Method for predicting operation and maintenance of urban shared electric bicycles
By analyzing the travel characteristics and urban inherent properties of shared electric motorcycles, using travel feature vector weighted regression to predict the rental volume and income of shared electric motorcycles, the problems of low prediction efficiency and low cost-effectiveness in the existing technology are solved, and the cost-effectiveness of operation and maintenance in the existing technology are achieved. Efficient operation and maintenance and resource utilization.
Patent Information
- Application Number
- CN202410997980.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-24
- Publication Date
- 2024-11-15
AI Technical Summary
The existing prediction methods of shared electric motorcycles are not efficient in scenarios with large data volume or large dimensions, and cannot achieve real-time prediction. The lack of theoretical support leads to low cost-effectiveness in operation and maintenance.
By analyzing the travel characteristics of shared electric motorcycles based on historical data, using travel feature vector weighted regression to predict the rental volume, and establishing an operation and maintenance data model based on the inherent properties of the city and influencing factors to conduct operation and maintenance prediction of shared electric motorcycles.
It realizes accurate prediction of the demand and revenue of shared electric motorcycles, provides theoretical basis for operation and maintenance, improves operational efficiency and utilization rate, and reduces resource waste.
Smart Images

Figure CN118966415A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of shared bicycle planning, and in particular to a prediction method for urban shared electric motorcycle operation and maintenance. Background Art
[0002] The current market demand for shared electric motorcycles is mostly predicted based on machine learning models. However, for scenarios with large data volumes or multiple data dimensions, the efficiency and scalability of machine learning methods are low, making it difficult to apply them in actual city-level shared electric motorcycle demand forecasts. In addition, it is impossible to achieve real-time forecasts for the next time period based on the current scenario, and it is impossible to effectively assist vehicle operation and maintenance. Therefore, there is little research on the current market forecasting of shared electric motorcycle revenue, lack of theoretical support for vehicle operation scenarios, and the actual problem of failing to improve the cost-effectiveness of shared electric motorcycle operations. A new shared electric motorcycle forecasting method is urgently needed.
[0003] Chinese patent document CN202311213665.6 discloses a "method for analyzing the self-circulation behavior of a shared bicycle mobile chain considering potential demand". The method adopted includes: (1) determining the inventory status of each grid based on the origin-destination (OD) trip data of the shared bicycle and calibrating the review label; (2) predicting the travel demand of each grid through the multiply censored Tobit model; (3) using the supply and demand matching method to form a new shared bicycle travel chain based on the minimum fleet size; (4) reconstructing the shared bicycle travel chain; (5) identifying the self-circulation chain in the reconstructed shared bicycle travel chain and analyzing various indicators. Furthermore, the steps for forming a new shared bicycle travel chain in step (3) are: traversing each trip, searching for the nearest available shared bicycle to assign to the trip, and if there is no shared bicycle within an acceptable walking distance, generating a new shared bicycle until all trips are completed; finally, a dictionary is generated to store all the trajectory points of the shared bicycle. Furthermore, the specific process of reconstructing the shared bicycle travel chain in step (4) is as follows: (41) Initialization and sorting: Initialize an empty dictionary to store the shared bicycle chain and sort the travel data according to the start time; (42) Calculate distance: For each trip in the travel data, calculate the distance between the current position and the last recorded position of each shared bicycle; (43) Select and update: If there is at least one shared bicycle within an acceptable walking distance, select the nearest shared bicycle, and then update the shared bicycle chain dictionary to assign the new position to the selected shared bicycle sequence; (44) Create a new shared bicycle: If there is no shared bicycle within an acceptable walking distance, create a new shared bicycle using the current location details and add it to the shared bicycle chain dictionary. Furthermore, the specific steps of identifying and reconstructing the self-circulating chain in the shared bicycle travel chain in step (5) are as follows: (51) obtaining the shared bicycle travel chain reconstructed in step (4), wherein each shared