CN111160948A - Multi-objective optimized shared network car-booking and car-sharing pricing method - Google Patents

Multi-objective optimized shared network car-booking and car-sharing pricing method Download PDF

Info

Publication number
CN111160948A
CN111160948A CN201911098257.4A CN201911098257A CN111160948A CN 111160948 A CN111160948 A CN 111160948A CN 201911098257 A CN201911098257 A CN 201911098257A CN 111160948 A CN111160948 A CN 111160948A
Authority
CN
China
Prior art keywords
passenger
ride
discount rate
detour
proportion
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN201911098257.4A
Other languages
Chinese (zh)
Inventor
曾伟良
林坤新
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Guangdong University of Technology
Original Assignee
Guangdong University of Technology
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Guangdong University of Technology filed Critical Guangdong University of Technology
Priority to CN201911098257.4A priority Critical patent/CN111160948A/en
Publication of CN111160948A publication Critical patent/CN111160948A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00—Commerce
    • G06Q30/02—Marketing; Price estimation or determination; Fundraising
    • G06Q30/0283—Price estimation or determination
    • G06Q30/0284—Time or distance, e.g. usage of parking meters or taximeters
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/40—Business processes related to the transportation industry

Landscapes

  • Business, Economics & Management (AREA)
  • Strategic Management (AREA)
  • Development Economics (AREA)
  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Business, Economics & Management (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Economics (AREA)
  • Marketing (AREA)
  • Accounting & Taxation (AREA)
  • Finance (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Game Theory and Decision Science (AREA)
  • Health & Medical Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Human Resources & Organizations (AREA)
  • Primary Health Care (AREA)
  • Tourism & Hospitality (AREA)
  • Devices For Checking Fares Or Tickets At Control Points (AREA)

Abstract

The application discloses a multi-objective optimized shared network car-appointment and car-ride combination pricing method, which comprises the following steps: determining the detour proportion and the common road section proportion of the passenger A and the passenger B in the process that the passenger A and the passenger B reach the respective destinations in the carpooling state; according to the detour proportion and the common road section proportion of the passenger A and the passenger B and the related discount rate of the passenger A and the passenger B; based on the charging expense calculation model, establishing an objective function, wherein the objective function comprises an objective function respectively representing the reduced expense proportion of the passenger A and the passenger B after the ride combination and an objective function of the income of a driver after the ride combination; setting a fitness function of the particle swarm algorithm according to the target function; calculating a reference discount rate to determine a value range of each relevant discount rate; and solving each relevant discount rate by using a particle swarm algorithm, solving an obtained optimal value according to the particle swarm algorithm, and determining the ride-share price of the passenger A and the ride-share price of the passenger B by using the charging fee calculation model.

