A personalized MaaS solution design and recommendation method and system

Abstract

The present application discloses a personalized MaaS solution design and recommendation method, including: obtaining historical travel information of residents in a target city, the historical travel information including travel attributes and attribute levels; based on the historical travel information, using a clustering algorithm to extract travel features to obtain the attribute levels of each travel attribute of different groups of residents in the city, wherein the group types of urban residents include positive, neutral and negative types; based on the attribute levels of each travel attribute of any group of urban residents, constructing a travel solution selection scenario for the corresponding group; obtaining the historical travel information of a target user, judging the group type to which the target user belongs based on the historical travel information of the target user, and recommending a corresponding travel solution selection scenario to the target user based on the group type to which the target user belongs. The present application can classify users into different feature types according to their travel habits, so as to provide users with accurate and personalized MaaS travel solutions.

Description

Technical Field

[0001] The present application relates to the field of transportation technology, and in particular to a personalized MaaS solution design and recommendation method and system. Background Art

[0002] With the development of information and communication technology, the emergence of MaaS (Mobility as a Service) provides a new strategy for travel management. By integrating transportation resources through digital platforms, integrated and integrated transportation can be realized, thus providing travelers with "door-to-door" travel services. The ideal MaaS platform breaks through the barriers of various transportation modes such as subways, buses, taxis, shared motorcycles, and shared bicycles. Through a mobile phone application, all transportation modes can be queried, ordered, and paid, and it has functions such as itinerary planning and travel reservations, providing users with travel services.

[0003] However, the existing technology does not fully consider factors such as the user's personal travel habits, and the travel solutions provided to users are too general and cannot provide users with personalized travel strategies. Summary of the invention

[0004] In order to solve the deficiencies of the prior art, this application adopts the following technical solutions:

[0005] In a first aspect, the present application provides a personalized MaaS solution design and recommendation method, the method comprising the following steps:

[0006] Step S11, obtaining historical travel information of residents of the target city, wherein the historical travel information includes travel attributes and attribute levels;

[0007] Step S12, based on the historical travel information, a clustering algorithm is used to extract travel characteristics to obtain the attribute levels of various travel attributes of different groups of urban residents, wherein the group types of urban residents include positive, neutral and negative types;

[0008] Step S13, based on the attribute levels of each travel attribute of any group of urban residents, constructing a travel plan selection scenario corresponding to the group;

[0009] Step S14, obtaining historical travel information of the target user, determining the type of group to which the target user belongs based on the historical travel information of the target user, and recommending a corresponding travel plan selection scenario to the target user based on the type of group to which the target user belongs.

[0010] In summary, the personalized MaaS solution design and recommendation method of the present application is based on the historical travel information of resident users, and divides resident users into three types of groups according to their travel characteristics: active, neutral, and passive. Based on the travel attribute levels of the three types of user groups, travel solution selection scenarios for the corresponding user groups are constructed. According to the historical travel information of the target users, the travel habits of the target users are analyzed, and the type of group to which the target users belong is determined, so as to recommend travel solution selection scenarios corresponding to the type of group to which the target users belong, thereby providing users with accurate and personalized travel solutions.

[0011] Further, an attribute level point set of any travel attribute in any group is determined, and an attribute level description of the travel attribute of the corresponding group is determined based on the attribute level point set, wherein the attribute level description includes three attribute level point values ​​selected from the attribute level point set, and the attribute level descriptions of the same travel attribute of different groups are set differently;

[0012] Constructing a travel plan selection scenario corresponding to the group based on a combination of attribute level point values ​​of each travel attribute;

[0013] Among them, the attribute level description of determining any travel attribute in any group includes:

[0014] Based on the attribute level of any travel attribute of the group, determining the distribution form of the attribute level,

[0015] In response to the attribute level being approximately normally distributed, selecting a quartile method to determine the attribute level point value to obtain the attribute level point set;

[0016] In response to the attribute level distribution being skewed, a quintile or smaller quantile is selected to determine the attribute level point value to obtain the attribute level point set.

