CN106257504A - A kind of BRT passenger based on Equilibrium Assignment Model goes on a journey benefit optimization method - Google Patents

Abstract

The invention discloses a kind of BRT passenger based on Equilibrium Assignment Model to go on a journey benefit optimization method, including step: 1) selected object of study, research range, gather relevant rudimentary data；2) analyze passenger demand characteristic: proposing to meet the occupant classification concept of passenger's OD demand, division can distribute passenger flow and can not distribute passenger flow, analyze in station passenger's passenger flow characteristic；3) distribution principle is determined: consider the characteristic of dissimilar passenger, determine distribution principle；4) optimizing under constraints: consider passenger's riding time cost, delay time at stop cost, congestion costs and relevant constraint, determine object function, and design corresponding optimizing algorithm, exports optimal case.The present invention is on the basis of analyzing passenger's trip characteristics, take into full account passenger's riding time cost, be delayed waiting time cost and total congestion costs, determine bus traveler assignment scheme according to passenger's maximizing the benefits of going on a journey, effectively promote passenger and go on a journey satisfaction and the service quality of BRT system.

Description

A kind of BRT passenger based on Equilibrium Assignment Model goes on a journey benefit optimization method

Technical field

The present invention relates to bus rapid transit (BRT) passenger go on a journey the technical field of benefit, particularly relate to a kind of based on all The BRT passenger of weighing apparatus Assignment Model goes on a journey benefit optimization method.

Background technology

The most at home in the operation and management of urban bus rapid transit, the trip benefit of passenger is increasingly by everybody Concern.BRT platform is provided with the information of vehicles that the information equipments such as information of vehicles platform are correlated with transmission at present, such as arrival time Deng, but passenger still lacks and optimizes guiding accordingly, especially arrives vehicle in passage the most, can take vehicle more, And front and back workshop is when shorter, passenger often selects circuit by bus by experience of taking in the past, i.e. clicks according to oneself point of destination Select circuit, select the circuit vehicle first arrived at a station to take in up to circuit, have ignored line conditions etc., and having a plurality of circuit to arrive When the time that reaches is close, the circuit of the maximizing the benefits that not do not selects to make oneself to ride, reduce passenger's benefit of going on a journey public with BRT The captivation of friendship system.

BRT, as emerging public transport mode, receives the favor of numerous passenger with its efficient efficiency of operation, for Current Situation, by rational method to public transit system the volume of the flow of passengers carry out equilibrium assignment, promoting passenger's benefit of going on a journey becomes The center of gravity of future development.At present about Equilibrium Assignment problem, domestic and international achievement in research is more, and main performance is as follows: Kim exists Under Wardrop user equilibrium principle, have studied the Complete Information impact on Traffic Net flow, it is believed that Complete Information is to public affairs In friendship system, passenger's choice for traveling line influence is bigger；Decea and Femande considers vehicle logistics capacity factor, establishes The equilibrium model of public transit system, it is achieved to the equilibrium assignment flowed standee；In affecting passenger's cost factor, Huang, LamHK The crowded impact on passenger etc. in have studied car, establishes the uncomfortable function of reaction passenger's impression, with crowded in this reacting vehicle Degree, and think the monotonic increasing function that passenger's congestion costs in car is number of people in car and time, weigh passenger with this crowded Cost；Kraus and Yoshida queue waiting time reacts tardiness cost, and establishes economics mould based on queuing time Type；Gao Ziyou etc. have studied the crowding phenomenon of queuing station, uses plan dynamic approach to construct the Equilibrium Assignment under crowed condition Model, also can realize assignment of traffic under crowed condition；Four soldier peaks, He Xuesheng, Xie Wei victory etc. have studied the city under static conditions Passenger's equalization problem on city's public traffic network, establishes the Equilibrium Assignment Model under the conditions of public traffic network, and Tian Qiong etc. have studied Passenger's travel behaviour in regular public traffic network, proposes to go on a journey dynamic equilibrium model based on passenger, and heterogeneous according to passenger Property, establish the equilibrium model under the conditions of peak period；Zhang Yujie etc. discuss the passenger type that crowded sensitivity is different, by the time Including tardiness cost and congestion cost in carriage are also contemplated for, establish the equilibrium model heterogeneous based on passenger.

But Equilibrium Assignment Model is studied the fewest in BRT field, healthy have studied information-based bar by intelligent modeling method Under part, the cognitive style of passenger and travel behaviour, establish phantom；Zheng Xiaofeng etc. propose based on the induction by bus of BRT platform Public transit vehicle carrying equilibrium model, the cabin factor being reached BRT circuit by Equilibrium Assignment is always equalized, but not by passenger's benefit Take into account, by traversal solving model, in the case of data volume is less, have certain suitability, but under big data qualification Lack enough solution efficiencies.

Summary of the invention

The technical problem to be solved in the present invention is, it is provided that a kind of BRT passenger based on Equilibrium Assignment Model goes on a journey benefit Optimization method, it is achieved BRT system passenger humanized, intelligentized go on a journey scheme recommend, it is ensured that BRT system service quality.

