CN107944611A - Towards the cross-module formula combined travel planing method of personalized trip requirements - Google Patents
Towards the cross-module formula combined travel planing method of personalized trip requirements Download PDFInfo
- Publication number
- CN107944611A CN107944611A CN201711142887.8A CN201711142887A CN107944611A CN 107944611 A CN107944611 A CN 107944611A CN 201711142887 A CN201711142887 A CN 201711142887A CN 107944611 A CN107944611 A CN 107944611A
- Authority
- CN
- China
- Prior art keywords
- travel
- data
- time unit
- time
- trip
- 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.)
- Granted
Links
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
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
- G06Q10/047—Optimisation of routes or paths, e.g. travelling salesman problem
Landscapes
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Engineering & Computer Science (AREA)
- Strategic Management (AREA)
- Economics (AREA)
- Marketing (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Development Economics (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Navigation (AREA)
Abstract
Towards the cross-module formula combined travel planing method of personalized trip requirements, belong to field of intelligent traffic information service.The trip scheme that the prior art provides is relatively fixed, can not according to user to time, economic dispatch different demands the characteristics of the personalized scheme of generation.The trip process of existing trip mode is carried out blocking decomposition by the present invention first, some measuring and calculating modes are introduced in haploidization and obtain accurately trip cell data.To conceptive setting node of the Regional Road Network in supernet, section, the comprehensive network network layers that different modes of transportation are formed are built, node ID coding and the time data for unit of going on a journey are matched, generate path operator.The data such as the level of comfort of various trip modes, cost, distance time are established as Adaptability Evaluation standard by this method from the demand of user.According to path operator, Adaptability Evaluation standard is demarcated, is solved using genetic algorithm, finally providing to the user has the characteristics that economic, time saving, comfortable, low-carbon trip scheme.
Description
Technical Field
The invention relates to a cross-mode combined travel planning method combined with user travel characteristics, and belongs to the field of intelligent traffic information services.
Background
The travel unit decomposition is to divide the data into different travel time units by combining the characteristics of different travel modes, divide the existing travel data into four types of historical data, dynamic data, static attribute data and personalized travel data, respectively combine with different travel time units to obtain the attributes of the travel units, and carry out relevant coding. The method comprises the steps of carrying out network building on an actual road network based on a super network model, carrying out path planning and solving by adopting a genetic algorithm, combining genetic operators of paths and time unit data, setting a cross-mode combination restrictive condition, carrying out solving on the genetic algorithm, and finally providing an optimal short-time, economic, low-carbon and comfortable travel scheme according to the demand characteristics of users on traffic modes. The existing patent technologies in China mostly concentrate on a navigation positioning module, a map browser, an interface module, a local map database, a map engine, a technical module for processing network access points, establishing or disconnecting a network, transmitting data, caching and managing local data and converting, the provided travel scheme is relatively fixed, a personalized scheme cannot be generated according to the characteristics of users on time, economy and different requirements, and only comprehensive travel planning services of ground buses, subways and walking can be provided. The method starts from user requirements, establishes a genetic operator coding mode based on a super network model, balances the individual travel requirements of the user in the aspects of time, economy, comfort and carbon emission, and realizes more comprehensive travel planning services of public transportation.
Disclosure of Invention
The method is based on the concept of the super network, and the super network formed by road networks of different traffic modes is built. The method comprises the steps of dividing time units of the existing trip mode, marking trip mode unit codes, combining the time units of different trip modes with history data, dynamic data, static attribute data and personalized trip data respectively, generating a path operator, calibrating an adaptability evaluation standard, solving by adopting a genetic algorithm, and finally providing different trip schemes which are time-saving, economical, comfortable and low-carbon.
In order to achieve the purpose, the method adopts the following technical processes:
step 1, trip unitized decomposition;
the travel unit decomposition is to divide the travel time units into different travel time units by combining the characteristics of different travel modes; firstly, respectively marking 7 different transportation modes of ground buses, subways, network rental cars, airport buses, business buses, riding and walking as-1 to-7, wherein the marking is called as an outgoing mode code; then according to the travel characteristics of 7 different travel modes, each travel process is sequentially decomposed into one or more of a waiting time unit, a taking time unit, a transfer time unit, an arrival time unit, an departure time unit, a calling time unit, a car searching time unit, a riding time unit and a walking time unit according to the characteristics of the travel process;
step 2, acquiring traffic data;
analyzing corresponding historical data, dynamic data, static data and personalized outgoing data aiming at time units of different outgoing modes, and acquiring outgoing characteristic values of different outgoing units;
step 3, establishing a trip network layer;
building network layers with different travel modes for the existing road network based on the concept of a super network; the combination of the nodes of the comprehensive network layer is completed through a two-dimensional coordinate system according to the following steps, and the simplification of the coordinates of the nodes is completed:
(1) Combining all the network layers to form an integrated layer, wherein all the network layer nodes are embodied in different colors in the integrated layer;
(2) Establishing a two-dimensional plane rectangular coordinate system in the integration layer, taking the east-west direction as the positive direction of an X axis, taking the north-south direction as the positive direction of a Y axis, coinciding the Y axis with the westest node in the integration layer, coinciding the X axis with the soutest node in the integration layer
(3) The unit lengths of the X axis and the Y axis of the rectangular coordinate system of the two-dimensional plane are meters; numbering longitudinal axes of all nodes in sequence from 0 in the sequence of positive half axes of an X axis, and numbering transverse axes of all nodes in sequence from 0 in the sequence of positive half axes of a Y axis;
