CN110986991A

CN110986991A - Same-starting-point vehicle sharing method

Info

Publication number: CN110986991A
Application number: CN201911386398.6A
Authority: CN
Inventors: 李佳佳; 刘昊; 夏秀峰; 宗传玉
Original assignee: Shenyang Aerospace University
Current assignee: Shenyang Aerospace University
Priority date: 2019-12-29
Filing date: 2019-12-29
Publication date: 2020-04-10
Anticipated expiration: 2039-12-29
Also published as: CN110986991B

Abstract

The invention discloses a vehicle sharing method with the same departure point, which comprises the steps of preprocessing a spatial road network, establishing a grid space index GSI, then pre-calculating the individual optimal driving path of a driver and the length of the optimal driving path, arranging the drivers in a descending order according to the length of the optimal driving path, matching the driver and the passengers by adopting a three-level matching strategy, processing the conflicted passengers by utilizing a priority strategy, and supplementing the passengers by expanding the grid area of the driver. The vehicle sharing method with the same starting point can quickly provide an optimal matching scheme for the driver and the passenger with the same starting point, and the total driving cost is minimized.

Description

Same-starting-point vehicle sharing method

Technical Field

The invention belongs to the technical field of bus sharing, and particularly provides a method for sharing vehicles with the same starting point, aiming at planning an optimal matching scheme for drivers and passengers with the same starting point and minimizing the overall total driving cost.

Background

With the improvement of the economic living standard of citizens, the quantity of private vehicles kept by commuters is increasing day by day, so that the traffic situation is worse. Under the condition that the car purchasing and maintaining costs are high, the demand of the interior of a large company and staff thereof for shared riding services is increased sharply, and more people pay attention to and receive car sharing travel modes. The car sharing travel can reduce the travel cost, can relieve traffic pressure, improves the transportation efficiency, reduces the energy consumption, and is an effective solution for solving the urban traffic problem. Nowadays, bus sharing is a preferred travel mode for more and more office workers due to the characteristics of low cost and strong flexibility. However, there are many bottlenecks to its development: (1) the explosion of location-based services has witnessed the arrival of a big data era. In large-scale data, how to provide satisfactory ride share quality of service (QoS) for each user in time has become a major challenge. (2) After receiving the passenger request, the driver delays the travel time because of unfamiliarity with the passenger position information, so that the driver can select the passenger who is familiar with and is willing to carry, and the enthusiasm of the downwind driver is eliminated. (3) Some units or the interior areas of houses are large, commuters are far away from stations, and the problem of the last kilometer of passengers cannot be solved by a traditional car sharing mode. Therefore, how to fully utilize the huge resource of the private car to relieve traffic pressure, solve the potential safety hazard of network car sharing, and improve the car sharing efficiency and the comfort level of journey has become a hot research problem in the current riding sharing field.

The riding share the existing main research work mainly comprises two parts: (1) a carpooling matching scheme; (2) and (4) global optimization. At present, many researches for solving a car sharing matching scheme are available, for example, designs of T-share, RSTR, SHAREK and the like are mainly used for helping passengers find optimal taxis, and a commonly adopted mode is to reduce a candidate set meeting constraint conditions through a pruning technology and accelerate quick matching between the passengers and drivers, but the researches are only suitable for taxi sharing scenes.

Those skilled in the art do not consider the following co-multiplication method: the driver and the passenger have a common departure point, and the driver has a destination, how to match the driver and the passenger to reduce the detour distance when the driver serves the passenger and reduce the possibility that the passenger goes out independently, and the problem of sharing vehicles with the departure point is effectively solved by minimizing the total driving cost of the driver and the passenger who is not matched.

Disclosure of Invention

In view of the above, the present invention provides a method for sharing vehicles with a same departure point, so as to solve the problem of the existing bus sharing method.