For solving above-mentioned technical problem, the present invention provides following technical scheme: a kind of BRT based on Equilibrium Assignment Model takes advantage of Visitor's trip benefit optimization method, comprises the following steps:

S1, choose object of study and research range, gather relevant rudimentary information；

S2, according to relevant rudimentary information, analyze passenger demand characteristic: the occupant classification proposing to meet passenger's OD demand is general Read, divide the characteristic of dissimilar passenger, including can distribution of passengers with can not distribution of passengers；

S3, by analyze dissimilar passenger characteristic, calculate the number of dissimilar passenger, determine distribution principle；

S4, according to distribution principle, calculate associated passenger cost, including passenger ride cost, be delayed waiting time cost with And total congestion costs；

S5, determine object function according to associated passenger cost, and determine the constraint of Equilibrium Assignment Model according to object function Condition；

S6, Equilibrium Assignment Model is solved, arrange the result solved, obtain rational bus traveler assignment scheme.

Further, the relevant rudimentary information in described step S1, including BRT system intrinsic information, each line tower foundation letter Breath and passenger's trip information, wherein

Described BRT system intrinsic information, stops public bus network numbering including site name, website quantity and website in road；

Described each line tower foundation information, including numbering, seating capacity, design capacity, greatest physical capacity, average operation of often standing Time and mean delay waiting time, also include stopping the seating capacity of public bus network of targeted sites, number of getting off, sailing From number；

Described passenger's trip information includes the travelling OD demand of passenger.

Further, the analysis passenger demand characteristic in described step S2, it concretely comprises the following steps:

In S21, analysis BRT express passway, the substitutability of occupant ride plan, if OD is interval in passage by bus, then Having a plurality of circuit can select to take, the circuit that its plan is taken can be substituted；If OD is interval outside passage by bus, then may select To take circuit unique；

S22, general OD by bus are interval in passage, and the passenger having a plurality of circuit to take is considered as can distribution of passengers；And take advantage of Car OD is interval outside passage, and taking that the unique passenger of circuit is considered as can not distribution of passengers.

Further, the distribution principle in described step S3 particularly as follows:

S31, determine station passenger's total number of persons:

$Q_j=q_j+\underset{j}{\Σ}\underset{i=n}{\overset{m}{\Σ}}{o}_{ij},\∀i\∈L$

Wherein, Q_jRepresent jth station station passenger's total number of persons；q_jRepresent jth station can distribution of passengers total number of persons；o_ijRepresent What i-th line road was taken in the plan of jth station can not distribute patronage；L represents the circuit vehicle set arrived at a station；

S32, according to different types of passenger's characteristic, pay the utmost attention to can not the demand by bus of distribution of passengers, in order to maximum limit Degree reduces total passenger delay cost；

S33, consideration vehicle heap(ed) capacityWith vehicle carrying situation, determine that the i-th line road vehicles at jth station can accommodate people Number e_ij:

$e_{ij}=c_i^m+d_{ij}-P_{ik}$

Wherein, d_ijRepresent the number of getting off on the i-th line road at jth station；P_ikRepresent that the sailing out of of i-th line road at kth station is taken advantage of Guest's number, and the upper website that kth station is jth station；

S34, the vehicle seating capacity comparing the i-th line road at jth station and the plan of jth station take i-th line road not Can the size of distribution of passengers number；Determine jth station actual take i-th line road can not distribute patronageAnd jth station I-th line road can not the delay number of distribution of passengers

$o_{ij}^p=\left\{\begin{array}{ccc}{o}_{ij},& if& o_{ij}\≤e_{ij}\\ e_{ij},& if& o_{ij}>e_{ij}\end{array}\right\}$

$o_{ij}^d=\left\{\begin{array}{ccc}0,& if& o_{ij}\≤e_{ij}\\ o_{ij}-e_{ij},& if& o_{ij}>e_{ij}\end{array}\right\}.$

Further, in described step S4, described passenger riding time cost TC₁Computational methods particularly as follows:

${\mathrm{TC}}_1=\underset{j}{\Σ}\underset{i=n}{\overset{m}{\Σ}}{t}_i^p(x_{ij}+o_{ij}^p),\∀i\∈L$

In formula, x_ijRepresent i-th line road in jth station can the number of getting on the bus of distribution of passengers；Represent i-th line road vehicles Average often stand the service time；

To can be considered as delay time at stop of distribution of passengers the mean delay time of selected circuit, and can not the prolonging of distribution of passengers For the delay time at stop of fixing circuit between mistaking, accordingly, it is determined that described delay waiting time cost TC₂Computational methods particularly as follows:

${\mathrm{TC}}_2=t^D\[q_j-\underset{j}{\Σ}\underset{i=n}{\overset{m}{\Σ}}{x}_{ij}\]+\underset{j}{\Σ}\underset{i=n}{\overset{m}{\Σ}}{t}_i^d{o}_{ij}^d,\∀i\∈L$