(4) Measuring and calculating the distance of all road network nodes on a coordinate axis, combining right-side points with left-side points to delete the horizontal and vertical coordinate axes of the right-side points, and sequentially subtracting 1 from the upper-side coordinates of the X-axis and the Y-axis;
(5) Repeating the step (4) until all points do not have the points with similar distances, and completing simplification of the coordinate axes of the road network; taking the number of the maximum digits in the final road network coordinate axis as a reference, and automatically supplementing zero before when the number of the rest coordinates is insufficient; the serial numbers of the X axis and the Y axis where all the nodes are located are the abscissa and the ordinate of the point, and the coordinate is called as a node ID; finally, matching the coordinates of the two nodes of the upper integration layer with the coordinates of the two nodes of the upper integration layer according to the codes of the travel mode to form codes of different network layer paths;
carrying out association of data and node chromosome codes by using the node ID codes, and generating a path decision of a path operator applied to a genetic algorithm;
step 4, solving by adopting a genetic algorithm;
1) Determining an initial population
Setting travel starting and ending points and genetic algorithm parameters; the ID of the starting point node is 0000, the ID of the end point node is 9999, a threshold value is set for the path, and when a certain index of the path exceeds the threshold value, the path is directly discarded; the specific threshold values are: the maximum distance of the walking path is less than 2000 meters, the maximum distance of the riding path is less than 5000 meters, the minimum distance of the network car rental path is greater than 2000 meters, and the total number of times of transfer of the path is less than 5 times; setting evaluation standards as a total price for completing travel, total travel time, riding comfort and total carbon emission for travel; under each single standard, namely the standard with the minimum price, the shortest route time, the highest comfort degree or the lowest carbon emission, more than 20 initial paths with the optimal single standard are generated from the starting point to the end point of the trip;
2) Cross mutation by genetic algorithm
Setting the inter-species variation probability of 0.2, the intra-species variation probability of 0.15, the inter-species cross probability of 0.35 and the intra-species cross probability of 0.3 in the genetic algorithm parameters; when the operation occurs, the individuals in the population are operated through the four operators to obtain new individuals, the old individuals are reserved, and the operation is repeated until all the old individuals are operated;
3) Standardizing adaptability evaluation indexes;
converting data of each station in a super road network into path segment data, wherein the method comprises the following steps:
(1) A time data calculation method; judging a starting point and a finishing point and a used vehicle through chromosome coding, and obtaining the distance between the starting point and the finishing point through the starting point and finishing point coding; the network car booking passing distance is divided by the average speed of the road section, the average speed is dynamic data, and the current road condition congestion degree and the road section control factors are comprehensively obtained; the public bicycle renting time can be obtained by dividing the passing distance by the average riding speed of the user; the ground bus time is obtained by predicting historical bus arrival time data, and how to set and dynamically update the arrival time every 5 minutes according to real-time GPS data and road condition information of the bus; the subway time can be obtained by inquiring a subway arrival schedule and a subway in-and-out station transfer schedule;
(2) A price data calculation method; judging a starting point and a destination and a vehicle through chromosome coding, and inquiring a corresponding price calculation table of public transport to obtain the public transport; the bicycle renting method comprises the following steps that network renting bicycles and public renting bicycles are calculated through an integrated pricing rule;
(3) A comfort level data calculation method; calculating after determining a starting point, a finishing point and a vehicle; firstly, carrying out pre-investigation and calculation to obtain comfort degree calculation formulas corresponding to different vehicles per kilometer, and then calculating the comfort degree of a corresponding path according to the distance;
(4) A carbon emission data calculation method; obtaining carbon emission of different vehicles per kilometer by the previous integrated data, and obtaining the carbon emission by multiplying the carbon emission by the corresponding distance;
after obtaining each item of path data, standardizing the path data, and calculating the path data according to the following method:
(1) The method of standardizing travel time, carbon emissions and price data is as follows:
μ 1 (x i )＝α 1 (x i )/α 1 (max)
μ 2 (x i )＝α 2 (x i )/α 2 (max)
μ 3 (x i )＝α 3 (x i )/α 3 (max)
wherein x is i Is an evaluated road section; mu.s 1 (x i ) For normalized travel time evaluation value, μ 2 (x i ) Mu for normalized evaluation value of carbon emission 3 (x i ) The normalized price evaluation value is obtained; alpha is alpha 1 (x i ) As initial evaluation value of travel time, α 2 (x i ) As an initial evaluation value of carbon emission, α 3 (x i ) An initial evaluation value is set for the price data; alpha is alpha 1 (max) is the maximum value in the travel time evaluation index, α 2 (max) is the maximum value among the carbon emission evaluation indices, α 3 (max) is the maximum value in the price evaluation index;
(2) The comfort data is normalized as follows:
β(x i )＝α 4 (x i )/α 4 (max)
μ 4 (x i )＝1-β(x i )
wherein mu 4 (x i ) For normalized comfort evaluation value, α 4 (x i ) As an initial value of the travel comfort evaluation value, α 4 (max) is the maximum value in the comfort level evaluation index; beta (x) i ) Is a comfort index; because the percentage definition represented by the comfort index number value is opposite to the other three definitions, the comfort data and the other three data are unified under the same standard by adopting the method;
4) Evaluating adaptability and judging evolution algebra;
firstly, deleting individuals of which one or more evaluation criteria exceed a threshold value in the obtained population to obtain a new population; because the obtained individuals in the population have indexes of four dimensions, and the comparison result of the multidimensional vectors is a Pareto optimization solution set; after adaptive evaluation and algorithm, the selected paths form a new population so as to carry out the next round of calculation; stopping when the iteration times reach more than 2000 times, and entering the next step;
step 5, decoding the final population;
screening out the individuals with the respectively optimal four indexes in the final population, and decoding the individuals; firstly, determining a travel mode through a single mode marking area, respectively decoding each chromosome in the marking area, wherein each code should represent a road section, and the road section in each marking area comprises road section length, road section crowding degree, road section travel price, comfort level and carbon emission information of the corresponding travel mode; finally, integrating the single mode marking areas to obtain the final optimal travel scheme.