[0017] Furthermore, before determining the distribution form of the attribute level, the attribute level is first standardized, and the distribution form is determined based on the attribute level after the standardization process.

[0018] The quantiles are determined based on the distribution form, and the standardized values ​​corresponding to each quantile are de-standardized and rounded to obtain the attribute level point values.

[0019] Furthermore, the travel attributes include public transportation and subway; the travel plan selection scenarios constructed for the active group include at least three active travel plans, and in any of the active travel plans, at least one of the public transportation and subway is not limited in the number of times it can be used.

[0020] Furthermore, the travel attributes also include shared bicycles and shared electric motorcycles; travel attribute descriptions of the shared bicycles and the shared electric motorcycles are provided, an attribute level point value is selected from the travel attribute descriptions of the shared bicycles and the shared electric motorcycles respectively for combination, and at least one of the bus and the subway is set to have unlimited usage times to construct the active travel plan.

[0021] Furthermore, the travel attributes include car reservation; the travel plan selection scenario constructed for the passive group includes at least three passive travel plans, and in each of the passive travel plans, only available travel distances for car reservation at different levels are provided.

[0022] Furthermore, a travel attribute description of the reserved car is provided, and an attribute level point value is selected from the travel attribute description of the reserved car to construct the passive travel plan.

[0023] Furthermore, the travel attributes include public transportation, subway, shared bicycles, shared electric motorcycles and reserved cars; the travel plan selection scenario constructed for the neutral group includes at least three neutral travel plans, providing travel attribute descriptions for each of the travel attributes, and selecting an attribute level point value combination from the travel attribute descriptions of each of the travel attributes to construct the neutral travel plan.

[0024] Furthermore, the price of any travel plan is expressed by the following formula:

[0025] M=(A*2+B*4+C*1.5+D*2.5+E*2.5)*F;

[0026] In the formula, A represents the attribute level point value of public transportation. If the public transportation is unlimited in the travel plan, then A=7.5; B represents the attribute level point value of subway. If the subway is unlimited in the travel plan, then B=6.25; C represents the attribute level point value of shared bicycles; D represents the attribute level point value of shared electric motorcycles; E represents the attribute level point value of reserved car use; and F represents the discount coefficient.

[0027] In a second aspect, the present application also provides a personalized MaaS solution design and recommendation system, which adopts the above-mentioned personalized MaaS solution design and recommendation method. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 A flowchart of the steps of a personalized MaaS solution design and recommendation method provided in one embodiment of the present application;

[0029] Figure 2 A specific flow chart of step S12 in the personalized MaaS solution design and recommendation method provided in one embodiment of the present application;

[0030] Figure 3 A specific flow chart of step S13 in the personalized MaaS solution design and recommendation method provided in one embodiment of the present application;

[0031] Figure 4 A flowchart of the steps of the RSC algorithm in the personalized MaaS solution design and recommendation method provided in one embodiment of the present application;

[0032] Figure 5 A schematic diagram of a neutral travel solution generated by a personalized MaaS solution design and recommendation method provided by an embodiment of the present application;

[0033] Figure 6 A schematic diagram of an active travel plan generated by a personalized MaaS plan design and recommendation method provided by an embodiment of the present application;

[0034] Figure 7 A schematic diagram of a passive travel plan generated by a personalized MaaS plan design and recommendation method provided in an embodiment of the present application. DETAILED DESCRIPTION

[0035] The present application will be described in detail below in conjunction with the specific implementation modes shown in the accompanying drawings, but these implementation modes do not limit the present application. Structural, methodological, or functional changes made by ordinary technicians in the field based on these implementation modes are included in the protection scope of the present application.