In formula, t^DExpression can mean delay waiting time of distribution of passengers；When representing the delay wait on i-th line road Between；

Described total congestion costs TC₃Calculation method particularly as follows:

${\mathrm{TC}}_3=\underset{i=n}{\overset{m}{\Σ}}{G}_i\[P_{ij}\]$

Wherein, congestion costs function G_i[P_(ij)] expression formula is as follows:

$G_i\&lsqb;P_{ij}\&rsqb;=\left\{\begin{array}{ccc}0,& if& P_{ij}\&le;c_i^s\\ \mathrm{\&alpha;}{(P_{ij}-c_i^s)}^{{\mathrm{\&gamma;}}_1},& if& c_i^d<P_{ij}\&le;c_i^s\\ \mathrm{\&alpha;}{(c_i^d-c_i^s)}^{{\mathrm{\&gamma;}}_1}+\mathrm{\&beta;}{(P_{ij}-c_i^m)}^{{\mathrm{\&gamma;}}_2}& if& c_i^s<P_{ij}\&le;c_i^m\end{array}\right\}$

In formula, P_ijRepresent jth station i-th line road sail out of patronage, be calculated as follows:

$P_{ij}=P_{ik}+x_{ij}+o_{ij}^p-d_{ij};$

Represent the seating capacity of i-th line road vehicles；Represent the design capacity of i-th line road vehicles；Represent i-th The greatest physical capacity of bar circuit vehicle；α、β、γ₁、γ₂Represent the congestion costs coefficient under different condition respectively.

Further, described step S5 determine object function particularly as follows: according to passenger ride cost, be delayed the waiting time become This, total congestion costs and the value weight of all kinds of cost, determine object function:

MinTC=λ₁TC₁+λ₂TC₂+λ₃TC₃

λ in formula₁、λ₂、λ₃Correspond to the value weight of each cost.

Further, described value weight, its value is by investigating decision, in investigation, uses by the pairing ratio of expert estimation Relatively method and the normal and method of scales given a mark by passenger, two kinds of methods combine and determine that each is worth weight, or referred to by public transport Draw the passenger's wish collected in APP and comprehensively determine to be worth weight.

Further, described step S5 determining, the constraints of Equilibrium Assignment Model is as follows:

1. decision variable x_ijSpan by vehicle capacity with preferentially distribute the patronage that can not distribute got on the bus and limited, And it is defined to set of integers:

$0\&le;x_{ij}\&le;e_{ij}-o_{ij}^p,\&ForAll;x_{ij}\&Element;Z;$

The always number of getting on the bus of the most all types of passengers should be less than being equal at station total number of persons, and ensure at station passenger's total benefit simultaneously Optimum, it is allowed to the existence of delay number:

$0<\underset{j}{\&Sigma;}\underset{i=n}{\overset{m}{\&Sigma;}}{x}_{ij}+\underset{j}{\&Sigma;}\underset{i=n}{\overset{m}{\&Sigma;}}{o}_{ij}^p\&le;Q_j;$

The cabin factor of the most each bar circuit is equal to the ratio sailing out of patronage and vehicle maximum logistics capacity of this circuit:And 0 ＜ p_ij≤1。

Further, Equilibrium Assignment Model is solved in described step S6, use Revised genetic algorithum to all Weighing apparatus Assignment Model solves, and comprises the steps:

S61, chromosome coding: use real coding method, by the floating-point table in individual each gene position given range Showing, this scope depends on that the span that decision variable arrives, individual code length depend on the number of decision content；Wherein, compile The scope of code is x_ijSpan；

S62, design fitness function: genetic algorithm distinguishes individual quality by fitness function, by passenger's totle drilling cost Inverse as individual fitness value, the individuality that i.e. passenger's cost is the least, fitness is the biggest, and individuality is the most excellent；If individual k's is suitable Response function is F_k[TC(x_ij)], then have:

$F_k\&lsqb;TC\left(x_{ij}\right)\&rsqb;=\frac{1}{TC\left(x_{ij}\right)}$

S63, selection operation: from old population, form new population with certain probability selection excellent individual, to Equilibrium Assignment mould Type uses roulette method, i.e. ratio selection method, individual selected probability to be directly proportional to fitness value, the biggest individual of fitness Body, selected probability is the biggest；Then to individual Probability p selected for i_kHave:

$p_k=\frac{F_k\&lsqb;TC\left(x_{ij}\right)\&rsqb;}{\underset{l=1}{\overset{z}{\&Sigma;}}{F}_l\&lsqb;TC\left(x_{ij}\right)\&rsqb;};$

S64, intersection and mutation operation: in order to heredity has the individuality of the outstanding feature of previous generation and keeps the various of population Property and the operation taked；Realize intersecting and mutation operation by the genetic operator that Calling MATLAB is built-in；

S65, judge whether to meet end condition: when evolutionary generation meets setting value, or algorithm is in continually evolving certain generation After number, when the fitness of solution does not significantly improve, i.e. export optimum results..

Using after technique scheme, the present invention at least has the advantages that by taking into full account in BRT passage not With passenger flow characteristic and the choice for traveling of type passenger, science determines riding time cost, is delayed delay cost and total congestion costs Selection preference in passenger's choice for traveling, is embodied selection preference by Optimized model, seeks passenger's totle drilling cost minimum Choice for traveling, contribute to substituting selection scheme single in existing riding guidance, humanized, intelligentized taking can be realized Take advantage of scheme to recommend, improve passenger and go on a journey satisfaction and the captivation of BRT system.

Accompanying drawing explanation

Fig. 1 is that a kind of BRT passenger based on Equilibrium Assignment Model of the present invention goes on a journey the flow chart of benefit optimization method；

Fig. 2 is that a kind of BRT passenger based on Equilibrium Assignment Model of the present invention goes on a journey the described case of benefit optimization method Conspectus；

Fig. 3 is that a kind of BRT passenger based on Equilibrium Assignment Model of the present invention goes on a journey the something lost of embodiment of benefit optimization method Pass evolution graph.

Detailed description of the invention

It should be noted that in the case of not conflicting, the embodiment in the application and the feature in embodiment can phases Combine mutually, with specific embodiment, the application is described in further detail below in conjunction with the accompanying drawings.

BRT passenger based on Equilibrium Assignment Model of the present invention goes on a journey benefit optimization method, from analyzing passenger demand Characteristic is set out, it is considered to affect passenger go on a journey benefit passenger's time cost, be delayed delay cost and congestion costs, go on a journey with passenger Total benefit maximum turns to target, establishes Optimized model.

The benefit optimization method as it is shown in figure 1, the BRT passenger based on Equilibrium Assignment Model described in the present embodiment goes on a journey, bag Include following steps:

1) selected object of study, research range, gather relevant rudimentary data, including BRT system intrinsic information, each circuit base Plinth information, passenger's trip information, wherein, described BRT system intrinsic information refer in site name, website quantity, website stop Public bus network is numbered；Described each line tower foundation information includes numbering, seating capacity, design capacity, greatest physical capacity, averagely often stands Service time, the mean delay waiting time and stop the seating capacity of public bus network of targeted sites, number of getting off, sail out of Number；Described passenger's trip information refers to the travelling OD demand of passenger；

2) analyze passenger demand characteristic: propose to meet the occupant classification concept of passenger's OD demand, division can distribute passenger flow with Passenger flow can not be distributed, analyze in station passenger's passenger flow characteristic；

3) distribution principle is determined: consider the characteristic of dissimilar passenger, determine distribution principle；

4) optimizing under constraints: consider passenger's riding time cost, delay time at stop cost, congestion costs and be correlated with Constraints, determines object function, and designs corresponding optimizing algorithm, exports optimal case.

In step 2) in, the key step analyzing passenger demand characteristic has:

2.1) analyze in BRT express passway, the substitutability of occupant ride plan, if the most interval in passage, then have A plurality of circuit can select to take, and its circuit of taking of plan can be substituted, otherwise the most selectable to take circuit unique；

2.2) according to passenger's passenger flow feature, it is judged that passenger's classification, in passage, a plurality of circuit will be had permissible in OD interval by bus The passenger taken be considered as can distribution of passengers, and OD is interval outside passage by bus, takes the unique passenger of circuit and is considered as distributing Passenger；

2.3) determine at station passenger's total number of persons:

$Q_j=q_j+\underset{j}{\&Sigma;}\underset{i=n}{\overset{m}{\&Sigma;}}{o}_{ij},\&ForAll;i\&Element;L$

Q_j: jth station station passenger's total number of persons；

q_j: jth station can distribution of passengers total number of persons；

o_ij: what i-th line road was taken in the plan of jth station can not distribute patronage；

L: represent the circuit vehicle set arrived at a station.

In step 3) in, determine that the key step of distribution principle has:

3.1) according to step 2) different types of passenger's characteristic of determining, draw pay the utmost attention to can not distribution of passengers by bus Demand, to reduce total passenger delay cost to greatest extent；

3.2) vehicle heap(ed) capacity is consideredWith vehicle carrying situation, determine that the i-th line road vehicles at jth station can accommodate people Number:

$e_{ij}=c_i^m+d_{ij}-P_{ik}$

d_ij: the number of getting off on the i-th line road at jth station；

P_ik: the i-th line road at kth station sail out of patronage, wherein kth station is a upper website at jth station；

3.3) compare the i-th line road vehicles seating capacity at jth station and i-th line road is taken in the plan of jth station can not The size of distribution of passengers number；Determine jth station actual take i-th line road can not distribute patronage and the i-th of jth station Bar circuit can not the delay number of distribution of passengers:

$o_{ij}^p=\left\{\begin{array}{ccc}{o}_{ij},& if& o_{ij}\&le;e_{ij}\\ e_{ij},& if& o_{ij}>e_{ij}\end{array}\right\}$

$o_{ij}^d=\left\{\begin{array}{ccc}0,& if& o_{ij}\&le;e_{ij}\\ o_{ij}-e_{ij},& if& o_{ij}>e_{ij}\end{array}\right\}$

Jth station is actual take i-th line road can not distribute patronage；

I-th line road, jth station can not the delay number of distribution of passengers.