Drawings
FIG. 1 block coding diagram of trip mode
FIG. 2 is a table of cell attribute acquisition modes;
FIG. 3 schematic diagram of the metro layer of the super network
FIG. 4 is a schematic diagram of a public transportation layer of a super network
FIG. 5 is a schematic diagram of a super network car rental layer
FIG. 6 is a schematic diagram of a super network step and ride layer
FIG. 7 is a diagram of a super network framework;
FIG. 8 super network node ID code graph
FIG. 9 software flow diagram
FIG. 10 is a generated path of the Oncology Hospital of Chinese medical sciences, beijing university of industry
FIG. 11 is an in-pattern crossover operator
FIG. 12 is an inter-mode crossover operator
FIG. 13 is an intra-pattern mutation operator
FIG. 14 is a diagram of inter-mode mutation operators
Detailed Description
Through test experiments on peripheral areas of Beijing industry university, a research area from east-west two rings to east four rings and from south-north great-expectation road to SongYunan road in Beijing city is selected, a super network in the area range is constructed, traffic data of the area are obtained, the super network is combined with ground buses, subways, network taxi booking, public rental bicycles and walking travel units, a travel scheme is solved by adopting a genetic algorithm, different travel schemes with time saving, economy, comfort and low carbon are provided, and personalized prompts are provided.
This embodiment comprises the steps of:
step 1, performing modular decomposition on a trip;
the travel unit decomposition is to divide the travel time units into different travel time units by combining the characteristics of different travel modes. Firstly, 7 different transportation modes of ground public transport, subway, network taxi, airport bus, business class bus, riding and walking are respectively marked as-1 to-7, and the marks are called as trip mode codes. Then according to the travel characteristics of 7 different travel modes, each travel process is sequentially decomposed into a waiting time unit, a taking time unit, a transfer time unit, an arrival time unit, an departure time unit, a calling time unit, a searching time unit, a riding time unit and a walking time unit according to the characteristics.
Step 2, acquiring traffic data;
and analyzing corresponding historical data, dynamic data, static data and individualized outgoing data aiming at the time units with different outgoing modes, and acquiring the outgoing characteristic values of different outgoing units. The characteristic values comprise real-time driving speed data of buses, subways, network car renters, airport buses and business buses, subway transfer, arrival time data and departure time data, a bus historical arrival schedule, individual speed data of riders and pedestrians, and network car renting and sharing single-vehicle distribution data. And calculating to obtain travel time corresponding to different travel units.
Wherein, when unitizing the public transit trip, for making data more accurate, this patent has designed following mechanism:
(1) An arrival time reliability evaluation mechanism; the bus arrival time mined by historical data is influenced by road conditions and shows the phenomenon that the arrival time is unreliable, so that the travel time predicted by the scheme is inaccurate; therefore, the daily arrival time and the historical arrival time value are analyzed in the following mode, a travel scheme is generated, meanwhile, the user is informed of the unreliable influence of the arrival time in a star evaluation mode, and the accuracy of the predicted travel time of the scheme is improved;
the buses arriving at the station A on a certain day are set to be N classes, and the arrival time of the station A on the same day is calculated by historical data Actual arrival time during peak time of day of acquisitionThen, the average relative error value STAB (t) is obtained and substituted with the departure interval DEPA (t) to obtain the uncertainty
Wherein: STAB (t)% DEPA (t) represents the remainder of dividing STAB (t) by DEPA (t), and the units of STAB (t) and DEPA (t) are minutes;
degree of unreliabilityThe lower the value of (A), the higher the evaluation is, the more accurate the trip time predicted by the scheme is; degree of unreliabilityIs a numerical value which is continuously iterated and fed back according to historical data;
(2) A dynamic update mechanism of bus transit time; the method is characterized in that the predicted travel time is inaccurate due to unreliable bus arrival time, the predicted time of arriving at a destination is updated every 5 minutes by combining real-time road condition data with a bus GPS and real-time transport speed, and whether a travel route is to be selected again or not can be selected according to travel requirements.
Step 3, establishing a super network;
selecting a research area from east-east two-ring to east-four-ring, from south-north wide-channel to SongYu south-channel in sunward areas in Beijing to construct a super network in the area range and network layers in different travel modes.