[0036] In order to solve the deficiencies of the prior art, on the one hand, the present application provides a personalized MaaS solution design and recommendation method, such as Figure 1 As shown, the method comprises the following steps:

[0037] Step S11, obtaining historical travel information of residents in the target city, where the historical travel information includes travel attributes and attribute levels;

[0038] Step S12, based on the historical travel information, a clustering algorithm is used to extract travel characteristics to obtain the attribute levels of various travel attributes of different groups of urban residents, wherein the group types of urban residents include positive, neutral and negative types;

[0039] Step S13, based on the attribute levels of each travel attribute of any group of urban residents, construct a travel plan selection scenario for the corresponding group;

[0040] Step S14, obtaining historical travel information of the target user, determining the type of group to which the target user belongs based on the historical travel information of the target user, and recommending a corresponding travel plan selection scenario to the target user based on the type of group to which the target user belongs.

[0041] The personalized MaaS solution design and recommendation method of this application is based on the historical travel information of resident users, and divides resident users into three groups according to their travel characteristics: active, neutral, and passive. Based on the travel attribute levels of the three user groups, travel solution selection scenarios for the corresponding user groups are constructed. According to the historical travel information of the target user, the travel habits of the target user are analyzed, and the type of group to which the target user belongs is determined, so as to recommend the travel solution selection scenario corresponding to the type of group to which the target user belongs, so as to provide the user with accurate and personalized travel solutions.

[0042] The personalized MaaS solution design and recommendation method of this embodiment is described in detail below.

[0043] In step S11, the historical travel information of urban residents includes: travel attributes and attribute levels. Among them, travel attributes include: public transportation, subway, shared bicycles, shared motorcycles, and reserved cars, a total of five categories, and the attribute level refers to the number of trips and / or travel distances corresponding to the above five types of travel attributes of the target user. In the embodiment of the present application, the four types of attributes of public transportation, subway, shared bicycles, and shared motorcycles in the travel attributes are recorded in the form of the number of trips, and the reserved car is recorded in the form of the travel distance.

[0044] In step S12, based on the historical travel information of urban residents, a clustering algorithm is used to extract the travel characteristics of resident users from the historical travel data, thereby obtaining the attribute levels of various travel attributes of different groups of urban residents. Specifically, Figure 2 As shown, the following steps are included:

[0045] Step S121, pre-process the historical travel information of urban residents to generate a data set containing urban residents' travel attribute data. The attribute level of residents' travel is used as the coordinate of the sample point, and the attribute level includes the number of bus trips, subway trips, shared bicycle trips, shared electric motorcycle trips, and reserved vehicle travel distance within a preset time range.

[0046] Step S122, setting clustering algorithm parameters: the neighborhood radius a between data points and the minimum number of samples MinPts required within the neighborhood radius a, to identify areas where the data density reaches and / or exceeds a preset density.

[0047] Step S123, set clustering rules: when the number of neighborhood sample points of a data point within the neighborhood radius a is greater than or equal to the minimum sample number MinPts, the data point is marked as a core point and a new cluster is expanded; when the number of neighborhood samples of a data point is less than the minimum sample number MinPts, the data point is marked as a noise point and no cluster expansion is performed.

[0048] Step S124: Based on the clustering rules, for each core point, all sample points in its neighborhood are traversed. If the points in these neighborhoods are not marked as any cluster, they are added to the current cluster. For each data point newly added to the cluster, the data points in its neighborhood are continuously checked, and the qualified data points are expanded to the current cluster until the neighborhoods of all data points in the current cluster have been traversed.

[0049] Step S125, repeatedly select the core points that have not been visited, check whether they meet the data density conditions, and repeat step S123 for the data points whose data density reaches and / or exceeds the preset density to form a new cluster. If all core points have been visited and controlled, it means that all areas with data density reaching and / or exceeding the preset density have formed clusters, and the remaining points are marked as noise points.

[0050] Step S126, outputting the clustering results, the clustering results include clusters and points marked as noise, the clusters in the clustering results represent the main travel modes of urban residents in the travel survey data.