bicycle chain node contains latitude and longitude information; (52) matching the latitude and longitude to the grid ID, using the latitude and longitude information, matching the latitude and longitude of each shared bicycle chain node with the grid information to determine the grid ID corresponding to the shared bicycle chain node; (53) traversing each shared bicycle chain node, calculating the number of times the grid ID appears, and for each shared bicycle chain node, counting the number of times its corresponding grid ID appears in the entire bicycle chain, which can be achieved by using a counter or dictionary during the traversal process; (54) if a grid ID appears n times in the entire shared bicycle chain, it is considered that the grid forms n-1 self-circulating chains, where n≥2. However, the patent does not take into account the prediction and calculation of the shared bicycle revenue part, and cannot provide assistance to the shared bicycle operation and maintenance. Summary of the invention
[0004] The present invention mainly solves the technical problem that the original shared electric motorcycles cannot be effectively and reasonably operated and maintained, and provides a prediction method for the operation and maintenance of urban shared electric motorcycles. The present invention obtains the historical data of shared electric motorcycles in the tested area. The above technical problem of the present invention is mainly solved by the following technical solutions: The present invention includes S1 analyzes the travel characteristics of shared electric motorcycles based on historical data; S2 predicts the rental volume of shared electric motorcycles based on the weighted regression of travel feature vectors; S3 calculates the operation and maintenance data of urban shared electric motorcycles based on travel characteristics to complete the prediction of urban shared electric motorcycle operation and maintenance.
[0005] Preferably, the specific steps of analyzing the travel characteristics of shared electric motorcycles in step S1 are as follows: S11 analyzes historical riding and revenue data of shared electric bikes; S12 analyzes the inherent attribute data of the city; S13 filters out row vector features.
[0006] Preferably, the specific steps of predicting the demand for shared electric motorcycles in step S2 are as follows: S21 calculates the benchmark rental data of shared electric bikes; S22 analyzes the cycling probability of different age groups; S23 analyzes the relationship between urban influencing factors and the rental volume of shared electric motorcycles; S24 builds a shared electric motorcycle rental volume prediction model.
[0007] Preferably, the specific steps of estimating the income of the city's shared electric motorcycles in step S3 are as follows: S31 analyzes the distribution of cycling time in different age groups based on the least squares model; S32 Establish a model for estimating the average income of a single trip of a shared electric motorcycle; S33 Establish a model for estimating the revenue of shared electric motorcycles; S34 completes the forecast of shared electric motorcycle demand and revenue based on the shared electric motorcycle single-trip average revenue estimation model and the shared electric motorcycle revenue estimation model.
[0008] Preferably, the expression of the shared electric motorcycle rental volume prediction model is as follows: R XY =(P XY *G a1 +P Xa2 *G a2 )*W t(Y) *W r(Y) *W h(Y) Where X represents the predicted region, Y represents the predicted day, a1 and a2 represent the age segments of two different renters, and R represents the rental amount. Then R XY It represents the rental volume in area X for Y days, G represents the riding probability, and G a1 represents the riding probability of the renting population in age segment a1, P represents the population, P Xa1 represents the population of age group a1 in region X, W represents the weight, W t(Y) W represents the weight of the impact of the temperature on the rental volume in region X on day Y. r(Y) W represents the weight of the impact of rainfall on the rental volume in region X on day Y. h(Y) Indicates the weight of whether day Y in region X is a working day on the rental volume.
[0009] Preferably, the expression of the model for estimating the average income of a single trip of a shared electric motorcycle in step S3 is as follows: in, represents the average one-way fare in region X, n represents the n time periods of the fare standard in region X, and f X is the charging standard of area X, L is the time period, then f X (L) is the charge in area X when the riding time is in time period L, and p(L) is the probability of riding time in time period L.
[0010] Preferably, the prediction formula for predicting the rental volume of shared electric motorcycles within a period of time in step S24 is as follows: Where R X[i,j] is the forecast result of the rental volume in region X from day i to day j, and m represents the number of days from day i to day j.
