CN111667086A

CN111667086A - Vehicle co-riding path optimizing method and system

Info

Publication number: CN111667086A
Application number: CN201910176782.7A
Authority: CN
Inventors: 邹难; 陈爽; 杨坤鸿
Original assignee: Shandong University
Current assignee: Shandong University
Priority date: 2019-03-08
Filing date: 2019-03-08
Publication date: 2020-09-15
Anticipated expiration: 2039-03-08
Also published as: CN111667086B

Abstract

The invention provides a vehicle ride-sharing path optimizing method and a system, which collect real-time data of vehicles, roads and passengers, screen and eliminate the data outside the area range, temporarily store the data in the area range, and take the highest driver income, the shortest vehicle driving distance, the lowest vehicle idle time and the highest passenger satisfaction as optimization targets to establish a multi-vehicle ride-sharing matching and path optimizing model with constraint conditions, and utilize an improved genetic algorithm and a simulated annealing algorithm to solve the multi-vehicle ride-sharing matching and path optimizing model according to the real-time data of vehicles, roads and passengers temporarily stored in a server module to obtain an optimal target function value, establish the relations among passengers, passengers and vehicles, and between vehicles, and effectively realize the ride-sharing matching among passengers and vehicles, the route planning is efficiently realized, the contradiction between passengers and drivers is balanced, the vehicle resources are saved, and the energy waste is reduced.

Description

Vehicle co-riding path optimizing method and system

Technical Field

The disclosure relates to a method and a system for optimizing a vehicle co-path.

Background

The statements in this section merely provide background information related to the present disclosure and may not necessarily constitute prior art.

At present, the quantity of motor vehicles is continuously increased, the demand of travel is rapidly increased, a series of traffic problems such as traffic jam, energy consumption and environmental pollution are caused, and travelers are often troubled by the problems of difficult driving, difficult parking and the like. However, with the rapid development of technologies such as global positioning system, mobile internet, social network, and the like, the travel together becomes a new travel mode.

The vehicle-carrying is beneficial to reducing the trip cost of travelers, can relieve traffic jam and reduce traffic pollution, and although the vehicle-carrying has a plurality of advantages, the vehicle-carrying is still not a universal traffic mode because an efficient and reasonable method for coordinating the trip routes and the trip time of different travelers is lacked.

Therefore, the problems to be solved by the present inventors are as follows: (1) there is no effective method to implement the planning of the traveler's route and travel time; (2) the current co-taking system cannot meet the individual riding requirements of passengers, the co-taking satisfaction is low, and the co-taking desire is not strong; (3) the vehicle is often empty or only carries one passenger for a long distance, so that resources are seriously wasted; (4) the waste of vehicle resources brings about a larger problem of exhaust emission, and the environmental pollution is aggravated; (5) the contradiction of finding a car by a person and finding a person by a car easily occurs in the peak hours of commuting, so that the traffic burden and the energy consumption are greatly increased, and the travel time of passengers is wasted.

Disclosure of Invention

In order to solve the defects of the prior art, the disclosure provides a vehicle sharing path optimizing method and system, which establish the relations between passengers, between passengers and vehicles, and between vehicles, effectively realize the sharing matching between passengers and between vehicles, efficiently realize the path planning, improve the economic income of drivers, improve the traveling efficiency of passengers, balance the contradictions between passengers and drivers, save vehicle resources, reduce energy waste, objectively reduce the tail gas emission, and reduce the environmental pollution.

In order to achieve the purpose, the following technical scheme is adopted in the disclosure:

in a first aspect, the present disclosure provides a method for optimizing a vehicle co-ride path;

a method for optimizing a vehicle co-riding path comprises the following steps:

101, establishing a data acquisition module, a data preprocessing module and a server module, wherein the data acquisition module is used for acquiring real-time data of vehicles, roads and passengers, the data preprocessing module is used for eliminating data outside an area range acquired by the data acquisition module according to a preset area, and the server module is used for summarizing and temporarily storing the real-time data processed by the data preprocessing module;

102, establishing a multi-vehicle ride-sharing matching and path optimization model with constraint conditions respectively aiming at four passenger types of an economic preference type, a speed preference type, a social preference type and a safety preference type by taking the highest driver income, the shortest vehicle driving distance, the least vehicle idle time and the highest passenger satisfaction as optimization targets;

103 is also provided with a processor module, which solves the multi-vehicle ride-sharing matching and path optimization model by using an improved genetic algorithm and a simulated annealing algorithm according to the vehicle, road and passenger riding real-time data temporarily stored in the server module to obtain an optimal objective function value, and matches and plans the vehicles and the people.

As some possible implementations, the area may be divided according to a region range of an administrative division, or may be divided according to a range of a certain region to be studied, and data outside the area may be removed.

As some possible implementations, in the step 101, the passenger satisfaction includes a non-co-riding mode travel satisfaction and a co-riding mode travel satisfaction,

the travel satisfaction degree of the non-co-multiplication mode is as follows:

the travel satisfaction degree of the co-taking mode is as follows:

wherein, when 1, the multiplying mode is shown as the multiplying mode, and when 2, the multiplying mode is shown as the single multiplying mode;

representing the utility of the passenger s traveling in mode 1,

representing the utility of the passenger s traveling in mode 2.

As some possible implementation manners, in the step 101, the constraint conditions include a node constraint, a pair constraint, a response constraint, a neighborhood constraint, a time constraint, an order constraint, a capacity constraint, and a value range constraint.

As some possible implementations, in step 101, the multi-vehicle ride-sharing matching and path optimization model is:

where k is the number of the vehicle, s is the passenger number, M is the set of all serviced vehicles, n is the total number of reserved passengers, W is the set of all road network nodes,

in order to be a travel fee for the ride together,

mu is the cost per unit distance of vehicle transportation,

representing vehicle k passing through road network node i to road network node j,

representing that vehicle k does not pass through road network node i to road network node j, d_ijRepresenting the distance from road network node i to road network node j,

the satisfaction is shared by the passengers s,

the passenger s is given non-ride satisfaction.

As some possible implementations, in step 103, the improved genetic algorithm includes a coding design of chromosomes, generation of an initial population, and an improved genetic operator design, where the coding design of chromosomes is coded according to the road network node numbers, that is, the transportation routes are used as chromosomes, and the road network nodes are used as genes on the chromosomes to code.

