Multi-scale carbon emission reduction estimation of shared bicycles and its key factor identification methods, equipment and media

Abstract

The present invention discloses a multi-scale shared bicycle carbon emission reduction estimation and key factor identification method, equipment and medium, including: dividing an area to be analyzed into a plurality of sections and grids; calculating a shared bicycle substitution rate based on a Bayesian formula; calculating the carbon emission reduction for each trip according to the travel distance of each transportation mode used for each trip and the shared bicycle substitution rate; calculating the shared bicycle carbon emission reduction corresponding to each section and each grid; inputting input factors into a first and a second carbon emission reduction prediction model respectively, using the shared bicycle carbon emission reduction corresponding to each grid as a training label, so as to predict the shared bicycle carbon emission reduction corresponding to peak and off-peak periods; calculating the SHAP value corresponding to each input factor, and using the first K1 and K2 input factors as key factors affecting the shared bicycle carbon emission reduction during peak and off-peak periods; performing dimensionality reduction and clustering on all SHAP values, and extracting the shared bicycle emission reduction mode.

Description

Technical Field

[0001] The present invention belongs to the technical field of transportation carbon emission reduction, and in particular relates to a multi-scale shared bicycle carbon emission reduction estimation and key factor identification method, equipment, and medium. Background Art

[0002] Although there has been a large amount of research on carbon reduction strategies for motor vehicles (such as vehicle electrification and eco-driving), there is still a significant gap in accurately estimating the carbon reduction potential of active transportation modes (such as shared bicycles). Shared bicycle services not only facilitate short-distance travel, but more importantly, they can be seamlessly integrated with public transportation systems. This integration significantly enhances the attractiveness of public transportation, reduces reliance on private cars, and thus reduces transportation-related carbon emissions. Therefore, it is particularly important to accurately estimate the carbon reduction of shared bicycles and reveal its key factors and reduction patterns.

[0003] Estimation of carbon emission reductions from shared bicycles is a basic task to identify key factors and emission reduction patterns. Estimation of carbon emission reductions from shared bicycles involves macro-level analysis (using macro data such as the size of shared bicycle fleets and station information) as well as micro-level analysis (such as using individual travel data). In addition, the life cycle assessment (LCA) framework expands the scope of carbon emission reduction estimates from shared bicycles to include indirect emissions from the production, maintenance, balancing, and scrapping stages of shared bicycles. However, the main focus of research on carbon emission reductions from shared bicycles is still on qualitative estimates at the macro level, and there are few studies on quantitative estimates at multiple scales.

[0004] The factors that affect carbon emissions reductions from bike-sharing are multifaceted, both at the macro and micro scales. At the macro level, these factors include bicycle lane infrastructure, the level of integration of bike-sharing systems with public transportation, population size, and urban functions. At the micro level, factors such as age, educational background, cost sensitivity, and environmental awareness play a key role in influencing carbon emissions reductions from bike-sharing. Despite this, few studies have considered bicycle networks, public transportation, urban functions, urban perception, and urban vitality factors simultaneously. In addition, existing studies have not identified key factors that have a significant impact on carbon emissions reductions from bike-sharing among the many factors. Summary of the invention

[0005] In view of the shortcomings of the existing technology, the present invention provides a multi-scale shared bicycle carbon emission reduction estimation and key factor identification method, equipment and medium.

[0006] In a first aspect, an embodiment of the present invention provides a multi-scale method for estimating carbon emission reduction of shared bicycles and identifying key factors thereof, the method comprising:

[0007] Obtain the area to be analyzed and divide it into several sections and grids;

[0008] The replacement rate of shared bicycles is calculated based on the Bayesian formula; the shared bicycle replacement rate includes the full-trip replacement rate of shared bicycles for different transportation modes, the partial replacement rate of shared bicycles and public transportation for car trips, and the new travel replacement rate after shared bicycles are introduced into the transportation system;

[0009] Obtain the travel trajectory of each mode of transportation to calculate the travel distance of each mode of transportation used for each trip; calculate the carbon emission reduction of each trip based on the travel distance of each mode of transportation used for each trip and the replacement rate of shared bicycles;

[0010] Allocate the carbon emission reduction of each trip to each road section according to the travel distance of the shared bicycle, and obtain the carbon emission reduction of the shared bicycle corresponding to each road section;

[0011] Calculate the carbon emission reduction of shared bicycles corresponding to each grid based on the carbon emission reduction of shared bicycles corresponding to each road section and the ratio of the length of the road section to the length of the road section in the grid;

[0012] Extract bicycle network factors, public transportation factors, urban function factors, urban perception factors and urban vitality factors; input them into the first carbon emission reduction prediction model and the second carbon emission reduction prediction model respectively, and use the carbon emission reduction of shared bicycles corresponding to each grid as the training label, so as to predict the carbon emission reduction of shared bicycles corresponding to peak hours and off-peak hours respectively;

[0013] By removing the input factors of the first carbon emission reduction prediction model in sequence, calculating the SHAP value corresponding to each input factor and sorting them, the first K1 input factors are used as the key factors affecting the carbon emission reduction of shared bicycles during peak hours; by removing the input factors of the second carbon emission reduction prediction model in sequence, calculating the SHAP value corresponding to each input factor and sorting them, the first K2 input factors are used as the key factors affecting the carbon emission reduction of shared bicycles during off-peak hours;

[0014] The SHAP values ​​corresponding to all input factors are reduced in dimension and clustered to extract the emission reduction pattern of shared bicycles.