[0051] Based on the output results of the clustering algorithm, the group types of urban residents are divided into three categories according to the attribute levels of each travel attribute of urban residents, including: positive type, neutral type, and negative type. As shown in Table 1, the positive type indicates that the resident group portrait is a high-level public transportation user; the negative type indicates that the resident group portrait is a low-level public transportation user and a high-level reservation car user; the neutral type indicates that the resident group portrait is a multi-modal transportation user, that is, a medium-level user of various transportation modes.

[0052]

[0053] Table 1

[0054] As an implementation method, in step S13, the attribute level point set of any travel attribute in any group is determined, and the attribute level description of the travel attribute of the corresponding group is determined based on the attribute level point set, wherein the attribute level description includes three attribute level point values ​​selected from the attribute level point set, and the attribute level descriptions of the same travel attribute of different groups are set differently. And based on the combination of the attribute level point values ​​of each travel attribute, the travel plan selection scenario corresponding to the group is constructed.

[0055] Among them, determining the attribute level point set of any travel attribute in any group includes: determining the distribution form of the attribute level based on the attribute level of any travel attribute of the group; in response to the attribute level being normally distributed, selecting the quartile method to determine the attribute level point value to obtain the attribute level point set; in response to the attribute level distribution being skewed, selecting the quintile or smaller quantile to determine the attribute level point value to obtain the attribute level point set.

[0056] Furthermore, as an implementation method, before determining the distribution form of the attribute level, the attribute level is first standardized, and the distribution form is judged based on the standardized attribute level; the quantile is determined based on the distribution form, and the standardized value corresponding to each quantile is de-standardized and rounded to obtain the attribute level point value.

[0057] Specifically, before determining the distribution form of the attribute level, the Z-score standardization method is used to standardize the attribute level so that the data is aggregated on a relatively uniform scale and the differences in the number of trips of different users are eliminated. The specific Z-score standardization method includes: determining the original value of the number of public transportation trips of resident users in a given period as X, and calculating the average value of the overall sample as µ , the standard deviation is σ, and the standardized value is Z=(X- µ ) / σ.

[0058] Based on the attribute level after standardization, the distribution form is determined. If the distribution form of the attribute level is approximately normal, the quartile method is selected to determine the attribute level point value to obtain the attribute level point set; if the distribution form of the attribute level is skewed, the quintile or smaller quantile is selected to determine the attribute level point value to obtain the attribute level point set.

[0059] Furthermore, after determining the quantiles based on the distribution of the attribute levels, the standardized values ​​corresponding to each quantile are denormalized and converted into the actual number of trips. The specific method can be expressed as: original number of trips = standardized quantile value * σ + µ ; and round the de-standardized results to adjust each representative level to an integer to obtain the point value of each attribute level.

[0060] Based on the above description, in specific application scenarios, such as Figure 3 As shown, step S13 may include the following steps:

[0061] Step S131, obtain the main travel modes of urban residents and their corresponding overall horizontal distribution, the overall horizontal distribution includes: the distribution of bus trips, subway trips, shared bicycle trips, shared electric motorcycle trips, and reserved car travel distance distribution within a given time range.

[0062] Step S132, based on the pre-processing of the travel data, the Z-score standardization method (mean 0, standard deviation 1) is used to convert each travel frequency value into a standardized value, so that the data is clustered on a relatively uniform scale and the differences in travel frequencies of different users are eliminated.

[0063] Step S133, based on the standardization process, the skewness and kurtosis of the data are calculated to determine the distribution form of the data, and the appropriate quantile is selected according to the distribution form. If the data is approximately normally distributed, the quartiles (25%, 50%, 75%, 100%) are selected; if the data is skewed, the quintile or smaller quantiles (20%, 40%, 60%, 80%) are selected to improve representativeness.