In step 4) in, under constraints, the key step of optimizing has:

4.1) step 1 is utilized) average often standing the service time of collecting, determine passenger riding time cost TC₁:

${\mathrm{TC}}_1=\underset{j}{\&Sigma;}\underset{i=n}{\overset{m}{\&Sigma;}}{t}_i^p(x_{ij}+o_{ij}^p),\&ForAll;i\&Element;L$

x_ij: in jth station, i-th line road can the number of getting on the bus of distribution of passengers；

The average of i-th line road vehicles often stands the service time (unit: min)；

In jth station, i-th line road can not the number of getting on the bus of distribution of passengers；

4.2) according to step 2) and step 3) analysis result, determine being considered as delay time at stop of distribution of passengers selected The mean delay time of circuit, and can not delay time at stop of distribution of passengers be the delay time at stop of fixing circuit, it is thus determined that be delayed Waiting time cost is:

${\mathrm{TC}}_2=t^D\&lsqb;q_j-\underset{j}{\&Sigma;}\underset{i=n}{\overset{m}{\&Sigma;}}{x}_{ij}\&rsqb;+\underset{j}{\&Sigma;}\underset{i=n}{\overset{m}{\&Sigma;}}{t}_i^d{o}_{ij}^d,\&ForAll;i\&Element;L$

t^D: can mean delay waiting time (unit: min) of distribution of passengers；

The delay waiting time (unit: min) on i-th line road；

Jth station do not take i-th line road can not distribute number.

4.3) consider patronage on car, determine total congestion costs:

${\mathrm{TC}}_3=\underset{i=n}{\overset{m}{\&Sigma;}}{G}_i\&lsqb;P_{ij}\&rsqb;$

Wherein congestion costs function G_i[P_(ij)] expression formula is as follows:

$G_i\&lsqb;P_{ij}\&rsqb;=\left\{\begin{array}{ccc}0,& if& P_{ij}\&le;c_i^s\\ \mathrm{\&alpha;}{(P_{ij}-c_i^s)}^{{\mathrm{\&gamma;}}_1},& if& c_i^d<P_{ij}\&le;c_i^s\\ \mathrm{\&alpha;}{(c_i^d-c_i^s)}^{{\mathrm{\&gamma;}}_1}+\mathrm{\&beta;}{(P_{ij}-c_i^m)}^{{\mathrm{\&gamma;}}_2}& if& c_i^s<P_{ij}\&le;c_i^m\end{array}\right\}$

In formula:

P_ij: represent jth station i-th line road sail out of patronage

$P_{ij}=P_{ik}+x_{ij}+o_{ij}^p-d_{ij};$

The seating capacity of i-th line road vehicles；

The design capacity of i-th line road vehicles；

The greatest physical capacity of i-th line road vehicles；

α and β, γ₁、γ₂Represent the congestion costs coefficient under different condition respectively.

4.4) consider passenger ride cost, be delayed waiting time cost, total congestion costs and the valency of all kinds of cost Value weight, determines object function:

MinTC=λ₁TC₁+λ₂TC₂+λ₃TC₃

λ in formula₁(unit/min), λ₂(unit/min), λ₃(first) is the value weight of each cost, and its value can be determined by investigation, Investigation will use the Paired comparison method by expert estimation and often combined and determine each valency with method of scales by what passenger gave a mark Value weight, it is possible to guided the passenger's wish collected in APP comprehensively to determine by public transport.

4.5) according to set objective function and practical situation, determine that the constraints of model is as follows:

1. decision variable x_ijSpan by vehicle capacity with preferentially distribute the patronage that can not distribute got on the bus and limited, And it is defined to set of integers；

$0\&le;x_{ij}\&le;e_{ij}-o_{ij}^p,\&ForAll;x_{ij}\&Element;Z$

The always number of getting on the bus of the most all types of passengers should be less than equal at station total number of persons, simultaneously for ensureing at station passenger's total benefit Optimum, it is allowed to be delayed the existence of number；

$0<\underset{j}{\&Sigma;}\underset{i=n}{\overset{m}{\&Sigma;}}{x}_{ij}+\underset{j}{\&Sigma;}\underset{i=n}{\overset{m}{\&Sigma;}}{o}_{ij}^p\&le;Q_j$

The cabin factor of the most each bar circuit is equal to the ratio sailing out of patronage and vehicle maximum logistics capacity of this circuit.

And 0 ＜ p_ij≤1

4.6) using Revised genetic algorithum to solve model, key step is as follows:

1. chromosome coding

Use real coding method herein, real coding by the floating point representation in individual each gene position given range, This scope depends on that the span that decision variable arrives, individual code length depend on the number of decision content.Encode model herein Enclose and be x_ijSpan.