[0015] In a second aspect, an embodiment of the present invention provides an electronic device, comprising a memory and a processor, wherein the memory is coupled to the processor; wherein the memory is used to store program data, and the processor is used to execute the program data to implement the above-mentioned multi-scale shared bicycle carbon emission reduction estimation and key factor identification method.

[0016] In a third aspect, an embodiment of the present invention provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the present invention achieves the following beneficial effects compared with the prior art:

[0017] The present invention provides a method for estimating the carbon emission reduction of shared bicycles at a multi-scale and identifying its key factors, which realizes the accurate estimation of the carbon emission reduction of shared bicycles at a multi-scale from individual, road section, grid to region, and comprehensively considers the nonlinear influencing factors of bicycle network, public transportation system, urban functional zoning, urban perception and urban vitality. By applying interpretable machine learning technology, the key factors affecting the carbon emission reduction potential of shared bicycles in cities are identified, and the SHAP values ​​corresponding to the influencing factors of shared bicycle carbon emission reduction are reduced in dimension and clustered to extract the emission reduction pattern of shared bicycles. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative labor.

[0019] Figure 1 is a schematic diagram of a multi-scale shared bicycle carbon emission reduction estimation method and a key factor identification method thereof provided by an embodiment of the present invention;

[0020] Figure 2 is a schematic diagram of the distribution results of carbon emission reduction of shared bicycles at the road section and grid level provided by an embodiment of the present invention;

[0021] Figure 3 is a schematic diagram of the distribution results of carbon emission reduction of shared bicycles at the regional level provided by an embodiment of the present invention;

[0022] Figure 4 is a result schematic diagram of the relationship between the carbon emission reduction and demand of shared bicycles provided by an embodiment of the present invention;

[0023] Figure 5 is a schematic diagram of the key factor identification results of carbon emission reduction of shared bicycles provided by an embodiment of the present invention;

[0024] Figure 6 It is a schematic diagram of the nonlinear dependency results of key factors provided by an embodiment of the present invention.

[0025] Figure 7 It is a schematic diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0026] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0027] It should be noted that, in the absence of conflict, the features in the following embodiments and implementations may be combined with each other.

[0028] like Figure 1 As shown, an embodiment of the present invention provides a multi-scale shared bicycle carbon emission reduction estimation method and its key factor identification method, the method specifically includes the following steps:

[0029] Step S1, obtaining the area to be analyzed and dividing it into several sections and grids.

[0030] Specifically, in this example, the area to be analyzed is divided into several road sections, grids, and shared bicycle riding areas; wherein the size of the grid is set to 1km×1km.

[0031] Step S2, based on the survey data, obtain the full-trip substitution rate of shared bicycles for different modes of transportation, the partial substitution rate of shared bicycles and public transportation for car trips, and the new travel substitution rate after the introduction of shared bicycles into the transportation system; obtain the travel trajectory of each mode of transportation, so as to calculate the travel distance of each mode of transportation used for each trip; calculate the carbon emission reduction of each trip based on the travel distance of each mode of transportation used for each trip and the three shared bicycle substitution rates; distribute the carbon emission reduction of each trip to each road section according to the travel distance of the shared bicycle, and obtain the shared bicycle carbon emission reduction corresponding to each road section; calculate the shared bicycle carbon emission reduction corresponding to each grid based on the shared bicycle carbon emission reduction corresponding to each road section and the ratio of the length of the road section to the length of the road section in the grid.

[0032] It should be noted that in this example, the survey data used is the survey data disclosed in the paper "Zhou Y, Yu Y, Wang Y, et al. Mode substitution and carbon emission impacts of electric bike sharing systems[J]. Sustainable Cities and Society, 2023, 89: 104312".

[0033] Specifically, step S2 includes the following sub-steps:

[0034] Step S201: Based on the survey data, taking into account the influence of departure time and travel distance, the Bayesian formula is used to calculate the full-trip substitution rate. , Partial trip replacement rate and new trip replacement rate , the specific formula is as follows:

[0035]

[0036]

[0037]

[0038]

[0039]

[0040]