[0064] Step S134, finally determine the attribute level point set, including four quantiles: P1 (such as the 25th or 20th percentile), P2 (median or 40th percentile), P3 (75th or 60th percentile), P4 (such as the 100th or 80th percentile) as representative quantiles to ensure that the division levels are uniform and fit the data distribution.

[0065] Step S135 , restore the quantiles to the integer values ​​of the original number of trips by denormalization, ensuring that the representative level is an integer that is easy to interpret.

[0066] Step S136, outputting four representative integers as key level point values ​​corresponding to each attribute, so as to facilitate direct application in travel plans.

[0067] After determining the travel attributes of the resident groups and the point sets of various attribute levels, considering the differences between resident groups, and the different attribute levels of different resident users in the same group, the level range of each travel attribute in different schemes (i.e., the attribute level description of differentiated settings) is determined based on the number / distance distribution of each travel mode used by residents in the target area within a given time range, as a basis for providing personalized scheme design for resident groups.

[0068] As a way of implementation, the price of any travel option is expressed by the following formula:

[0069] M=(A*2+B*4+C*1.5+D*2.5+E*2.5)*F;

[0070] In the formula, A represents the attribute level point value of public transportation. If the public transportation is unlimited in the travel plan, then A=7.5; B represents the attribute level point value of subway. If the subway is unlimited in the travel plan, then B=6.25; C represents the attribute level point value of shared bicycles; D represents the attribute level point value of shared electric motorcycles; E represents the attribute level point value of reserved car use; and F represents the discount coefficient.

[0071] It should be noted that in addition to the travel attributes and attribute levels of the resident group, the discount coefficient is an important factor affecting the user's choice of travel plan. The discount coefficient affects the travel price of the recommended plan, thereby affecting the user's choice of travel plan. In this application, the discount coefficient is set to three levels of [0.8, 1.0, 1.2].

[0072] The following describes the design and recommendation methods for the personalized MaaS solutions provided in this application, and how to formulate personalized solutions for three resident groups.

[0073] As an implementation method, for the neutral group among the residents, the travel attributes of the neutral group include public transportation, subway, shared bicycles, shared motorcycles and reserved cars (reserved cars include online car-hailing and taxis). The travel plan selection scenario constructed for the neutral group includes at least three neutral travel plans, providing travel attribute descriptions of each travel attribute, and selecting an attribute level point value combination from the attribute level description of each travel attribute to construct a neutral travel plan.

[0074] Specifically, according to the above description, the neutral scheme is aimed at the medium-level multimodal transportation user group. As shown in Table 2, in the neutral scheme, the 25% quantile trip times, average trip times, and 75% quantile trip times of buses, subways, shared bicycles, and shared electric motorcycles are counted respectively. For the neutral scheme, the lower level attribute level point values ​​are selected from the attribute level point set to generate the attribute level description. Specifically, the usage range of public buses is set to three levels: [P1-B, P2-B, P3-B]; the usage range of subways is set to three levels: [P1-M, P2-M, P3-M]; the usage range of shared bicycles is set to three levels: [P1-K, P2-K, P3-K]; the usage range of shared electric motorcycles is set to three levels: [P1-EK, P2-EK, P3-EK]; for reserved cars, the 25th percentile travel distance, average travel distance, and 75th percentile travel distance of reserved cars are counted, and the usage distance range of reserved cars is set to three levels: [P1-E, P2-E, P3-E].

[0075]

[0076] Table 2

[0077] In the personalized MaaS solution design and recommendation method provided in this application, the D-efficient design method is adopted to define the design matrix, that is, the combination of all attribute levels. As shown in Table 3, by selecting the attribute level combination, each row represents a different MaaS neutral solution. In this design matrix, each MaaS neutral solution has corresponding level selections for attributes such as buses, subways, shared bicycles, shared electric motorcycles, reserved cars, and discount coefficients.