2. fitness function is designed

Genetic algorithm distinguishes individual quality by fitness function, herein model be to solve for target be passenger's totle drilling cost Little, using the inverse of passenger's totle drilling cost as individual fitness value, the individuality that i.e. passenger's cost is the least, fitness is the biggest, individual The most excellent.If the fitness function of individual k is F_k[TC(x_ij)], then have:

$F_k\&lsqb;TC\left(x_{ij}\right)\&rsqb;=\frac{1}{TC\left(x_{ij}\right)}$

3. operation is selected

Selection is to form new population with certain probability selection excellent individual from old population, uses roulette method herein, i.e. Ratio selection method, individual selected probability is directly proportional to fitness value, the individuality that fitness is the biggest, and selected probability is more Greatly.Then to individual Probability p selected for i_kHave:

$p_k=\frac{F_k\&lsqb;TC\left(x_{ij}\right)\&rsqb;}{\underset{l=1}{\overset{z}{\&Sigma;}}{F}_l\&lsqb;TC\left(x_{ij}\right)\&rsqb;}$

4. cross and variation operation

Intersect and mutation operation is in order to heredity has the individuality of the outstanding feature of previous generation and keeps the multiformity of population And the operation taked, realize intersecting and mutation operation herein by the genetic operator that Calling MATLAB is built-in.

6. judge whether to meet end condition

When evolutionary generation meets setting value, or algorithm is after continually evolving certain algebraically, and the fitness of solution the most substantially changes When entering, i.e. export optimum results.

4.6) arrange the result of output, formulate rational bus traveler assignment scheme.

Analysis of cases

As described in Figure 2, it is located in a BRT express passway, is provided with Z₀, Z₁, Z₂, Z₃Website in totally four passages, Z is had outside BRT passage₄, Z₅, Z₆, Z₇, Z₈Website in five passages, with Z₀Targeted sites as this suboptimization；It is provided with circuit B1, B2, B3, B4, B5 are through this BRT express passway, and the most only B1 is uniquely through the outer website Z of passage₄, B2 is uniquely outside passage Website Z₅, B3 is uniquely through the outer website Z of passage₆, B4 is uniquely through the outer website Z of passage₇, B5 is uniquely through the outer website Z of passage₇。

The vehicle parameter such as following table of each bar circuit vehicle:

Table 1 circuit vehicle parameter

Assume known at Z₀Passenger's total number of persons of website is 130 people, and passenger's OD demand is as shown in the table:

Table 2 passenger's OD demand schedule

Assume during this suboptimization, arrive Z₀The circuit vehicle of website is B1, B2, B3, B4, B5, each bar regular vehicle Carrying situation such as following table:

Table 3 each circuit vehicle carrying situation

In practical problem, through visiting and investigating Guangzhou BRT platform passenger, about 76 passengers use 5 grades Riding time cost, delay waiting time cost, the value weight of total congestion costs are given a mark by appraisement system respectively, its In try to achieve each weighted mean and be respectively 2.89,2.02 and 3.87, for ease of verifying model further, value power is set in proportion It is heavily λ₁=1.5, λ₂=1, λ₃=2.Additionally, by analyzing the OD trip characteristics of passenger, drawing can distribution of passengers in present case Number q_j=80, number o can not be distributed_ijIt is respectively 10,10,15,10,5.Consult pertinent literature, take congestion quotiety α=0.5, β =1, γ₁=1, γ₂=1；Population scale is 200, and crossover probability, mutation probability take 0.8,0.01 respectively.By MatLab software Being programmed solving, genetic evolution figure is as it is shown on figure 3, result shows, optimal value converges on 917.75 yuan, and reality is always got on the bus number Being 128, wherein corresponding optimum each bar circuit distribution number is 18,17,12,9,22, and the number of getting on the bus of each bar circuit is 28, 27,27,19,27, delay number is 2 people, shown in table specific as follows:

The output data that table 4 is respectively stood

Although an embodiment of the present invention has been shown and described, for the ordinary skill in the art, permissible It is understood by, these embodiments can be carried out the change of multiple equivalence without departing from the principles and spirit of the present invention Changing, revise, replace and modification, the scope of the present invention is limited by claims and equivalency range thereof.

Claims (9)

1. a BRT passenger based on Equilibrium Assignment Model goes on a journey benefit optimization method, it is characterised in that comprise the following steps:

S1, choose object of study and research range, gather relevant rudimentary information；

S2, according to relevant rudimentary information, analyze passenger demand characteristic: propose to meet the occupant classification concept of passenger's OD demand, draw The characteristic of point dissimilar passenger, including can distribution of passengers with can not distribution of passengers；

S3, by analyze dissimilar passenger characteristic, calculate the number of dissimilar passenger, determine distribution principle；

S4, according to distribution principle, calculate associated passenger cost, including passenger ride cost, be delayed waiting time cost and total Congestion costs；

S5, determine object function according to associated passenger cost, and determine the constraints of Equilibrium Assignment Model according to object function；

S6, Equilibrium Assignment Model is solved, arrange the result solved, obtain rational bus traveler assignment scheme.