Description

Multi-objective optimized shared network car-booking and car-sharing pricing method
Technical Field
The invention relates to the technical field of computers and transportation, in particular to a multi-objective optimized shared network car-booking and car-sharing pricing method.
Background
With the improvement of living standard of people, the number of private cars is increased, and urban traffic jam is more and more serious. In order to improve the effective utilization of the transport capacity and save energy, sharing and sharing are gradually becoming a novel travel mode. Different from the traditional non-shared taxi mode, the shared-riding mode of shared-network taxi appointment is greatly shortened in waiting time of total passengers and saved in oil consumption and trip cost on the basis of meeting scheduled trip time of a plurality of passengers through a refined designated shared-riding route and a driver-riding matching strategy.
In order to attract more passengers to use shared network appointments, price advantage is a key factor affecting sharing. The goal of shared ride-share pricing is to achieve a win-win situation for the driver and passengers. Optimized ride rates may increase driver revenue over non-ride rides, but sharing rides by multiple people creates detour problems, thereby incurring detour costs.
The conventional network car-booking and car-sharing pricing method comprises the following steps:
1. a taxi co-taking pricing model considering an intermediate station setting mode is established in taxi co-taking mode pricing analysis based on a co-taking mode, wherein the minimum total cost of all passengers is taken as a target, the passenger cost is fixedly restrained, a driver is detoured and compensated, and the problem of taxi co-taking pricing of a fixed station is solved. The disadvantages are that: the model has fixity on the price of passengers, introduces a discount rate artificially specified by the passengers, takes the discount rate as a critical value of the expense discount rate, and is easy to generate unfair conditions of 'short detour and large discount' on partial car sharing states.
2. A pricing optimization model of a dynamic co-riding mode is provided in optimization research on a taxi co-riding mode and a pricing model, and a solution for the problem of unbalanced income of a driver and a passenger is discussed to a certain extent. The method considers the reduction range of the cost of the passengers and the increase range of the income of the driver, and benefits both sides by optimizing the optimal discount rate when different numbers of passengers are combined, thereby improving the enthusiasm of both sides participating in the combination. However, the disadvantages of this method are: the method only discounts the common road section, and the pricing of the global path is not considered; taking a fixed detour discount rate since the detour portions are also considered as common segments, the detoured passenger will pay more for the detour, which will result in more shared travel costs than for the sole ride.
3. Research on Taxi marking Model and Optimization for carpooling Detour Problem establishes a multi-objective Optimization Model for solving Taxi carpooling with a lane-winding Problem, and designs a genetic algorithm to optimize three objectives of cost saving for passengers, profit increment per kilometer for drivers and cost saving after passengers detour. The method comprises the following steps: the whole road section of the detour passenger is discounted in the model, the detour passenger is provided with more passengers and needs low price, meanwhile, the cost of the driver is increased, and if the detour passenger is discounted in the whole road section, the benefit of the driver is weakened to a certain extent, and the driver and passenger benefit is not balanced. Due to the aforementioned disadvantages, there are cases: if two passengers do not detour, the passenger getting on the bus firstly makes a-fold on the common road section, and the passenger getting on the bus later makes a-fold on the whole course, so that the system is not fair.
4. The thesis aims at minimizing average cost paid by passengers, maximizing daily average income of taxis and minimizing the difference between vehicle management cost and subsidy in a model, and constructs a particle swarm multi-objective optimization model based on an improved niche. The method comprises the following steps: the cost calculation of the scheme comprises detailed conditions such as gas cost, gas subsidy and the like, but the detailed conditions are lack of the bypass problem, wherein non-bypass clients normally charge, and the bypass clients discount the whole travel road section, so that the benefit of a driver is weakened to a certain extent. The fitness function in the algorithm is: the differences between the distances of the particle target values and the corresponding optimal particle target values are added, and the effect of simultaneous optimization on several targets is small.
Disclosure of Invention
The invention aims to provide a multi-objective optimization shared network car-appointment and car-ride combination pricing method, which guarantees the fairness of pricing by route compensation and seeks balance points of a pricing strategy from multiple optimization objectives such as passenger interests and driver interests.
In order to realize the task, the invention adopts the following technical scheme:
a multi-objective optimized shared network car-appointment and ride-sharing pricing method is used for determining the charging cost of a ride-sharing scheme for a passenger A and a passenger B at different starting points, and comprises the following steps:
determining the detour proportion and the common road section proportion of the passenger A and the passenger B in the process that the passenger A and the passenger B reach the respective destinations in the carpooling state;
according to the detour proportion and the common road proportion of the passenger A and the passenger B and the related discount rate of the passenger A and the passenger B, a charging fee calculation model of the passenger A and the passenger B during ride-together is established; the relevant discount rates comprise a detour discount rate of the passenger A, a detour discount rate of the passenger B and a common road section discount rate of the passenger A and the passenger B;