The technical scheme provided by the invention is as follows: a method for sharing vehicles with the same departure point comprises the following steps:

s1: preprocessing a spatial road network, and establishing a grid spatial index GSI: dividing a spatial road network into N-N grids with equal specifications by adopting a grid division strategy, defining a unique identifier grid number for each grid, and recording and storing query nodes in the grids, wherein the query nodes comprise a driver destination point and a passenger destination point;

s2: precalculating driver's individual optimum driving path and optimum driving path length, and arranging drivers in descending order according to optimum driving path length, defining the order as sort_v；

S3: essort (r) de_vThe sequence adds passengers with destinations and driver destinations in the same grid into the passenger candidate set of the corresponding driver and arranges the passengers in descending order according to Euclidean distance between the passenger destinations and the departure point, and the order is defined as sort_candiv；

S4: adding passengers with destinations on the optimal driving path of the driver into the passenger candidate set of the driver, arranging the passengers in a descending order according to Euclidean distances between the destinations and the departure of the passengers, and updating the sort_candivThen, passengers with destinations in grids passed by the optimal driving path of the driver are added into a passenger candidate set, the passengers are arranged in a descending order according to Euclidean distances between the destinations and the departure of the passengers, and the sort is updated_candiv；

S5: determining whether the intersection exists in the passenger candidate sets, if so, executing S6, if not, and the number of passengers in the passenger candidate set is less than or equal to the maximum passenger capacity of the vehicle, executing S7, and if not, and the number of passengers in the passenger candidate set is greater than the maximum passenger capacity of the vehicle, selecting sort according to the maximum passenger capacity of the vehicle_candivItem n, and executing S7, wherein n is equal to the maximum passenger capacity of the vehicle;

s6: processing the conflicted passengers according to a three-level priority processing strategy, selecting drivers for the conflicted passengers and updating a passenger candidate set, then arranging the passengers in the passenger candidate set in a descending order according to the Euclidean distance between a passenger destination point and a passenger departure point, and updating sort_candivWherein, the three-level priority processing strategy is as follows:

(1) the highest priority: selecting drivers for the conflicting passengers whose destination points are within the same grid as their destination points;

(2) secondary priority: drivers who select an optimal driving path for the conflicting passengers to pass through their destination points;

(3) the priority of the last stage: the grid through which the optimal driving path is selected for the conflict passenger comprises a driver of the grid where the destination point is located;

then, determining whether the number of passengers in the passenger candidate set is less than or equal to the maximum passenger capacity of the vehicle, if so, executing S7, otherwise, selecting sort according to the maximum passenger capacity of the vehicle_candivItem n, and executing S7, wherein n is equal to the maximum passenger capacity of the vehicle;

s7: passenger supplementation: updating the current driver and passenger set, expanding the grid area of the driver and redistributing the rest passengers;

s8: and determining the order of each vehicle to deliver the passengers by utilizing the 1NN neighbor query technology.

Preferably, in S2, the calculation of the driver' S individual optimal driving path based on the coordinate system oriented bi-directional shortest path query method includes the following steps:

s21: determining the grids G where the origin point S and the target point D are located_sAnd G_dThen, respectively with G_sAnd G_dCreating a rectangular coordinate system for the center and calculating G_dThe grid in the quadrant is defined as the target query grid TQG;

s22: from G_sAnd G_dTwo ends simultaneously initiate shortest path inquiry until the departure point S and the target point D are positioned in the same grid, and then adopt a direct heuristic method to inquire the shortest path to finally obtain the personal optimal driving path of the driver, wherein, the shortest path is carried out from the departure point S to the target point DThe query method comprises the following steps:

calculating the departure point S to the grid G_sThe shortest path lengths of all the boundary points are stored in a queue Pathlis1t, then, the nodes with the minimum path lengths in the Pathlis1 continue to expand, if the grid where the next expanded node of the node is located belongs to a target query grid TGQ, the next expanded node of the node is inserted into the Pathlis1, and if the next expanded node of the node does not belong to the target query grid TGQ and the expansion cost of the next expanded node of the Pathlis1 is greater than the expansion cost of the next node to be expanded, the next expanded node of the node is not expanded;

the method for inquiring the shortest path from the target point D to the departure point S comprises the following steps:

calculating a target point D to a grid G_dThe shortest path lengths of all the boundary points are stored in a queue Pathlist2, then, the nodes with the minimum path lengths in the Pathlist2 continue to expand, if the grid where the next expanded node of the node is located belongs to the target query grid TGQ, the next expanded node of the node is inserted into the Pathlist2, and if the next expanded node of the node does not belong to the target query grid TGQ and the expansion cost of the next expanded node of the Pathlist2 is greater than the expansion cost of the next node to be expanded, the next expanded node of the node is not expanded.