The combination of the nodes of the comprehensive network layer is completed through a two-dimensional coordinate system according to the following steps, and the simplification of the coordinates of the nodes is completed:
(1) Combining all the network layers to form an integrated layer, wherein all the network layer nodes are embodied in different colors in the integrated layer;
(2) Establishing a two-dimensional plane rectangular coordinate system in the integration layer, taking the east-west direction as the positive direction of an X axis, taking the north-south direction as the positive direction of a Y axis, coinciding the Y axis with the westest node in the integration layer, coinciding the X axis with the soutest node in the integration layer
(3) The unit lengths of the X axis and the Y axis of the two-dimensional plane rectangular coordinate system are meters; numbering longitudinal axes of all nodes in sequence from 0 in the sequence of positive half axes of an X axis, and numbering transverse axes of all nodes in sequence from 0 in the sequence of positive half axes of a Y axis;
(4) Measuring and calculating the distance of all road network nodes on a coordinate axis, combining right-side points with left-side points to delete the horizontal and vertical coordinate axes of the right-side points, and sequentially subtracting 1 from the upper-side coordinates of the X-axis and the Y-axis;
(5) Repeating the step (4) until all points do not have the points with similar distances, and completing simplification of the coordinate axes of the road network; taking the number of the maximum digits in the final road network coordinate axis as a reference, and automatically supplementing zero before when the number of the rest coordinates is insufficient; the serial numbers of the X axis and the Y axis where all the nodes are located are the abscissa and the ordinate of the point, and the coordinate is called as a node ID; finally, matching coordinates of two nodes of the upper integration layer with coordinates of the two nodes according to the codes of the travel mode to form codes of different network layer paths;
carrying out association of data and node chromosome codes by using the node ID codes, and generating a path decision of a path operator applied to a genetic algorithm;
step 4, solving by adopting a genetic algorithm;
1) Determining an initial population
Setting starting and ending points and genetic algorithm parameters. The ID of the starting node is 0000, the ID of the destination node is 9999, the maximum distance of a walking path is less than 2000 meters, the maximum distance of a riding path is less than 5000 meters, the minimum distance of a network car rental path is greater than 2000 meters, and the total transfer times of the path are less than 5 times; setting evaluation standards as a total price for completing travel, total travel time, riding comfort and total carbon emission for travel; under each single standard, namely the standard with the minimum price, the shortest travel time, the highest comfort level or the lowest carbon emission, more than 20 initial paths with the best single standard are generated from the starting point to the end point of the trip; meanwhile, the threshold values of the four standards are set to be that the total trip price is less than 60 yuan, the total trip time is less than 90 minutes, and the upper limit and the lower limit of the total carbon emission and the total trip comfort level are not set.
Example (c): FIG. 10 is a generated path of the Oncology Hospital of Chinese medical sciences, beijing university of industry
2) Cross mutation by genetic algorithm
Setting the inter-species variation probability of 0.2, the intra-species variation probability of 0.15, the inter-species cross probability of 0.35 and the intra-species cross probability of 0.3 in the genetic algorithm parameters; when the operation occurs, the individuals in the population are operated through the four operators to obtain new individuals, the old individuals are reserved, and the operation is repeated until all the old individuals are operated; in the patent, an operator in a mode does not introduce a new traffic mode on path planning, namely, the number of transfer times is controlled to be increased; the inter-mode operator can cause the increase of the transfer times through the operation of the operator; the crossover operator is that when the partial path code of a certain path is the same as the partial code of another path, the two paths start to exchange from the same position until the path is finished, and the two original path codes can be changed through crossover operation; the mutation operator generates a new local path code under the condition that the original path code is not changed, and replaces the part, which is in the original path code, with the same position as the starting and ending points of the local path code, so that each pair of path codes are operated by adopting the mutation operator, and the new path code is generated.
Note: p1 and P2 are paths before cross mutation, C1 and C2 are paths after cross mutation, marked as a start change point and an end change point.
FIG. 12 is an inter-mode crossover operator
FIG. 13 is a diagram of intra-mode mutation operators
FIG. 14 is a diagram of inter-mode mutation operators
3) Standardizing adaptability evaluation indexes;
the method for converting the data of each station in the super road network into the path segment data comprises the following steps:
(1) A time data calculation method; the starting and ending points and the used vehicles are judged through the chromosome codes, and the distance between the starting and ending points can be obtained through the starting and ending point codes. The average speed is dynamic data and is obtained by comprehensively calculating the congestion degree of the current road condition and the road section control factors. The public bicycle renting time is obtained by dividing the distance by the average riding speed of the user. The bus time is obtained by performing weight analysis and calculation on historical bus arrival data and comprehensively by the aid of a dynamic update mechanism of the bus transit time. The subway time can be obtained by inquiring a subway arrival schedule and a subway in-and-out transfer schedule.
(2) A price data calculation method; the starting and ending points and the transportation means are judged through the chromosome codes, and the public transportation is obtained through inquiring the corresponding price calculation table. The bicycle renting and bicycle renting are calculated by integrating the pricing rules.
(3) A comfort level data calculation method; the calculation is carried out after the starting and ending points and the vehicles are determined. Firstly, a comfort degree calculation formula corresponding to each kilometer of different transportation means is obtained through preliminary investigation and calculation, and then the comfort degree of a corresponding path is calculated according to the distance.
(4) A carbon emission data calculation method; the carbon emission of different vehicles per kilometer is obtained by the previous integration data, and the carbon emission is obtained by multiplying the corresponding distance.
After obtaining each item of path data, standardizing the path data, and calculating the path data according to the following method:
(1) The method of standardizing travel time, carbon emissions and price data is as follows:
μ 1 (x i )＝α 1 (x i )/α 1 (max)
μ 2 (x i )＝α 2 (x i )/α 2 (max)
μ 3 (x i )＝α 3 (x i )/α 3 (max)
wherein x is i Is the evaluated link. Mu.s 1 (x i ) For normalized travel time evaluation value, μ 2 (x i ) Mu for normalized evaluation value of carbon emission 3 (x i ) Is a normalized price evaluation value. Alpha is alpha 1 (x i ) As initial evaluation value of travel time, α 2 (x i ) As an initial evaluation value of carbon emission, α 3 (x i ) An initial rating value is the price data. Alpha (alpha) ("alpha") 1 (max) is the maximum value in the travel time evaluation index, α 2 (max) is the maximum value among the carbon emission evaluation indices, α 3 (max) is the maximum value of the price evaluation indices.
(2) The comfort data is normalized as follows:
β(x i )＝α 4 (x i )/α 4 (max)
μ 4 (x i )＝1-β(x i )
wherein mu 4 (x i ) For normalized comfort evaluation value, α 4 (x i ) As an initial value of the travel comfort evaluation value, α 4 (max) is the maximum value of the comfort level evaluation indices. Beta (x) i ) Is a comfort index. Number of indexes of comfort
The values are expressed in percentages defined as opposed to the remaining three, so the method is used to unify the comfort data with the remaining three to the same standard.