As some possible implementations, the generating of the initial population includes the following steps:

501, calculating a time continuity judgment matrix TC among all passengers;

502 for m vehicles, generating m + 20, and using the m + 20 as the mark bits of the beginning and the end of each vehicle path;

503 assigned l for each vehicle for the last time period_kEach reservation passenger inserts the reservation passenger into the corresponding vehicle route without changing the position of the route node;

504 randomly disorderly sorts the numbers of the new NA passengers which are not allocated and are willing to be shared, wherein the number sequence is NA; randomly disordering and sequencing the numbers of new non-distributed NB passengers who do not wish to share the ride, wherein the number sequence is NB;

505, inserting the origin-destination positions of the unallocated passengers willing to share into the path by adopting a paired insertion method;

506, inserting the origin-destination positions of the unallocated passengers who are not willing to ride together into a path by adopting a paired insertion method;

507, returning to step 505, determining the next vehicle route until all vehicle routes are arranged, if the passenger is not assigned, indicating that the passenger is rejected, placing the starting point number ln(s) and the end point number en(s) of the passenger in the last 0 zone bit in sequence, and connecting all vehicle routes to form an individual;

508 returns to step 504 until M individuals are formed that satisfy the population count.

As some possible implementations, the genetic operator refined design includes selection, intersection and variation of chromosomes, the selection step of chromosomes includes:

601 selecting 1/4 individuals with the best population from the parent generation according to fitness and putting the 1/4 individuals into a set U1;

602 selecting the optimal 1/4 individuals from the population after the parent-generation hybridization according to the fitness, and putting the individuals into a set U2;

603 selecting the optimal 1/4 individuals from the parent-generation mutated population according to fitness and putting the individuals into a set U3;

604 selecting the optimal 1/4 individuals from the population after the hybrid generation variation according to the fitness, and putting the individuals into a set U4;

605 the optimal individuals are combined to obtain the next generation population U ═ U₁∪U₂∪U₃∪U₄。

As some possible implementations, the step of crossing over chromosomes includes:

701 randomly selecting two chromosomes from the parent generation population, then randomly selecting a vehicle, and routing the vehicle corresponding to the two chromosomes;

702 exchanging the gene segments corresponding to the two parent chromosomes to generate two new chromosomes, wherein the two gene segments have the same part and different parts, so that the chromosomes have gene deletion and gene duplication;

703 deletion of duplicated genes and filling up the deleted genes by forward pair insertion.

As some possible implementations, the chromosome mutating step is:

801 randomly selecting two vehicle paths of 1 chromosome, wherein the current number of the vehicle persons of the two selected vehicles is 0;

802 the initial position of the vehicle and the number sequence of the passengers who have got on the vehicle and are initially assigned to reserve are unchanged, the riding nodes of the passengers who have not got on the vehicle and the passengers who have not been assigned are divided into a group, the serial numbers are disordered, and the riding paths of the next vehicle and the previous vehicle are arranged according to a reverse-order pair insertion method.

As some possible implementations, in step 103, the simulated annealing algorithm includes the following steps:

901 obtains the current solution, sets the current temperature value T_i；

902, generating a new solution for the current demodulation, and solving an objective function value a of the new solution, where the original objective function value is B, and △ f is used to represent the increment of the objective function value, that is, △ f is a-B, and when the increment of the objective function value is greater than 0, the probability of the system accepting the new solution is 1, otherwise, the probability is used to calculate the new solution, otherwise

Discarding the new solution;

903 is cooled according to the decay rate r, and the temperature value T is updated_i+1＝T_i·r；

904, judging a termination condition, and if the termination condition is met, turning to a step 905; otherwise, go to step 902;

905 output the current solution.

In a second aspect, the present disclosure provides a vehicle co-ride path optimization system;

a vehicle-sharing path optimizing system comprises a data acquisition module, a data preprocessing module, a server module and a processor module, wherein the data acquisition module comprises a plurality of data acquisition terminals and is used for acquiring riding real-time data of vehicles, roads and passengers;

the data preprocessing module screens and rejects the data outside the area range acquired by the data acquisition module in real time according to a preset area; the server module is used for summarizing and temporarily storing the real-time data processed by the data preprocessing module;

the processor module is provided with a plurality of data processing terminals, a multi-vehicle ride-sharing matching and path optimizing model with constraint conditions is embedded in the data processing terminals, and the data processing terminals solve the multi-vehicle ride-sharing matching and path optimizing model by using an improved genetic algorithm and a simulated annealing algorithm according to vehicle, road and passenger riding real-time data temporarily stored in the server module to obtain an optimal objective function value.

Compared with the prior art, the beneficial effect of this disclosure is:

1. the scheme disclosed by the invention takes the highest driver income, the shortest vehicle driving distance, the least vehicle idle time and the highest passenger satisfaction degree as optimization targets, establishes a multi-vehicle ride-sharing matching and path optimization model with constraint conditions, establishes the relations between passengers, between passengers and vehicles and between vehicles, effectively realizes ride-sharing matching between passengers, passengers and vehicles, improves the driver income, improves the passenger traveling efficiency, balances the contradiction between passengers and drivers, saves vehicle resources, reduces energy consumption, objectively reduces exhaust emission and lightens environmental pollution.

2. According to the scheme, the analysis on the passenger riding satisfaction is effectively realized by establishing the co-riding and non-co-riding satisfaction functions, so that a foundation is provided for achieving the highest passenger satisfaction, the passenger satisfaction can be improved through the analysis, efficient vehicle co-riding is further really realized, the traveling efficiency is improved, and the resource waste is reduced.

3. The model is provided with a plurality of constraint conditions, because the sharing of multiple vehicles is a complex matching scheduling optimization problem, if all factors are considered during modeling, the research problem is extremely complex and difficult to solve, therefore, the optimization of the sharing path can be effectively realized through reasonably setting the constraint conditions, the solution of the rapid realization model is also realized, and the rapid path planning is realized.

4. The scheme disclosed by the invention carries out model solution through an improved self-adaptive genetic algorithm and a simulated annealing algorithm, the genetic algorithm has high flexibility, can break through the limitation of an objective function and repeated constraint, has strong compatibility, can be fused with various optimization algorithms, but the local search energy of a single genetic algorithm is weak, the premature phenomenon easily occurs, the simulated annealing algorithm is simple to operate, the complex nonlinear problem is simplified, the robustness and the local search capability are strong, but the global search capability is poor, the scheme disclosed by the invention can effectively complement the genetic algorithm and the simulated annealing algorithm through fusion, the rapid solution of the model is realized, and the co-product matching and the path planning efficiency are greatly improved.

5. The content of the disclosure improves the genetic algorithm to a certain extent, the performance of the genetic algorithm is influenced by the change of control parameters to a certain extent, while the traditional standard genetic algorithm adopts fixed cross probability and variation probability, the selection of the probability is generally determined by subjective experience and can not be well adapted to the problem of complex change.