Further preferably, in S7, the current driver and passenger set is updated, and the method for expanding the grid area of the driver is as follows: updating the driver set, deleting the already fully loaded driver from the driver set, and updating the sort_v(ii) a Updating the passenger set, and calculating Euclidean distance D between the passenger destination point and the departure point_EAnd the remaining passengers are arranged according to D_EDescending order, defined as sort_p(ii) a And then, expanding the grid area of the driver, and recording the grid number of the grid to which each driver belongs.

Further preferably, the method for expanding the grid area of the driver is as follows:

if the grid to which the driver destination point belongs and the grid to which the departure point belongs are located in different rows and different columns, expanding a middle square area with the row number and the column number of the grid to which the driver destination point belongs and the row number and the column number of the grid to which the departure point belongs as boundaries;

if the grid to which the driver destination point belongs and the grid to which the departure point belongs are located in the same row, expanding two adjacent row areas of the row; similarly, if the columns are in the same column, the neighbor column area of the column is expanded.

Further preferably, in S7, the method for reassigning the remaining passengers is as follows: sort according to descending sequence of remaining passengers_pThe method comprises the steps of sequentially allocating passengers, firstly, determining grid numbers of grids to which destination points of the passengers belong and drivers corresponding to the grids, adding the grid numbers into a passenger candidate set of the corresponding drivers, if the grid numbers of the grids to which the destination points of the passengers belong correspond to a plurality of drivers, calculating Euclidean distances between the destination points of the passengers and the destination points of the drivers, taking the passengers nearby, and updating a final matching scheme.

Further preferably, the method for sharing vehicles with the same departure point further includes S9: calculating total distance traveled using a two-way shortest path query technique based on coordinate system orientation

Wherein D is_minAs the total travel distance, the distance of travel,

for all vehicle driving distances, ∑ spath_pThe distance that the passenger who is not carried in the passenger set travels to the destination independently.

The vehicle co-riding method with the same starting point can quickly provide an optimal matching scheme for a driver and a passenger with the same starting point, and minimizes the total driving cost.

Drawings

The invention is described in further detail below with reference to the following figures and embodiments:

FIG. 1 is a flowchart of a method for co-multiplying vehicles having the same departure point provided by the present invention;

FIG. 2 is a schematic diagram of a bi-directional shortest path query based on coordinate system orientation;

fig. 3 is a spatial road network graph, which is an undirected graph composed of sets N, E, and W, and is denoted by G ═ N (E, W), where N ═ N₁,n₂,…,n_nIs a finite set of nodes; e ═ n (n)_i,n_j)|n_i,n_jE.g. N, i ≠ j }, which is a finite set of edges connecting two different nodes in N; w ═ W (n)_i,n_j)|(n_i,n_j) E, representing the weight of each edge of the spatial road network, namely the length of an actual road section, wherein a point S represents a common starting point, v represents a driver, and p represents a passenger.

Detailed Description

In order to make the objects, technical solutions and advantages of the present invention more clear, the present invention is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

As shown in fig. 1, the present invention provides a method for sharing vehicles with a same departure point, including the following steps:

s1: preprocessing a spatial road network, and establishing a grid spatial index GSI: the method comprises the following steps of dividing a spatial road network into N-N grids with equal specifications by adopting a grid division strategy, defining a unique identifier grid number (GID) for each grid, recording and storing query nodes in the grids, namely: driver destination place dest (v) and passenger destination place dest (p), as shown in fig. 3, the spatial road network is divided into grids G₁,G₂,…,G₉Recording the nodes stored in each grid, G₁(v₁,p₁)，G₂(p₅)，G₃(v₂,p₆)，G₄(p₂)，G₅(S,p₃)，G₆(p₇)，G₈(v₃,p₄)；

S2: precalculating out individual optimum driving path spath of driver_vAnd an optimal travel path spath_vLength of sd (spath)_v) And make the driver follow the optimal driving path spath_vLength of sd (spath)_v) Sort in descending order, defining the order as sort_vIn the embodiment shown in FIG. 3, sort_v＝(v₁,v₂,v₃)；