[0078]

[0079] Table 3

[0080] Furthermore, based on the design matrix obtained by using the D-efficient design method, the personalized MaaS solution design and recommendation method provided in this application also adopts a column-based RSC algorithm to ensure the effectiveness of the recommended solution.

[0081] Specifically, Figure 4 As shown, the specific steps are as follows:

[0082] Step S21 : creating different columns for each travel attribute, and generating all possible combinations based on all travel attributes and their attribute levels.

[0083] Step S22, generating a design matrix by combining attribute columns, and using a column-based RSC algorithm to select three optimal combinations from the above possible combinations.

[0084] Step S23, calculate the efficiency error, evaluate the D-error value of each possible combination, and select the combination with the smallest D-error value as the final design solution.

[0085] The formula for calculating the D-error value is as follows:

[0086] ;

[0087] in, is the prior parameter The joint probability density function of is calculated using the Halton sequence sampling method. Perform simulation calculations.

[0088] According to the above description, three personalized travel plans are generated for the MaaS neutral user group through the D-efficient design method. The column-based RSC algorithm is used to screen effective designs and calculate the D-error value to ensure the effectiveness and efficiency of the plan, and to provide accurate and personalized MaaS recommendations for different users.

[0089] For example, in combination with the travel plan price formula, the personalized MaaS plan design and recommendation method provided by this application provides a MaaS neutral plan for a neutral user group as follows: Figure 5 As shown, three different neutral travel plans are provided for users. Based on the differences among the three travel plans, the prices of the three travel plans are different, providing users with a variety of personalized travel plans.

[0090] As an implementation method, for the active group among the residents, the travel attributes of the active group include buses and subways; the travel plan selection scenario constructed for the active group includes at least three active travel plans, and, in any active travel plan, at least one of the bus and the subway has no limit on the number of times it can be used.

[0091] Furthermore, as an implementation method, the travel attributes of the active group also include shared bicycles and shared electric motorcycles; a travel attribute description of shared bicycles and shared electric motorcycles for the active group is provided, an attribute level point value is selected from the travel attribute descriptions of shared bicycles and shared electric motorcycles for combination, and at least one of the bus and subway is set to have unlimited usage times to construct an active travel plan.

[0092] Specifically, according to the above description, the active scheme is aimed at the high-level public transportation user group. As shown in Table 4, in the active scheme, for buses and subways, combined with the current bus and subway package mode, the number of times of use of buses and subways is set to two levels of [0, unlimited times]. Among them, when recommending personalized schemes, the rule is: the bus and subway appear at least once at the [unlimited times] level to avoid the situation where all the buses or subways are 0. For the active scheme, a higher level of attribute level point value is selected from the attribute level point set to generate an attribute level description. For example, for shared bicycles and shared electric motorcycles, a range of times of use is set higher than the attribute level of the neutral scheme. Specifically, the range of the number of times of use of shared bicycles is set to three levels of [P2-K, P3-K, P4-K]. The range of the number of times of use of shared electric motorcycles is set to three levels of [P2-EK, P3-EK, P4-EK]. In the active scheme, there is no reservation car attribute.

[0093]

[0094] Table 4

[0095] Similarly, referring to the construction of neutral solutions, based on the attribute level description of the active group, three personalized travel solutions are generated for the MaaS active user group through the D-efficient design method, and the effective design is screened and the D-error value is calculated through the column-based RSC algorithm to ensure the effectiveness and efficiency of the solution, so as to provide accurate and personalized MaaS recommendations for different users. The specific process of constructing an active solution is similar to that of constructing a neutral solution. The specific implementation process has been described above and will not be repeated here.

[0096] For example, in combination with the travel plan price formula, the personalized MaaS plan design and recommendation method provided by the present application provides a MaaS active plan for active user groups as follows: Figure 6 As shown, three different active travel plans are provided for users. Based on the differences between the three travel plans, the prices of the three travel plans are different, providing users with a variety of personalized travel plans.