4) Evaluating adaptability and judging evolution algebra;
let the total number of path codes of the generation be G, wherein the total number of path codes that do not exist in the previous generation is G 1 The number of the individuals with any one of the four indexes failing to reach the standard in the path coding is M, if G is 1 > = M, all substandard path codes are removed, otherwise G is randomly selected 1 And deleting the unqualified codes.
As the obtained population has indexes of four dimensions in all individuals, and the comparison result of the multidimensional vectors is a Pareto optimization solution set. In the vector comparison, the concept of "dominance" is introduced for expression. That is, there are individuals A, B, if all four indices of A are less than B, then B dominates A.
f i (x 0 )≤f i (x 1 )i＝1，2，...，m
The ranking number of the individuals is equal to the number of the individuals dominating the individuals in the current population plus 1, and all non-inferior individual ranking values are 1 and serve as individual adaptation values.
The operator selection operation is based on the defined population size G and the current number of individuals G of the new population 1 Determining the number of selected superior individuals X = G-G 1 。G 1 The method for selecting excellent individuals comprises the following steps: counting the number M of non-repetitive individuals with an evaluation value of 1, and randomly selecting the non-repetitive individuals with the current evaluation value if X is less than or equal to MThe multiple individuals enter a next generation of population; if X&And M, the current non-repeated optimal individuals enter the next generation, and excellent individuals are determined to enter the next generation of population according to the binary tournament method until the selection number is reached.
And defining the eliminated new population as a child to replace the original population. If the maximum evolution generation number is 2000, if the fashion does not reach 2000 times, repeating the step 2. The number of new individuals generated from parent to child is defined as the number of replacements. The replacement number is the expression of the situation of replacing and eliminating individuals by the parent population. If the number of substitutions is 0, the offspring does not produce a more optimal individual. Definition, if the number of successive 20 substitutions is 0, the evolution is considered to be stable, and the evolution is stopped. The final offspring is defined as the final population.
Step 5, decoding the final population;
screening out the individuals with the respectively optimal four indexes in the final population, and decoding the individuals; firstly, determining a travel mode through a single mode marking area, respectively decoding each chromosome in the marking area, wherein each code should represent a road section, and the road section in each marking area comprises road section length, road section crowding degree, road section travel price, comfort level and carbon emission information of the corresponding travel mode; finally, integrating the single mode marking areas to obtain the final optimal travel scheme.
The method comprises the steps of obtaining time and cost values given by three modes of riding, public transportation and renting of existing travel software, and automatically calculating carbon emission values and comfort values of the three modes of riding, public transportation and renting of the travel software according to the characteristics of vehicles; and (3) taking time, cost, carbon emission and comfort values given by the travel mode combination scheme of the method. The travel time of the combined scheme is saved by 33%, the cost is saved by 80% on the premise of similar comfort level, and the carbon emission is reduced by 90%. Table 1 travel scheme characteristic attribute comparison table
Claims (7)
1. A cross-mode combined travel planning method for personalized travel demands is characterized by comprising the following steps: the method comprises the following steps of,
step 1, trip unitized decomposition;
the travel unit decomposition is to divide the travel time units into different travel time units by combining the characteristics of different travel modes; firstly, 7 different transportation modes of ground public transport, subway, network taxi booking, airport bus, business class bus, riding and walking are respectively marked as-1 to-7, and the marks are called as trip mode codes; then according to the travel characteristics of 7 different travel modes, each travel process is sequentially decomposed into one or more of a waiting time unit, a taking time unit, a transfer time unit, an inbound time unit, an outbound time unit, a calling time unit, a seeking time unit, a riding time unit and a walking time unit according to the characteristics;
step 2, acquiring traffic data;
analyzing corresponding historical data, dynamic data, static data and personalized outgoing data aiming at time units of different outgoing modes, and acquiring outgoing characteristic values of different outgoing units;
step 3, establishing a trip network layer;
building network layers with different travel modes for the existing road network based on the concept of a super network; the combination of the nodes of the comprehensive network layer is completed through a two-dimensional coordinate system according to the following steps, and the simplification of the coordinates of the nodes is completed:
(1) Combining all the network layers to form an integrated layer, wherein all the network layer nodes are embodied in different colors in the integrated layer;
(2) Establishing a two-dimensional plane rectangular coordinate system in the integration layer, taking the east-west direction as the positive direction of an X axis, taking the north-south direction as the positive direction of a Y axis, coinciding the Y axis with the westest node in the integration layer, coinciding the X axis with the soutest node in the integration layer
(3) The unit lengths of the X axis and the Y axis of the two-dimensional plane rectangular coordinate system are meters; numbering longitudinal axes of all nodes in sequence from 0 in the sequence of positive half axes of an X axis, and numbering transverse axes of all nodes in sequence from 0 in the sequence of positive half axes of a Y axis;
(4) Measuring and calculating the distance of all road network nodes on a coordinate axis, combining right-side points with left-side points to delete the horizontal and vertical coordinate axes of the right-side points, and sequentially subtracting 1 from the upper-side coordinates of the X-axis and the Y-axis;
(5) Repeating the step (4) until all the points do not have the points with similar distances, and completing simplification of the coordinate axes of the road network; taking the digit number of the maximum digit in the coordinate axis of the final road network as a reference, and automatically filling zero in the front when the digit numbers of the rest coordinates are insufficient; the serial numbers of the X axis and the Y axis where all the nodes are located are the abscissa and the ordinate of the point, and the coordinate is called as a node ID;
finally, matching coordinates of two nodes of the upper integration layer with coordinates of the two nodes according to the codes of the travel mode to form codes of different network layer paths;
carrying out association of data and node chromosome codes by using the node ID codes, and generating a path decision of a path operator applied to a genetic algorithm;