based on the charging expense calculation model, establishing an objective function, wherein the objective function comprises an objective function respectively representing the reduced expense proportion of the passenger A and the passenger B after the ride combination and an objective function of the income of a driver after the ride combination;
setting a fitness function of the particle swarm algorithm according to the target function;
calculating a reference discount rate to determine a value range of each relevant discount rate;
and solving each relevant discount rate by using a particle swarm algorithm, solving an obtained optimal value according to the particle swarm algorithm, and determining the ride-share price of the passenger A and the ride-share price of the passenger B by using the charging fee calculation model.
Further, in the process that the passenger a and the passenger B reach their respective destinations in the ride-sharing state, the detour proportion and the common road section proportion of the passenger a and the passenger B include:
detour proportion of passenger A
Figure BDA0002269035280000031
Proportion of common route section of passenger A
Figure BDA0002269035280000032
Detour proportion of passenger B
Figure BDA0002269035280000033
Proportion of common road section with passenger B
Figure BDA0002269035280000034
Wherein L isaIndicating the arrival of passenger a at destination D from the starting pointaDistance of (L)bIndicating the arrival of passenger B at destination D from the starting pointbDistance ofFrom, LxIndicating arrival at passenger a at passenger D after passenger BbThe distance of (d); l isabRepresents the shortest distance from passenger a to passenger B; l isrThe distance that the passenger B still walks after the passenger A gets off the vehicle is shown; LDaIndicating the distance, LD, traveled by passenger A after the detourb＝Lx+Lr-Lb；LDbIndicating the distance, LD, traveled by passenger B after the detourb＝Lx+Lr-Lb。
Further, the model for calculating the charging fee of the passenger a and the passenger B during the combined ride comprises:
Figure BDA0002269035280000035
Figure BDA0002269035280000036
wherein the relevant discount rates comprise a detour discount rate α of the passenger A, a detour discount rate β of the passenger B, a common road section discount rate η of the passenger A and the passenger B, and cost (L)a)、cost(Lb) Indicating passenger a to his destination DaFor one ride, passenger B to his destination DbThe unique cost of.
Further, the passenger A arrives at his destination DaFor one ride, passenger B to his destination DbThe calculation method of the unique cost comprises the following steps:
Figure BDA0002269035280000037
wherein a, b and c represent constants, and L represents mileage; wherein a is the starting price of the bicycle, c is the kilometers included in the starting price, and b is the cost per kilometer after exceeding c kilometers; respectively mixing La、LbThe cost (L) can be obtained by substituting the parameter L into the above formulaa)、cost(Lb)。
Further, based on the charging fee calculation model, an objective function is established, which comprises an objective function representing the reduced fee proportion of the passenger A and the passenger B after the ride combination respectively, and an objective function representing the income of the driver after the ride combination, and comprises the following steps:
the objective function representing the proportion of the reduced cost of passenger a after a ride pool is:
Figure BDA0002269035280000041
the objective function representing the proportion of the cost of passenger B reduction after a ride pool is:
Figure BDA0002269035280000042
the objective function of the driver's income after the ride combination is:
Figure BDA0002269035280000043
further, the fitness function of the particle swarm algorithm is expressed as:
Figure BDA0002269035280000044
wherein:
Figure BDA0002269035280000045
in the above formula, globalxFor a corresponding objective function MaxZxX is a, b, d.
Further, the calculating a reference discount rate includes:
α0＝1-kaθa
β0＝1-kbθb
Figure BDA0002269035280000046
α therein0、β0、η0To reference discount rate, thetaa、θbBypass ratios for passenger A and passenger B, ηa、ηbThe proportions of the common road sections of the passenger A and the passenger B respectively; k is a radical ofa,kb,kηIs a proportional parameter;
the value range of each relevant discount rate is represented as:
0＜α＜α0
0＜β＜β0
0＜η＜η0
further, k isa,kb,kηThe values of (A) are as follows: k is a radical ofa,kbThe value of 0.5, kηThe value is 0.3.
Compared with the prior art 1 mentioned in the background art, the method aims to improve the discount rate problem, and the discount rate and the detour degree of the method have a great relationship and are not artificially fixed. The detour is large and is often corresponding to the preferential price, otherwise, the detour is not much different from the original price. Compared with the existing method 2, the method aims to improve the rationality of the discount rate and the reasonable charging problem, the discount rate is automatically optimized according to the benefits of drivers and conductors, the passenger cost is divided into a common road section part and a detour part, and the opening cost is divided. The method is suitable for complex situations that two passengers bypass or do not bypass and the like. Compared with the prior art 3, the scheme aims at improving the diversity of the car sharing state and solving the problem of reasonable charging, and the scheme assumes that a passenger A gets on the car firstly and gets on the car after B, the car sharing state is divided into 4 conditions that the passenger A bypasses the road B, both passengers detour the road, the passenger A bypasses the road B, and both passengers do not detour the road. The charging is divided into the same road section and the detour road section, and the charging is divided, so that the purpose of fairness among passengers is achieved. Compared with the existing method 4, the method aims to improve the rationality of the fitness function, the method carries out the absolute value minimization on the fitness function of the particle swarm, achieves the aim of optimizing the target function at the same time, and prevents the situation that the target is unbalanced because the fitness function in the method 4 adopts an addition form, so that the size of the addition in the algorithm cannot be controlled. It can be seen that the present application has the following technical features compared to the prior art:
1. the application provides a driver and passenger interest balancing strategy, so that the interests of a driver and passengers are similar and do not tend to a certain party; the realization method is to use the difference of the riding interests to calculate the minimum value, so that the riding interests are balanced.
2. Simplification of the ride-sharing model, the model is divided into 5 parts, namely (L)a,Lb,Lab,Lx,Lr) The five variables are respectively: the car sharing state can summarize the car sharing states of most double-person combined riding modes.
3. And simultaneously optimizing the expectations of the three targets and the model by adopting a fitness function of the particle swarm. The three goals are: the proportion of cost savings to passengers A, B and the proportion of increased driver revenue; the expectation of the model is that the driver and passenger benefits are balanced on the premise that the driver and passenger benefits are maximized.
4. The passenger sectional charging mode is to divide the car-sharing road section into a detour part and a common road section part, and separately charge according to the detour road section and the common road section.
Drawings
FIG. 1 is a schematic flow chart of a multi-objective optimized shared network car-appointment and car-ride combination pricing method according to the present application;
FIG. 2 is a schematic diagram of a passenger A without detour and a passenger B with detour;
FIG. 3 is a schematic diagram of both passenger A and passenger B detouring;
FIG. 4 is a schematic diagram of a bypass by passenger A and a bypass by passenger B;
FIG. 5 is a schematic view of passenger A and passenger B both not detouring;
FIG. 6 is a schematic flow chart of a particle swarm algorithm;
fig. 7 is a diagram illustrating a variation process of the fitness of the optimization process in the specific example of the present application.
Detailed Description
The multi-objective optimized shared network car-booking and car-sharing pricing method mainly aims at the situation that when a passenger A, B is at different starting points, a car first carries the passenger A, then the car first reaches the destination of the passenger A after carrying the passenger B, and then the car reaches the destination of the passenger B; in this process, there are generally 4 kinds of ride-sharing situations, which are:
the first method comprises the following steps: the vehicle-mounted passenger A sends out, gets on the way without detour to the passenger B, then takes the passenger A to the destination by detour, and finally sends the passenger B to the destination, as shown in figure 2.
And the second method comprises the following steps: the vehicle-mounted passenger A sends out, bypasses the passenger B on the way, then bypasses the passenger A to carry to the destination, and finally sends the passenger B to the destination, as shown in figure 3.
And the third is that: the vehicle-mounted passenger A sends out, takes a detour on the way to the passenger B, then carries the passenger A to the destination without detour, and finally sends the passenger B to the destination, as shown in figure 4.
And fourthly: the vehicle-mounted passenger A sends out, gets on the passenger B without detour on the way, then carries the passenger A to the destination without detour, and finally sends the passenger B to the destination, as shown in figure 5.
The conditions followed by the above-described co-multiplication are three:
condition 1: if the detour limit is not more than 40% of the shortest path (the value is calculated by adopting a large amount of data by using the method), the method changes to a single-multiplication mode to carry out single-multiplication charging.
Condition 2; it is assumed that the passenger's path is arranged as the relatively shortest path in the road network.
Condition 3: it is assumed that the driver gains more revenue than a single ride when driving the same trip within the set discount rate.
The above condition 1 can be determined when the pool plan is generated, and if the detours of the passenger a and the passenger B are greater than 40% of the distances from the passenger a to the destination and from the passenger B to the destination, the pool plan is not adopted. Condition 2 above, when the co-multiplication scheme is generated, existing path planning algorithms may be employed to determine that the paths scheduled for passenger A, B are all relatively shortest paths. The above condition 3 is a condition generally adopted in each algorithm and related software at present, and the premise of the present scheme is that the condition is satisfied.
The carpooling state in the scheme refers to that two passengers A, B are in a carpooling modeThe state of the path under, this scheme will be (L)a,Lb,Lab,Lx,Lr) Represented as a carpool state, the vector covers the 4 ride-sharing scenarios described above.
The following is an explanation of the basic variables:
name of variable Description of the invention
DaAnd Db Destinations of passenger A and passenger B
A and B Starting points of passenger A and passenger B
Lab Shortest distance from passenger A to passenger B, i.e. shortest distance from point A to point B
La Distance of passenger A after planning path from starting point to destination, namely from point A to point DaIs a distance of
Lb Distance of passenger B after planning path from starting point to destination, i.e. from point B to point DbIs a distance of
Lx Arriving at passenger A at passenger D after passenger BbDistance (common road section)
Lr Distance that passenger A still needs to travel after getting off
LDb Distance, LD, traveled by passenger B after detourb＝Lx+Lr-Lb
LDa Distance, LD, traveled by passenger A after detoura＝Lx+Lab-La
α Detour discount rate for passenger a
η Common road segment discount rate for passenger A and passenger B
β Bypass discount rate for passenger B
The application provides a multi-objective optimized shared network car-booking and car-sharing pricing method, which is used for pricing a car-booking scheme of a passenger A and a passenger B which are located at different starting points, and as shown in fig. 1, the method comprises the following steps:
step 1, determining the detour proportion and the common road section proportion of the passenger A and the passenger B in the process that the passenger A and the passenger B reach the respective destinations in the carpooling state.