[0097] As an implementation method, for the passive group among the residents, the travel attributes of the passive group include booking a car, and booking a car includes online car-hailing and taxis; the travel plan selection scenario constructed for the passive group includes at least three passive travel plans, and, in each passive travel plan, only available travel distances for booking a car at different levels are provided.

[0098] Furthermore, as an implementation method, a travel attribute description of the reserved car is provided, and an attribute level point value is selected from the travel attribute description of the reserved car to construct a passive travel plan.

[0099] Specifically, according to the above description, the passive scheme is aimed at the high-level reservation car user group. As shown in Table 5, in the passive scheme, there are no bus, subway, shared bicycle and shared electric motorcycle attributes. For the passive scheme, a higher level of attribute level point value is selected from the attribute level point set to generate an attribute level description. For example, for reservation cars, a usage distance range higher than the attribute level of the neutral scheme is set. Therefore, the usage distance range of reservation cars is set to three levels [P2-E, P3-E, P4-E].

[0100]

[0101] Table 5

[0102] Similarly, referring to the construction of neutral solutions, based on the attribute level description of the passive group, three personalized travel solutions are generated for the MaaS passive user group through the D-efficient design method, and the effective design is screened and the D-error value is calculated through the column-based RSC algorithm to ensure the effectiveness and efficiency of the solution, so as to provide accurate and personalized MaaS recommendations for different users. The specific process of constructing a passive solution is similar to that of constructing a neutral solution. The specific implementation process has been described above and will not be repeated here.

[0103] For example, in combination with the travel plan price formula, the personalized MaaS plan design and recommendation method provided by the present application provides a MaaS passive plan for the passive user group as follows: Figure 7 As shown, three different passive travel plans are provided for users. Based on the differences among the three travel plans, the prices of the three travel plans are different, thus providing users with a variety of personalized travel plans.

[0104] According to the above description, the personalized MaaS solution design and recommendation method of the present application is based on the historical travel information of resident users, and divides resident users into three types of groups according to their travel characteristics: active, neutral, and passive. Based on the historical travel information of the three types of user groups, the attribute level description of any travel attribute of any user group is determined. Through the D-efficient design method, personalized travel plans for user groups are constructed, and the effectiveness and efficiency of the plans are ensured through the RSC algorithm. Based on the historical travel information of the target user, the type of group to which the target user belongs is analyzed, so as to provide users with accurate and personalized travel plans.

[0105] On the second aspect, the present application also provides a personalized MaaS solution design and recommendation system, which adopts the personalized MaaS solution design and recommendation method described above to recommend travel solutions that are suitable for users' travel habits.

[0106] It will be appreciated that the word "exemplary" as used herein means "serving as an example, instance, or illustration". Any embodiment described as "exemplary" is not necessarily preferred or superior to other embodiments and / or does not exclude the combination of features of other embodiments. It will be appreciated that certain features of the present application described in the context of separate embodiments for the sake of clarity may also be provided in a single embodiment by combination. Conversely, various features of the present application described in the context of a single embodiment for the sake of clarity may also be provided individually or in any suitable combination or as any other described embodiment of the present application.

[0107] The above disclosure is only the preferred embodiment of the present application, but it is not intended to limit the scope of rights of the present application. A person of ordinary skill in the art can understand that without departing from the spirit and scope of the present application and the appended claims, changes, modifications, substitutions, combinations, and simplifications should all be equivalent replacement methods and still fall within the scope of the invention.