6. The content disclosed by the disclosure is formed by optimizing individual selection, namely the partial optimal population of the parent population, the partial optimal population of the crossed population and the partial optimal population after mutation, so that the operation efficiency is effectively improved, the optimal solution in the population must enter the next generation, and the problem of difficult solution caused by complex models is effectively solved.

7. The traditional crossing operation comprises single-point crossing and double-point crossing, but the visiting sequence of nodes on the driving path of each vehicle in the model disclosed by the disclosure is divided in sequence, and the starting point and the stopping point of the same passenger need to be visited by the same vehicle at the same time, if a traditional crossing mode is adopted, the sequence of solutions is disordered, a large number of infeasible solutions are formed, the content disclosed by the disclosure ensures the new calculation efficiency with excellent feasibility in the crossing process by adopting a crossing operator based on vehicle path segmentation, and the problem of the sequence disorder of the solutions brought by the traditional crossing method is effectively solved.

8. The simulated annealing algorithm has strong deep searching capability, overcomes the defect of poor local searching capability of the genetic algorithm, and greatly improves the accuracy of the obtained objective function value.

Drawings

Fig. 1 is a flow chart of a data preprocessing module according to embodiment 1 of the present disclosure.

Fig. 2 is a flowchart of a multiplicative path optimizing process described in embodiment 1 of the present disclosure.

FIG. 3 is a flow chart of the genetic algorithm for fusion simulated annealing according to example 1 of the present disclosure.

Detailed Description

It should be noted that the following detailed description is exemplary and is intended to provide further explanation of the disclosure. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure belongs.

It is noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of example embodiments according to the present disclosure. As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, and it should be understood that when the terms "comprises" and/or "comprising" are used in this specification, they specify the presence of stated features, steps, operations, devices, components, and/or combinations thereof, unless the context clearly indicates otherwise.

Example 1:

as shown in fig. 1 to 3, embodiment 1 of the present disclosure provides a method for optimizing a shared-vehicle path, where a shared-vehicle is a novel traveling mode, and the mode can not only alleviate the problem that a traveler is difficult to take a bus at a peak time, but also improve the passenger carrying rate of a vehicle, reduce the cost of one-person traveling, improve traffic congestion, reduce traffic pollution, achieve multiple purposes, and have a significant application value, and this embodiment comprehensively analyzes the economy, timeliness, comfort, responsiveness, and safety in passenger riding experience, and constructs a passenger riding satisfaction function; and simultaneously, considering various constraint conditions such as passenger shared willingness, luggage carrying and the like, respectively carrying out vehicle matching and path optimization modeling on three passenger types aiming at the on-vehicle passenger, the reserved and allocated passenger who is not on the vehicle and the reserved and allocated passenger who is not allocated with the lowest vehicle carrying cost and the highest passenger satisfaction as optimization targets.

The multi-vehicle sharing matching problem researched by the embodiment is a multi-to-multi optimization problem aiming at multiple vehicles, multiple passengers and time windows, namely, on one hand, the passengers issue required information such as the on-off places, expected time, sharing willingness, luggage carrying and the like of the passengers to a vehicle dispatching center through reservation modes such as telephones or taxi-taking platforms and the like before going out, and on the other hand, a certain number of vehicles run in a limited trip range and send the positions and states of the vehicles to the dispatching center in real time, and at this time, the passengers may be available on the vehicles, or the passengers who have successful reservation but have not been available on the vehicles, and simultaneously, a plurality of new reserved passengers wait for distribution; therefore, the dispatching center determines the optimal matching scheme of the vehicle and the passengers and plans the vehicle path under the condition of meeting various constraints according to the position and the state of the current vehicle, the requirement information of the passengers and the externally connected social network information, thereby improving the carrying efficiency of the vehicle.

The method for optimizing the co-multiplication path comprises the following steps:

the method comprises the steps that a data acquisition module, a data preprocessing module and a server module are established, wherein the data acquisition module is used for acquiring real-time data of vehicles, roads and passengers during riding, the data preprocessing module is used for removing data outside an area range acquired by the data acquisition module according to a preset area, and the server module is used for summarizing and temporarily storing the real-time data processed by the data preprocessing module;

the method comprises the steps that a multi-vehicle ride-sharing matching and path optimization model with constraint conditions is established respectively for three passenger types, namely, on-vehicle passengers, reserved and unallocated passengers and passenger satisfaction with the purposes of highest driver income, shortest vehicle driving distance, least vehicle idle time and highest passenger satisfaction as optimization targets;

the system is also provided with a processor module, and the processor module utilizes an improved genetic algorithm and a simulated annealing algorithm to solve a multi-vehicle ride-sharing matching and path optimization model according to the vehicle, road and passenger riding real-time data temporarily stored in the server module to obtain an optimal objective function value.

The multiple vehicles are shared together to form a complex matching scheduling optimization problem, and if all factors are considered during modeling, the research problem is extremely complex and difficult to solve, so that on the premise of ensuring reasonability, a model is simplified through some basic assumptions, and the assumptions of the model are as follows:

(1) the method comprises the following steps of assuming that the current position of a vehicle and the boarding and alighting nodes of passengers are randomly distributed in a bounded road network;

(2) neglecting the influence of road conditions, and assuming that the shortest distance between two nodes in a road network is the straight-line distance between the two nodes;

(3) neglecting the influence of traffic conditions in the running process of the vehicle, and assuming that the vehicle runs at a constant speed;

(4) assuming that all vehicles are the same in type, namely the maximum passenger capacity is the same and the average running speed is the same;

(5) assuming that the riding requests of passengers are generated randomly, the riding requests of the passengers comprise riding time, riding position, riding together willingness and the like;

(6) the time of getting on and off the bus of the passenger is ignored, and the operation time of updating the state of the dispatching system is not considered;

(7) in order to realize system optimization, the riding request of the passenger is allowed to be rejected;

(8) the system performs status updates at a fixed time, once a passenger's reservation request is assigned, the passenger must service the assigned vehicle, no further changes are allowed, and for passengers who are not successfully assigned, the next reassignment can be waited for within an acceptable time frame.

Parameters related to the vehicle-sharing path optimization model mainly comprise information variables such as vehicles, passengers, riding requests, request responses and the like, and the specific model parameters are described as follows:

description of the symbols

O₁In-vehicle passenger collection

O₂Reserved passenger collection distributed but not on board

O₃Unassigned reserved passenger collection

Set of all road network nodes

M set of all serviced vehicles

m total number of vehicles

n total number of reserved passengers

Node in i, j road network

s number of passenger

Number of k vehicles

c_kMaximum passenger capacity of vehicle k, excluding driver

f_kMaximum load capacity of vehicle k, mainly large luggage

v_kAverage speed of vehicle k

Cost per unit distance for mu vehicle transport

h_sh_s1, the passenger s is willing to ride together

h_sWhen 0, the passenger s is unwilling to ride together

g_sThe amount of large pieces of luggage carried by the passenger s

k(s) passenger s has been assigned to vehicle k at the beginning of the current assignment

b (k) current vehicle position

e(s) coordinates of starting point of reserved passenger s

l(s) destination coordinates of reserved passengers s

Earliest time of getting on of passenger s acceptable at starting point

Latest time of getting on a vehicle that passenger s can accept at starting point

Earliest departure time for passenger s to be accepted at terminal

Latest alighting time that passenger s can accept at the end point

Time of arrival of vehicle k at road network node i

Waiting time for vehicle k to arrive at road network node i

d_ijActual shortest distance between node i and node j

t_ijTravel time of vehicle from node i to node j

Total number of vehicles k when leaving node i

G is a positive infinite number

Shared riding road section participated by passenger s

Non-shared riding road section participated by passenger s

Time for passenger to send out bus taking reservation request

Time to which passenger's ride request is assigned

Vehicle k gets on at node i

Vehicle k gets on or off at node i

Passenger s is assigned to vehicle k

Passenger s not assigned to vehicle k

Vehicle k passes through road network node i to road network node j

The vehicle k does not pass through the road network node i to the road network node j

Passenger s carries vehicle k from road network node i to road network node j

Passenger s carries vehicle k from road network node i to road network node j

The method for optimizing the car ride-sharing path specifically comprises the following aspects:

first, passenger trip satisfaction analysis

1. Travel mode selection influence factor analysis

The traveling mode researched by the embodiment is mainly directed to a non-ride-sharing mode and a ride-sharing mode, when a general passenger selects the traveling mode, the general passenger is influenced by various factors, most of the existing problems of optimization of the path shared by vehicles mostly only pay attention to the influence of the detour distance shared by vehicles and the riding cost, and the embodiment comprehensively considers five factors of economy, timeliness, comfort, responsiveness and safety of the ride-sharing service.

(1) Economy of use

The riding cost is an important driving force for stimulating the selection of passengers to ride together, the passengers often need to be forced to bypass, generally, a driver can charge 70% of the fee displayed by the fare meter for the riding-together road section based on certain economic compensation of the passengers in the taxi charging rule of the Jinan city. In order to simplify the model, in terms of charging, only the transportation charge is considered in the embodiment, according to the charging standard of taxis in the Jinan city, the starting time is 3 kilometers and 9 yuan, the exceeding part is 1.5 yuan per kilometer, and the more the passenger wants to travel, the better the saved cost is.

The expression (1-1) represents the charging standard of the bus in the conventional non-common-ride mode, and the expression (1-2) represents the charging pricing method in the common-ride mode.

(2) Aging property

The passengers all want to spend as short time as possible in the riding experience, the time consumption of the passengers is mainly reflected in waiting time and riding time, and the sum of the waiting time and the riding time is equal to the time difference between the waiting time of the passengers at the riding point and the getting-off time of the passengers. When a passenger takes a car, the passenger has an expected time window for the taking time

I.e. acceptable earliest time to get on

Latest time of arrival

The earliest boarding time of the passenger in the embodiment is the waiting time of the passenger at the riding point, and expressions (1-3) and (1-4) are the same and respectively represent the time consumption of the passenger in the non-ride-sharing and ride-sharing modes.

Wherein,

denotes the average human GDP, h_workThe working time of the average person year is represented; according to the survey of Jinan 2016, the GDP is 36394 yuan per person, and the working time per person is 2200h, because the value theta of unit time is 16.54 yuan.

(3) Sociability

The social interaction is an internal motivation for people to participate in collective activities, people can meet new friends through carpooling and get fun in talking with people, the social interaction researched by the disclosure means that the more people of the same age share topics in a relatively closed carriage, the higher the comfortable feeling of carpooling, and the comfort level in the embodiment is mainly related to the minimum age difference of passengers in a carpooling section.

Route(s){i，j…f} (1-6)

Wherein, (1-7) the node sequence set passed by the vehicle in the driving path of the passenger s is assumed that the driving path of the passenger s is divided into lm sections, the number of passengers in each section is different,

the age difference grade of the passenger s in the carpooling section l is shown, phi represents the conversion rate of the average age difference grade and the riding cost, and 0.3 is taken here, and expressions (1-8) and (1-9) respectively represent the social benefits of the passenger in the carpooling mode and the non-carpooling mode.

(4) Safety feature

Passenger safety concerns over riding patterns arise largely from distrust of strangers. Thus, to reduce the impact of this sense of insecurity caused by this type of distrust, the present disclosure improves the degree of matchmaking primarily through an assessment of the level of intimacy between the passenger and the passenger's level of confidence, studies have shown that sharing with relatives and friends may increase the willingness and tolerance of the passenger. It is assumed that the system can obtain the data of the intimacy grade and the credit grade between persons by connecting to other social networks or analyzing the previous riding records of passengers, and if the social relationship between persons is divided into five grades, the data is-2, -1, 0,1, 2. The level of the immediate relatives is defined as 2, the level of the common friends is defined as 1, and the level of the strangers is defined as 0. The higher the credit rating of the pool, the higher the security, assuming that the credit rating of the pool is divided into five levels of-2, -1, 0,1, 2. Likewise, passengers may also influence the level of intimacy and credit between passengers based on feedback from the shared experience.

Wherein,

an intimacy level coefficient indicating a person with whom the passenger is relatively most intimated in the section l;

representing the credit coefficient of the passenger with the highest credit rating in the section i. L is_lRepresents the path length of the link i; phi represents the conversion rate of the average safety level and the riding cost, and is 0.2; expressions (1-10) and (1-11) represent the social benefits of the passenger in the pool mode and the non-pool mode, respectively.

Comprehensively analyzing influence factors in passenger riding experience, and respectively defining the traveling utilities of passengers in a ride combination mode and a ride non-ride combination mode as (1-12) and (1-13):

wherein,

determined by the individual preferences of the passengers.

2. Individual preference demand index matrix

Because the passengers have heterogeneity, different passengers have different preferences, for example, female passengers mostly pay more attention to riding safety and riding cost, while male passengers prefer to pursue speed; it is difficult to meet the needs of all passengers with the same set of criteria, and thus in order to meet the individual needs of more passengers, the present disclosure adjusts the ride-sharing service strategy by establishing an individual preference requirement matrix.

For n un-reserved passengers, the demand degree of each passenger on the economic, timeliness, sociality and safety indexes of riding is different, so that the passenger individual preference demand index matrix is as follows:

each row in the X individual preference requirement matrix represents the preference degree of each passenger for the four indexes of economy, timeliness, sociality and safety, and the calculation of the travel utility weight in (1-12) and (1-13) is determined by the individual preference of the passenger, namely,

the matrix can be evaluated autonomously by the passenger at the time of a ride appointment, but requires x_ij∈[0,1]At the same time

When x is_i1＝max(x_i1,x_i2,x_i3,x_i4) Defining the passenger preference as an economic preference type;

when x is_i2＝max(x_i1,x_i2,x_i3,x_i4) Defining the passenger preference as a speed preference type;

when x is_i3＝max(x_i1,x_i2,x_i3,x_i4) Then, defining the passenger preference as a social preference type;

when x is_i4＝max(x_i1,x_i2,x_i3,x_i4) The passenger preference is defined as a security preference type.

3. Travel satisfaction function construction

The satisfaction degree of passengers in riding is related to important indexes such as economy, timeliness, sociality and safety in riding experience, and is also influenced by various random factors such as weather conditions and individual moods. None of these random factors can be quantified, and thus the present disclosure employs a random utility theory to estimate passenger ride-share satisfaction, equating the probability of theoretically selecting a ride-share mode to the passenger satisfaction level.

The upsilon represents a mode of taking a taxi, represents a co-taking mode when the upsilon is 1, and represents a single-taking mode when the upsilon is 2;

representing the utility of the passenger s traveling in the mode v,

representing the maximum travel utility of the two travel modes. The passenger satisfaction in the ride-on and ride-off mode can be found according to the formulas (3-15), as shown in the formulas (1-16) and (1-17).

Second, model construction

1. Optimizing target analysis

(1) Operating revenue

The operation income refers to the direct income obtained by all drivers for picking up and sending passengers, and the operation income is the support for maintaining the dispatching operation of the system and is also the key for attracting more drivers to join the platform. The operating revenue consists of passenger spending for all ride-sharing and non-ride-sharing modes.

(2) Cost of transportation

The transportation cost of the vehicle mainly refers to the fuel charge generated by the fuel consumed by the vehicle during driving, and generally, the farther the driving distance is, the more fuel charge is generated, so that the transportation cost is proportional to the driving distance of the vehicle, and can be expressed as:

(3) cost of idleness

The vehicle arrives in advance at the earliest time that the passenger can accept, and the vehicle waits to be idle, which means that the resources are not fully utilized, and certain loss is generated. More ride demands can be serviced if the waiting time is utilized. The idle cost is as follows:

(4) satisfactory reduction of loss costs

The service satisfaction of the passengers to the system is reduced, the passenger loss is caused, the development of the platform is not facilitated in the long term, and the average satisfaction of the passengers in the system is expressed as:

the analysis is to balance the benefits of both sides of the driver and the passengers of the vehicle, simplify the model and convert the multi-target problem into a single target.

2. Constraint analysis

(1) Node equalization constraints

Node balancing means that the traffic flow of each node is equal. Wherein, the expression (1-23) ensures that all vehicles start from the current initial position, (1-24) (1-25) ensures that passengers reserved successfully before are bound to be served by the vehicle from the reserved starting point to the corresponding terminal point, (1-26) and (1-27) represent node flow conservation, and the inlet flow and the outlet flow of any node are equal.

(2) Paired constraints

Paired constraints mean that the boarding and disembarking points of any passenger need to be serviced in pairs by the same vehicle.

(3) Response constraints

The response indicates that the reservation request for a passenger is serviced by at most one vehicle response, i.e., without limitation requiring that all passengers' reservations must be satisfied, allowing the vehicle to reject the passenger.

(4) Adjacent constraint

The adjacent constraint represents a condition that any two nodes can be continuously accessed by the same vehicle, and the latest getting-on time and the earliest arriving time which meet the condition are firstly obtained from the earliest getting-on time and the latest arriving time which can be accepted by the departure point and the destination in the reservation information of the passenger, such as formulas (1-/30) and (1-31). If in the optimal situation, the vehicle does not arrive before the acceptable latest arrival time of the next node without any detour delay from the earliest arrival time of the previous node in the middle, the two nodes cannot be adjacent, and the situation is represented as (1-32).

(5) Time constraints

The time constraint indicates that the vehicle should service the passenger at an acceptable time period for the passenger's reservation and the expressions (1-33) (1-34) indicate that the vehicle reaches the start of boarding and the end of disembarking of the passenger within the time window.

(7) Order constraints

The order constraint means that the visiting sequence of the same vehicle to the getting-on place of the same passenger is prior to the getting-off place;

(8) capacity constraints

The capacity constraint mainly refers to passenger carrying constraint and carrying constraint. The passenger carrying constraint is divided into a passenger carrying constraint willing to carry passengers and a passenger carrying constraint unwilling to carry passengers, which are respectively expressed as (1-36), (1-37) and (1-38) to indicate that the number of the carried objects of the vehicle can not exceed the maximum carrying number of the vehicle at any time.

(9) Value range constraint

The decision variables represent the relation between people, vehicles and roads, and are all 0,1 variable and are represented as (1-39).

Third, algorithm optimization design

The genetic algorithm has high flexibility, can break through the limitations of a target function and heavy constraint, has strong compatibility, and can be fused with various optimization algorithms. Although the genetic algorithm has strong global search capability, the local search capability is weak, the premature phenomenon is easy to occur, the simulated annealing algorithm is simple to operate, the complicated nonlinear problem is simplified, and the genetic algorithm has strong robustness and local search capability, but the global search capability is poor. The genetic algorithm and the simulated annealing algorithm can be effectively complemented by fusion, but the algorithm needs to be improved aiming at specific problems aiming at the problem of dynamic vehicle co-multiplication.

1. Genetic algorithm element design

In order to adapt to the high update frequency of the co-multiplication information state, the genetic algorithm with the parallel computing mode is selected in the embodiment. Although genetic algorithms are widely applied to optimization problems, traditional genetic algorithms have difficulty in obtaining ideal effects in complex situations due to relatively fixed genetic strategies. In order to further improve the performance of the algorithm, a genetic algorithm needs to be further improved for the problem of multi-vehicle co-multiplication.

2. Coding and decoding design of chromosome

Since the vehicle-sharing path scheme is a problem that a plurality of vehicles carry a plurality of passenger lines, and the starting position of each vehicle is different, the departure point, the destination, the service time range received, the sharing intention and the number of carried luggage of each passenger are also different. For the characteristics of the model, the present embodiment adopts coding according to the number of the road network nodes, that is, coding is performed by using the transportation route as a chromosome and using the road network nodes as genes on the chromosome.

The coding rule is implemented by adding zero, and each time zero is added, a vehicle driver is added. For one route, the nodes of the road network are arranged according to a certain sequence, a plurality of 'zeros' are added to divide the whole route into a plurality of route segments, and each route segment is distributed by a driver. Except for the vehicle occupant in the initial state, the getting-on point and the getting-off point of any other occupant must be in the path of the same vehicle; besides 0, the number of each node contains a set of attribute values such as position coordinate information, time window information, the number of passengers at the node, and riding will. The order of different nodes in a single vehicle path characterizes different paths, and changes in the order of nodes within a single path can cause changes in the solution. When the sequence of the path nodes of the single vehicle is not changed, the sequence of the vehicles is changed, and the value of the solution is not influenced.

For example, the chromosome 1-10 represents the current position node of the vehicle, and the path number of the vehicle 2 in the chromosome is 21-13-14-22, wherein 2 is the current position of the vehicle, 13 represents the getting-on position of the passenger 2, and 14 represents the getting-off position of the passenger 2. The last 0 represents a passenger riding node which is not allocated for waiting for allocation in the next period temporarily in the current period, and the nodes behind the last 0 represent riding nodes of passengers with refused reservation requests.

Parent 1: 0111120212131422 … 0353601718

3. Generation of an initial population

In the genetic algorithm, the initial population is generally generated randomly, but in the problem of temporary shared riding of the vehicle, the boarding and disembarking nodes of passengers have strict sequential access sequence, and the vehicle needs to meet paired constraints, time window constraints, capacity constraints and riding desire constraints. If the random generation method is adopted, a large number of invalid solutions can be generated, and the operation efficiency of the algorithm is greatly reduced. Therefore, the method of pairwise insertion of the passenger origin-destination points is adopted on the basis of randomly ordering the passenger numbers, a plurality of possible solutions are generated, and the initial population is constructed; the specific generation steps of the population are as follows:

the first step is as follows: calculating a time continuity judgment matrix TC among all passengers;

the second step is that: for m vehicles, generating m + 20 vehicles as the mark bits of the beginning and the end of each vehicle path;

the third step: assigned l for each vehicle for the last time period_kIndividual reservation passengers (including passengers who have reached a destination, passengers who are in a car, and passengers who are assigned but not boarding a car) do not change the path node positions. Inserting it into a corresponding vehicle path;

the fourth step: randomly disordering and sequencing the numbers of new NA passengers which are not distributed and wish to be shared, wherein the number sequence is NA; randomly disordering and sequencing the numbers of new non-distributed NB passengers who do not wish to share the ride, wherein the number sequence is NB;

the fifth step: inserting the origin-destination positions of the unassigned passengers willing to share into the path by adopting a paired insertion method

I. Starting from the route of the first vehicle, selecting a starting point number en(s) and an end point number ln(s) of a passenger s according to the NA number sequence when the current vehicle number k is 1;

attempt to insert en(s) into the initial path of the vehicle k from the current position of the vehicle (i.e., node k), first determine whether the node has a willingness to ride together. If not, moving backward by 2 bits, and retrying insertion; if so, judging whether the point meets the time constraint according to the value of the time continuity judgment matrix TC. If yes, the point is temporarily placed at the position, and the next step is carried out; if the current position does not meet the requirement, sequentially moving 1 position backwards until the position is inserted to the tail end of the vehicle path, and if the position is not inserted to the tail end successfully, repeating the step to try to insert the en (s +1) of the next passenger;

III.first, it is determined whether the node has a joint will, by attempting to insert ln(s) into the path of the vehicle k from the position immediately after en(s). If not, moving backward by 2 bits, and retrying insertion; if yes, judging whether the point meets time constraint after insertion, if yes, temporarily placing the point at the position, returning to the step II, and selecting the boarding and disembarking node of the next passenger for insertion; if not, the passenger is shifted backwards by 1 bit in sequence until the end of the vehicle path is inserted, and if the passenger is not successfully inserted until the end, the passenger's ln(s) and en(s) are together shifted from the vehicle path R_kDeleting, returning to the step II, and selecting the boarding and disembarking nodes of the next passenger to insert;

judging the capacity constraint of the generated path after traversing all the passengers in the NA, if any one of en(s) and ln(s) is not satisfied, then the ln(s) and en(s) of the passenger are together transmitted from the vehicle path R_kThereby determining the passenger set B served by the vehicle k_k；

V. will gather B_kThe passenger number in (1) is deleted from the sequence N, and NA is updated.

And a sixth step: inserting the origin-destination positions of unassigned passengers who are unwilling to ride together into a path using pairwise insertion

I. Starting from the route of the first vehicle, selecting a starting point number en(s) and an end point number ln(s) of a passenger s according to the NB number sequence when the current vehicle number k is 1;

II, trying to insert en(s) into an initial path of the vehicle k from the current position of the vehicle (namely the node k), firstly judging whether the number of occupants of the current node is 0, and if not, moving 1 bit backwards to retry insertion; if the time constraint is 0, judging whether the node and en(s) meet the time constraint according to the value of the time continuity judgment matrix TC, if so, judging whether ln(s) can be adjacent to the next node of the node in time, if both the front and the back meet the time constraint, inserting en(s) and ln(s) into the node in pair, then selecting the upper node and the lower node of the next passenger in the NB to try to insert, and repeating the steps. If the current position does not meet the requirement, moving the vehicle backwards by 1 bit to retry insertion, if the vehicle moves to the end and the insertion is still not successful, selecting the upper node and the lower node of the next passenger in the NB to try insertion, and repeating the step;

and III, after traversing all passengers in the NB, inserting the successful passenger s number, deleting the passenger s number from the NB, and updating the NB.

The seventh step: and returning to the step five when k is k +1, and determining the next vehicle path until all vehicle paths are scheduled. If the passengers are not allocated at last, indicating that the passengers are rejected, placing the ln(s) and the en(s) of the passengers in sequence at the last 0 mark position, and connecting the paths of all vehicles to form an individual.

Eighth step: and returning to the fourth step until M individuals meeting the population number are formed.

4. Calculation of fitness value

The fitness value of the genetic algorithm is used for representing the adaptability of the individual to the environment, the value of the fitness value of the individual directly influences the survival chance of the individual, and the probability of the gene being selected and inherited to the next generation. According to the multi-vehicle path optimization model, the objective function value of the model is selected as the fitness value, the larger the objective function value in the model is, the larger the total income is represented, the better the route optimization effect is, the larger the fitness value at the moment is, and the higher the probability that the gene of the individual is inherited to the next generation is.

5. Genetic operator improved design

Genetic manipulation mainly involves selection, crossing and mutation of chromosomes.

(1) Individual selection

And selecting chromosomes with high fitness in the population to enter the next genetic operation by taking the advantages and the disadvantages as a selection mechanism, wherein the selection mechanism is a key operator for ensuring that the overall fitness of the population is continuously improved. In the traditional genetic algorithm, only the offspring is selected, due to the complexity of the model, the generation of a feasible solution is difficult, most of the running time of the algorithm needs to be consumed, and the genetic operation of the parent chromosome in each step of crossing and mutation generates a new individual, namely a new feasible solution. In order to improve the operation efficiency of the algorithm and ensure that the optimal solution in the population must enter the next generation, the composition of the next generation individuals comprises a part of optimal population of the parent population, a part of optimal population of the crossed population and a part of optimal population after mutation. The method comprises the following specific steps:

the first step is as follows: selecting 1/4 individuals with the best population from the parent generation according to the fitness person, and putting the 1/4 individuals into the set U₁；

The second step is that: selecting the optimal 1/4 individuals from the parent hybridized population according to the fitness person, and putting the individuals into a set U₂；

The third step: selecting the optimal 1/4 individuals from the parent-generation mutated population according to the fitness of the fitness person, and putting the individuals into a set U₃；

The fourth step: selecting the optimal 1/4 individuals from the population after the hybrid generation variation according to the fitness, and putting the individuals into a set U₄；

Combining the optimal individuals to obtain the next generation population U ═ U₁∪U₂∪U₃∪U₄。

(2) Crossing of individuals

The individual crossing aims to enable the filial generation to inherit excellent genes of the parent generation as much as possible to form an individual with higher fitness, and the idea simulates the genetic evolution of organisms. Conventional crossover operations include single-point crossover, double-point crossover, and the like. However, in the model, the access sequence of the nodes on the driving path of each vehicle is divided successively, and the start point and the stop point of the same passenger need to be accessed by the same vehicle at the same time, so that if a traditional intersection mode is adopted, the sequence of solutions is disordered, and a large number of infeasible solutions are formed. Harbaoui et al solve the infeasible problem of solution by modifying the dot-to-dot sequence after crossing, but this method not only reduces the computation speed, but also reduces the genetic efficiency of the superior characteristics in the offspring because the crossing destroys the structural characteristics of the parent gene. For the problem, the present embodiment adopts a crossover operator based on vehicle path segmentation to ensure the calculation efficiency of generating new excellent feasibility in the crossover process. The steps of interleaving are as follows:

the first step is as follows: randomly selecting two chromosomes from the father generation population, then randomly selecting a vehicle, and selecting the paths of the vehicle corresponding to the two chromosomes, wherein the gray shaded font numbers represent the respective selected gene segments;

parent 1: 0116170212131422 … 0353601718

Parent 2: 01111314120212161722 … 001718

The second step is that: and exchanging the gene segments corresponding to the two parent chromosomes to generate two new chromosomes. Since the two gene fragments have the same part and different parts, the deletion and duplication of the gene occur in the chromosome, and the preliminary transformation chromosome is as follows:

progeny 1: 0116170212161722 … 0353601718

And (3) progeny 2: 01111314120212131422 … 001718

The third step: deleting repeated genes, wherein the parent 2 and the parent 1 are different genes (1617), and for the offspring 1, the genes which belong to the repeated genes need to be deleted, and the formation mechanism of the repeated genes of the offspring 2 is the same as that of the offspring 1;

progeny 1: 010212161722 … 0353601718

And (3) progeny 2: 0111120212131422 … 001718

The fourth step: filling up the deleted gene, wherein the parent 1 and the parent 2 are different genes (1314), for the offspring 1, the deleted gene needs to be filled up, and the forming mechanism of the deleted gene of the offspring 2 is the same as that of the offspring 1; aiming at a filling method of a deleted gene, a forward pair insertion method is adopted; because the gene coding sequence of the offspring is disordered, the rejected passenger node genes in the parent can be served, and therefore, the gene number behind the last 0 marker bit in the parent and the deleted gene are reinserted into the chromosome.

Progeny 1: 010212161722 … 0353601718

Gene 13143536 to be inserted into offspring 1

And (3) progeny 2: 0111120212131422 … 001718

Gene 1617 to be inserted into progeny 2

Through the cross operation of the mode, the chromosome is still a feasible solution after the cross, and excellent genes of parents are inherited as much as possible.

(3) Variation of individuals

The mutation operator is used for generating genes different from parents of other children, so that the diversity of the population is increased, the local searching capability of the algorithm is enhanced, and the algorithm is prevented from getting premature. Common mutation operations include single-point mutation, transposition mutation and the like, but the conventional method is likely to generate an infeasible solution, and the embodiment adopts a single-vehicle path overall interchange mutation method, which specifically comprises the following steps:

the first step is as follows: randomly selecting two vehicle paths of 1 chromosome, wherein the current number of the vehicles to the vehicle is 0;

parent 1: 01161702122219141320 … 001718

The second step is that: the initial position of the vehicle and the number sequence of passengers who are assigned to get on the vehicle in the initial reservation are unchanged, the riding nodes of passengers who are not getting on the vehicle and passengers who are not assigned are divided into a group, the serial numbers are disordered, and the riding paths of the next vehicle and the previous vehicle are arranged according to a reverse-order paired insertion method, so that the number of the passengers of the vehicle with the front number is prevented from being far larger than that of the vehicle with the back number.

Progeny 1: 0102122219141320 … 001718

Gene 161719141320 needs to be inserted into progeny 1.

5. Adaptive control parameter design

The performance of the genetic algorithm is influenced by the change of control parameters to a certain extent, while the traditional standard genetic algorithm adopts fixed cross probability and variation probability, the selection of the probability is generally determined by subjective experience, and the problem of complex change cannot be well adapted. According to the self-adaptive control method, individuals in the population are associated with the whole by self-adaptive adjustment of the cross probability and the variation probability, individual change is influenced according to the state change of the whole population, and the self-adjustment of self-adaptive control parameters of the algorithm can avoid damage to the structure of the better individuals or premature phenomenon to a greater extent.

(1) Adaptive adjustment of cross probability

The magnitude of the crossover probability directly affects the velocity of the new individual. The greater the probability, the faster the population changes. However, if the cross probability value is too large, the better individual structure may be damaged, and the optimal solution is difficult to obtain; however, if the crossover probability is too small, the optimization process will be slowed down. Therefore, in order to obtain a solution with a faster convergence rate and higher quality, the fixed cross probability needs to be changed into a value capable of adaptively adjusting the value according to the population change.

In the iterative optimization process, in order to enable the population to rapidly evolve towards a better direction, an individual with a lower fitness value needs to be subjected to higher cross probability; less crossover probability is taken for individuals with higher fitness values. In order to reflect the change state of the population in the genetic evolution process. The present disclosure first introduces an inverse coefficient of variation ζ to characterize the degree of difference between the population of each generation as a whole and the superior individuals.

Wherein N represents the number of population individuals, f_maxRepresenting the maximum fitness value among the i-th generation population individuals. The greater the individual difference from the maximum fitness value, the smaller the zeta. The cross probability usually ranges from p_c∈[0.4,0.9](ii) a I.e. p_cmaxIs 0.4, p_cmaxIs 0.9. In order to prevent the situation that the iteration is carried out for multiple times and then the local optimum is achieved, the upper limit value and the lower limit value of the cross probability are changed by adjusting zeta. Denoted as (4-2) and the improved cross probability formula is (4-3).

Wherein,

respectively representing the value upper and lower limits of the cross probability after the j generation population is adjusted;

expressing the crossover probability of chromosome i of the adjusted jth generation population;

the fitness value of chromosome i representing the j-th generation population,

the average fitness value of the j generation population is shown, and the value of A is 9.803438.

(2) Adaptive adjustment of mutation probability

Although the probability of variation is small, it also plays a crucial role. The variation operation can not only increase the diversity of the population, but also avoid the algorithm from falling into local optimum, but if the variation probability is too large, the better individual structure can be damaged; however, if the cross probability is too small, the genetic algorithm becomes too random and lacks directionality. Therefore, the mutation probability also needs to be adaptively adjusted to the size of the self value. The adaptive adjustment principle is like cross probability variation.

The mutation probability is usually 0.001-0.1, the upper and lower limit values of the cross probability are changed by adjusting zeta, and the value is expressed as (4-4), and the improved mutation probability formula is (4-5).

Wherein,

respectively representing the value upper and lower limits of the variation probability after the j generation population is adjusted;

expressing the variation probability of the chromosome i of the adjusted jth generation population;

the fitness value of chromosome i representing the j-th generation population,

6. Improved design of simulated annealing algorithm

The simulated annealing algorithm has strong deep searching capability and can make up for the defect of poor local searching capability of the genetic algorithm. In this embodiment, a new individual is obtained through a genetic algorithm, an objective function value is calculated for the new individual, and finally a new solution acceptance determination is performed through a Metropolis criterion. The simulated annealing algorithm comprises the following steps:

the first step is as follows: obtaining the current solution, and setting the current temperature value T_i；

The second step is that: and generating a new solution Nesolution for the current demodulation and integration, solving an objective function value A of the new solution, wherein the original objective function value is B, and the increment of the objective function value is represented by using delta f, namely delta f is A-B, when the Metropolis criterion indicates that the increment of the objective function value is more than 0, the probability of the system accepting the new solution is 1, otherwise, the probability is used as the probability

The new solution is discarded.

The third step: cooling according to the decay rate r, and updating the temperature value T_i+1＝T_i·r；

The fourth step: judging a termination condition, and if the termination condition is met, turning to the fifth step; otherwise, turning to the second step;

the fifth step: and outputting the current solution.

The embodiment 1 of the present disclosure further provides a vehicle-sharing path optimizing system, which includes a data acquisition module, a data preprocessing module, a server module and a processor module, wherein the data acquisition module includes a plurality of data acquisition terminals, and is used for acquiring real-time data of vehicles, roads and passengers during riding;

The above description is only a preferred embodiment of the present disclosure and is not intended to limit the present disclosure, and various modifications and changes may be made to the present disclosure by those skilled in the art. Any modification, equivalent replacement, improvement and the like made within the spirit and principle of the present disclosure should be included in the protection scope of the present disclosure.

Claims

1. A method for optimizing a vehicle-sharing path is characterized by comprising the following steps:

101, collecting real-time data of vehicles, roads and passengers, presetting a data collection area, screening and removing the collected data outside the area range according to the preset area, and temporarily storing the data within the area range;

102, establishing a multi-vehicle ride-sharing matching and path optimization model with constraint conditions respectively aiming at three passenger types, namely, on-vehicle passenger, reserved and unallocated passenger and reserved and unallocated passenger, by taking the highest driver income, the shortest vehicle driving distance, the least vehicle idle time and the highest passenger satisfaction as optimization targets;

103, solving the multi-vehicle co-riding matching and path optimization model by using an improved genetic algorithm and a simulated annealing algorithm according to the temporary vehicle, road and passenger riding real-time data to obtain an optimal objective function value, and matching and path planning the vehicles.

2. The method for optimizing vehicle-sharing path according to claim 1, wherein in step 101, the passenger satisfaction includes a non-co-riding mode trip satisfaction and a co-riding mode trip satisfaction, and the non-co-riding mode trip satisfaction is:

the travel satisfaction degree of the co-taking mode is as follows:

representing the utility of the passenger s traveling in mode 1,

representing the utility of the passenger s traveling in mode 2.

3. The method for optimizing vehicle-sharing path according to claim 1, wherein in step 101, the multi-vehicle sharing matching and path optimizing model is:

where k is the number of the vehicle, s is the passenger number, M is the set of all serviced vehicles, N is the set of all reserved passengers, N is the total number of reserved passengers, W is the set of all road network nodes,

in order to be a travel fee for the ride together,

mu is the cost per unit distance of vehicle transportation,

the satisfaction is shared by the passengers s,

the passenger s is given non-ride satisfaction.

4. The method for optimizing vehicle-sharing path according to claim 1, wherein in step 103, the improved genetic algorithm includes a coding design of chromosome, generation of initial population, and improvement of genetic operator, the coding design of chromosome is coded according to road network node number, that is, the transportation route is used as chromosome, and the road network node is used as gene on the chromosome.

5. The method for optimizing a vehicle-sharing path according to claim 4, wherein the generating of the initial population comprises the steps of:

501, calculating a time continuity judgment matrix TC among all passengers;

503assigned l for each vehicle for the last time period_kEach reservation passenger inserts the reservation passenger into the corresponding vehicle route without changing the position of the route node;

6. The method for optimizing vehicle-borne paths according to claim 4, wherein the genetic operator improvement design comprises selection, intersection and mutation of chromosomes, and the step of selecting chromosomes comprises:

7. The method for optimizing vehicle-borne paths according to claim 6, wherein the step of crossing chromosomes comprises:

8. The method for optimizing the vehicle-borne path according to claim 6, wherein the chromosome mutation step is:

9. The method for optimizing vehicle-sharing path according to claim 1, wherein in step 103, the simulated annealing algorithm comprises the following steps:

901 obtains the current solution, sets the current temperature value T_i；

Discarding the new solution;

905 output the current solution.

10. A vehicle-sharing path optimizing system comprises a data acquisition module, a data preprocessing module, a server module and a processor module, wherein the data acquisition module comprises a plurality of data acquisition terminals and is used for acquiring riding real-time data of vehicles, roads and passengers;