CN102737356A - Intelligent bus scheduling calculation method - Google Patents

Abstract

The invention discloses an intelligent bus scheduling calculation method, which comprises the following steps of: constructing a math model of a genetic algorithm; and then designing specific solution steps of the genetic algorithm by the math model, wherein the termination condition of the algorithm is that a genetic algebra Gen is equal to the set maximum genetic algebra. By the intelligent bus scheduling calculation method, the comprehensive benefit between a bus company and passengers can be comprehensively considered; and the shortcomings about solution for multi-target non-linear problems in the conventional bus scheduling algorithm are overcome.

Description

The intelligent public transportation dispatching computing method of arranging an order according to class and grade

Technical field

The invention belongs to city intelligence public transportation system technical field, be specifically related to a kind of intelligent public transportation dispatching computing method of arranging an order according to class and grade.

Background technology

It is the basic assurance that guarantees the normal operation of public transit system that public transit system scheduling is arranged an order according to class and grade, and the public transport smart shift scheduling is the transportation of the quasi-representative combinatorial optimization problem of arranging an order according to class and grade, and it is found the solution has certain complicacy, is a forward position research topic to its research always.Traditional public transport research method of arranging an order according to class and grade mainly adopts methods such as mathematical analysis, analog simulation and mathematical programming, but along with the increase of problem solving difficulty and problem solving scale, classic method has run into greatly to be challenged, and can not guarantee the accuracy that target is separated.

Summary of the invention

The purpose of this invention is to provide a kind of intelligent public transportation dispatching computing method of arranging an order according to class and grade, can take all factors into consideration the comprehensive profit between public transport company and passenger, solved the traditional public transport deficiency of algorithm in multiple goal, nonlinear problem are found the solution of arranging an order according to class and grade.

The technical scheme that the present invention adopted is, a kind of intelligent public transportation dispatching computing method of arranging an order according to class and grade is characterized in that, specifically may further comprise the steps:

At first, set up the mathematical model of genetic algorithm:

Step 1, to formulate assumed condition following:

Same circuit vehicle adopts same model, and bus does not have passenger's trapping phenomena through later, and vehicle at the uniform velocity goes and to pass in and out the station time certain, and every circuit is certain working time, and the passenger arrives at a station to obey evenly and distributes;

Step 2, define following systematic parameter:

M: public transport operation period collection (m=1,2 ..., M), M is a circuit section working time number,

S: the circuit Station XXX (S=1 ... S), s is a public bus network website number,

B: circuit public transport collection (B=1,2 ... B),

Δ t _m: the departure interval of m period,

T _m: the time span of m period,

K: the number of times of always dispatching a car,

r _MS: the m period, s station passenger's arrival rate,

Get on the bus the stand-by period of passenger when taking B car in

S station

P: one day average handling capacity of passengers,

L: total line length,

Q: the ridership that bus can carry;

Step 3, confirm each parameter and basic departure interval Δ t _mBetween funtcional relationship:

The number of times of dispatching a car in the m period can be calculated by (formula 1):

K _m=T _m/ Δ t _m(formula 1),

Total number of times of dispatching a car in service can be calculated by (formula 2) in one day:

$K=\underset{m=1}{\overset{M}{\mathrm{\Σ}}}{K}_m$ (formula 2),

The m period, S station passenger's arrival rate can be calculated by (formula 3):

$r_{\mathrm{MS}}=\frac{M_{\mathrm{MS}}}{T_m}$ (formula 3),

Wherein, M _MSBe m period S station passengers quantity statistics,

The passenger at a time in the section S station take and be expressed as (formula 4) B averaging time that car spends:

$t_{\mathrm{BS}}^m=r_{\mathrm{MS}}*{\mathrm{\&Delta;\; <figure-callout\; id="2"\; label="t\; m"\; filenames="HDA00001761480600011.png,HDA00001761480600021.png"\; state="\{\{state\}\}">t</figure-callout>}}_m^{}/2$ (formula 4),

The economic loss of passenger's every day on spending in and waiting for bus can be expressed as (formula 5):

(formula 5),

Wherein, η is the wage income level of the average per minute in civic,

Every day, public transport company's running cost was (formula 6):

(formula 6),

Wherein, μ is the cost of public transport each run public transport company,

Two constraint conditions of above-mentioned (formula 5) and (formula 6) multiply by different weighted factor respectively, and β is converted into single goal optimization with multiple-objection optimization, and the objective function of final problem solving is:

(formula 7);

Secondly, utilize the concrete solution procedure of mathematical model design genetic algorithm that obtains, the algorithm end condition is for reaching the maximum genetic algebra of regulation as genetic algebra Gen:

Step 4, acquisition circuit passenger flow historical statistics information; It is expressed as matrix A; The A matrix is the two-dimensional matrix of M*s, and the row vector representation is the volume of the flow of passengers data of interior each website of section sometime, and column vector is represented the passenger flow statistics information of same website in different time sections;

Step 5, initial population coding, coding adopts the binary coding mode of being convenient to operate to encode, and choosing length PRECI is 4, and then chromosome coding length is PRECI*M, and choosing population scale POP is 20～200, genetic algebra Gen is chosen for 50～100:

A, generation population scale are POP, and chromosome coding length is the initial binary coding population InitPopBinary of PRECI*M,

B, utilize the conversion formula between binary coding and real number:

is converted into real number array InitPop with binary coding population InitPopBinary; Its size is POP*M

C, find the solution matrix A column vector and, be designated as array Num, array size is 1*M, representes the passenger flow statistics in each time period,

D, with array Num ordering, adjust initial InitPop and InitPopBinary with its big minispread sequence number, its size is arranged according to the size order of array Num;

Step 6, utilize formula

to calculate the fitness function value f (t) of each coded sequence;

The selection probability of step 7, calculation code sequence; Computing formula for

wherein Fint (t) be the chromosomal selection probability of t bar, f (t) is the chromosomal fitness of t bar;

Step 8, utilize the suitable selection operator with certain probability P _SelSelect the high individuality of fitness, form progeny population SubPop;

Step 9, utilize less variation probability P _MigSubPop carries out mutation operation to the filial generation population;

Step 10, carry out the reorganization of population, the individuality that the fitness in the progeny population is high is inserted in the initial population;

Step 11, utilize the c in the above-mentioned steps 5, two steps of d are carried out ordering again to chromosome;

Whether step 12, check reach the maximum genetic algebra of algorithm end condition, if reach then carry out next step, otherwise return step 7;

Step 13, obtain the highest chromosome coding of fitness; And be translated into real number through formula

, finally predict the outcome as algorithm;

Step 14, formulate the departure time-table in bus dispatching with this departure interval, finish.

The beneficial effect of the inventive method is: under the prerequisite of taking all factors into consideration constraint aspect public transport company and the passenger two; The current bus dispatching plan of historical passenger flow data prediction with public transport operation; Make the public transport frequency can adapt to the variation of passenger flow, realize the utilization of public transport reasonable resources.

Description of drawings

Fig. 1 is the present invention proof of algorithm figure as a result under the MATLAB platform;

Wherein: (a) be passenger flow statistics hum pattern in each time period, (b) be each station passenger flow statistics hum pattern in each time period, (c) be interior public transport of each time period number of times figure that dispatches a car, (d) be the public transport departure interval figure that predicts the outcome.

Embodiment

The intelligent public transportation dispatching of the present invention computing method of arranging an order according to class and grade at first, are set up the mathematical model of genetic algorithm, secondly, utilize the concrete solution procedure of mathematical model design genetic algorithm that obtains, and the algorithm end condition is for reaching the maximum genetic algebra of regulation as genetic algebra Gen.Specifically may further comprise the steps:

Step 1, for simplifying the problem solving difficulty, actual public transport operation is made the hypothesis of following more satisfactoryization.Formulation assumed condition is following:

Same circuit vehicle adopts same model, and bus does not have passenger's trapping phenomena through later, and vehicle at the uniform velocity goes and to pass in and out the station time certain, and every circuit is certain working time, and the passenger arrives at a station to obey evenly and distributes.

Step 2, calculate, define following systematic parameter for convenient:

M: public transport operation period collection (m=1,2 ..., M), M is a circuit section working time number,

S: the circuit Station XXX (S=1 ... S), s is a public bus network website number,

B: circuit public transport collection (B=1,2 ... B),

Δ t _m: the departure interval of m period,

T _m: the time span of m period,

K: the number of times of always dispatching a car,

r _MS: the m period, s station passenger's arrival rate,

Get on the bus the stand-by period of passenger when taking B car in

S station

P: one day average handling capacity of passengers,

L: total line length,

Q: the ridership that bus can carry.

Step 3, confirm each parameter and basic departure interval Δ t _mBetween funtcional relationship:

The number of times of dispatching a car in the m period can be calculated by (formula 1):

K _m=T _m/ Δ t _m(formula 1),

Total number of times of dispatching a car in service can be calculated by (formula 2) in one day:

$K=\underset{m=1}{\overset{M}{\mathrm{\&Sigma;}}}{K}_m$ (formula 2),

Because the passenger arrives at a station and is assumed to be even distribution, the m period, S station passenger's arrival rate can be calculated by (formula 3):

$r_{\mathrm{MS}}=\frac{M_{\mathrm{MS}}}{T_m}$ (formula 3),

Wherein, M _MSBe m period S station passengers quantity statistics,

The passenger at a time in the section S station take and be expressed as (formula 4) B averaging time that car spends:

$t_{\mathrm{BS}}^m=r_{\mathrm{MS}}*{\mathrm{\&Delta;\; <figure-callout\; id="2"\; label="t\; m"\; filenames="HDA00001761480600011.png,HDA00001761480600021.png"\; state="\{\{state\}\}">t</figure-callout>}}_m^{}/2$ (formula 4),

The economic loss of passenger's every day on spending in and waiting for bus can be expressed as (formula 5):

(formula 5),

Wherein, η is the wage income level of the average per minute in civic,

Every day, public transport company's running cost was (formula 6):

(formula 6),

Wherein, μ is the cost of public transport each run public transport company,

Two constraint conditions of above-mentioned (formula 5) and (formula 6) multiply by different weighted factor respectively, and β is converted into single goal optimization with multiple-objection optimization, and the objective function of final problem solving is:

(formula 7).

Consider that simultaneously the passenger waits for that the public transport time can influence the satisfaction of public transport service, so be necessary the public transport departure interval is made certain constraint, rule of thumb knowledge can be limited in Δ t with it _mIn the interval of ∈ [3,15].

Step 4, acquisition circuit passenger flow historical statistics information; It is expressed as matrix A; The A matrix is the two-dimensional matrix of M*s, and the row vector representation is the volume of the flow of passengers data of interior each website of section sometime, and column vector is represented the passenger flow statistics information of same website in different time sections.

Step 5, initial population coding, coding adopts the binary coding mode of being convenient to operate to encode, and choosing length PRECI is 4, and then chromosome coding length is PRECI*M, and choosing population scale POP is 20～200, genetic algebra Gen is chosen for 50～100:

A, generation population scale are POP, and chromosome coding length is the initial binary coding population InitPopBinary of PRECI*M,

B, utilize the conversion formula between binary coding and real number:

is converted into real number array InitPop with binary coding population InitPopBinary; Its size is POP*M

C, find the solution matrix A column vector and, be designated as array Num, array size is 1*M, representes the passenger flow statistics in each time period,

D, with array Num ordering, adjust initial InitPop and InitPopBinary with its big minispread sequence number, its size is arranged according to the size order of array Num.

Step 6, utilize formula

to calculate the fitness function value f (t) of each coded sequence.

The selection probability of step 7, calculation code sequence; Computing formula for

wherein Fint (t) be the chromosomal selection probability of t bar, f (t) is the chromosomal fitness of t bar.

Step 8, utilize the suitable selection operator with certain probability P _SelSelect the high individuality of fitness, form progeny population SubPop.

Step 9, utilize less variation probability P _MigSubPop carries out mutation operation to the filial generation population.

Step 10, carry out the reorganization of population, the individuality that the fitness in the progeny population is high is inserted in the initial population.

Step 11, utilize the c in the above-mentioned steps 5, two steps of d are carried out ordering again to chromosome.

Whether step 12, check reach the maximum genetic algebra of algorithm end condition, if reach then carry out next step, otherwise return step 7.

Step 13, obtain the highest chromosome coding of fitness; And be translated into real number through formula

, finally predict the outcome as algorithm.

Step 14, formulate the departure time-table in bus dispatching with this departure interval, finish.

As shown in Figure 1, it is result of under the MATLAB platform, utilizing genetic algorithm tool box function to verify, can prove absolutely the inventive method rationality and practicality.In the present embodiment, each algorithm parameter is: population scale POP=100, maximum genetic algebra Gen=100, selection operation utilize roulette to select, and crossover operator is that single-point intersects, and crossover probability is P _Sel=0.9, the variation probability is P _Mig=0.1, the population regrouping process utilizes the big offspring individual of fitness to replace the little parent of fitness individual.Public bus network website along the line is counted s=26 in the checking primary data, and moving the total time period number of period 6:00 ~ late 24:00 morning altogether is M=18.Passenger flow statistics figure and the number of times figure that dispatches a car can find out that this bus dispatching algorithm of arranging an order according to class and grade can accomplish and make the requirement of public transport departure interval along with passenger flow information variation.