[0011] Preferably, the shared electric motorcycle revenue estimation model in step S3 includes the total revenue of shared electric motorcycles in the predicted area from the i-th day to the j-th day and the daily average revenue of each vehicle in the time interval from the i-th day to the j-th day. The total revenue of shared electric motorcycles in the predicted area from the i-th day to the j-th day is expressed as follows: Among them I X[i,j] That is, the total revenue of shared electric motorcycles in region X from day i to day j.
[0012] Preferably, the expression of the average daily income per vehicle in the time interval from the i-th day to the j-th day in the tested area is as follows: in is the average daily income of each vehicle in region X from day i to day j, V Xis the number of vehicles deployed in area X from day i to day j.
[0013] Preferably, in step S3, the total rental income for a period of time in the test area and the average daily income of shared electric motorcycles are obtained based on the known influence of different inherent city attributes and city influencing factors on riding demand, and combined with the shared electric motorcycle single-trip average income estimation model and the shared electric motorcycle income estimation model. The demand and income forecast is completed through the relevant parameter indicators of the above-mentioned shared electric motorcycles, and the vehicle dispatch of shared bicycles in the test area is assisted.
[0014] The beneficial effects of the present invention are as follows: the present invention provides quantitative support for slow traffic planning by predicting the demand and income of shared electric motorcycles in the city. Based on the travel characteristics and demand prediction model of shared electric motorcycles, the demand for rental and return of electric motorcycles by urban residents can be known. Through the travel characteristics and income estimation model of shared electric motorcycles, a theoretical basis can be provided for the deployment and operation units of electric motorcycles, assisting in the deployment of shared electric motorcycles, and then enabling effective operation and maintenance of shared electric motorcycles, ultimately reducing resource waste and improving the utilization rate of shared electric motorcycles. BRIEF DESCRIPTION OF THE DRAWINGS
[0015] Figure 1 It is a flow chart of the travel characteristics analysis of shared electric motorcycles based on historical data of the present invention.
[0016] Figure 2 It is a flow chart of the present invention for predicting the demand for shared electric motorcycles by weighted regression of travel feature vectors.
[0017] Figure 3 It is a flow chart of the present invention for estimating the income of urban shared electric motorcycles based on travel characteristics. DETAILED DESCRIPTION
[0018] The technical solution of the present invention is further specifically described below through embodiments and in conjunction with the accompanying drawings.
[0019] Example 1: A prediction method for the operation and maintenance of a city-shared electric motorcycle in this example, such as Figure 1 As shown, before predicting the demand and income of shared electric motorcycles in any region, it is necessary to first understand the relevant historical data of the region. Only by analyzing the relevant data can we more accurately grasp the demand for shared electric motorcycles in the region, and based on these historical data, we can improve the accuracy of the prediction for the region. Therefore, it is necessary to analyze the travel characteristics of shared electric motorcycles in the tested area based on historical data. This step includes analyzing the historical riding and income data of shared electric motorcycles. In this embodiment, four typical related cities A, B, C, and D are selected, and the basic data of each city is counted, including the average daily income and ranking of a single electric motorcycle in the city, to obtain the following typical city electric motorcycle income data table.
[0020] By calculating the number of vehicles and total vehicle income in different cities in the above table over a period of time, we can get the average daily income of vehicles in different cities within the statistical time range, and then we can know that the average daily income of vehicles in cities A and B is relatively good. In the form of the above chart, when we need to understand the travel characteristics of shared electric motorcycles in any test area, we can intuitively feedback the riding demand status of shared bicycles by combining the historical statistical data of different test cities within a certain statistical time range, which is convenient for the selection of subsequent travel vector features.
[0021] After initially understanding the usage status of shared bicycles in the cities to be tested through the historical data of shared bicycle revenue, it is necessary to analyze the inherent attribute data of the cities to be tested. The differences in riding demand in different cities to be tested are largely subject to the inherent attributes of the cities. In this embodiment, based on historical data, 10 factors that affect the rental volume of shared electric motorcycles in cities are selected, and combined with the selected typical cities, the influencing factors are listed to obtain the following typical city influencing factor table.