step 4, solving by adopting a genetic algorithm;
1) Determining an initial population
Setting travel starting and ending points and genetic algorithm parameters; the ID of the starting point node is 0000, the ID of the end point node is 9999, a threshold value is set for the path, and when a certain index of the path exceeds the threshold value, the threshold value is directly discarded; the specific threshold values are: the maximum distance of the walking path is less than 2000 meters, the maximum distance of the riding path is less than 5000 meters, the minimum distance of the network car rental path is more than 2000 meters, and the total transfer times of the path is less than 5 times; setting evaluation standards as a total price for completing travel, total travel time, riding comfort and total carbon emission for travel; under each single standard, namely the standard with the minimum price, the shortest journey time, the highest comfort level or the lowest carbon emission, more than 20 initial paths with the optimal single standard are generated from the starting point to the end point of the trip;
2) Cross mutation by genetic algorithm
Setting the inter-species variation probability of 0.2, the intra-species variation probability of 0.15, the inter-species cross probability of 0.35 and the intra-species cross probability of 0.3 in the genetic algorithm parameters; when the operation occurs, the individuals in the population are operated through the four operators to obtain new individuals, the old individuals are reserved, and the operation is repeated until all the old individuals are operated;
3) Standardizing adaptability evaluation indexes;
converting data of each station in a super road network into path segment data, wherein the method comprises the following steps:
(1) A time data calculation method; judging a start point and an end point and a used vehicle through the chromosome codes, and obtaining the distance between the start point and the end point through the start point and the end point codes; the network car booking is obtained by dividing the passing distance by the average speed of the road section, the average speed is dynamic data, and the current road condition congestion degree and the road section control factors are comprehensively obtained; the public bicycle renting time is obtained by dividing the distance by the average riding speed of the user; the ground bus time is obtained by estimating historical bus arrival time data, and how to set the arrival time which is dynamically updated every 5 minutes according to the real-time GPS data and road condition information of the bus; the subway time is obtained by inquiring a subway arrival schedule and a subway station entering and exiting transfer schedule;
(2) A price data calculation method; judging a starting point and a finishing point and a vehicle through chromosome coding, and inquiring a corresponding price calculation table by public transport to obtain; the bicycle renting method comprises the following steps that network renting bicycles and public renting bicycles are calculated through integrated pricing rules;
(3) A comfort level data calculation method; calculating after determining a starting point, a finishing point and a vehicle; firstly, carrying out pre-investigation and calculation to obtain comfort degree calculation formulas corresponding to different vehicles per kilometer, and then calculating the comfort degree of a corresponding path according to the distance;
(4) A carbon emission data calculation method; obtaining carbon emission of different vehicles per kilometer by the previous integrated data, and obtaining the carbon emission by multiplying the carbon emission by the corresponding distance;
after obtaining each item of path data, standardizing the path data, and calculating the path data according to the following method:
(1) The method for standardizing travel time, carbon emissions and price data is as follows:
μ 1 (x i )＝α 1 (x i )/α 1 (max)
μ 2 (x i )＝α 2 (x i )/α 2 (max)
μ 3 (x i )＝α 3 (x i )/α 3 (max)
wherein x is i Is an evaluated road section; mu.s 1 (x i ) For normalized travel time evaluation value, μ 2 (x i ) Mu for normalized evaluation value of carbon emission 3 (x i ) Is a normalized price evaluation value; alpha (alpha) ("alpha") 1 (x i ) As an initial evaluation value of the travel time, α 2 (x i ) As an initial evaluation value of carbon emission, α 3 (x i ) An initial evaluation value for the price data; alpha is alpha 1 (max) is the maximum value in the travel time evaluation index, α 2 (max) is the maximum value of the evaluation indexes for carbon emission, α 3 (max) is the maximum value in the price evaluation index;
(2) The comfort data is normalized as follows:
β(x i )＝α 4 (x i )/α 4 (max)
μ 4 (x i )＝1-β(x i )
wherein mu 4 (x i ) For normalized comfort evaluation value, α 4 (x i ) As an initial value of the travel comfort evaluation value, α 4 (max) is the maximum value in the comfort level evaluation index; beta (x) i ) Is a comfort index; because the percentage definition represented by the comfort level index value is opposite to the definitions of the rest three items, the comfort level data and the rest three items of data are unified to the same standard by adopting the method;
4) Evaluating adaptability and judging evolution algebra;
firstly, deleting individuals of which one or more evaluation criteria exceed a threshold value in the obtained population to obtain a new population; because the obtained individuals in the population have indexes of four dimensions, and the comparison result of the multidimensional vectors is a Pareto optimization solution set; after adaptive evaluation and algorithm, the selected paths form a new population so as to carry out the next round of calculation; stopping when the iteration times reach more than 2000 times, and entering the next step;
step 5, decoding the final population;
screening out the individuals with the respectively optimal four indexes in the final population, and decoding the individuals; firstly, determining a travel mode through a single mode marking area, respectively decoding each chromosome in the marking area, wherein each code should represent a road section, and the road section in each marking area comprises the road section length, the road section congestion degree, the road section travel price, the comfort level and the carbon emission information of the corresponding travel mode; finally, integrating the single mode marking areas to obtain the final travel scheme.
2. The cross-mode combined travel planning method for personalized travel demands according to claim 1, characterized in that the steps 2 are decomposed into a waiting time unit, a taking time unit, a transfer time unit, an inbound time unit, an outbound time unit, a calling time unit, a searching time unit, a riding time unit and a walking time unit according to their characteristics; the method specifically comprises the following steps: the ground bus comprises a waiting time unit and a taking time unit; the subway comprises an inbound time unit, a waiting time unit, a riding time unit, a transfer time unit and an outbound time unit; the network taxi booking comprises a taxi calling time unit and a taxi taking time unit; the airport bus comprises a waiting time unit and a riding time unit; the business class bus comprises a waiting time unit and a riding time unit; the riding comprises a vehicle searching time unit and a riding time unit; walking includes walking time units.