Preferably, in S2, the calculation of the driver' S individual optimal driving path based on the coordinate system oriented bi-directional shortest path query method includes the following steps: (the query diagram is shown in FIG. 2)

s22: from G_sAnd G_dTwo ends simultaneously initiate shortest path query until the departure point S and the target point D are positioned in the same grid, and then the shortest path is queried by adopting a direct heuristic method to finally obtain the personal optimal driving path of the driver, wherein the method for querying the shortest path from the departure point S to the target point D is as follows:

After the shortest path query method is oriented based on a coordinate system, the expansion mode is not a divergent mode any more, but direction judgment is carried out. If the grid to be expanded is the grid in TQG, the expansion is performed, and if the grid is not in TQG and the expansion cost is greater than the grid in expansion TQG, the expansion is not performed. The shortest path query based on coordinate system orientation can reduce the expansion amount of nodes, so that the expansion direction of the path query approaches to a target point, unnecessary expansion grids are reduced, and the path query efficiency is improved.

S3: essort (r) de_vThe passenger whose destination and driver' S destination are in the same grid is added into the passenger candidate set of corresponding driver and according to Euclidean distance D between passenger destination place Dest (p) and departure point S_EAnd (3) descending order arrangement: determining a grid GID where a driver destination place dest (v) is located, calling a grid space index GSI to obtain passenger destination places dest (p) stored in the grid GID where the driver destination place dest (v) is located, and adding the passengers into a passenger candidate set candi of the corresponding driver_vThen, the candidate set candi of passengers for each driver is individually selected_vAccording to the Euclidean distance D between the passenger destination point dest (p) and the departure point S_EDescending order, defining the order as sort_candiv；

In the embodiment shown in FIG. 3, G₁(p₁)，G₃(p₆)，G₈(p₄) Hence, sort_candiv1＝{p₁}，sort_candiv2＝{p₆}，sort_candiv3＝{p₄And if so, the current matching result is shown in table 1.

Table 1: results based on driver destination grid matching

Driver ID_v	Passenger candidate set candi_v
		v₁	p₁
v₂	p₆
		v₃	p₄

S4: the passengers with the destinations on the optimal driving path of the driver are added into the passenger candidate set of the driver and according to the Euclidean distance D between the destination point dest (p) of the passenger and the starting point S_ESort in descending order, update sort_candiv: query optimal driving path spath_vThe passenger destination place dest (p) and adds it to the corresponding candi_v(ii) a Then, the passengers with the destinations in the grids passed by the optimal driving path of the driver are added into a passenger candidate set, and the passengers are arranged according to Euclidean distance D between the destination points dest (p) of the passengers and the departure point S_ESort in descending order, update sort_candiv: determining optimal driving path spath_vThrough the grid GID, inquiring the grid space index GSI to obtain the destination points dest (p) of the passengers therein, adding the passengers into the passenger candidate set candi of the driver_vAnd update sort_candiv；

In the embodiment shown in FIG. 3, spath_v1：S→p₂→p₁→v₁，spath_v2：S→p₅→p₆→v₂，spath_v3：S→p₃→v₃。

spath_v1Passing through the grid: g₅，G₄，G₁；spath_v2Passing through the grid: g₅，G₂，G₃；spath_v3Passing through the grid: g₅，G₈Hence, sort_candiv1＝{p₁,p₂,p₃}，sort_candiv2＝{p₆,p₅,p₃}，sort_candiv3＝{p₄,p₃And f, the current matching result is shown in table 2.

Table 2: based on optimal driving path and result of grid matching

Driver ID_v	Passenger candidate set candi_v
		v₁	p₁，p₂，p₃
v₂	p₆，p₅，p₃
		v₃	p₄，p₃

S5: determining individual passenger candidate sets candi_vIf there is an intersection, if so, then S6 is executed, if not, and the number count of passengers in the passenger candidate set is counted_vIf the passenger capacity of the vehicle is less than or equal to the maximum passenger capacity of the vehicle, the step S7 is executed, if the passenger number count of the passenger candidate set does not exist_vIf the maximum passenger capacity of the vehicle is larger than the preset threshold, selecting sort according to the maximum passenger capacity of the vehicle_candivItem n, and executing S7, wherein n is equal to the maximum passenger capacity of the vehicle;

in the embodiment shown in FIG. 3, as can be seen from Table 2, passenger p₃At the same time by₁，v₂，v₃Loading, if the execution is finished, executing S6;

s6: processing the conflicted passenger according to a three-level priority processing strategy, selecting a driver for the conflicted passenger, updating a passenger candidate set, and then processing the passenger in the passenger candidate set according to the Euclidean distance D between the passenger destination point dest (p) and the departure point S_ESort in descending order, update sort_candivWherein the three levels of priorityThe processing strategy is as follows:

(1) the highest priority: selecting drivers for the conflicting passengers whose destination points dest (v) are located in the same grid as their destination points dest (p);

(2) secondary priority: selecting an optimal travel path spat for conflicting passengers_vDrivers passing their destination point dest (p);

(3) the priority of the last stage: selecting an optimal travel path spat for conflicting passengers_vThe passing grid comprises drivers of the grid with the destination point dest (p);

in the embodiment shown in FIG. 3, p₃Is located at driver v₃Spath of_v3Upper, the priority is highest, so p will be₃From v₁，v₂Candi of_vIs deleted, therefore, sort_candiv1＝{p₁,p₂}，sort_candiv2＝{p₆,p₅}，sort_candiv3＝{p₄,p₃The current matching result is shown in table 3.

Table 3: processed results of conflicting passengers

Driver ID_v	Passenger candidate set candi_v
		v₁	p₁，p₂
v₂	p₆，p₅
		v₃	p₄，p₃

Then, determining the number count of passengers in the passenger candidate set_vIf the maximum passenger capacity of the vehicle is less than or equal to the maximum passenger capacity of the vehicle, executing S7, otherwise, selecting sort according to the maximum passenger capacity of the vehicle_candivItem n, and executing S7, wherein n is equal to the maximum passenger capacity of the vehicle;

as can be seen from Table 3, v is the embodiment shown in FIG. 3₁，v₂，v₃Are all smaller than 4 (assuming that the maximum passenger capacity of the vehicle is 4), so S7 is executed;

the method for expanding the grid area of the driver comprises the following steps of updating the current driver and passenger set: updating the driver set V, deleting the fully loaded driver from the driver set V, and updating the sort_v(ii) a Updating the passenger set P, and calculating the Euclidean distance D between the destination point dest (P) and the departure point S of the passenger_EAnd the remaining passengers are arranged according to D_EDescending order, defined as sort_p(ii) a Then, expanding the grid area of the driver, and recording the GID of the grid to which each driver belongs;

example shown in FIG. 3, in combination with Table 3, sort_v＝(v₁,v₂,v₃)，sort_p＝(p₇)；

The grid area expanding method of the driver comprises the following steps:

if the grid to which the driver destination place dest (v) belongs and the grid to which the departure point S belongs are located in different rows and different columns, expanding a middle square area with the number of rows and columns of the grid to which the driver destination place dest (v) belongs and the number of rows and columns of the grid to which the departure point S belongs as boundaries;

if the grid to which the driver destination place dest (v) belongs and the grid to which the departure point S belongs are positioned in the same row, expanding two neighbor row areas of the row; similarly, if the columns are in the same column, the neighbor column area of the column is expanded.

In the embodiment shown in FIG. 3, v₁And S are respectivelyLocated in grid G₁And G₅，G₁In the first row and the first column, G₅In the second row and second column, so v₁The expanded grid area is (G)₁,G₂,G₄,G₅) Similarly, v can be obtained₂，v₃(iii) expansion area of (v)₂The expanded grid area is (G)₂,G₃,G₅,G₆)，v₃The expanded grid area is (G)₄,G₅,G₆,G₇,G₈,G₉)。

Wherein the method of reassigning the remaining passengers is as follows: sort according to descending sequence of remaining passengers_pSequentially distributing passengers, firstly, determining the grid number GID of the grid to which the destination point dest (p) of the passenger belongs and the driver corresponding to the grid, and adding the grid number GID and the driver into the passenger candidate set candi of the corresponding driver_vIf the grid number GID of the grid to which the passenger destination point dest (p) belongs corresponds to a plurality of drivers, calculating the Euclidean distance between the passenger destination point dest (p) and the driver destination point dest (v), taking the passenger nearby, and updating the final matching scheme;

in the embodiment shown in FIG. 3, p₇Located in grid G₆，G₆Corresponding to driver v at the same time₂，v₃And Dest (p)₇) And Dest (v)₂) A close distance, so that p₇Is assigned to v₂At this point, the matching process ends and the final matching result is updated as shown in table 4.