[0041] Where, t represents the departure time, when t=1, it represents the peak period, when t=2, it represents the off-peak period; i represents the travel mode, when i=1, it represents walking, when i=2, it represents private bicycle, when i=3, it represents bus, when i=4, it represents subway, when i=5, it represents car; d represents the travel distance, when d=1, it represents the travel distance is less than 1 km, when d=2, it represents the travel distance is 1-2 km, when d=3, it represents the travel distance is 2-3 km, when d=4, it represents the travel distance is 3-4 km, when d=5, it represents the travel distance is greater than 4 km; Q represents the total number of valid questionnaires collected, q represents the valid questionnaires collected for the qth time; P(Dist d ∣M i T t ) is the probability that the user travels a distance d given the travel mode i and the departure time t, P(M i ∣T t ) is the probability that the user uses travel mode i under the condition of departure time t, P(T t ) is the probability that the user’s departure time is t, P(B|Dist d M i T t ) is the probability that the user is willing to use a shared bicycle to replace the original travel mode under the conditions of departure time t, travel distance d and travel mode i, P(I|Dist d M5T t ) is the probability that the user is willing to use a combination of shared bicycles and public transportation to partially replace the car trip under the conditions of departure time t and travel distance d, P(Dist d ∣T t) is the probability that the user travels a distance d under the condition that the departure time is t, P(New|Dist d T t ) is the probability of additional trips made by users under the conditions of travel time t and travel distance d after the introduction of shared bicycles into the transportation system.

[0042] The substitution rate of shared bikes for public transport and cars (part of the trip) is set to zero in the following two cases: (1) when the starting point or the end point of the trip is not within the public transport service area (the bus service area is defined as a radius of 400 meters and the subway service area is defined as a radius of 800 meters), or (2) when the trip occurs during the non-operating hours of public transport. Afterwards, the remaining shared bike substitution rates are normalized to ensure that the sum of the substitution rates is 1.

[0043] Step S202: Obtain the travel trajectory of each mode of transportation, so as to calculate the travel distance of each mode of transportation used in each trip. Calculate the carbon emission reduction of each trip based on the travel distance of each mode of transportation used in each trip and the shared bicycle replacement rate.

[0044] Specifically, due to the lack of original travel route information, in this example, the travel distance of each transportation mode is calculated using the shortest path algorithm.

[0045] Furthermore, based on the Life Cycle Assessment (LCA) model, the carbon emission reduction (BCER) of the k-th shared bicycle trip is k ) is calculated as follows:

[0046]

[0047] Among them, EF i Represents the carbon emission factor of the i-th travel mode. dist k,i is the travel distance of the kth trip using transportation mode i. carpartial Represents the carbon emission factor of automobiles, dist k,carpartial is the distance traveled by shared bicycles instead of cars in the kth trip, EF bike represents the carbon emission factor of shared bicycles based on the full life cycle model, dist k,bike represents the travel distance of the k-th trip using a shared bicycle.

[0048] Carbon emission factor EF of shared bicycles based on the full life cycle model bike The calculation process includes:

[0049] EF bike =EF bike,pro +EF bike,op +EF bike,dp

[0050] Among them, the expression of carbon emission factor in the production stage is as follows:

[0051] EF bike,pro =(CE bike,pro ) / (SL bike ×DUR bike ×ADT bike )

[0052] Among them, CE bike,pro is the total carbon emissions from the production phase of a single shared bicycle, SL bike is the service life of a shared bicycle, DUR bike is the average daily usage frequency of each shared bicycle, ADT bike is the average travel distance of shared bicycle trips.

[0053] The expression of carbon emission factor in the operation phase is as follows:

[0054] EF bike,op =EF bike,mt +EF bike,rb

[0055] Among them, EF bike,mt and EF bike,rb They are the carbon emission factors of shared bicycles during the maintenance and rebalancing stages respectively.

[0056] Among them, the expression of the carbon emission factor in the rebalancing stage is as follows:

[0057]

[0058] Among them, FS is the fleet size of shared bicycles in the study area, RBR is the average daily rebalancing ratio of shared bicycles, RBD is the average daily rebalancing distance of each shared bicycle, and DR is the average daily rebalancing distance of each shared bicycle. truck and DR e-trike are the proportion of dispatch trucks and electric tricycles, EF truck and EF e-trike are the carbon emission factors of dispatch trucks and electric tricycles, AL truck and AL e-trike They are the average cargo capacity of dispatched trucks and electric tricycles during the rebalancing process, ADU bike is the average daily frequency of shared bike usage.

[0059] The expression of carbon emission factor in the maintenance phase is as follows:

[0060] EF bike,mt =(CE bike,mt ) / (SL bike ×DUR bike ×ADTbike )

[0061] Among them, CE bike,mt It is the total carbon emissions of a single shared bicycle during the maintenance phase.

[0062] The carbon emissions in the disposal phase are calculated as follows:

[0063] EF bike,dp =(CE bike,dp ) / (SL bike ×DUR bike ×ADT bike )

[0064] Among them, CE bike,dp It is the total carbon emissions of a single shared bicycle during the disposal stage.

[0065] Step S203, allocating the carbon emission reduction of each trip to each road section according to the travel distance of the shared bicycle, and obtaining the carbon emission reduction of the shared bicycle corresponding to each road section.