1.1 bypass ratio of passenger A is noted as θaThat is, the detour distance ratio of the passenger a, the calculation method is the distance that the passenger a has to walk after detour/the original total distance of the passenger a, and is expressed as:
Figure BDA0002269035280000071
1.2 proportion of the common stretch of the passenger A ηaThat is, the proportion of the common road segment of the passenger a and the passenger B to the original road segment of the passenger a is expressed as follows:
Figure BDA0002269035280000081
1.3 bypass ratio of passenger B is marked as thetabThat is, the detour distance ratio of the passenger B, the calculation method is the distance that the passenger B has to walk after detour/the original total distance of the passenger B, and is expressed as:
Figure BDA0002269035280000082
1.4 proportion of the common stretch of the passenger B ηbThe proportion of the common road section of the passenger A and the passenger B in the original road section of the passenger B is expressed as follows:
Figure BDA0002269035280000083
the above four proportional variables determine the passenger's payment price, where θa、θbCalled degree of detour, ηa、ηbThe charging rate of the scheme is divided into a detour part, a common road section part and the rest part for charging.
And 2, establishing a charging fee calculation model for the passengers A and B during ride-together according to the detour proportion and the common road section proportion of the passengers A and B and the related discount rates of the passengers A and B, wherein the related discount rates comprise the detour discount rate α of the passenger A, the detour discount rate β of the passenger B and the common road section discount rate η of the passengers A and B.
Wherein, the charging fee calculation model is as follows:
Figure BDA0002269035280000084
Figure BDA0002269035280000085
wherein, CPa、CPbRespectively representing calculation models of charging fees for determining passenger A and passenger B during ride-through, α being the discount rate of detour of passenger A, η being the discount rate of common road section of passenger A and passenger B, β being the discount rate of detour of passenger B, cost (L)a)、cost(Lb) Indicating passenger a to his destination DaFor one ride, passenger B to his destination DbThe individual cost (original price) of (c) is expressed as:
Figure BDA0002269035280000091
in the above formula, a, b and c represent constants, and L represents mileage; wherein a is the starting price of the bicycle, c is the kilometers included in the starting price, and b is the cost per kilometer after exceeding c kilometers. The values of the several constants can be set according to the actual conditions of different regions, for example, in a certain city G, the starting price a is 12 yuan, the number c of kilometers is 3 kilometers, and the cost b of each kilometer after exceeding 3 kilometers is 3 yuan.
Respectively mixing La、LbThe cost (L) can be obtained by substituting the parameter L into the above formulaa)、cost(Lb)。
In CPa、CPbThe method comprises the following steps:
the discounted object of the discount rate α is θ in the road sectionaThe part, i.e. the passenger A detour part, and the discounted object of the discount rate β is theta in the road sectionbThe section, i.e. the passenger B detour section, and the discounted object of discount rate η is η in the road sectiona、ηbAnd a section, i.e., a section of a road segment common to two passengers. Wherein, thetaa、θb、ηa、ηbAll decimal fractions less than 1:
when ηi+θiWhen the price is more than 1, the passenger fee is (the original price theta part is the detour discount rate + the original price η part is the common road section discount rate) × 1/(η)i+θi)。i＝a,b。
When ηi+θiWhen the passenger fee is less than or equal to 1, the passenger fee is equal to the original price theta part, the detour discount rate + the original price η part, the common road section discount rate + the original price (1- η)i-θi) And (4) partial.
Wherein, the coefficient addition of the discount rate is ensured to be equal to 1 in the process, namely, the charging is ensured to be carried out according to the original price after the division and then the discount is carried out.
ηi+θi1/(theta) of > 1 parti+ηi) To make the discount rate coefficient sum equal to 1, the same scale reduction is performed without changing the scale degree of the original road section ηi+θi1 or less of (1-theta)i-ηi) The remaining portion, except for the detours and common road segments, is charged, and the discount rate is 1 since this portion has no significant loss to passengers.
And 3, establishing an objective function based on the charging fee calculation model, wherein the objective function comprises an objective function respectively representing the reduced fee proportion of the passenger A and the passenger B after the ride combination and an objective function of the income of a driver after the ride combination.
The objective function is the objective of the solution, which is to expect the most passenger savings and the more earned drivers, and certainly, the two objectives are very small and may reach the maximum at the same time, but it is desirable to make them as large as possible. In this scheme, the established objective function includes:
Figure BDA0002269035280000092
Figure BDA0002269035280000101
the two above objective functions MaxZa、MaxZbThe proportion of the cost reduced by the passenger A and the passenger B after the ride combination is shown, namely, a relative numerical method is adopted, so that the cost saving of the passengers with different prices is comparable; in the scheme, the values of the two objective functions should be as large as possible, and passengers can obtain more favorable prices. In addition to passengers, there is a need to establish driver profits after a ride combinationTarget function MaxZd：
Figure BDA0002269035280000102
The objective function of the above equation is the proportion of increase in driver revenue, i.e., the increase in the co-ride versus mono-ride revenue, CP, for the same distance traveleda+CPbCost (L) is the income of the driver when the passenger A and the passenger B jointly take the rideab+Lx+Lr) The income of the driver when the passenger A and the passenger B take alone. This goal indirectly increases the unit gain of the driver. Therefore, in the present scheme, the objective function MaxZdThe value of (A) should be as large as possible, so that the driver has better profit, thereby promoting the enthusiasm of the driver.
And 4, setting a fitness function of the particle swarm algorithm according to the target function.
Three objective functions MaxZ established from step 3a、MaxZbAnd MaxZdAs can be seen, MaxZa+MaxZbOverall and MaxZdIn a negative correlation relationship. As the driver's revenue increases, the cost of the passenger increases, MaxZa+MaxZbThe value of (c) becomes small and even negative. Vice versa, the objective function MaxZ is used when the money saved by the passenger increasesdThe value of (c) is decreased and may be negative. The case where both objective functions are negative at the same time does not exist.
In the scheme, the form of the fitness function is set as the absolute value of the difference between half of the first two objective functions and the last objective function, namely:
Figure BDA0002269035280000103
wherein:
Figure BDA0002269035280000104
in the above formula, x is a, b, d. Wherein, the globalxFor a corresponding objective function MaxZxGlobal optimum of (3), globalaAs an objective function MaxZaGlobal optimum of (3), globalbAs an objective function MaxZbGlobal optimum of (3), globaldAs an objective function MaxZdThe global optimum value of (c). When the value of some objective function is greater than its corresponding global optimum, f (MaxZ)x) Is equal to 1, otherwise equal to positive infinity, i.e. fitness is positive infinity, an optimal data update cannot be performed.
Figure BDA0002269035280000111
Partially representing the degree of the proportion of the cost saved by passenger A, passenger B, MaxZdAnd (3) representing the profit increment of the driver, taking the difference between the profit increment and the profit increment, and then taking the absolute value to obtain the fitness function of the particle swarm algorithm, and solving the minimum fitness value to obtain the closest value of the reduction of the passenger cost and the increase of the driver benefit, namely achieving the goal of balanced driver and passenger benefits.
And 5, calculating reference discount rates to determine a value range of each relevant discount rate, wherein the value range comprises a detour discount rate α of the passenger A, a detour discount rate β of the passenger B and a value range of a common road section discount rate η of the passenger A and the passenger B.
The discount rate and the detour degree have close relation, the value of the discount rate is limited according to the detour proportion or the common road section proportion, when the discount rate is larger than a certain value, the discount rate is not increased any more, and the discount rate is assigned to the following limit value, which is called as the reference discount rate in the scheme:
α0＝1-kaθa
β0＝1-kbθb
Figure BDA0002269035280000112
α therein0、β0、η0To reference discount rate, thetaa、θbBypass ratios for passenger A and passenger B, ηa、ηbProportion of common road section for passenger A and passenger BSee step 1 for the calculation formula. For the proportional parameter k in the above formulaa,kbThe value of 0.5, kηThe value of the ratio parameter is 0.3, and the values of the ratio parameters are as follows:
in the scheme, the common road section degree or the detour degree is used as a rough discount rate, and the idea is as follows: what percentage of time is consumed by the passenger (journey is used in the scheme to indicate time because under good vehicle conditions, journey is positively correlated with time, and the time ratio, i.e. journey ratio), should be compensated for the same percentage of price. Since detour is in the interest of the driver, the detour proportion is taken as 50% as a reference, the fixed detour proportion is multiplied by 0.5, the discount rate of the common road section is not high according to the rule, 30% is taken as a reference, the fixed detour proportion is multiplied by 0.3, and k is obtaineda,kb,kηThe values of (a) are as follows:
0＜α＜α0
0＜β＜β0
0＜η＜η0
wherein, α0、β0、η0For reference, α is the detour discount rate of passenger a, β is the detour discount rate of passenger B, and η is the common link discount rate of passenger a and passenger B.
And 6, solving each relevant discount rate (the detour discount rate of the passenger A, the detour discount rate of the passenger B and the common road section discount rate of the passenger A and the passenger B) by using a particle swarm algorithm, outputting an optimal value of each relevant discount rate, and determining the ride-by price of the passenger A and the ride-by price of the passenger B by using the calculation model in the step 2 according to the optimal values. As shown in fig. 6, the particle swarm algorithm is as follows:
6.1 setting inertia weight and learning factor in the particle swarm optimization.
6.2, the iteration times and the population number of the particle swarm are set, and the car sharing state is input.
6.3 initializing the population, and randomly generating the discount rate of the corresponding individual within the discount rate range.
And 6.4, performing iterative updating, searching an optimal value, and performing iterative updating on the global optimal value and the individual optimal value according to the fitness function.
6.5 get the optimal discount rate.
Embodiments of the present solution are further described below by specific examples.
For the ride share model of fig. 2, the ride share state (8,6,3,6,2) is instantiated, passenger a detours, and passenger B detours as well. Discount rate optimized by algorithm:
α＝0.891316907800047，η＝0.8583333333333334，β＝0.7238621352236142。
CPa＝24.08319456382516，CPb＝17.15610918144489。
passenger a independent ride price: 27, passenger B ride alone: 21.
the driver's income increases: 5.23930374527005.
the proportion of increase in the driver's income: 0.14553621514639037.
price of one ride Ride-sharing price
Passenger A 27 24.08319456382516
Passenger B 21 17.15610918144489
The process of variation of fitness in the optimization process is shown in fig. 7.
The method covers most of double-person carpooling conditions, particle swarm optimization is carried out on the randomly generated discount rate by adopting a particle swarm multi-target optimization model according to the conditions of detours and common road sections of the carpooling state, the overall optimal discount value is finally obtained, the maximum value of driver benefits and passenger cost saving is optimized to the greater extent, and the purpose of balancing driver benefits and riding benefits is achieved.

Claims (8)

1. A multi-objective optimized shared network car-booking and car-sharing pricing method is used for determining the charging cost of a car-sharing scheme for a passenger A and a passenger B at different starting points, and comprises the following steps:
determining the detour proportion and the common road section proportion of the passenger A and the passenger B in the process that the passenger A and the passenger B reach the respective destinations in the carpooling state;
according to the detour proportion and the common road proportion of the passenger A and the passenger B and the related discount rate of the passenger A and the passenger B, a charging fee calculation model of the passenger A and the passenger B during ride-together is established; the relevant discount rates comprise a detour discount rate of the passenger A, a detour discount rate of the passenger B and a common road section discount rate of the passenger A and the passenger B;
based on the charging expense calculation model, establishing an objective function, wherein the objective function comprises an objective function respectively representing the reduced expense proportion of the passenger A and the passenger B after the ride combination and an objective function of the income of a driver after the ride combination;
setting a fitness function of the particle swarm algorithm according to the target function;
calculating a reference discount rate to determine a value range of each relevant discount rate;
and solving each relevant discount rate by using a particle swarm algorithm, solving an obtained optimal value according to the particle swarm algorithm, and determining the ride-share price of the passenger A and the ride-share price of the passenger B by using the charging fee calculation model.
2. The multi-objective optimized shared network car-booking and carpooling pricing method according to claim 1, wherein in the process that the passenger A and the passenger B reach the respective destinations in the carpooling state, the detour proportion and the common road section proportion of the passenger A and the passenger B comprise:
detour proportion of passenger A
Figure FDA0002269035270000011
Proportion of common route section of passenger A
Figure FDA0002269035270000012
Detour proportion of passenger B
Figure FDA0002269035270000013
Proportion of common road section with passenger B
Figure FDA0002269035270000014
Wherein L isaIndicating the arrival of passenger a at destination D from the starting pointaDistance of (L)bIndicating the arrival of passenger B at destination D from the starting pointbDistance of (L)xIndicating arrival at passenger a at passenger D after passenger BbThe distance of (d); l isabRepresents the shortest distance from passenger a to passenger B; l isrThe distance that the passenger B still walks after the passenger A gets off the vehicle is shown; LDaIndicating the distance, LD, traveled by passenger A after the detourb＝Lx+Lr-Lb；LDbIndicating the distance, LD, traveled by passenger B after the detourb＝Lx+Lr-Lb。
3. The multi-objective optimized shared network car-booking and ride-sharing pricing method according to claim 1, wherein the model for calculating the charging fee of the passenger A and the passenger B during ride-sharing comprises:
Figure FDA0002269035270000021
Figure FDA0002269035270000022
wherein the relevant discount rates comprise a detour discount rate α of the passenger A, a detour discount rate β of the passenger B, a common road section discount rate η of the passenger A and the passenger B, and cost (L)a)、cost(Lb) Indicating passenger a to his destination DaFor one ride, passenger B to his destination DbThe unique cost of.
4. The multi-objective optimized shared network car-appointment and ride-share pricing method according to claim 1, wherein the passenger A gets to the destination D thereofaFor one ride, passenger B to his destination DbThe calculation method of the unique cost comprises the following steps:
Figure FDA0002269035270000023
wherein a, b and c represent constants, and L represents mileage; wherein a is the starting price of the bicycle, c is the kilometers included in the starting price, and b is the cost per kilometer after exceeding c kilometers; respectively mixing La、LbThe cost (L) can be obtained by substituting the parameter L into the above formulaa)、cost(Lb)。
5. The multi-objective optimized share-network car-appointment and car-pool pricing method according to claim 1, wherein establishing an objective function based on the charging fee calculation model, the objective function comprising an objective function representing the reduced fee proportions of a passenger A and a passenger B after car pool and an objective function of a driver's income after car pool comprises:
the objective function representing the proportion of the reduced cost of passenger a after a ride pool is:
Figure FDA0002269035270000024
the objective function representing the proportion of the cost of passenger B reduction after a ride pool is:
Figure FDA0002269035270000025
the objective function of the driver's income after the ride combination is:
Figure FDA0002269035270000026
6. the multi-objective optimized shared network car-reduction car-ride-combination pricing method according to claim 1, wherein a fitness function of the particle swarm optimization is expressed as:
Figure FDA0002269035270000031
wherein:
Figure FDA0002269035270000032
in the above formula, globalxFor a corresponding objective function MaxZxX is a, b, d.
7. The multi-objective optimized shared network car-booking and car-ride-sharing pricing method according to claim 1, wherein the calculating of the reference discount rate comprises:
α0＝1-kaθa
β0＝1-kbθb
Figure FDA0002269035270000033
α therein0、β0、η0To reference discount rate, thetaa、θbBypass ratios for passenger A and passenger B, ηa、ηbThe proportions of the common road sections of the passenger A and the passenger B respectively; k is a radical ofa,kb,kηIs a proportional parameter;
the value range of each relevant discount rate is represented as:
0＜α＜α0
0＜β＜β0
0＜η＜η0
8. the multi-objective optimized shared network car-booking and car-sharing pricing method according to claim 7, wherein k is the ka,kb,kηThe values of (A) are as follows: k is a radical ofa,kbThe value of 0.5, kηThe value is 0.3.
CN201911098257.4A 2019-11-12 2019-11-12 Multi-objective optimized shared network car-booking and car-sharing pricing method Pending CN111160948A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201911098257.4A CN111160948A (en) 2019-11-12 2019-11-12 Multi-objective optimized shared network car-booking and car-sharing pricing method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201911098257.4A CN111160948A (en) 2019-11-12 2019-11-12 Multi-objective optimized shared network car-booking and car-sharing pricing method

Publications (1)

Publication Number Publication Date
CN111160948A true CN111160948A (en) 2020-05-15

Family

ID=70555946

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201911098257.4A Pending CN111160948A (en) 2019-11-12 2019-11-12 Multi-objective optimized shared network car-booking and car-sharing pricing method

Country Status (1)

Country Link
CN (1) CN111160948A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114239891A (en) * 2021-12-20 2022-03-25 首约科技(北京)有限公司 Method for reducing destination modification of user

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114239891A (en) * 2021-12-20 2022-03-25 首约科技(北京)有限公司 Method for reducing destination modification of user

Similar Documents

Publication Publication Date Title
Ma et al. The morning commute problem with ridesharing and dynamic parking charges
Li et al. A path-based equilibrium model for ridesharing matching
CN110458332B (en) Electric vehicle rapid charging demand scheduling method based on load space transfer
CN107564269B (en) A kind of half flexible bus dispatching method based on willingness to pay
CN107092976B (en) Method for cooperatively optimizing departure intervals of multiple bus routes by multi-objective model
Sayarshad et al. Non-myopic dynamic routing of electric taxis with battery swapping stations
CN108922169B (en) Differential charging strategy method based on equilibrium theory under parking sharing environment
Ulusoy et al. Optimal all-stop, short-turn, and express transit services under heterogeneous demand
Li et al. A dynamic pricing method for carpooling service based on coalitional game analysis
CN111667086A (en) Vehicle co-riding path optimizing method and system
CN108806249B (en) Passenger trip optimization method based on bus APP software
CN109308574A (en) A kind of flexible bus dispatching method in internet of real-time response half
CN106682759B (en) Battery supply system for electric taxi and network optimization method
Jiao et al. Incentivizing shared rides in e-hailing markets: Dynamic discounting
CN105006026B (en) A kind of taxi expense allocation Talmud methods
Abualola et al. A matching game-based crowdsourcing framework for last-mile delivery: Ground-vehicles and unmanned-aerial vehicles
Li et al. Stochastic ridesharing equilibrium problem with compensation optimization
CN114936666A (en) Electric vehicle charging navigation method and system based on vehicle-station-platform system
CN111160948A (en) Multi-objective optimized shared network car-booking and car-sharing pricing method
de Palma et al. Ride-sharing with inflexible drivers in the Paris metropolitan area
CN108898437A (en) Collaboration share-car cost sharing method based on Dynamic Uncertain demand under a kind of car networking environment
Li et al. Collaborative matching of ridesplitting and connection in the ridesourcing market
CN111127076A (en) Pricing method in taxi sharing mode
CN116468219A (en) Method for matching taxi sharing schedule by junction station
CN115169669A (en) Taxi sharing method based on track big data support

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
RJ01 Rejection of invention patent application after publication

Application publication date: 20200515

RJ01 Rejection of invention patent application after publication