Table 4: post-passenger augmented final matching results

Driver ID_v	Passenger candidate set candi_v
		v₁	p₁，p₂
v₂	p₆，p₅，p₇
		v₃	p₄，p₃

And S8, determining the order of each vehicle to deliver passengers by utilizing a 1NN neighbor query technology.

Preferably, the method further comprises S9: calculating total distance traveled using a two-way shortest path query technique based on coordinate system orientation

Wherein D is_minAs the total travel distance, the distance of travel,

for all vehicle driving distances, ∑ spath_pThe distance traveled by the passenger who is not loaded in the passenger set to the destination by himself is obtained, wherein the vehicle corresponds to the driver.

As an example shown in fig. 3, by the above steps, it is possible to obtain: plan_v1:S→p₂→p₁→v₁，plan_v2:S→p₅→p₆→p₇→v₂，plan_v3:S→p₃→p₄→v₃；sd(plan_v1)＝16，sd(plan_v2)＝29，sd(plan_v3)＝8；

Sigma burst because there is no passenger going out alone_p0. Thus, D is 53.

The common-departure-point vehicle sharing method can facilitate the query and management of spatial data information through a grid spatial index technology in the early data preprocessing stage, and provides a three-level matching strategy in the matching scheme determining stage, wherein the three-level matching strategy comprises a destination grid matching strategy, a driving route and grid matching strategy and a passenger supplement strategy. Meanwhile, the three-level strategy can reduce the possibility of the passenger going out independently. By the method, the global travel cost (total travel distance) can be minimized, and therefore the co-riding problem of vehicles with the same departure point can be effectively solved.

The embodiments of the present invention have been described in detail with reference to the drawings, but the present invention is not limited to the above embodiments, and various changes can be made within the knowledge of those skilled in the art without departing from the gist of the present invention.

Claims

1. The method for sharing vehicles with the same starting point is characterized by comprising the following steps:

S4: passengers whose destinations are located on the driver's optimal travel path are added to the driver's passenger candidate set and exit according to the passenger destination pointsSorting the Euclidean distances between the transmission points in a descending order and updating the sort_candivThen, passengers with destinations in grids passed by the optimal driving path of the driver are added into a passenger candidate set, the passengers are arranged in a descending order according to Euclidean distances between the destinations and the departure of the passengers, and the sort is updated_candiv；

2. Method for co-multiplying vehicles having the same origin according to claim 1, characterized in that: in S2, the calculation of the driver' S individual optimal driving route based on the coordinate system oriented bidirectional shortest path query method includes the following steps:

3. Method for co-multiplying vehicles having the same origin according to claim 1, characterized in that: in the step S7, the first step,the current driver and passenger set is updated, and the method for expanding the grid area of the driver is as follows: updating the driver set, deleting the already fully loaded driver from the driver set, and updating the sort_v(ii) a Updating the passenger set, and calculating Euclidean distance D between the passenger destination point and the departure point_EAnd the remaining passengers are arranged according to D_EDescending order, defined as sort_p(ii) a And then, expanding the grid area of the driver, and recording the grid number of the grid to which each driver belongs.

4. Method for co-multiplying vehicles with a same starting point according to claim 3, characterized in that: the method for expanding the grid area of the driver comprises the following steps:

5. Method for co-multiplying vehicles with a same starting point according to claim 3, characterized in that: in S7, the method of reassigning the remaining passengers is as follows: sort according to descending sequence of remaining passengers_pThe method comprises the steps of sequentially allocating passengers, firstly, determining grid numbers of grids to which destination points of the passengers belong and drivers corresponding to the grids, adding the grid numbers into a passenger candidate set of the corresponding drivers, if the grid numbers of the grids to which the destination points of the passengers belong correspond to a plurality of drivers, calculating Euclidean distances between the destination points of the passengers and the destination points of the drivers, taking the passengers nearby, and updating a final matching scheme.

6. Method for co-multiplying vehicles having the same origin according to claim 1, characterized in that: further comprising S9: calculating total distance traveled using a two-way shortest path query technique based on coordinate system orientation

Wherein D is_minAs the total travel distance, the distance of travel,