[0066] Among them, the carbon emission reduction of shared bicycles corresponding to road section s is BCER s The expression is as follows:

[0067]

[0068] Among them, N sample is the sample size of shared bicycle trips in the area to be analyzed, N sample,city is the sample size of shared bicycle trips in the city where the area to be analyzed is located, N real,city is the total number of shared bicycle trips in the city where the area to be analyzed is located. This data can be obtained from the city’s monthly or annual traffic report. s,k Represents the distance of the k-th trip through road segment s.

[0069] Step S204, based on the carbon emission reduction of shared bicycles corresponding to each road section and the ratio of the length of the road section to the length of the road section in the grid, calculate the carbon emission reduction of shared bicycles corresponding to each grid.

[0070] The expression for calculating the carbon emission reduction of shared bicycles corresponding to grid g is as follows:

[0071]

[0072] Where BCER g represents the carbon emission reduction of shared bicycles in grid g, N s is the total number of road sections in the area to be analyzed, l s represents the length of the road segment s, l s,gRepresents the length of road section s in grid g. Considering that the carbon emission reduction and demand of shared bicycles follow the power-law distribution, a log-log heat map is used in this example to illustrate their coupling relationship.

[0073] The step S2 further comprises:

[0074] Step S205, calculating the carbon emission reduction of shared bicycles in the shared bicycle riding area, the expression is as follows:

[0075]

[0076] Among them, BCER r represents the carbon emission reduction of shared bicycles in region r, BCER s represents the carbon emission reduction of shared bicycles corresponding to road section s, l s represents the length of the road segment s, l s,r Represents the length of road segment s in region r.

[0077] Figure 2 A schematic diagram of the distribution results of carbon emission reductions of shared bicycles at the road section and grid levels is shown. The emission reductions during peak hours are significantly higher than those during off-peak hours. Figure 3 The results of the regional distribution of carbon emission reductions from shared bikes are shown. The carbon emission reductions in regions 1, 2, and 3 are significantly higher than those in other regions. In addition, the total carbon emission reductions from shared bikes on a single working day are as high as 377.2 tons of CO2. In the areas served by ground buses and subways, the carbon emission reductions from shared bikes account for 80.1% and 52.7% of the total emission reductions, respectively. Figure 4 The schematic diagram of the relationship between carbon emission reduction and demand of shared bicycles shows that the two show a linear trend on the power law scale, and their relationship is described by the equation log(y)=2.06121+1.0006×log(x). However, the R² value is 0.732, indicating that the model fails to fully capture the changes in carbon emission reduction of shared bicycles, because carbon emission reduction of shared bicycles is the result of the joint action of travel demand, OD distribution, mode substitution and path selection.

[0078] Step S3, extract bicycle network factors, public transportation factors, urban function factors, urban perception factors and urban vitality factors; input them into the first carbon emission reduction prediction model and the second carbon emission reduction prediction model respectively, and use the carbon emission reduction of shared bicycles corresponding to each grid as the training label, so as to predict the carbon emission reduction of shared bicycles corresponding to peak hours and off-peak hours respectively.

[0079] Furthermore, the process of extracting bicycle network factors, public transportation factors, urban function factors, urban perception factors and urban vitality factors includes:

[0080] The bicycle network factors include degree centrality, betweenness centrality and bicycle lane density.

[0081] Public transportation factors include public transportation passenger flow and public transportation accessibility; among them, the public transportation passenger flow includes ground bus passenger flow during peak hours, ground bus passenger flow during off-peak hours, subway passenger flow during peak hours, and subway passenger flow during off-peak hours; the public transportation accessibility includes: ground bus stop density, ground bus stop distance, ground bus line density, subway station density, subway station distance, and subway line density.

[0082] Specifically, in this example, the gravity model is used to distribute passenger flow, and the formula is as follows:

[0083]

[0084] Where m represents the public transportation mode (ground bus and subway), s is the public transportation station, and i is the POI in the bus service area. In this example, the service radius of the ground bus and subway is set to 400 meters and 800 meters respectively. m,s,i is the passenger flow of POI i assigned by site s, PF m,s is the passenger flow of station s, edist s,i is the Euclidean distance from site s to POI i, It is the sum of the reciprocals of the squares of the Euclidean distances from site s to all POIs in the service area.

[0085] Passenger flow in grid g (PF m,g ) is calculated as follows:

[0086]

[0087] Among them, N s,m Represents the total number of stops for public transportation mode m.

[0088] Density of public transportation stations within grid g PTSD m,g The definition is as follows:

[0089]

[0090] Among them, N s,m,g represents the number of stations of public transportation mode m in grid g, A g Represents the area of ​​the grid g.

[0091] The urban function factors include work POI density, residential POI density, service POI density, education POI density, medical POI density, and functional mixed entropy;

[0092] The city perception factors include: greenery perception, openness perception, walkability index, normalized vegetation index, safety perception, and beauty perception;

[0093] The city vitality factors include: population density, GDP density, social media activity intensity during peak hours, and social media activity intensity during off-peak hours.

[0094] The factor results are summarized in Table 1:

[0095] Table 1: Summary of factors

[0096]

[0097] Furthermore, the bicycle network factor, public transportation factor, urban function factor, urban perception factor and / or urban vitality factor are respectively input into the first carbon emission reduction prediction model and the second carbon emission reduction prediction model, and the carbon emission reduction of shared bicycles corresponding to each grid is used as a training label, so as to predict the carbon emission reduction of shared bicycles corresponding to peak hours and off-peak hours respectively. The process includes:

[0098] The first carbon emission reduction prediction model includes a first neural network and a first XGBoost model (denoted as NeXGBoost);

[0099] Input bicycle network factors, public transportation factors, urban function factors, urban perception factors and / or urban vitality factors into the first neural network for feature extraction to obtain first intermediate features; input the first intermediate features into the first XGBoost model, and use the carbon emission reduction of shared bicycles corresponding to each grid as a training label to predict the carbon emission reduction of shared bicycles corresponding to peak hours;

[0100] The second carbon emission reduction prediction model includes a second neural network and a second XGBoost model;

[0101] The bicycle network factors, public transportation factors, urban function factors, urban perception factors and / or urban vitality factors are input into the second neural network for feature extraction to obtain the second intermediate features. The second intermediate features are input into the second XGBoost model, and the carbon emission reduction of shared bicycles corresponding to each grid is used as the training label to predict the carbon emission reduction of shared bicycles corresponding to the off-peak period.

[0102] It should be noted that the first carbon emission reduction prediction model and the second carbon emission reduction prediction model may also adopt support vector machine (SVM), random forest (RF), lightweight gradient boosting machine (LightGBM), extreme gradient boosting machine (XGBoost) and neural network (NN) and the like.

[0103] Furthermore, in this example, the root mean square error (RMSE), mean absolute error (MAE) and coefficient of determination (R²) are used to evaluate the performance of the model, and the results are shown in Table 2:

[0104] Table 2: Model performance results

[0105]

[0106] Step S4: Remove the input factors of the first carbon emission reduction prediction model in sequence, calculate the SHAP value corresponding to each input factor, and sort them, and use the first K1 input factors as the key factors affecting the carbon emission reduction of shared bicycles during peak hours; remove the input factors of the second carbon emission reduction prediction model in sequence, calculate the SHAP value corresponding to each input factor, and sort them, and use the first K2 input factors as the key factors affecting the carbon emission reduction of shared bicycles during off-peak hours.

[0107] Furthermore, the expression for calculating the SHAP value corresponding to each input factor is as follows:

[0108]

[0109] Among them, SHAP(i) represents the SHAP value corresponding to input factor i, F is the set of all input factors, S is the subset of input factors except i, |S| is the number of input factors in set S, |F| is the total number of input factors, f(S) is the carbon emission reduction output of the model when only considering the input factor set S, and f(S∪{i}) is the carbon emission reduction output of the model when considering the input factor set S and input factor i.

[0110] In order to capture the relationship between these key factors and carbon emission reduction, a generalized additive model (GAM) is used to smooth the SHAP dependency graph. The key factors of carbon emission reduction of shared bicycles during off-peak and peak hours are shown in Figure 5 During peak hours, the key factors are work POI density, ground bus passenger flow, and bicycle lane density, which contribute 20.5%, 9.5%, and 7.2% to carbon emission reduction, respectively. During off-peak hours, ground bus passenger flow, work POI density, and social media activity intensity are key factors, which contribute 21.5%, 13.7%, and 9.2% to carbon emission reduction, respectively. Figure 6 GAM-SHAP dependency plot showing key factors, Figure 6 (A) shows the GAM-SHAP dependency graph of the working POI density. Figure 6 (B) shows the GAM-SHAP dependency diagram of ground bus passenger flow. Figure 6 (C) shows the GAM-SHAP plot of bike lane density. Figure 6(D) in Figure 1 shows the GAM-SHAP dependency graph of social media activity intensity. Specifically, when the work POI density exceeds 300 (1 / km²), it has a significant positive effect on the carbon emission reduction of shared bicycles during peak hours, and this effect exceeds that during non-peak hours. When the ground bus passenger flow exceeds 700 (pax / h), it also has a positive effect on the carbon emission reduction of shared bicycles during peak hours, but this effect is non-peak hours.

[0111] Step S5, perform dimension reduction and clustering on the SHAP values ​​corresponding to all input factors to extract the emission reduction pattern of shared bicycles.

[0112] Furthermore, the step S5 specifically includes:

[0113] Dimensionality reduction of SHAP values ​​corresponding to all input factors was performed using principal component analysis (PCA), t-distributed neighborhood embedding (t-SNE), autoencoder, or unified manifold approximation and projection method (UMAP);

[0114] K-means clustering, Gaussian mixture model (GMM), self-organizing map (SOM) or density clustering method (DBSCAN) are used to cluster all SHAP values ​​after dimension reduction to obtain the emission reduction model of shared bicycles;

[0115] The extracted emission reduction patterns of shared bicycles were statistically verified by two-tailed Wilcoxon rank sum test, Kruskal-Wallis test and multivariate analysis of variance. Specifically, the two-tailed Wilcoxon rank sum test was used to evaluate the significance of the emission reduction pattern at the local level, the Kruskal-Wallis test was used to evaluate the significance of the emission reduction pattern at the global univariate level, and the multivariate analysis of variance was used to evaluate the significance of the emission reduction pattern at the global multivariate level.

[0116] Among them, the Kruskal-Wallis test statistic H is as follows:

[0117]

[0118] The two-tailed Wilcoxon rank sum test statistic U is as follows:

[0119]

[0120] Where N is the total number of regional grids, n p is the number of carbon reduction modes, and is the sample size of modes i and j, and is the sum of the ranks of the total samples of patterns i and j.

[0121] During peak hours, the UMAP-DBSCAN algorithm was selected as the best dimensionality reduction and clustering method, and three unique carbon emission reduction modes were successfully identified, which were mainly distributed in the fifth ring area, the urban-rural transition area, and the suburbs. Mode I showed significant positive attribution in terms of work-related POI density and ground bus passenger flow, which made it significantly different from Mode II and Mode III. In contrast, Mode III showed significant negative attribution in terms of work-related POI density, ground bus passenger flow, and subway passenger flow. These findings are helpful for formulating more precise shared bicycle management policies, such as promoting the integration of shared bicycles with public transportation systems in urban areas to achieve synergistic emission reduction benefits. The results of the two-tailed Wilcoxon rank sum test, Kruskal-Wallis test, and MAOVA test further showed that there were significant differences in the carbon emission reduction and SHAP values ​​of key factors among different carbon emission reduction modes, which statistically verified the rationality of these mode divisions.

[0122] Accordingly, the present application also provides an electronic device, including: one or more processors; a memory for storing one or more programs; when the one or more programs are executed by the one or more processors, the one or more processors implement the above-mentioned multi-scale shared bicycle carbon emission reduction estimation and key factor identification method. Figure 7 As shown, it is a hardware structure diagram of any device with data processing capability where the multi-scale shared bicycle carbon emission reduction estimation and key factor identification method provided by the embodiment of the present invention is located, except Figure 7 In addition to the processor, memory and network interface shown, any device with data processing capability in which the apparatus in the embodiment is located may also include other hardware, generally based on the actual functions of the device with data processing capability, which will not be described in detail.

[0123] Accordingly, the present application also provides a computer-readable storage medium on which computer instructions are stored, and when the instructions are executed by the processor, the multi-scale shared bicycle carbon emission reduction estimation and key factor identification method as described above are implemented. The computer-readable storage medium can be an internal storage unit of any device with data processing capabilities described in any of the aforementioned embodiments, such as a hard disk or memory. The computer-readable storage medium can also be an external storage device, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), an SD card, a flash card (Flash Card), etc. equipped on the device. Furthermore, the computer-readable storage medium can also include both an internal storage unit and an external storage device of any device with data processing capabilities. The computer-readable storage medium is used to store the computer program and other programs and data required by any device with data processing capabilities, and can also be used to temporarily store data that has been output or is to be output.

[0124] The above embodiments are only used to illustrate the design ideas and features of the present invention, and their purpose is to enable those skilled in the art to understand the content of the present invention and implement it accordingly. The protection scope of the present invention is not limited to the above embodiments. Therefore, any equivalent changes or modifications made based on the principles and design ideas disclosed by the present invention are within the protection scope of the present invention.

Claims

1. A multi-scale method for estimating carbon emission reductions of shared bicycles and identifying key factors, characterized in that: The method comprises: Obtain the area to be analyzed and divide it into several sections and grids; The replacement rate of shared bicycles is calculated based on the Bayesian formula; the shared bicycle replacement rate includes the full-trip replacement rate of shared bicycles for different transportation modes, the partial replacement rate of shared bicycles and public transportation for car trips, and the new travel replacement rate after shared bicycles are introduced into the transportation system; Obtain the travel trajectory of each mode of transportation to calculate the travel distance of each mode of transportation used for each trip; calculate the carbon emission reduction of each trip based on the travel distance of each mode of transportation used for each trip and the replacement rate of shared bicycles; Allocate the carbon emission reduction of each trip to each road section according to the travel distance of the shared bicycle, and obtain the carbon emission reduction of the shared bicycle corresponding to each road section; Calculate the carbon emission reduction of shared bicycles corresponding to each grid based on the carbon emission reduction of shared bicycles corresponding to each road section and the ratio of the length of the road section to the length of the road section in the grid; Extract bicycle network factors, public transportation factors, urban function factors, urban perception factors and urban vitality factors; input them into the first carbon emission reduction prediction model and the second carbon emission reduction prediction model respectively, and use the carbon emission reduction of shared bicycles corresponding to each grid as the training label, so as to predict the carbon emission reduction of shared bicycles corresponding to peak hours and off-peak hours respectively; By removing the input factors of the first carbon emission reduction prediction model in sequence, calculating the SHAP value corresponding to each input factor and sorting them, the first K1 input factors are used as the key factors affecting the carbon emission reduction of shared bicycles during peak hours; by removing the input factors of the second carbon emission reduction prediction model in sequence, calculating the SHAP value corresponding to each input factor and sorting them, the first K2 input factors are used as the key factors affecting the carbon emission reduction of shared bicycles during off-peak hours; Perform dimension reduction and clustering on the SHAP values ​​corresponding to all input factors to extract the emission reduction pattern of shared bicycles; Among them, the carbon emission reduction of shared bicycles for the kth trip is calculated based on the full life cycle model, and the expression is as follows: ； Among them, EF i represents the carbon emission factor of the i-th travel mode; dist k,i is the travel distance of the kth trip using transportation mode i; EF carpartial Represents the carbon emission factor of automobiles, dist k,carpartial is the distance traveled by shared bicycles instead of cars in the kth trip, EF bike represents the carbon emission factor of shared bicycles based on the full life cycle model, dist k,bike represents the travel distance of the kth trip using a shared bicycle; The calculation process of the carbon emission factor of shared bicycles includes: IF bike =IF bike,pro +EF bike,op +EF bike,dp ; Among them, the expression of carbon emission factor in the production stage is as follows: EF bike,pro =(CE bike,pro ) / (SL bike ×DURABLE bike ×ADT bike ); Among them, CE bike,pro is the total carbon emissions from the production phase of a single shared bicycle, SL bike is the service life of a shared bicycle, DUR bike is the average daily usage frequency of each shared bicycle, ADT bike is the average trip distance of shared bike trips; The expression of carbon emission factor in the operation phase is as follows: IF bike,op =IF bike,mt +EF bike,rb ; Among them, EF bike,mt and EF bike,rb They are the carbon emission factors of shared bikes during the maintenance and rebalancing stages; Among them, the expression of the carbon emission factor in the rebalancing stage is as follows: ； Among them, FS is the fleet size of shared bicycles in the study area, RBR is the average daily rebalancing ratio of shared bicycles, RBD is the average daily rebalancing distance of each shared bicycle, and DR is the average daily rebalancing distance of each shared bicycle. truck and DR e-trike are the proportion of dispatch trucks and electric tricycles, EF truck and EF e-trike are the carbon emission factors of dispatch trucks and electric tricycles, AL truck and AL e-trike They are the average cargo capacity of dispatched trucks and electric tricycles during the rebalancing process, ADU bike is the average daily frequency of shared bike usage; The expression of carbon emission factor in the maintenance phase is as follows: EF bike,mt =(CE bike,mt ) / (SL bike ×DURABLE bike ×ADT bike ); Among them, CE bike,mt is the total carbon emissions of a single shared bike during the maintenance phase; The carbon emissions in the disposal phase are calculated as follows: EF bike,dp =(CE bike,dp ) / (SL bike ×DURABLE bike ×ADT bike ); Among them, CE bike,dp It is the total carbon emissions of a single shared bicycle during the disposal stage.

2. According to claim 1, a multi-scale shared bicycle carbon emission reduction estimation and key factor identification method is characterized in that: The process of calculating the replacement rate of shared bicycles based on the Bayesian formula includes: Calculate the full-trip replacement rate of shared bikes for different modes of transportation , the expression is as follows: ； Calculate the partial substitution rate of car trips by combining bike-sharing and public transportation , the expression is as follows: ； Calculate the new travel substitution rate after shared bicycles are introduced into the transportation system , the expression is as follows: ； in, ； ； ； Where, t represents the departure time, when t=1, it represents the peak period, when t=2, it represents the off-peak period; i represents the travel mode, when i=1, it represents walking, when i=2, it represents private bicycle, when i=3, it represents public transportation, when i=4, it represents subway, when i=5, it represents car; d represents the travel distance, when d=1, it represents the travel distance is less than 1 km, when d=2, it represents the travel distance is 1-2 km, when d=3, it represents the travel distance is 2-3 km, when d=4, it represents the travel distance is 3-4 km, when d=5, it represents the travel distance is greater than 4 km; Q represents the total number of valid questionnaires collected, q represents the valid questionnaires collected for the qth time; P(Dist d ∣M i T t ) is the probability that the user travels a distance d given the travel mode i and the departure time t, P(M i ∣T t ) is the probability that the user uses travel mode i under the condition of departure time t, P(T t ) is the probability that the user’s departure time is t, P(B|Dist d M i T t ) is the probability that the user is willing to use a shared bicycle to replace the original travel mode under the conditions of departure time t, travel distance d and travel mode i, P(I|Dist d M5T t ) is the probability that the user is willing to use a combination of shared bicycles and public transportation to partially replace the car trip under the conditions of departure time t and travel distance d, P(Dist d ∣T t ) is the probability that the user travels a distance d under the condition that the departure time is t, P(New|Dist d T t ) is the probability of additional trips made by users under the conditions of travel time t and travel distance d after the introduction of shared bicycles into the transportation system.