Claims (1)

1. intelligent public transportation dispatching computing method of arranging an order according to class and grade is characterized in that, specifically may further comprise the steps:

At first, set up the mathematical model of genetic algorithm:

Step 1, to formulate assumed condition following:

Same circuit vehicle adopts same model, and bus does not have passenger's trapping phenomena through later, and vehicle at the uniform velocity goes and to pass in and out the station time certain, and every circuit is certain working time, and the passenger arrives at a station to obey evenly and distributes;

Step 2, define following systematic parameter:

M: public transport operation period collection (m=1,2 ..., M), M is a circuit section working time number,

S: the circuit Station XXX (S=1 ... S), s is a public bus network website number,

B: circuit public transport collection (B=1,2 ... B),

Δ t _m: the departure interval of m period,

T _m: the time span of m period,

K: the number of times of always dispatching a car,

r _MS: the m period, s station passenger's arrival rate,

Get on the bus the stand-by period of passenger when taking B car in

S station

P: one day average handling capacity of passengers,

L: total line length,

Q: the ridership that bus can carry;

Step 3, confirm each parameter and basic departure interval Δ t _mBetween funtcional relationship:

The number of times of dispatching a car in the m period can be calculated by (formula 1):

K _m=T _m/ Δ t _m(formula 1),

Total number of times of dispatching a car in service can be calculated by (formula 2) in one day:

$K=\underset{m=1}{\overset{M}{\mathrm{\&Sigma;}}}{K}_m$ (formula 2),

The m period, S station passenger's arrival rate can be calculated by (formula 3):

$r_{\mathrm{MS}}=\frac{M_{\mathrm{MS}}}{T_m}$ (formula 3),

Wherein, M _MSBe m period S station passengers quantity statistics,

The passenger at a time in the section S station take and be expressed as (formula 4) B averaging time that car spends:

$t_{\mathrm{BS}}^m=r_{\mathrm{MS}}*{\mathrm{\&Delta;\; t}}_m^2/2$ (formula 4),

The economic loss of passenger's every day on spending in and waiting for bus can be expressed as (formula 5):

(formula 5)

Wherein, η is the wage income level of the average per minute in civic,

Every day, public transport company's running cost was (formula 6):

(formula 6)

Wherein, μ is the cost of public transport each run public transport company,

Two constraint conditions of above-mentioned (formula 5) and (formula 6) multiply by different weighted factor respectively, and β is converted into single goal optimization with multiple-objection optimization, and the objective function of final problem solving is:

(formula 7);

Secondly, utilize the concrete solution procedure of mathematical model design genetic algorithm that obtains, the algorithm end condition is for reaching the maximum genetic algebra of regulation as genetic algebra Gen:

Step 4, acquisition circuit passenger flow historical statistics information; It is expressed as matrix A; The A matrix is the two-dimensional matrix of M*s, and the row vector representation is the volume of the flow of passengers data of interior each website of section sometime, and column vector is represented the passenger flow statistics information of same website in different time sections;

Step 5, initial population coding, coding adopts the binary coding mode of being convenient to operate to encode, and choosing length PRECI is 4, and then chromosome coding length is PRECI*M, and choosing population scale POP is 20～200, genetic algebra Gen is chosen for 50～100:

A, generation population scale are POP, and chromosome coding length is the initial binary coding population InitPopBinary of PRECI*M,

B, utilize the conversion formula between binary coding and real number:

is converted into real number array InitPop with binary coding population InitPopBinary; Its size is POP*M

C, find the solution matrix A column vector and, be designated as array Num, array size is 1*M, representes the passenger flow statistics in each time period,

D, with array Num ordering, adjust initial InitPop and InitPopBinary with its big minispread sequence number, its size is arranged according to the size order of array Num;

Step 6, utilize formula

to calculate the fitness function value f (t) of each coded sequence;

The selection probability of step 7, calculation code sequence; Computing formula for

wherein Fint (t) be the chromosomal selection probability of t bar, f (t) is the chromosomal fitness of t bar;

Step 8, utilize the suitable selection operator with certain probability P _SelSelect the high individuality of fitness, form progeny population SubPop;

Step 9, utilize less variation probability P _MigSubPop carries out mutation operation to the filial generation population;

Step 10, carry out the reorganization of population, the individuality that the fitness in the progeny population is high is inserted in the initial population;

Step 11, utilize the c in the above-mentioned steps 5, two steps of d are carried out ordering again to chromosome;

Whether step 12, check reach the maximum genetic algebra of algorithm end condition, if reach then carry out next step, otherwise return step 7;

Step 13, obtain the highest chromosome coding of fitness; And be translated into real number through formula , finally predict the outcome as algorithm;

Step 14, formulate the departure time-table in bus dispatching with this departure interval, finish.

Images