square errors of all new parking demand prediction values ​​are compared, and the smoothing coefficient corresponding to the new parking demand prediction value with the smallest mean square error is taken as the optimal value of the smoothing coefficient.

[0139] The formula for calculating the mean square error is: N is the total number of samples, and z represents the zth sample.

[0140] The other steps and parameters are the same as those in Specific Embodiments 1 to 8.

[0141] Specific implementation method ten: This implementation method is different from any one of specific implementation methods one to nine in that, in step six, the relative deviation of the parking demand prediction value obtained in step five is calculated. If the relative deviation is less than a threshold value, the prediction value is accepted, otherwise the prediction value is not accepted;

[0142] The specific process is:

[0143] The parking demand forecast value Y is obtained by using the quadratic exponential smoothing method t+T , which needs to be evaluated.

[0144] Using the parking demand forecast value Y t+T The actual observed value y t+T The relative deviation σ k As an evaluation criterion, if the relative deviation σ k If it is less than the threshold value of 15%, the predicted value is accepted, otherwise the predicted value is not accepted;

[0145] Relative deviation σ k The calculation formula is as follows:

[0146]

[0147] The other steps and parameters are the same as those in Specific Embodiments 1 to 9.

[0148] Example

[0149] In order to verify the effect of the present invention, the embodiment selects the central urban area of ​​Beijing as an actual case for analysis, and verifies the constructed "a method for predicting urban parking demand based on road network traffic carrying capacity".

[0150] 1. Analysis of parking supply and demand

[0151] When selecting the central urban area of ​​Beijing as an actual case for analysis, parking needs can be divided into commuting, commuting, official business and others. According to the sample survey, Beijing residents are more inclined to "green travel" and therefore are more inclined to choose public transportation. The structure of transportation travel for different purposes is shown in Table 1.

[0152] Table 1 Composition of travel modes for different travel purposes in the central urban area of ​​Beijing

[0153] Public Transportation Private Car Cycling walk taxi other Commuting 43％ 22％ 17％ 14％ 1％ 3％ Going to and from school 37％ 5％ 26％ 21％ 2％ 9％ Business trip 45％ 16％ 17％ 13％ 8％ 1％ other 40％ 15％ 7％ 8％ 20％ 10％

[0154] 2. Analysis of regional road network traffic composition

[0155] Refer to the annual traffic development report of Beijing over the years, and take the data of Beijing's central urban area in 2020 as an example to demonstrate the analysis process of road network traffic composition. The traffic in Beijing's central urban area during peak hours in 2020 has the form of local tides. Figure 2 During the morning rush hour from 8:00 to 9:00, the traffic volume reached 18,157 pcu / h. During this period, the traffic volume heading towards the central city totaled 10,890 pcu / h, while the traffic volume leaving the central city was 7,267 pcu / h. Figure 3 During the evening rush hour from 18:00 to 19:00, the traffic volume was 16706pcu, of which the traffic volume into the central urban area was 6984pcu / h, while the traffic volume out of the city was as high as 9722pcu / h.

[0156] In 2020, the traffic pressure in the central urban area of ​​Beijing during peak hours was significant, and the annual average traffic index reached 5.07. In detail, the annual average traffic index during the morning peak period was 4.87, while the annual average traffic index during the evening peak period was 5.26, showing the differences and characteristics of traffic congestion during the morning and evening peaks. Based on the above data, the road traffic composition of the central urban area can be calculated, and the specific values ​​are shown in Table 2. From the analysis of the road network traffic composition, it is known that the traffic volume Q in the evening peak area of ​​the central urban area is 区内 and attract traffic volume Q 吸 It accounts for 52.4% of the total traffic volume, so the evening peak period is selected for calculation in the central urban area.

[0157] Table 2 Traffic composition of the road network in the central urban area during the morning and evening peak hours

[0158]

[0159] See also Figure 4 In 2020, commuting trips in the central urban area accounted for 46.5% of all trips, while life trips accounted for 52.6% of all trips. In this category, trips for activities such as shopping, leisure and entertainment, and fitness accounted for a high proportion. This shows that in addition to work and study, people's lives are becoming more and more colorful. According to the composition table of all-day travel purposes of residents in the central urban area of ​​Beijing and the composition table of travel modes for different travel purposes, the proportion of various purposes of residents in the central urban area of ​​Beijing using private cars for travel can be calculated, as shown in Table 3.

[0160] Table 3 Proportion of residents' private car travel for different purposes during peak hours in the central urban area

[0161] Purpose of travel Commuting Going to and from school Business trip Lifestyle travel Proportion(%) 54.1 0.8 0.8 44.3

[0162] 3. Determination of road network traffic carrying capacity

[0163] According to formulas (2), (3) and (4), the traffic carrying capacity of the road network in the central urban area of ​​Beijing in 2020 is 1579778 pcu. Combined with the traffic composition data of the road network in the morning and evening peak hours in the central urban area of ​​Beijing, the private car traffic volume Q′ in the area can be obtained: 区内 and attracted traffic volume Q′ 吸 Combined with the proportion of different purposes of private car travel by residents in the central urban area, the traffic volume of private cars in the area and the traffic volume of attracted private cars when residents travel for different purposes can be obtained, as shown in Table 4.

[0164] Table 4 Traffic volume shared by different residents during peak hours in the central urban area of ​​Beijing

[0165]

[0166]

[0167] 4. Relationship between road network traffic carrying capacity and parking demand

[0168] According to basic data, the average parking generation rate γ for each type of travel purpose is 0.715. Therefore, according to formula (7), the peak hour parking demand in the central urban area of ​​Beijing in 2020 under the constraint of road network traffic carrying capacity can be obtained as follows:

[0169] D′ d =(604531+223272)×0.715=591879 (pieces)

[0170] After determining the peak hour parking demand under the constraints of the central urban road network traffic carrying capacity, combined with the proportion of different purposes of residents in the central urban area using private cars for travel, the parking demand quantity in the central urban area of ​​Beijing based on the road network traffic carrying capacity can be obtained, see Table 5.

[0171] Table 5 Parking demand for different purposes of vehicle travel during peak hours in central Beijing

[0172] Purpose of travel Commuting Going to and from school Business trip other Parking demand (units) 320207 4735 4735 262202

[0173] 5. Parking Demand Forecasting Based on Quadratic Exponential Smoothing

[0174] Referring to the above steps, the peak hour parking demand in the central urban area of ​​Beijing under the constraints of road network traffic carrying capacity in 2018 and 2019 can be calculated in the same way, which are 1,065,490 and 905,096 respectively. Through the quadratic exponential smoothing method, according to formulas (10), (11), (12), (13), and (14), it is determined that the mean square error (MSE) is the smallest when the optimal smoothing parameter α is 0.64. Then, using the above series of data from 2018 to 2020, the number of parking demands in the central urban area of ​​Beijing from 2021 to 2023 can be predicted, which are 591,538 in 2021, 615,236 in 2022, and 623,415 in 2023.

[0175] According to Beijing's annual traffic development reports, in 2020, there were 76,000 electronic parking spaces on the roads, 2,579 commercial public parking lots, and 707,977 parking spaces, of which 73% were concentrated in the central urban area, that is, there are 572,303 parking spaces in the central urban area of ​​Beijing. Similarly, it can be determined that the current parking spaces in the central urban area of ​​Beijing in 2021, 2022, and 2023 will be 612,478, 649,459, and 678,948 respectively. By comparing the actual and predicted values ​​of the number of parking demands in the central urban area of ​​Beijing from 2021 to 2023, we can get Figure 5 . See Figure 5 The performance of the model on historical data shows a high degree of accuracy. According to formula (15), the average prediction error is 5.7%. If the relative deviation is within 15%, the fit is acceptable, which shows that the model can effectively capture and predict changes in parking demand. In 2020, parking demand fell sharply, and data collection was difficult. There were many types of variables to be determined. Currently, only a series of traffic data since 2018 can be obtained from the official website, so the amount of verification data is small. We look forward to the official website publishing more data to further verify the effectiveness of the model.

[0176] The present invention may also have many other embodiments. Without departing from the spirit and essence of the present invention, those skilled in the art may make various corresponding changes and modifications based on the present invention, but these corresponding changes and modifications should all fall within the scope of protection of the claims attached to the present invention.

Claims

1. A method for predicting urban parking demand based on road network traffic carrying capacity, characterized by: The specific process of the method is: Step 1: Obtain the total traffic volume of the road network in the study area during peak hours; The peak time is divided into morning peak time and evening peak time; Step 2: Construct a road network traffic carrying capacity model; Step 3: Make the total traffic volume of the road network in the study area during peak hours obtained in step 1 meet the traffic carrying capacity of the road network; Step 4: Calculate the parking demand in the study area during peak hours based on the total traffic volume of the study area road network during peak hours that meets the road network traffic carrying capacity; Step 5: Based on the parking demand in the study area during peak hours and the quadratic exponential smoothing prediction model, the parking demand prediction value is obtained; Step 6: Calculate the relative deviation of the parking demand prediction value obtained in step 5. If the relative deviation is less than the threshold, the prediction value is accepted, otherwise the prediction value is not accepted.

2. The method for predicting urban parking demand based on road network traffic carrying capacity according to claim 1, characterized in that: In the step 1, the total traffic volume of the road network in the study area during peak hours is obtained; The peak time is divided into morning peak time and evening peak time; The specific process is: Vehicles traveling on the road network in the study area are divided into four categories: intra-regional traffic in the study area, transit traffic in the study area, attracted traffic in the study area, and generated traffic in the study area; Traffic volume within the study area refers to the volume of motor vehicles that start and end within the study area; The transit traffic volume in the study area refers to the motor vehicle flow whose origin and destination are not within the study area; The traffic volume attracted by the study area refers to the motor vehicle flow whose starting point is not in the study area but whose end point is in the study area; Traffic volume generated in the study area refers to the motor vehicle flow that starts in the study area but ends outside the study area; The total traffic volume of the road network in the study area during peak hours is expressed as: Q=Q 产 +Q 吸 +Q 区内 +Q 过境 (1) Where: Q is the total traffic volume in the study area; Q 产 Generate traffic volume for the study area, Q 吸 To attract traffic to the study area, Q 区内 is the traffic volume in the study area, Q 过境 The transit traffic volume in the study area.

3. The method for predicting urban parking demand based on road network traffic carrying capacity according to claim 2, characterized in that: In the step 2, a road network traffic carrying capacity model is constructed; The specific process is: The calculation formula of the road network traffic carrying capacity model is expressed as: The constraints are: Where: A is the traffic carrying capacity of the road network; A u is the traffic carrying capacity of the u-th basic unit, u=1,2,…,B; is the equivalent lane traffic carrying capacity in the u-th basic unit; is the traffic carrying capacity of the non-equivalent lane in the u-th basic unit; is the traffic carrying capacity of the stacked release zone in the u-th basic unit; is the traffic carrying capacity of the free travel area in the u-th basic unit; · is the multiplication sign; B is the set of basic traffic units in the road network; is the traffic proportion of non-equivalent lanes in the u-th basic unit; is the traffic proportion of the equivalent lane in the u-th basic unit; p w is the phase sequence of the wth signalized intersection, w∈W; W is the set of signalized intersections in the road network; p u represents the total number of phases contained in the u-th basic unit; L u is the length of the basic unit u; is the expected average travel speed of the road network; is the expected average travel speed of the u-th basic unit; Indicates that in the u-th basic unit, g pw The saturated headway of the traffic flow on the road section ij within the time period; g pw is the green light time of the pth phase of the wth signalized intersection; E[] is the mathematical expectation; In the u-th basic unit, g is pw The original number of lanes allocated; λ pw is the green-to-signal ratio of the p-th phase at the w-th signalized intersection; C w is the period of the wth signalized intersection; is the traffic proportion of the τth direction in the uth basic unit, τ = {right, through, left}, right is right turn, through is straight, left is left turn; The basic traffic unit is the road between the upstream and downstream intersection stop lines; The equivalent lane is a lane that converts the intermittent traffic flow controlled by the signal on the import lane into a continuous release traffic flow; The non-equivalent lane is a lane on the import lane that continuously releases traffic flow without signal control; Free travel means that vehicles travel at free flow speed on the roads within sections and intersections, and the corresponding area is called the free travel area; Cascading release means that traffic is released in a cascading manner before the stop line, and the corresponding area is called the cascading release driving area.

4. The method for predicting urban parking demand based on road network traffic carrying capacity according to claim 3, characterized in that: In the step 3, the total amount of road network traffic in the study area during peak hours obtained in step 1 satisfies the road network traffic carrying capacity; The expression is: Q≤A (5).

5. The method for predicting urban parking demand based on road network traffic carrying capacity according to claim 4, characterized in that: In the step 4, the parking demand in the study area during peak hours is calculated based on the total traffic volume of the study area during peak hours that meets the traffic carrying capacity of the road network; The specific process is: Step 4.

1. Parking demand in the study area during peak hours is calculated by the traffic volume Q attracted by the study area. 吸 and the traffic volume Q in the study area 区内 Composition, its expression is: Where: D d is the parking demand in the research scope during peak hours; γ is the average parking generation rate of various travel modes and travel purposes during peak hours; k is the traffic mode classification; c is the travel purpose classification; S t-1 The regional parking demand generated by private car trips for category C travel purposes in the study area during peak hours; Step 42: Calculate the parking demand for private cars in the study area during peak hours; the expression is: D′ d =(Q′ 区内 +Q′ 吸 )×γ′ (7) in: Where: D′ d Q′ is the parking demand for private cars in the study area during peak hours; 吸 The traffic volume attracted by private car travel in the study area; Q′ 区内 is the traffic volume in the private car travel area of ​​the study area; γ′ is the average parking generation rate of private cars for various travel purposes; μ c The proportion of private cars for category C trips in the total number of trips; γ c The rate of private car parking generation for type c travel purposes; Step 4.3: Study the parking demand D′ of private cars in the area based on peak hours d , calculate the regional parking demand S generated by private car travel for type c travel purpose in the study area t-1 ; The calculation formula is:

6. The method for predicting urban parking demand based on road network traffic carrying capacity according to claim 5, characterized in that: In the step 5, based on the parking demand in the study area during peak hours and the quadratic exponential smoothing prediction model, a parking demand prediction value is obtained; The specific process is: The quadratic exponential smoothing forecasting model is as follows: Y t+T =a t +b t T (10) Where: Y t+T is the parking demand forecast value for the t+T period; a t 、b t is the coefficient of the quadratic exponential smoothing method; T is the time period, which is a positive integer; t is the time.

7. The method for predicting urban parking demand based on road network traffic carrying capacity according to claim 6, characterized in that: The quadratic exponential smoothing coefficient a t 、b t The expression is: Where: for The first exponential smoothing value of ; for The double exponential smoothing value of ; a is the smoothing coefficient, 0≤a≤1.

8. The method for predicting urban parking demand based on road network traffic carrying capacity according to claim 7, characterized in that: The first exponential smoothing value The expression is: Where: y t is the actual value in period t; The regional parking demand S generated by private car travel for type C travel purpose in the study area t-1 An exponential smoothing of The quadratic exponential smoothing value The expression is: Where: The regional parking demand S generated by private car travel for type C travel purpose in the study area t-1 The second exponential smoothing value of .

9. The method for predicting urban parking demand based on road network traffic carrying capacity according to claim 8, characterized in that: The determination process of the value of the smoothing coefficient a is as follows: Substituting a=0.1 into formula (13) and formula (14), we can get the new exponential smoothing value and the new quadratic exponential smoothing value The new exponential smoothing value and the new quadratic exponential smoothing value Substitute into formula (11) and formula (12) respectively to obtain the new quadratic exponential smoothing coefficient a t 、b t ; Set the new quadratic exponential smoothing coefficient a t 、b t Substitute into formula (10) to obtain the new parking demand forecast value; Substituting a=0.2 into formula (13) and formula (14), we can get the new exponential smoothing value and the new quadratic exponential smoothing value The new exponential smoothing value and the new quadratic exponential smoothing value Substitute into formula (11) and formula (12) respectively to obtain the new quadratic exponential smoothing coefficient a t 、b t ; Set the new quadratic exponential smoothing coefficient a t 、b t Substitute into formula (12) to obtain the new parking demand forecast value; Substituting a=0.3 into formula (13) and formula (14), we can get the new exponential smoothing value and the new quadratic exponential smoothing value The new exponential smoothing value and the new quadratic exponential smoothing value Substitute into formula (11) and formula (12) respectively to obtain the new quadratic exponential smoothing coefficient a t 、b t ; Set the new quadratic exponential smoothing coefficient a t 、b t Substitute into formula (10) to obtain the new parking demand forecast value; Substituting a=0.4 into formula (13) and formula (14), we can obtain the new exponential smoothing value and the new quadratic exponential smoothing value The new exponential smoothing value and the new quadratic exponential smoothing value Substitute into formula (11) and formula (12) respectively to obtain the new quadratic exponential smoothing coefficient a t 、b t ; Set the new quadratic exponential smoothing coefficient a t 、b t Substitute into formula (10) to obtain the new parking demand forecast value; Substituting a=0.5 into formula (13) and formula (14), we can get the new exponential smoothing value and the new quadratic exponential smoothing value The new exponential smoothing value and the new quadratic exponential smoothing value Substitute into formula (11) and formula (12) respectively to obtain the new quadratic exponential smoothing coefficient a t 、b t ; Set the new quadratic exponential smoothing coefficient a t 、b t Substitute into formula (10) to obtain the new parking demand forecast value; Substituting a=0.6 into formula (13) and formula (14), we can get the new exponential smoothing value and the new quadratic exponential smoothing value The new exponential smoothing value and the new quadratic exponential smoothing value Substitute into formula (11) and formula (12) respectively to obtain the new quadratic exponential smoothing coefficient a t 、b t ; Set the new quadratic exponential smoothing coefficient a t 、b t Substitute into formula (10) to obtain the new parking demand forecast value; Substituting a=0.7 into formula (13) and formula (14), we can get the new exponential smoothing value and the new quadratic exponential smoothing value The new exponential smoothing value and the new quadratic exponential smoothing value Substitute into formula (11) and formula (12) respectively to obtain the new quadratic exponential smoothing coefficient a t 、b t ; Set the new quadratic exponential smoothing coefficient a t 、b t Substitute into formula (10) to obtain the new parking demand forecast value; Substituting a=0.8 into formula (13) and formula (14), we can obtain the new exponential smoothing value and the new quadratic exponential smoothing value The new exponential smoothing value and the new quadratic exponential smoothing value Substitute into formula (11) and formula (12) respectively to obtain the new quadratic exponential smoothing coefficient a t 、b t ; Set the new quadratic exponential smoothing coefficient a t 、b t Substitute into formula (10) to obtain the new parking demand forecast value; Substituting a=0.9 into formula (13) and formula (14), we can get the new exponential smoothing value and the new quadratic exponential smoothing value The new exponential smoothing value and the new quadratic exponential smoothing value Substitute into formula (11) and formula (12) respectively to obtain the new quadratic exponential smoothing coefficient a t 、b t ; Set the new quadratic exponential smoothing coefficient a t 、b t Substitute into formula (10) to obtain the new parking demand forecast value; The mean square errors of all new parking demand prediction values ​​are compared, and the smoothing coefficient corresponding to the new parking demand prediction value with the smallest mean square error is taken as the optimal value of the smoothing coefficient.

10. The method for predicting urban parking demand based on road network traffic carrying capacity according to claim 9, characterized in that: In step 6, the relative deviation of the parking demand prediction value obtained in step 5 is calculated. If the relative deviation is less than a threshold, the prediction value is accepted; otherwise, the prediction value is not accepted; The specific process is: Using the parking demand forecast value Y t+T The actual observed value y t+T The relative deviation σ k As an evaluation criterion, if the relative deviation σ k If it is less than the threshold value of 15%, the predicted value is accepted, otherwise the predicted value is not accepted; Relative deviation σ k The calculation formula is as follows:

Images