3. The multi-scale shared bicycle carbon emission reduction estimation and key factor identification method according to claim 1 is characterized in that: The carbon emission reduction of each trip is distributed to each road section according to the travel distance of the shared bicycle, and the carbon emission reduction of the shared bicycle corresponding to each road section is obtained; among which, the carbon emission reduction BCER of the shared bicycle corresponding to road section s is s The expression is as follows: ； Among them, N sample is the sample size of shared bicycle trips in the area to be analyzed, N sample,city is the sample size of shared bicycle trips in the city where the area to be analyzed is located, N real,city is the total number of shared bicycle trips in the city where the area to be analyzed is located; l s,k represents the distance of the k-th trip through road segment s; Based on the carbon emission reduction of shared bicycles corresponding to each road section and the ratio of the length of the road section to the length of the road section in the grid, the carbon emission reduction of shared bicycles corresponding to each grid is calculated; the expression for calculating the carbon emission reduction of shared bicycles corresponding to grid g is as follows: ； Where BCER g represents the carbon emission reduction of shared bicycles in grid g, N s is the total number of road sections in the area to be analyzed, l s represents the length of the road segment s, l s,g Represents the length of the road segment s within the grid g.

4. The multi-scale shared bicycle carbon emission reduction estimation and key factor identification method according to claim 1 is characterized in that: The method further comprises: Divide the area to be analyzed into several shared bicycle riding areas; Calculate the carbon emission reduction of shared bicycles in the shared bicycle riding area. The expression is as follows: ； Among them, BCER r represents the carbon emission reduction of shared bicycles in region r, BCER s represents the carbon emission reduction of shared bicycles corresponding to road section s, l s represents the length of the road segment s, l s,r Represents the length of road segment s in region r.

5. The multi-scale shared bicycle carbon emission reduction estimation and key factor identification method according to claim 1 is characterized in that: The bicycle network factors include degree centrality, betweenness centrality, and bicycle lane density; Public transportation factors include public transportation passenger flow and public transportation accessibility; the public transportation passenger flow includes ground bus passenger flow during peak hours, ground bus passenger flow during off-peak hours, subway passenger flow during peak hours, and subway passenger flow during off-peak hours; the public transportation accessibility includes: ground bus station density, ground bus station distance, ground bus line density, subway station density, subway station distance, and subway line density; The urban function factors include work POI density, residential POI density, service POI density, education POI density, medical POI density, and functional mixed entropy; The city perception factors include: greenery perception, openness perception, walkability index, normalized vegetation index, safety perception, and beauty perception; The city vitality factors include: population density, GDP density, social media activity intensity during peak hours, and social media activity intensity during off-peak hours.

6. The multi-scale shared bicycle carbon emission reduction estimation and key factor identification method according to claim 1 is characterized in that: Inputting bicycle network factors, public transportation factors, urban function factors, urban perception factors and / or urban vitality factors into the first carbon emission reduction prediction model and the second carbon emission reduction prediction model respectively, taking the carbon emission reduction of shared bicycles corresponding to each grid as the training label, and thus predicting the carbon emission reduction of shared bicycles corresponding to peak hours and off-peak hours respectively, the process includes: The first carbon emission reduction prediction model includes a first neural network and a first XGBoost model; Input bicycle network factors, public transportation factors, urban function factors, urban perception factors and / or urban vitality factors into the first neural network for feature extraction to obtain first intermediate features; input the first intermediate features into the first XGBoost model, and use the carbon emission reduction of shared bicycles corresponding to each grid as a training label to predict the carbon emission reduction of shared bicycles corresponding to peak hours; The second carbon emission reduction prediction model includes a second neural network and a second XGBoost model; The bicycle network factors, public transportation factors, urban function factors, urban perception factors and / or urban vitality factors are input into the second neural network for feature extraction to obtain the second intermediate features. The second intermediate features are input into the second XGBoost model, and the carbon emission reduction of shared bicycles corresponding to each grid is used as the training label to predict the carbon emission reduction of shared bicycles corresponding to the off-peak period.

7. The multi-scale shared bicycle carbon emission reduction estimation and key factor identification method according to claim 1 is characterized in that: The process of reducing the dimension and clustering the SHAP values ​​corresponding to all input factors and extracting the emission reduction pattern of shared bicycles includes: The SHAP values ​​corresponding to all input factors are reduced in dimension using principal component analysis, t-distributed neighborhood embedding, autoencoder, or unified manifold approximation and projection methods; K-means clustering, Gaussian mixture model, self-organizing map or density clustering method are used to cluster all SHAP values ​​after dimension reduction to obtain the emission reduction model of shared bicycles; The extracted emission reduction pattern of shared bicycles was verified by Kruskal-Wallis test, two-tailed Wilcoxon rank sum test and multivariate analysis of variance test.

8. An electronic device, comprising a memory and a processor, characterized in that: The memory is coupled to the processor; wherein the memory is used to store program data, and the processor is used to execute the program data to implement the multi-scale shared bicycle carbon emission reduction estimation and key factor identification method as described in any one of claims 1-7 above.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by the processor, the multi-scale shared bicycle carbon emission reduction estimation and key factor identification method as described in any one of claims 1-7 are implemented.

Images