Claims

1. A personalized MaaS solution design and recommendation method, characterized in that: The method comprises the following steps: Acquire historical travel information of residents in the target city, wherein the historical travel information includes travel attributes and attribute levels; Based on the historical travel information, a clustering algorithm is used to extract travel characteristics to obtain the attribute levels of various travel attributes of different groups of urban residents, wherein the group types of urban residents include positive, neutral and negative types; Based on the attribute levels of each travel attribute of any group of urban residents, construct a travel plan selection scenario corresponding to the group; Acquire historical travel information of a target user, determine the type of group to which the target user belongs based on the historical travel information of the target user, and recommend a corresponding travel plan selection scenario to the target user based on the type of group to which the target user belongs; The method further comprises: Determine an attribute level point set of any travel attribute in any group, and determine an attribute level description of the travel attribute of the corresponding group based on the attribute level point set, wherein the attribute level description includes three attribute level point values ​​selected from the attribute level point set, and the attribute level descriptions of the same travel attribute of different groups are set differently; Constructing a travel plan selection scenario corresponding to the group based on a combination of attribute level point values ​​of each travel attribute; Among them, the attribute level point set for determining any travel attribute in any group includes: Based on the attribute level of any travel attribute of the group, determining the distribution form of the attribute level, In response to the attribute level being approximately normally distributed, selecting a quartile method to determine the attribute level point value to obtain the attribute level point set; In response to the attribute level distribution being skewed, a quintile or smaller quantile is selected to determine the attribute level point value to obtain the attribute level point set.

2. The personalized MaaS solution design and recommendation method according to claim 1, characterized in that: The method further comprises: Before determining the distribution form of the attribute level, the attribute level is first standardized, and the distribution form is determined based on the standardized attribute level. The quantiles are determined based on the distribution form, and the standardized values ​​corresponding to each quantile are de-standardized and rounded to obtain the attribute level point values.

3. The personalized MaaS solution design and recommendation method according to claim 1, characterized in that: The method further comprises: The travel attributes include bus and subway; The travel plan selection scenarios constructed for the active group include at least three active travel plans, and in any of the active travel plans, at least one of the bus and the subway is not limited in the number of times it can be used.

4. The personalized MaaS solution design and recommendation method according to claim 3, characterized in that: The method further comprises: The travel attributes also include shared bicycles and shared electric motorcycles; Provide travel attribute descriptions of the shared bicycles and the shared electric motorcycles, select an attribute level point value from the travel attribute descriptions of the shared bicycles and the shared electric motorcycles respectively for combination, and set unlimited usage times for at least one of the bus and the subway to construct the active travel plan.

5. The personalized MaaS solution design and recommendation method according to claim 1, characterized in that: The method further comprises: The travel attributes include car reservation; The travel plan selection scenario constructed for the passive group includes at least three passive travel plans, and in each of the passive travel plans, only available travel distances for reserved vehicles of different levels are provided.

6. The personalized MaaS solution design and recommendation method according to claim 5, characterized in that: The method further comprises: A travel attribute description of the reserved car is provided, and an attribute level point value is selected from the travel attribute description of the reserved car to construct the passive travel plan.

7. The personalized MaaS solution design and recommendation method according to claim 1, characterized in that: The method further comprises: The travel attributes include public transportation, subway, shared bicycles, shared electric motorcycles and reserved vehicles; The travel plan selection scenario constructed for the neutral group includes at least three neutral travel plans. A travel attribute description of each of the travel attributes is provided, and an attribute level point value combination is selected from the travel attribute description of each of the travel attributes to construct the neutral travel plan.

8. The personalized MaaS solution design and recommendation method according to any one of claims 2 to 7, characterized in that: The method further comprises: The price of any travel option is expressed by the following formula: M=(A*2+B*4+C*1.5+D*2.5+E*2.5)*F; In the formula, A represents the attribute level point value of public transportation. If the public transportation is unlimited in the travel plan, then A=7.5; B represents the attribute level point value of subway. If the subway is unlimited in the travel plan, then B=6.25; C represents the attribute level point value of shared bicycles; D represents the attribute level point value of shared electric motorcycles; E represents the attribute level point value of reserved car use; and F represents the discount coefficient.

9. A personalized MaaS solution design and recommendation system, characterized in that: The system adopts the personalized MaaS solution design and recommendation method as described in any one of claims 1 to 8.

Images