A kind of BRT passenger based on Equilibrium Assignment Model goes on a journey benefit optimization method, and its feature exists In, the relevant rudimentary information in described step S1, go on a journey letter including BRT system intrinsic information, each line tower foundation information and passenger Breath, wherein

Described BRT system intrinsic information, stops public bus network numbering including site name, website quantity and website in road；

Described each line tower foundation information, including numbering, seating capacity, design capacity, greatest physical capacity, averagely often stands the service time And the mean delay waiting time, also include stopping the seating capacity of public bus network of targeted sites, number of getting off, sailing out of people Number；

Described passenger's trip information includes the travelling OD demand of passenger.

A kind of BRT passenger based on Equilibrium Assignment Model goes on a journey benefit optimization method, and its feature exists In, the analysis passenger demand characteristic in described step S2, it concretely comprises the following steps:

In S21, analysis BRT express passway, the substitutability of occupant ride plan, if OD is interval in passage by bus, then have many Bar circuit can select to take, and the circuit that its plan is taken can be substituted；If OD is interval outside passage by bus, the most selectable take Take advantage of circuit unique；

S22, general OD by bus are interval in passage, and the passenger having a plurality of circuit to take is considered as can distribution of passengers；And OD by bus Interval outside passage, taking that the unique passenger of circuit is considered as can not distribution of passengers.

A kind of BRT passenger based on Equilibrium Assignment Model goes on a journey benefit optimization method, and its feature exists In, distribution principle in described step S3 particularly as follows:

S31, determine station passenger's total number of persons:

$Q_j=q_j+\underset{j}{\&Sigma;}\underset{i=n}{\overset{m}{\&Sigma;}}{o}_{ij},\&ForAll;i\&Element;L$

Wherein, Q_jRepresent jth station station passenger's total number of persons；q_jRepresent jth station can distribution of passengers total number of persons；o_ijRepresent jth station What i-th line road was taken in plan can not distribute patronage；L represents the circuit vehicle set arrived at a station；

S32, according to different types of passenger's characteristic, pay the utmost attention to can not the demand by bus of distribution of passengers, in order to subtract to greatest extent Few total passenger delay cost；

S33, consideration vehicle heap(ed) capacityWith vehicle carrying situation, determine the i-th line road vehicles seating capacity at jth station e_ij:

$e_{ij}=c_i^m+d_{ij}-P_{ik}$

Wherein, d_ijRepresent the number of getting off on the i-th line road at jth station；P_ikRepresent kth station i-th line road sail out of passenger people Number, and the upper website that kth station is jth station；

The inseparable of i-th line road is taken in S34, the vehicle seating capacity comparing the i-th line road at jth station and the plan of jth station Join the size of patronage；Determine jth station actual take i-th line road can not distribute patronageAnd the of jth station I bar circuit can not the delay number of distribution of passengers

$o_{ij}^p=\left\{\begin{array}{ccc}{o}_{ij},& if& o_{ij}\&le;e_{ij}\\ e_{ij},& if& o_{ij}>e_{ij}\end{array}\right\}$

$o_{ij}^d=\left\{\begin{array}{ccc}0,& if& o_{ij}\&le;e_{ij}\\ o_{ij}-e_{ij},& if& o_{ij}>e_{ij}\end{array}\right\}.$

A kind of BRT passenger based on Equilibrium Assignment Model goes on a journey benefit optimization method, and its feature exists In, in described step S4, described passenger riding time cost TC₁Computational methods particularly as follows:

${\mathrm{TC}}_1=\underset{j}{\&Sigma;}\underset{i=n}{\overset{m}{\&Sigma;}}{t}_i^p(x_{ij}+o_{ij}^p),\&ForAll;i\&Element;L$

In formula, x_ijRepresent i-th line road in jth station can the number of getting on the bus of distribution of passengers；Represent the flat of i-th line road vehicles The most often stand the service time；

To can be considered as delay time at stop of distribution of passengers the mean delay time of selected circuit, and can not the delay of distribution of passengers time Between for delay time at stop of fixing circuit, accordingly, it is determined that described delay waiting time cost TC₂Computational methods particularly as follows:

${\mathrm{TC}}_2=t^D\&lsqb;q_j-\underset{j}{\&Sigma;}\underset{i=n}{\overset{m}{\&Sigma;}}{x}_{ij}\&rsqb;+\underset{j}{\&Sigma;}\underset{i=n}{\overset{m}{\&Sigma;}}{t}_i^d{o}_{ij}^d,\&ForAll;i\&Element;L$

In formula, t^DExpression can mean delay waiting time of distribution of passengers；Represent the delay waiting time on i-th line road；

Described total congestion costs TC₃Calculation method particularly as follows:

${\mathrm{TC}}_3=\underset{i=n}{\overset{m}{\&Sigma;}}{G}_i\&lsqb;P_{ij}\&rsqb;$

Wherein, congestion costs function G_i[P_(ij)] expression formula is as follows:

$G_i\&lsqb;P_{ij}\&rsqb;=\left\{\begin{array}{ccc}0,& if& P_{ij}\&le;c_i^s\\ \mathrm{\&alpha;}{\left(P_{ij}-c_i^s\right)}^{{\mathrm{\&gamma;}}_1},& if& c_i^d<P_{ij}\&le;c_i^s\\ \mathrm{\&alpha;}{\left(c_i^d-c_i^s\right)}^{{\mathrm{\&gamma;}}_1}+\mathrm{\&beta;}{\left(P_{ij}-c_i^m\right)}^{{\mathrm{\&gamma;}}_2},& if& c_i^s<P_{ij}\&le;c_i^m\end{array}\right\}$

In formula, P_ijRepresent jth station i-th line road sail out of patronage, be calculated as follows:

$P_{ij}=P_{ik}+x_{ij}+o_{ij}^p-d_{ij};$

Represent the seating capacity of i-th line road vehicles；Represent the design capacity of i-th line road vehicles；Represent i-th line The greatest physical capacity of road vehicles；α、β、γ₁、γ₂Represent the congestion costs coefficient under different condition respectively.

A kind of BRT passenger based on Equilibrium Assignment Model goes on a journey benefit optimization method, and its feature exists In, described step S5 determine object function particularly as follows: according to passenger ride cost, be delayed waiting time cost, total congestion costs And the value weight of all kinds of cost, determine object function:

MinTC=λ₁TC₁+λ₂TC₂+λ₃TC₃

λ in formula₁、λ₂、λ₃Correspond to the value weight of each cost.

A kind of BRT passenger based on Equilibrium Assignment Model goes on a journey benefit optimization method, and its feature exists In, described value weight, its value is by investigating decision, in investigation, uses by the Paired comparison method of expert estimation and by passenger Normal and the method for scales of marking, two kinds of methods combine and determine that each is worth weight, or guided that collects in APP to take advantage of by public transport Visitor wish and comprehensively determine be worth weight.

A kind of BRT passenger based on Equilibrium Assignment Model goes on a journey benefit optimization method, and its feature exists In, described step S5 determining, the constraints of Equilibrium Assignment Model is as follows:

1. decision variable x_ijSpan by vehicle capacity with preferentially distribute the patronage that can not distribute got on the bus and limited, and limit It is set to set of integers:

$0\&le;x_{ij}\&le;e_{ij}-o_{ij}^p,\&ForAll;x_{ij}\&Element;Z;$

The always number of getting on the bus of the most all types of passengers should be less than equal at station total number of persons, and ensure at station passenger's total benefit the most simultaneously Excellent, it is allowed to the existence of delay number:

$0<\underset{j}{\&Sigma;}\underset{i=n}{\overset{m}{\&Sigma;}}{x}_{ij}+\underset{j}{\&Sigma;}\underset{i=n}{\overset{m}{\&Sigma;}}{o}_{ij}^p\&le;Q_j;$

The cabin factor of the most each bar circuit is equal to the ratio sailing out of patronage and vehicle maximum logistics capacity of this circuit:And 0 ＜ p_ij≤1。

A kind of BRT passenger based on Equilibrium Assignment Model goes on a journey benefit optimization method, and its feature exists In, Equilibrium Assignment Model is solved in described step S6, use Revised genetic algorithum that Equilibrium Assignment Model is carried out Solve, comprise the steps:

S61, chromosome coding: use real coding method, by the floating point representation in individual each gene position given range, should Scope depends on that the span that decision variable arrives, individual code length depend on the number of decision content；Wherein, the model of coding Enclose and be x_ijSpan；

S62, design fitness function: genetic algorithm distinguishes individual quality by fitness function, falling passenger's totle drilling cost Number is as individual fitness value, and the individuality that i.e. passenger's cost is the least, fitness is the biggest, and individuality is the most excellent；If the fitness of individual k Function is F_k[TC(x_ij)], then have:

$F_k\&lsqb;TC\left(x_{ij}\right)\&rsqb;=\frac{1}{TC\left(x_{ij}\right)}$

S63, selection operation: from old population, form new population with certain probability selection excellent individual, Equilibrium Assignment Model is adopted With roulette method, i.e. ratio selection method, individual selected probability is directly proportional to fitness value, the individuality that fitness is the biggest, quilt The probability chosen is the biggest；Then to individual Probability p selected for i_kHave:

$p_k=\frac{F_k\&lsqb;TC\left(x_{ij}\right)\&rsqb;}{\underset{l=1}{\overset{z}{\&Sigma;}}{F}_l\&lsqb;TC\left(x_{ij}\right)\&rsqb;};$

S64, intersection and mutation operation: in order to heredity have the outstanding feature of previous generation individuality and keep population multiformity and The operation taked；Realize intersecting and mutation operation by the genetic operator that Calling MATLAB is built-in；

S65, judge whether to meet end condition: when evolutionary generation meets setting value, or algorithm is after continually evolving certain algebraically, When the fitness solved does not significantly improves, i.e. export optimum results.