[0022] Analyzing the past data of the typical cities selected in the above table, we can see that the population aged 18-59 is the main cycling users, accounting for about 98% of the total cycling population; the temperature of 10-30 degrees is the most suitable climate for cycling. Therefore, the analysis results of the typical city influencing factor table can assist relevant persons in charge to clearly identify the factors affecting different cities and thus improve the accuracy of the analysis results of the correlation between the cities to be tested and the rental volume and income of shared electric motorcycles.
[0023] After completing the analysis of the historical riding and income data of the above-mentioned shared electric motorcycles and the inherent attributes of the city, the travel vector characteristics of shared electric motorcycles in different cities to be tested are finally screened out according to the analysis results. During the screening, according to the correlation analysis of the influencing factors in the above-mentioned statistical table and the income of shared electric motorcycles, the number of popular areas in the city to be tested, such as the number of business districts, colleges and universities and transportation hubs; the proportion of the population suitable for riding is concentrated in which part; the number of days with suitable weather for riding is specifically concentrated in which time period of the year and other data results. Therefore, it can be clearly understood that in this embodiment, the number of popular areas in the city, the proportion of the population suitable for riding, and the number of days with suitable weather for riding are the main factors affecting the overall rental volume and income of electric motorcycles in the city. And the time factor is an important factor affecting whether users choose shared electric motorcycles as a means of travel. Further, the temperature, rainfall conditions and whether it is a holiday on the day when considering whether to choose a shared bicycle are the three most important factors affecting the rental volume. Therefore, through the above steps, the specific travel vector characteristics of different cities to be tested can be accurately known, thereby providing effective data support for the subsequent prediction of the demand for shared electric motorcycles to be tested in the subsequent timeline of different cities.
[0024] like Figure 2 As shown, the specific steps of predicting the demand for shared electric motorcycles based on the weighted regression of the travel feature vector are as follows: the benchmark rental data of shared electric motorcycles in the city to be tested needs to be calculated in advance; in this embodiment, combined with the results of the analysis of the influencing factors in the step of analyzing the travel characteristics of shared electric motorcycles based on historical data, the 15-day rental volume data under the benchmark conditions of City A and City B are selected respectively, and it can be known that the average value of the benchmark rental volume in City A is 13407, and the average value of the benchmark rental volume in City B is 9725, which are used as the daily rental volume benchmark data of City A and City B respectively. Then, it is necessary to analyze the riding probability of the population in different age groups. Since the rental volume demand of different age groups is different, the riding probability of the population aged 18-59 and over 60 in City A and City B under the benchmark conditions is analyzed. The riding probability is obtained by calculating the number of shared electric motorcycles rented by the population in this age group / the number of people in this age group. After calculation, the following comprehensive riding probability table of people in the same age group is obtained. Age Group Average value of city A Average value of city B Comprehensive average value (G) Comprehensive median 18-59 years old 2.86% 3.06% 2.96% 2.95% Over 60 years old 0.14% 0.11% 0.125% 0.124%
[0025] At the same time, it is also necessary to analyze the relationship between urban influencing factors and the rental volume of shared electric motorcycles. Urban influencing factors include temperature, rainfall, and whether it is a weekday. When analyzing the relationship between temperature and rental volume, since the rental volume is different under different temperature weather, in this embodiment, City A and City B are compared and analyzed under other identical external conditions, such as weekdays or no rain; the temperature settings include different temperatures such as below 10 degrees, 10-30 degrees (suitable riding temperature), and above 30 degrees. Under the above conditions, the rental volume of shared electric motorcycles in the two cities is obtained, and the data of the two cities are combined to obtain the following weight table of the impact of different temperatures on the rental volume of shared electric motorcycles.
[0026] When analyzing the relationship between rainfall and rental volume, since the rental volume under different rainfall weather conditions is different, in this embodiment, the rental volume of city A and city B under rainfall and no rainfall conditions with the same external conditions is compared and analyzed, and the following weight table of the impact of different rainfall weather on the rental volume of the two cities is obtained: When analyzing the relationship between whether it is a weekday and the rental volume, due to the different rental demand on weekdays and holidays, a comparative analysis was conducted on the rental volume on weekdays and holidays in the two cities under the same other external conditions to obtain the following weight table of the impact of different date properties on the rental volume in the two cities.
[0027] After completing the above data table statistics, a shared motorcycle rental volume prediction model is constructed based on the data in the table. In this embodiment, based on the population of area A, the weather conditions of the day, and whether it is a holiday, the motorcycle rental volume in area A on a certain day can be estimated, as shown in the following formula (1): R AB = (P Aa1 * G a1 +P Aa2 *G a2 )*W t(B) *W r(B) *W h(B) (1) Where A represents region A, B represents day B, a1 and a2 represent the age groups of 18-59 and over 60 respectively; R represents the rental amount, then R AB It represents the rental volume in area A on day B. G represents the riding probability. a1 It represents the cycling probability of people aged 18-59 years old. P represents the population, P Aa1 W represents the population aged 18-59 in region A, and W represents the weight. t(B) represents the weight of the impact of the temperature on the rental volume in area A on day B, W r(B) represents the weight of the impact of rainfall on the rental volume in area A on day B, W h(B) Indicates the influence weight of whether the Bth day in region A is a working day on the rental volume. If you need to calculate the rental volume over a period of time, you can get the following formula (2), which is the rental volume estimation formula for region A from day i to day j: The above-mentioned shared electric motorcycle rental volume prediction model can accurately know the shared electric motorcycle rental volume in different tested cities in a certain period of time in the future, which is convenient for relevant personnel to formulate operation and maintenance configuration plans for shared electric motorcycles in the tested cities within this time period.
[0028] The output of the shared electric motorcycle rental volume prediction model can be used to further estimate the income status of shared electric motorcycles in the city to be tested. Figure 3 As shown, the specific steps for estimating the income of shared electric motorcycles in cities are as follows: first, the least squares model is used to analyze the distribution of riding time of different age groups. Since the single-trip charging situation is related to the user's riding time, in this embodiment, 16 days of travel data in cities A and B under different weather, weekdays and holidays are selected, and the distribution ratio of riding time is analyzed. Through the comprehensive analysis of the data of cities A and B, a fitting curve of the riding time distribution ratio will be obtained, and the charging probability p(t) of different time periods will be calculated. The specific calculation method is shown in the following formula (3): P(y)=5.243*10 -10 *y 8-2015*10 -8 *y 7 -4.432*10 -6 *y 6 -3.861*10 -5 *y 5 -0.0008944*y 4 +0.01002*y 3 -0.0559*y 2 +0.1209y+0.04339 (3) Among them, t is the time interval, that is, 16.
[0029] Combined with the fitting results of the riding time distribution ratio fitting curve, a model for estimating the average income of a single trip of a shared electric motorcycle is established. When establishing, since the income of a single trip is related to the charging standard of the region, in this embodiment, the trip data of City A is used as the standard, and after fitting the riding time distribution probability curve of shared electric motorcycles, the average single trip charge in the region can be estimated. The specific estimation formula is as follows: In the formula represents the average one-way fare in area A, n represents the fare standard in the area divided into n time periods, and f A It indicates the charging standard of area A, y indicates the time period, and f A (y) represents the charge for riding in area A within the time period y, p represents the probability, and p(y) is the probability of riding in the time period y obtained according to the above fitting formula.
[0030] At the same time, it is also necessary to establish a model for estimating the income of shared electric motorcycles. When establishing it, since the total income is the product of the rental volume and the average income of a single trip, the above formulas (3) and (4) can be used to obtain the formula for calculating the total income of area A from the i-th day to the j-th day. The specific calculation method is shown in the following formula (5): And according to formula (4), the average daily income of each vehicle in area A from day i to day j can be obtained, as shown in formula (6): Among them, V A is the number of vehicles deployed in area A from day i to day j. In this embodiment, it is assumed that the number of deployed vehicles remains unchanged and is within a reasonable deployment range. The above formula is used to complete the establishment of the shared electric motorcycle single-trip average income estimation model and the shared electric motorcycle income estimation model. Finally, the shared electric motorcycle demand and income prediction is completed based on the shared electric motorcycle single-trip average income estimation model and the shared electric motorcycle income estimation model.
[0031] With the acceleration of urbanization and the continuous growth of population, urban traffic problems are becoming increasingly complex. Slow traffic refers to a mode of transportation based on walking or non-motorized vehicles, which can greatly alleviate problems such as urban traffic congestion. The present invention provides quantitative support for slow traffic planning by using a prediction method for the operation and maintenance of urban shared electric motorcycles. Through the electric motorcycle travel characteristics and demand prediction model proposed in this embodiment, it is possible to preliminarily evaluate the demand for rental and return of electric motorcycles by urban residents, assist the government and transportation planning departments in formulating appropriate slow traffic planning policies, and promote the healthy development of urban transportation. At the same time, through the user's use of electric motorcycle travel characteristics and income estimation model proposed in this patent, it can provide a theoretical basis for the deployment and operation of electric motorcycles, assist in vehicle deployment, site construction, charging rules formulation, etc., reduce resource waste, improve vehicle utilization, and more targetedly help citizens alleviate traffic difficulties, provide a basis for the site selection and construction of public electric motorcycle transportation infrastructure, and maximize revenue. In addition, through the shared electric motorcycle travel demand prediction model proposed in this patent, the vehicle demand at different times in the corresponding city, region or station can be calculated in real time, and the operation and maintenance personnel can be assisted to transfer vehicles between stations, improve the operation and maintenance efficiency, meet the user's vehicle needs, and provide data support for the operation and maintenance of public electric motorcycles. Therefore, by using the prediction method of the demand and income of urban shared electric motorcycles in this embodiment, before the relevant managers adjust the deployment strategy of shared electric motorcycles in a certain city, they can preliminarily estimate the overall demand for electric motorcycle rental and return in the city, and then more reasonably deduce the vehicles that can be deployed in the city, thereby improving the operational cost-effectiveness of shared electric motorcycles. When the relevant managers are adjusting the deployment and maintenance strategy of shared bicycles in a certain city, with the help of the prediction method in the present invention, the demand for electric motorcycle rental and return at the station or region can be refined and estimated, and the staff can be arranged to transfer vehicles in a more targeted manner, improve operational efficiency, and reduce operating costs.
[0032] Embodiment 2: In this embodiment, a prediction method for the operation and maintenance of a city shared electric motorcycle is provided. Before the prediction, the travel characteristics of the shared electric motorcycle need to be analyzed based on historical data. The specific steps for analyzing the travel characteristics of the shared electric motorcycle are as follows: Analyze the historical riding and income data of the shared electric motorcycle. The expression of the shared electric motorcycle rental volume prediction model is as follows: R XY =(P XY *G a1 +P Xa2 *G a2 )*W t(Y) *W r(Y) *W h(Y) Where X represents the predicted region, Y represents the predicted day, a1 and a2 represent the age segments of two different renters, and R represents the rental amount. Then R XYIt represents the rental volume in area X for Y days, G represents the riding probability, and G a1 represents the riding probability of the renting population in age segment a1, P represents the population, P Xa1 represents the population of age group a1 in region X, W represents the weight, W t(Y) W represents the weight of the impact of the temperature on the rental volume in region X on day Y. r(Y) W represents the weight of the impact of rainfall on the rental volume in region X on day Y. h(Y) Indicates the weight of whether the Yth day in region X is a working day on the rental volume; analyzes the inherent attribute data of the city; and filters out the travel vector features. Then, based on the travel feature vector weighted regression, the demand for shared electric motorcycles is predicted. The specific steps for predicting the demand for shared electric motorcycles are as follows: calculate the benchmark rental data of shared electric motorcycles; analyze the riding probability of different age segments of the population; analyze the relationship between urban influencing factors and the rental volume of shared electric motorcycles. The prediction formula for predicting the rental volume of shared electric motorcycles within a period of time is as follows: Where R X[i,j] is the forecast result of the rental volume in region X from day i to day j, and m represents the number of days from day i to day j. Finally, a prediction model for the rental volume of shared electric motorcycles is constructed.
[0033] Finally, based on the travel characteristics, the income of shared electric motorcycles in the city is estimated to complete the prediction of shared electric motorcycle demand and income. The expression of the shared electric motorcycle single-trip average income estimation model is as follows: in, represents the average one-way fare in region X, n represents the n time periods of the fare standard in region X, and f X is the charging standard of area X, L is the time period, then f X (L) is the charge in the X region when the riding time is in the L time period, p(L) is the probability of the riding time being in the L time period, and the specific steps for estimating the income of shared electric motorcycles in the city are as follows: analyze the distribution of riding time of different age groups based on the least squares model; establish a model for estimating the average income of a single trip of shared electric motorcycles; establish a model for estimating the income of shared electric motorcycles; and complete the forecast of the demand and income of shared electric motorcycles based on the model for estimating the average income of a single trip of shared electric motorcycles and the model for estimating the income of shared electric motorcycles. The model for estimating the income of shared electric motorcycles includes the total income of shared electric motorcycles in the region to be predicted from the i-th day to the j-th day and the average daily income of each vehicle in the time interval from the i-th day to the j-th day. The expression for the total income of shared electric motorcycles in the region to be predicted from the i-th day to the j-th day is as follows: Among them I X[i,j]That is, the total revenue of shared electric motorcycles in region X from day i to day j. The expression of the average daily revenue of each vehicle in the time interval from day i to day j in the tested region is as follows: in is the average daily income of each vehicle in region X from day i to day j, V X is the number of vehicles deployed in area X from day i to day j.
[0034] Finally, according to the known impact of different inherent attributes of cities and city influencing factors on riding demand, and combined with the shared electric motorcycle single-trip average income estimation model and the shared electric motorcycle income estimation model, the total rental income in the tested area for a period of time and the daily income of the average shared electric motorcycle are obtained, and the demand and income forecast is completed through the above-mentioned related parameter indicators of shared electric motorcycles, and the vehicle dispatch of shared bicycles in the tested area is assisted. In this embodiment, when the income of shared electric motorcycles in the city is estimated based on travel characteristics to complete the forecast of shared electric motorcycle demand and income, in addition to the total income of all shared electric motorcycles and the average daily income of a single vehicle in the city, if the city is divided into large areas, it will be divided into several areas according to the geography of the city, and the total income and average daily income of shared electric motorcycles in different areas will be counted for different areas. In this way, the operation and maintenance status of shared electric motorcycles in different areas of the city can be more carefully grasped, and the efficiency and accuracy of shared electric motorcycle operation and maintenance can be improved.
Claims
1. A prediction method for urban shared electric motorcycle operation and maintenance, characterized in that: include S1 analyzes the travel characteristics of shared electric motorcycles based on historical data; S2 predicts the rental volume of shared electric motorcycles based on the weighted regression of travel feature vectors; S3 calculates the operation and maintenance data of urban shared electric motorcycles based on travel characteristics to complete the prediction of urban shared electric motorcycle operation and maintenance.
2. A prediction method for urban shared electric motorcycle operation and maintenance according to claim 1, characterized in that: The specific steps of analyzing the travel characteristics of shared electric motorcycles in step S1 are as follows: S11 analyzes historical riding and revenue data of shared electric bikes; S12 analyzes the inherent attribute data of the city; S13 filters out row vector features.
3. A prediction method for urban shared electric motorcycle operation and maintenance according to any one of claims 1 or 2, characterized in that: The specific steps of predicting the rental volume of shared electric motorcycles in step S2 are as follows: S21 calculates the benchmark rental data of shared electric bikes; S22 analyzes the cycling probability of different age groups; S23 analyzes the relationship between urban influencing factors and the rental volume of shared electric motorcycles; S24 builds a shared electric motorcycle rental volume prediction model.
4. A prediction method for urban shared electric motorcycle operation and maintenance according to claim 3, characterized in that: The specific steps of calculating the operation and maintenance data of the city shared electric motorcycle in step S3 are as follows: S31 analyzes the distribution of cycling time in different age groups based on the least squares model; S32 Establish a model for estimating the average income of a single trip of a shared electric motorcycle; S33 Establish a model for estimating the revenue of shared electric motorcycles; S34 completes the forecast of shared electric motorcycle demand and revenue based on the shared electric motorcycle single-trip average revenue estimation model and the shared electric motorcycle revenue estimation model.
5. The prediction method for urban shared electric motorcycle operation and maintenance according to claim 3, characterized in that: The expression of the shared electric motorcycle rental volume prediction model is as follows: R XY =(P XY *G a1 +P Xa2 *G a2 )*W t(Y) *W r(Y) *W h(Y) Where X represents the predicted region, Y represents the predicted day, a1 and a2 represent the age segments of two different renters, and R represents the rental amount. Then R XY It represents the rental volume in area X for Y days, G represents the riding probability, and G a1 represents the riding probability of the renting population in age segment a1, P represents the population, P Xa1 represents the population of age group a1 in region X, W represents the weight, W t(Y) W represents the weight of the impact of the temperature on the rental volume in region X on day Y. r(Y) W represents the weight of the impact of rainfall on the rental volume in region X on day Y. h(Y) Indicates the weight of whether day Y in region X is a working day on the rental volume.
6. A prediction method for urban shared electric motorcycle operation and maintenance according to claim 5, characterized in that: The expression of the model for estimating the average income of a single trip of a shared electric motorcycle in step S3 is as follows: in, represents the average one-way fare in region X, n represents the n time periods of the fare standard in region X, and f X is the charging standard of area X, L is the time period, then f X (L) is the charge in area X when the riding time is in time period L, and p(L) is the probability of riding time in time period L.
7. A prediction method for urban shared electric motorcycle operation and maintenance according to claim 4, characterized in that: The prediction formula for predicting the rental volume of shared electric motorcycles within a period of time in step S24 is as follows: Where R X[i,j] is the forecast result of the rental volume in region X from day i to day j, and m represents the number of days from day i to day j.
8. A prediction method for urban shared electric motorcycle operation and maintenance according to claim 7, characterized in that: The shared motorcycle revenue estimation model in step S3 includes the total revenue of shared motorcycles in the predicted area from the i-th day to the j-th day and the daily average revenue of each vehicle in the time interval from the i-th day to the j-th day. The total revenue of shared motorcycles in the predicted area from the i-th day to the j-th day is expressed as follows: Among them I X[i,j] That is, the total revenue of shared electric motorcycles in region X from day i to day j.
9. A prediction method for urban shared electric motorcycle operation and maintenance according to claim 8, characterized in that: The expression of the average daily income per vehicle in the time interval from the i-th day to the j-th day in the tested area is as follows: in is the average daily income of each vehicle in region X from day i to day j, V X is the number of vehicles deployed in area X from day i to day j.
10. A prediction method for urban shared electric motorcycle operation and maintenance according to claim 9, characterized in that: In step S3, the total rental income for a period of time in the test area and the average daily income of shared electric motorcycles are obtained based on the known impact of different city inherent attributes and city influencing factors on riding demand, and combined with the shared electric motorcycle single-trip average income estimation model and the shared electric motorcycle income estimation model. The demand and income forecast is completed through the above-mentioned shared electric motorcycle related parameter indicators, and the shared bicycle vehicle dispatch in the test area is assisted.
Citation Information
Cited By
Shared bicycle unsatisfied demand prediction method and system based on lower bound deletion learning
CN122635867A