3. The method for planning a trans-mode combined trip facing an individualized trip demand according to claim 1, wherein the historical data of step 2 analyzes the corresponding historical data, dynamic data, static data and individualized trip data for time units of different trip modes, and obtains trip characteristic values of different trip units; the time unit corresponding to the historical data comprises: a subway waiting time unit and a riding time unit; a network taxi booking and calling time unit; the airport bus waiting time unit and the boarding time unit are arranged; a waiting time unit and a riding time unit of the commercial bus.
4. The cross-mode combined travel planning method for personalized travel demands according to claim 1, wherein the historical data in step 2 is used for analyzing corresponding historical data, dynamic data, static data and personalized travel data for time units of different travel modes to obtain travel characteristic values of different travel units; the time unit corresponding to the dynamic data comprises: a bus waiting time unit and a bus taking time unit; and the network taxi booking and calling time unit.
5. The cross-mode combined travel planning method for personalized travel demands according to claim 1, wherein the historical data in step 2 is used for analyzing corresponding historical data, dynamic data, static data and personalized travel data for time units of different travel modes to obtain travel characteristic values of different travel units; the time unit corresponding to the static data comprises: the system comprises a subway station entering time unit, a subway waiting time unit, a transfer time unit and a subway station exiting time unit; a riding time unit of a commercial regular bus; and a riding and vehicle searching unit.
6. The cross-mode combined travel planning method for personalized travel demands according to claim 1, wherein the historical data in step 2 is used for analyzing corresponding historical data, dynamic data, static data and personalized travel data for time units of different travel modes to obtain travel characteristic values of different travel units; the time unit corresponding to the personalized outgoing data comprises: the riding and vehicle-searching time unit and the riding time unit are arranged; a walking time unit; and the personalized outgoing data is obtained through GPS positioning.
7. The cross-mode combined travel planning method for personalized travel needs according to claim 1, characterized in that the travel times corresponding to different travel units are calculated in step 2; when unitizing a bus trip, the following mechanism is designed:
(1) An arrival time reliability evaluation mechanism; the bus arrival time mined by historical data is influenced by road conditions and shows the phenomenon that the arrival time is unreliable, so that the travel time predicted by the scheme is inaccurate; therefore, the daily arrival time and the historical arrival time value are analyzed in the following mode, a trip scheme is generated, meanwhile, the user is informed of the unreliable influence of the station time in a star evaluation mode, and the accuracy of the predicted trip time of the scheme is improved;
setting N classes of buses arriving at the station A on a certain day, and respectively calculating the arrival time of the station A on the same day asActual arrival time during peak time of day of acquisition Then, the average relative error value STAB (t) is obtained and substituted with the departure interval DEPA (t) to obtain the unreliability
Wherein: STAB (t)% DEPA (t) represents the remainder of dividing STAB (t) by DEPA (t), and the units of STAB (t) and DEPA (t) are minutes;
degree of unreliabilityThe lower the value of (A), the higher the evaluation is, the more accurate the trip time predicted by the scheme is; degree of unreliabilityIs a root ofContinuously iterating the fed back numerical value according to historical data;
(2) A dynamic update mechanism of bus transit time; and under the condition that the predicted travel time is inaccurate due to unreliable bus arrival time, updating the predicted travel time of arriving at the destination every 5 minutes by combining real-time road condition data with a bus GPS and real-time transport speed, and selecting whether to reselect a travel route according to travel requirements.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711142887.8A CN107944611B (en) | 2017-11-17 | 2017-11-17 | Cross-mode combined travel planning method for personalized travel demand |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711142887.8A CN107944611B (en) | 2017-11-17 | 2017-11-17 | Cross-mode combined travel planning method for personalized travel demand |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107944611A true CN107944611A (en) | 2018-04-20 |
| CN107944611B CN107944611B (en) | 2021-11-26 |
Family
ID=61931655
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711142887.8A Active CN107944611B (en) | 2017-11-17 | 2017-11-17 | Cross-mode combined travel planning method for personalized travel demand |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107944611B (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108801280A (en) * | 2018-05-03 | 2018-11-13 | 杨靖旭 | A kind of paths planning method and delivery system dispensed by the way for campus life |
| CN109034451A (en) * | 2018-06-15 | 2018-12-18 | 杭州后博科技有限公司 | A kind of combination traffic plan recommended method and system based on expense |
| CN109460937A (en) * | 2018-11-23 | 2019-03-12 | 东南大学 | Evaluation urban railway station periphery Slow transport system is plugged into horizontal process and method |
| CN109918567A (en) * | 2019-03-05 | 2019-06-21 | 百度在线网络技术（北京）有限公司 | Trip mode recommended method and device |
| CN110288849A (en) * | 2019-07-29 | 2019-09-27 | 电子科技大学 | A kind of traffic path recommended method based on mixed traffic mode |
| WO2020106211A1 (en) * | 2018-11-19 | 2020-05-28 | Grabtaxi Holdings Pte. Ltd. | Communications server apparatus, method and communications system for managing request for transport-related services |
| CN112434116A (en) * | 2020-09-09 | 2021-03-02 | 北京交通发展研究院 | Low-carbon trip carbon emission reduction verification method and system based on trip chain big data |
| CN113656746A (en) * | 2021-07-21 | 2021-11-16 | 东南大学 | Travel mode chain selection method considering group heterogeneity under dynamic structure |
| WO2023098095A1 (en) * | 2021-12-02 | 2023-06-08 | 北京嘀嘀无限科技发展有限公司 | Method and apparatus for managing combined travel |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101853294A (en) * | 2010-05-21 | 2010-10-06 | 中国科学院地理科学与资源研究所 | Multi-mode multi-standard path search method based on genetic algorithm |
| CN103324982A (en) * | 2013-06-07 | 2013-09-25 | 银江股份有限公司 | Path planning method based on genetic algorithm |
| US20170059341A1 (en) * | 2015-08-31 | 2017-03-02 | Sap Se | Diversified route planning for public transportation network |
-
2017
- 2017-11-17 CN CN201711142887.8A patent/CN107944611B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101853294A (en) * | 2010-05-21 | 2010-10-06 | 中国科学院地理科学与资源研究所 | Multi-mode multi-standard path search method based on genetic algorithm |
| CN103324982A (en) * | 2013-06-07 | 2013-09-25 | 银江股份有限公司 | Path planning method based on genetic algorithm |
| US20170059341A1 (en) * | 2015-08-31 | 2017-03-02 | Sap Se | Diversified route planning for public transportation network |
Non-Patent Citations (2)
| Title |
|---|
| 于海璁 等: "一种基于遗传算法的多模式多标准路径规划方法", 《测绘学报》 * |
| 赵婷 等: "旅客视角下基于时变的多模式交通网络出行路径", 《科学技术与工程》 * |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108801280A (en) * | 2018-05-03 | 2018-11-13 | 杨靖旭 | A kind of paths planning method and delivery system dispensed by the way for campus life |
| CN109034451A (en) * | 2018-06-15 | 2018-12-18 | 杭州后博科技有限公司 | A kind of combination traffic plan recommended method and system based on expense |
| WO2020106211A1 (en) * | 2018-11-19 | 2020-05-28 | Grabtaxi Holdings Pte. Ltd. | Communications server apparatus, method and communications system for managing request for transport-related services |
| US11733051B2 (en) | 2018-11-19 | 2023-08-22 | Grabtaxi Holdings Pte. Ltd. | Communications server apparatus, method and communications system for managing request for transport-related services |
| CN109460937A (en) * | 2018-11-23 | 2019-03-12 | 东南大学 | Evaluation urban railway station periphery Slow transport system is plugged into horizontal process and method |
| CN109460937B (en) * | 2018-11-23 | 2021-06-15 | 东南大学 | Process and method for evaluating connection level of slow traffic system around track station |
| CN109918567A (en) * | 2019-03-05 | 2019-06-21 | 百度在线网络技术（北京）有限公司 | Trip mode recommended method and device |
| CN110288849A (en) * | 2019-07-29 | 2019-09-27 | 电子科技大学 | A kind of traffic path recommended method based on mixed traffic mode |
| CN112434116A (en) * | 2020-09-09 | 2021-03-02 | 北京交通发展研究院 | Low-carbon trip carbon emission reduction verification method and system based on trip chain big data |
| CN113656746A (en) * | 2021-07-21 | 2021-11-16 | 东南大学 | Travel mode chain selection method considering group heterogeneity under dynamic structure |
| CN113656746B (en) * | 2021-07-21 | 2022-06-17 | 东南大学 | Travel mode chain selection method considering group heterogeneity under dynamic structure |
| WO2023098095A1 (en) * | 2021-12-02 | 2023-06-08 | 北京嘀嘀无限科技发展有限公司 | Method and apparatus for managing combined travel |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107944611B (en) | 2021-11-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107944611B (en) | Cross-mode combined travel planning method for personalized travel demand | |
| CN109409599B (en) | Customized bus line opening optimization method based on real-time requirements | |
| Chen et al. | Solving the first‐mile ridesharing problem using autonomous vehicles | |
| CN109612488B (en) | Big data micro-service-based mixed travel mode path planning system and method | |
| CN110942198A (en) | Passenger path identification method and system for rail transit operation | |
| CN110704993A (en) | Customized bus route design method for relieving subway passenger flow pressure | |
| CN115409388A (en) | Multi-vehicle type customized bus operation optimization method | |
| CN110222912B (en) | Railway travel route planning method and device based on time dependence model | |
| CN113327424B (en) | Traffic demand prediction method and device and electronic equipment | |
| Yan et al. | A car pooling model and solution method with stochastic vehicle travel times | |
| CN112561249A (en) | Real-time demand-oriented city customized bus scheduling method | |
| Wei et al. | Optimal integrated model for feeder transit route design and frequency‐setting problem with stop selection | |
| Shu et al. | A novel demand-responsive customized bus based on improved ant colony optimization and clustering algorithms | |
| CN110570656A (en) | Method and device for customizing public transport line | |
| CN111291321A (en) | Urban traffic model construction method based on big data | |
| CN114358808A (en) | Public transport OD estimation and distribution method based on multi-source data fusion | |
| Zwick et al. | Impact of service design on urban ridepooling systems | |
| Lyu et al. | R-sharing: Rendezvous for personalized taxi sharing | |
| CN111882915A (en) | On-demand bus route planning method adopting composite algorithm and interactive model | |
| CN116468219A (en) | Method for matching taxi sharing schedule by junction station | |
| Hariz et al. | Mobility traffic model based on combination of multiple transportation forms in the smart city | |
| Sundt et al. | Heuristics for customer-focused ride-pooling assignment | |
| Afandizadeh Zargari et al. | Optimizing algorithm for allocating passengers in shared taxis | |
| Uehara et al. | Evaluation of a hierarchical cooperative transport system using demand responsive bus on a dynamic simulation | |
| CN114692960A (en) | Vehicle path planning method and system with cooperation of oil consumption and optimal path |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |