CN101266718A - Traffic optimization control method based on intersection group - Google Patents

Abstract

The invention discloses an optimal control method for intersection group traffic flow, which resolves the defect that the smallest unit of current urban traffic control is a single intersection, which cannot satisfy the needs of actual traffic control, can realize the optimization of local traffic flow, and provides helps for micro control of traffic flow. The steps of the method are:(1) real time acquiring actual traffic flow of ''intersection group'' on roads through a ground induction coil which has been paved in a city; (2) constructing an optimal control model with a smart approach and inferring and optimizing the parameters of a controller according to real-time traffic information;(3) the controller carrying out optimal control for traffic flow according to the control model.

Description

Traffic optimization control method based on the crossing group

Technical field

The present invention relates to the control method in a kind of traffic flow control technology field, specifically be based on the traffic optimization control method of crossing group.

Background technology

Traditional urban transportation current control minimum unit is single intersection, practice shows, under the less situation of vehicle flowrate, it is feasible that intelligent control algorithm is taked in an isolated crossing, yet, the mutual close together in a plurality of intersections in urban traffic network under the situation that vehicle flowrate is bigger in the network, still adopts identical algorithm just dying.At this situation, the present invention makes up the crossing in the transportation network effectively and divides, phase mutual edge distance intersection far away can be thought to isolate, two to three very near intersections of distance then combine them, do as a whole control, this integral body just is called " crossing group ".

Find through retrieval: Huang Huixian, Shi Zhongke, " real-time genetic algorithm optimal control that the oral sex of the single cross cross road is through-flow ", the system engineering theory and practice, 2001.3:102-106 prior art; Liu Zhiyong, Zhu Jing etc. " the leggy fuzzy control of single cross prong ", information and control, 1999.28 (6): 453-458; Chen Shuyan, Chen Senfa etc., " the real-time fuzzy control of the leggy of single intersection traffic ", the system engineering theory and practice, 2003.1:110-115; Fan Ailong, Marvin's pavilion etc., " prediction of single cross prong mixed traffic flow and signal timing dial research ", Liaoning University's journal, 2007.27 (4): 234-237.In the prior art of mentioning in the above-mentioned document, about single-point traffic flow control all with single intersection as minimum unit, do not consider the correlativity when the magnitude of traffic flow is big between the very near crossing of distance, therefore, a plurality of crossings combine as one " crossing group " be optimized control be characteristics of the present invention.

Summary of the invention

The objective of the invention is to overcome traditional cities traffic control minimum unit is the deficiency that actual traffic control needs can not be satisfied in single intersection, and a kind of traffic optimization control method based on " crossing group " is provided.At the actual needs of traffic engineering, two to three close crossings of distance on the city thoroughfare are combined and control, can realize the optimization of local traffic stream, the traffic flow micromanagement is offered help.

The present invention is achieved through the following technical solutions, and may further comprise the steps:

1. utilize existing information acquisition means such as ground inductive coil, gather " crossing group " real-time traffic flow information;

2. set up optimizing control models with intelligent method, according to the parameter of reasoning of real-time traffic stream information and optimal controller;

Step 1., concrete grammar is: gather the actual traffic vehicle flowrate of " crossing group " on the urban road in real time according to the ground inductive coil that modern city has laid.

Step 1., the described existing information acquisition means such as ground inductive coil of utilizing, gathering " crossing group " real-time traffic flow information is meant: according to the existing information acquisition means of modern city, particularly the traffic flow information of whole " the crossing group " of being gathered by the ground inductive coil gathers up.

Step 1., described " crossing group " is meant: with distance adjacent in the city thoroughfare closer (generally being to be no more than 800m) and two to three intersections with correlativity node unit (this paper is the example discussion with three crossings) as trunk roads traffic flow control, this node unit just is referred to as " crossing group ".Regulation: to be leftmost crossing be no more than 1600m to the distance at rightmost crossing to the inner distance of " crossing group ".In this case, can regard trunk roads in city as with " crossing group " be the set of a plurality of node units of node, as shown in Figure 1.

Step 2. in, describedly set up optimizing control models with intelligent method, parameter according to Real-time Traffic Information reasoning and optimal controller is meant: total vehicle flowrate in " crossing group " future has direct influence to the size of next signal control cycle T, if total vehicle flowrate in " crossing group " future is big, then require the length that the cycle tries one's best, if total vehicle flowrate in " crossing group " future is little, then the length requirement to the cycle is then opposite. based on this idea, come the T value of fuzzy reasoning next cycle with the wagon flow variable quantity in total vehicle flowrate of estimating and nearly two cycles, the reasoning block diagram as shown in Figure 2, simultaneously, foundation is the Optimization Model of target with vehicle mean delay minimum, adopt genetic algorithm optimization to obtain the optimal control parameter of " crossing group ", concrete step is as follows:

1. obfuscation

To estimate total flow q, fluctuations in discharge amount Δ q, that period T is divided into seven fuzzy subsets respectively is as follows:

Q={ seldom, and is less, few, medium, many, more, a lot }

Δ q={ is negative big, and is negative little in bearing, and zero, just little, the center, honest

T={ is very short, and is shorter, short, medium, long, longer, very long }

Its domain is divided as follows:

$q=\left\{\begin{array}{c}\mathrm{0,1,2,3,4,5,6,7,8,9,10,11},\\ \mathrm{12,13,14,15,16,17,18,19,20}\end{array}\right\}$ Δq＝{-5，-4，-3，-2，-1，0，1，2，3，4，5}

T＝{0，1，2，3，4，5，6，7，8}

In fact, the actual value of estimating total flow q, fluctuations in discharge amount Δ q, period T should be q ∈ [q _Min, q _Max], Δ q ∈ [20 ,+20], T ∈ [40,150]. quantizing factor K ₁, K ₂With scale factor K ₃Value be respectively:

$K_1=\frac{q_{\max }-q_{\min }}{20},$ $K_2=\frac{20-(-20)}{5-(-5)}=4,$ $K_3=\frac{150-40}{8-0}=13.75.$

Through repetition test, input, each fuzzy subset's of output variable subordinate function is all chosen triangular function such as Fig. 4 shows.

2. fuzzy rule, reasoning and reverse gelatinization

According to estimating total flow q, fluctuations in discharge amount Δ q through the fuzzy set after the obfuscation, we adopt the fuzzy inference rule of if x is A andy is B then z is C form to carry out reasoning, the fuzzy relation R that obtains synthesizing, thereby obtain fuzzy rule base. then, with reality estimate flow q, fluctuations in discharge amount Δ q value is input to controller and removes to look into fuzzy reasoning table, obtain the fuzzy subset of period T, carry out the reverse gelatinization, obtain corresponding accurately amount T by weighted mean ^*:

$T^{*}=\frac{\underset{j=1}{\overset{9}{\mathrm{\Σ}}}\mathrm{\μ}\left(T_j\right)*T_j}{\underset{j=1}{\overset{9}{\mathrm{\Σ}}}\mathrm{\μ}\left(T_j\right)}---\left(1\right)$

T ^*Value can't directly take back usefulness, need passing ratio factor conversion can be applied in the optimal control of back and go, in fact the T value is:

T＝40+K ₃*T ^* (2)

3. calculate optimum split

Is that target is set up Optimization Model to " crossing group " with vehicle mean delay minimum, and the split of choosing " crossing group " is a variable.For the simplification problem, with g ₃, g ₄, g ₅Be independent variable, (wherein, $g_1=1-\underset{m=2}{\overset{5}{\mathrm{\Σ}}}{g}_m,$ g ₂For known) be ternary minimizing problem with four variable minimizing problem reductions, the expression formula (3) of " crossing group " average vehicle delay can be expressed as:

d(t)＝P(g ₃，g ₄，g ₅)(3)

Then objective function and constraint condition are:

min?d(t)＝P(g ₃，g ₄，g ₅)

$s.t.\left\{\begin{array}{c}\frac{10s}{T}\≤g_3\≤\frac{T-t_2-30s}{T}\\ \frac{10s}{T}\≤g_4\≤\frac{T-t_2-30s}{T}\\ \frac{10s}{T}\≤g_5\≤\frac{T-t_2-30s}{T}\end{array}\right\}---\left(4\right)$

(4) formula is the minimizing problem, adopts genetic algorithm to be optimized and need convert maximum problem to, thereby exist objective function to the fitness function mapping problems, gets following conversion:

f＝c-αd(t)(5)

Wherein, f is a fitness function, and c makes f get a constant of positive number, and α is a conversion coefficient, and its value increases gradually along with the convergence of desired value.

Through the conversion of formula (5), ask objective function minimum problem to be converted into and ask fitness function f maximum problem, adopt genetic algorithm optimization to obtain optimal control parameter (algorithm block diagram is seen Fig. 3).

4. genetic algorithm optimization step:

The genetic algorithms use real coding is with string [c ₁, b ₁, a ₁] chromosome of expression, [a ₁], [b ₁], [c ₁] represent the timing time t of phase place 1, phase place 2, phase place 3 respectively ₁, t ₂, t ₃, and generating initial population and must consider to satisfy bound for objective function through intersection, the new part kind group time of mutation operator generation.

Adopt genetic algorithm as follows to the algorithm of this problem optimizing:

1) initialization, setting cycle value, the vehicle fleet size of this cycle and last one-period, population number, chromosome length, the total algebraically of iteration, the probability that duplicates, intersects, makes a variation;

2) according to the data in nearest two cycles, use discreet value and variable quantity fuzzy reasoning next cycle value.Adopt real number coding method, produce predetermined population number purpose chromosome at random;

3) randomly draw in candidate solution colony according to predetermined crossing-over rate and several candidate solution is carried out interlace operation;

4) extract according to predetermined hybridization, aberration rate that candidate solution intersects, mutation operation;

5) calculate the target function value of each candidate solution, eliminate the poorest person of target function value in the candidate solution according to selected mortality, generation from objective function is separated is preferably put in institute's omission.

6) judge whether predetermined iterations,, otherwise change (3) if then continue next step.

7) calculate each phase place timing by the optimal-adaptive value;

8) estimate vehicle flowrate of following one-period, change 2) enter next loop cycle.

The invention has the beneficial effects as follows: micromanagement has proposed a new thinking to urban transportation, and the pressure that alleviates present urban transportation is had practical significance.Single cross cross road mouth is the minimum basic unit of urban traffic control, can improve the traffic flow of local traffic unit and trunk roads by this method, improve the traffic capacity of whole piece trunk roads, reduce the stop frequency of vehicle on the trunk roads, effectively avoid blocking up of trunk roads, for the modern city traffic control provides the favorable service function.

Description of drawings

Fig. 1 is " crossing group " of the present invention synoptic diagram;

Fig. 2 is the schematic block diagram in fuzzy reasoning cycle of the present invention;

Fig. 3 is " crossing group " of the present invention traffic flow optimal control figure;

Fig. 4 is the division figure of actual " crossing group ";

Fig. 5 is that " crossing group " phase place is divided synoptic diagram.

Embodiment:

The invention will be further described below in conjunction with accompanying drawing and embodiment.

Present embodiment has provided detailed embodiment and process being to implement under the prerequisite with the technical solution of the present invention, but protection scope of the present invention is not limited to following embodiment.

A kind of traffic optimization control method based on " crossing group ", as Fig. 1, Fig. 2, Fig. 3, Fig. 4, shown in Figure 5, its step is:

(1) the ground inductive coil that utilizes the city to lay is gathered the actual traffic vehicle flowrate of crossing group on the urban road in real time;

(2) set up optimizing control models with intelligent method, according to the parameter of Real-time Traffic Information reasoning and optimal controller; Its method is: the total vehicle flowrate that utilizes crossing group future has direct influence to the size of next signal control cycle T, if total vehicle flowrate in crossing group future is big, then require the length that the cycle tries one's best, if total vehicle flowrate in crossing group future is little, then the length requirement to the cycle is then opposite; Come the T value of fuzzy reasoning next cycle with the wagon flow variable quantity in total vehicle flowrate of estimating and nearly two cycles; Simultaneously, setting up with vehicle mean delay minimum is the Optimization Model of target, adopts genetic algorithm optimization to obtain the optimal control parameter of crossing group;

(3) controller is optimized signal lamp control traffic flow according to controlling models.

In the described step (1), described crossing group is meant with distance adjacent in the city thoroughfare less than 800m and two to three intersections with a correlativity node unit as trunk roads traffic flow control; To be leftmost crossing be no more than 1600m to the distance at rightmost crossing to the inner distance of crossing group, thereby with the set as a plurality of node units of the trunk roads in city.

In the described step (2), the process of the T value of fuzzy reasoning next cycle is, by estimating total flow q, fluctuations in discharge amount Δ q, carries out obfuscation, fuzzy reasoning, sharpening successively under rule base instructs, and then controlled period T; Its concrete steps are:

The a obfuscation

To estimate total flow q, fluctuations in discharge amount Δ q, that the signal controlling period T is divided into seven fuzzy subsets respectively is as follows:

Q={ seldom, and is less, few, medium, many, more, a lot }

Δ q={ is negative big, and is negative little in bearing, and zero, just little, the center, honest

T={ is very short, and is shorter, short, medium, long, longer, very long }

Its domain is divided as follows:

$q=\left\{\begin{array}{c}\mathrm{0,1,2,3,4,5,6,7,8,9,10,11},\\ \mathrm{12,13,14,15,16,17,18,19,20}\end{array}\right\}$ Δq＝{-5，-4，-3，-2，-1，0，1，2，3，4，5}

T＝{0，1，2，3，4，5，6，7，8}

In fact, the actual value of estimating total flow q, fluctuations in discharge amount Δ q, period T should be q ∈ [q _Min, q _Max], Δ q ∈ [20 ,+20], T ∈ [40,150]. quantizing factor K ₁, K ₂With scale factor K ₃Value be respectively:

$K_1=\frac{q_{\max }-q_{\min }}{20},$ $K_2=\frac{20-(-20)}{5-(-5)}=4,$ $K_3=\frac{150-40}{8-0}=13.75.$

Input, each fuzzy subset's of output variable subordinate function is all chosen triangular function;

B fuzzy reasoning and sharpening

According to estimating total flow q, fluctuations in discharge amount Δ q through the fuzzy set after the obfuscation, adopt the fuzzy inference rule of if x is A and y isB then zis C form to carry out reasoning, the fuzzy relation R that obtains synthesizing, thus obtain fuzzy rule base; Then, with actual estimate flow q, fluctuations in discharge amount Δ q value is input to controller and removes to look into fuzzy reasoning table, obtains the fuzzy subset of period T, carries out the reverse gelatinization by weighted mean, obtains the corresponding T that accurately measures ^*:

$T^{*}=\frac{\underset{j=1}{\overset{9}{\mathrm{\Σ}}}\mathrm{\μ}\left(T_j\right)*T_j}{\underset{j=1}{\overset{9}{\mathrm{\Σ}}}\mathrm{\μ}\left(T_j\right)}$

T ^*Value can't directly take back usefulness, need passing ratio factor conversion can be applied in the optimal control of back and go, in fact the T value is:

T＝40+K ₃*T ^*。

In the described step (3), optimizing signal lamp control traffic flow process is, from certain a moment, write down in each control cycle wagon flow through the quantity of crossing group, calculate the changing value of preceding two cycle vehicle flowrates since the 3rd cycle, of the input of the wagon flow variable quantity in current vehicle flowrate and preceding two cycles as controller;

Be input in the model of mind of fuzzy reasoning gathering the vehicle flowrate that comes, obtain the numerical value in next cycle, adopt genetic algorithm optimization on average to prolong the function that minimum is a target, obtain optimum split, period allocated is gone down to realize control according to this split with vehicle.

Calculating optimum split method is, is that target is set up Optimization Model to the crossing group with vehicle mean delay minimum, and the split of choosing " crossing group " is a variable; For the simplification problem, with g ₃, g ₄, g ₅Be independent variable, (wherein, $g_1=1-\underset{m=2}{\overset{5}{\mathrm{\Σ}}}{g}_m,$ g ₂For known) be ternary minimizing problem with four variable minimizing problem reductions, the expression formula (3) of " crossing group " average vehicle delay can be expressed as:

d(t)＝P(g ₃，g ₄，g ₅)(3)

Then objective function and constraint condition are:

min?d(t)＝P(g ₃，g ₄，g ₅)

$s.t.\left\{\begin{array}{c}\frac{10s}{T}\≤g_3\≤\frac{T-t_2-30s}{T}\\ \frac{10s}{T}\≤g_4\≤\frac{T-t_2-30s}{T}\\ \frac{10s}{T}\≤g_5\≤\frac{T-t_2-30s}{T}\end{array}\right\}---\left(4\right)$

(4) formula is the minimizing problem, adopts genetic algorithm to be optimized and need convert maximum problem to, thereby exist objective function to the fitness function mapping problems, gets following conversion:

f＝c-αd(t) (5)

Wherein, f is a fitness function, and c makes f get a constant of positive number, and α is a conversion coefficient, and its value increases gradually along with the convergence of desired value;

Through the conversion of formula (5), ask objective function minimum problem to be converted into and ask fitness function f maximum problem, adopt genetic algorithm optimization to obtain optimal control parameter;

The genetic algorithm optimization step:

The genetic algorithms use real coding is with string [c ₁, b ₁, a ₁] chromosome of expression, [a ₁], [b ₁], [c ₁] represent the timing time t of phase place 1, phase place 2, phase place 3 respectively ₁, t ₂, t ₃, and generating initial population and must consider to satisfy bound for objective function through intersection, the new part kind group time of mutation operator generation.

Described genetic algorithm is as follows to the algorithm of this problem optimizing:

1) initialization, setting cycle value, the vehicle fleet size of this cycle and last one-period, population number, chromosome length, the total algebraically of iteration, the probability that duplicates, intersects, makes a variation;

2) according to the data in nearest two cycles, use discreet value and variable quantity fuzzy reasoning next cycle value; Adopt real number coding method, produce predetermined population number purpose chromosome at random;

3) randomly draw in candidate solution colony according to predetermined crossing-over rate and several candidate solution is carried out interlace operation;

4) extract according to predetermined hybridization, aberration rate that candidate solution intersects, mutation operation;

5) calculate the target function value of each candidate solution, eliminate the poorest person of target function value in the candidate solution according to selected mortality, generation from objective function is separated is preferably put in institute's omission;

6) judge whether predetermined iterations,, otherwise change (3) if then continue next step;

7) calculate each phase place timing by the optimal-adaptive value;

8) estimate vehicle flowrate of following one-period, change (2) and enter next loop cycle.

Embodiment 1:

Present embodiment, the Jingshi Road of choosing the Jinan City be as research object, and young East Road, three crossings that Qianfo Mount road and Lishan Mountain road and Jingshi Road intersect are divided into " crossing group " and implement control.Second " crossing group " of Jingshi Road, Jinan City comprises that crossing, Lishan Mountain, crossing, Qianfo Mount and Shun plough the crossing, and the distance between the crossing is respectively 700m and 400m in twos, meets the condition of " crossing group ", as shown in Figure 4.

Traffic flow model that should " crossing group " as shown in Figure 1, it comprises three single cross cross road mouths, uses C respectively ₁, C ₂, C ₃Represent that each single cross cross road mouth has east, south, west, north four direction in " crossing group ", each direction all exists right lateral, craspedodrome, three lane flows of left lateral, and wherein, east-west direction is the trunk roads direction.Every there are two inductive coils in the track, and a coil is embedded in the stop line place, is called the stop line inductive coil, and another is embedded in the place apart from stop line coil 100～160M, is called the upstream inductive coil.The stop line inductive coil is used for detecting and leaves this regional vehicle flowrate, and the upstream inductive coil is used for measuring the vehicle flowrate that enters zone between two coils, by detected car flow information, for the control of traffic flow provides data necessary.

The phase place of " crossing group " is divided with single intersection and is distinguished to some extent, and the phase place that contains " the crossing group " of three intersections is divided as shown in Figure 5.Provide five phase places among the figure, four phase places of the division of first and third, four, five phase places and single cross cross road mouth are duplicate, the thing of representing trunk roads respectively keep straight on mutually and the thing left lateral mutually, phase and north and south left lateral are kept straight on mutually in the north and south of branch road direction.Second phase place is owing to the characteristics that " crossing group " has divide out, expression craspedodrome left lateral phase, the situation of wagon flow be the 1. number crossing keep straight on from east to west and by east orientation south left lateral and the 3. number crossing keep straight on from west to east and by west left lateral northwards, and the 2. number access connection traffic flow not influenced by this.The vehicle of wishing single cross cross road mouth as us can be fully by the crossing after green light finishes, when wishing that also " crossing group " green light on thing craspedodrome direction finishes, the vehicle that enters " crossing group " can not have the ground of delay by whole " crossing group ", the bright clearance substantially to vehicle of green light that promptly guarantees the exit finishes, and is exactly second phase place during this period of time.

The 2. the intersection be a common single cross cross road mouth, it has four phase places is respectively thing craspedodrome phase, thing left-hand rotation phase, north and south keep straight on mutually with the north and south left lateral mutually.

The concrete steps of present embodiment are as follows:

1. from certain a moment, write down the quantity of the interior wagon flow process of each phase time " crossing group " in each control cycle, the wagon flow quantity of passing through in all phase times added up just obtain in the whole cycle wagon flow quantity by " crossing group ", calculate the changing value of preceding two cycle vehicle flowrates since the 3rd cycle, of the input of the wagon flow variable quantity in current vehicle flowrate and preceding two cycles as controller.

2. be input in the model of mind of fuzzy reasoning gathering the vehicle flowrate that comes, obtain the numerical value in next cycle, adopt genetic algorithm optimization on average to prolong the function that minimum is a target, obtain optimum split, period allocated is gone down to realize control according to this split with vehicle.

The The whole control experiment is controlled respectively at three crossings and " crossing group " control has been carried out respectively ten cycles, and the result who obtains as shown in Table 1 and Table 2.

Comparison by to two table data is very easy to find, and the volume of traffic under two kinds of situations of branch road direction changes little, almost is the same; The volume of traffic under two kinds of situations of trunk roads direction then has difference, and the volume of traffic during than three separate controls of belisha beacon has improved about 6.90% as " crossing group " volume of traffic when carrying out the signal lamp optimal control for three crossings.This has illustrated, three crossings are optimized signal lamp control traffic flow as one " crossing group ", will improve the traffic of trunk roads direction, helps dredging and circulating of trunk roads direction wagon flow, and is not very big to the vehicle flowrate influence of branch road direction.

Volume of traffic when table 1 is node with all intersections

Table 2 is with " crossing group " volume of traffic when being node

Embodiment 2

Present embodiment, the Jingshi Road of choosing the Jinan City is as research object, the whole piece Jingshi Road is divided into a plurality of " crossing groups " implements control, its division methods is 5 " crossing groups ", and promptly main road crossing, mountain Shi Donglu crossing to mountain, Shun plough crossing to crossing, Lishan Mountain road, Weiyi Road crossing to crossing, people's livelihood street, latitude 12 crossings to latitude five tunnel crossings and crossing, the Nan Xin village extremely through seven tunnel crossings (Fig. 4 represents second " crossing group ").

Add test carriage from the east porch of Jingshi Road, set out respectively in the above two kinds of cases 10 times, obtain the stop frequency result of table 3, table 4.

By the comparison of his-and-hers watches 3 and table 4 data, can see obviously that 10 average stop frequencies that set out are 3.9 times under first kind of situation, average stop frequency is 3.1 times under second kind of situation, second kind of situation reduced 0.8 time than first kind of average stop frequency of situation.This explanation has not only improved the volume of traffic of main line direction but also has widened green wave band to a certain extent with " crossing group " node unit as trunk roads, makes the operation of vehicle more unimpeded.

Stop frequency when table 3 is node with all intersections

Table 4 is with " crossing group " stop frequency when being node

Claims (6)

1. crossing group traffic optimization control method is characterized in that its step is:

(1) the ground inductive coil that utilizes the city to lay is gathered the actual traffic vehicle flowrate of crossing group on the urban road in real time;

(2) set up optimizing control models with intelligent method, according to the parameter of Real-time Traffic Information reasoning and optimal controller; Its method is: the total vehicle flowrate that utilizes crossing group future has direct influence to the size of next signal control cycle T, if total vehicle flowrate in crossing group future is big, then require the length that the cycle tries one's best, if total vehicle flowrate in crossing group future is little, then the length requirement to the cycle is then opposite; Come the T value of fuzzy reasoning next cycle with the wagon flow variable quantity in total vehicle flowrate of estimating and nearly two cycles; Simultaneously, setting up with vehicle mean delay minimum is the Optimization Model of target, adopts genetic algorithm optimization to obtain the optimal control parameter of crossing group;

(3) controller is optimized the control traffic flow according to controlling models.

2. crossing as claimed in claim 1 group traffic optimization control method, it is characterized in that, in the described step (1), described crossing group is meant with distance adjacent in the city thoroughfare less than 800m and two to three intersections with a correlativity node unit as trunk roads traffic flow control; To be leftmost crossing be no more than 1600m to the distance at rightmost crossing to the inner distance of crossing group, thereby with the set as a plurality of node units of the trunk roads in city.

3. crossing as claimed in claim 1 group traffic optimization control method, it is characterized in that, in the described step (2), the process of the T value of fuzzy reasoning next cycle is, by estimating total flow q, fluctuations in discharge amount Δ q, under instructing, rule base carries out obfuscation, fuzzy reasoning, sharpening successively, and then controlled period T; Its concrete steps are:

The a obfuscation

To estimate total flow q, fluctuations in discharge amount Δ q, that the signal controlling period T is divided into seven fuzzy subsets respectively is as follows:

Q={ seldom, and is less, few, medium, many, more, a lot }

Δ q={ is negative big, and is negative little in bearing, and zero, just little, the center, honest

T={ is very short, and is shorter, short, medium, long, longer, very long }

Its domain is divided as follows:

$q=\left\{\begin{array}{c}\mathrm{0,1,2,3,4,5,6,7,8,9,10,11},\\ \mathrm{12,13,14,15,16,17,18,19,20}\end{array}\right\}$ Δq＝{-5，-4，-3，-2，-1，0，1，2，3，4，5}

T＝{0，1，2，3，4，5，6，7，8}

In fact, the actual value of estimating total flow q, fluctuations in discharge amount Δ q, period T should be q ∈ [q _Min, q _Max], Δ q ∈ [20 ,+20], T ∈ [40,150]. quantizing factor K ₁, K ₂With scale factor K ₃Value be respectively:

$K_1=\frac{q_{\max }-q_{\min }}{20},$ $K_2=\frac{20-(-20)}{5-(-5)}=4,$ $K_3=\frac{150-40}{8-0}=13.75.$

Input, each fuzzy subset's of output variable subordinate function is all chosen triangular function;

B fuzzy reasoning and sharpening

According to estimating total flow q, fluctuations in discharge amount Δ q through the fuzzy set after the obfuscation, adopt the fuzzy inference rule of if x is A and y isB then z is C form to carry out reasoning, the fuzzy relation R that obtains synthesizing, thus obtain fuzzy rule base; Then, with actual estimate flow q, fluctuations in discharge amount Δ q value is input to controller and removes to look into fuzzy reasoning table, obtains the fuzzy subset of period T, carries out the reverse gelatinization by weighted mean, obtains the corresponding T that accurately measures ^*:

$T^{*}=\frac{\underset{j=1}{\overset{9}{\mathrm{\Σ}}}\mathrm{\μ}\left(T_j\right)*T_j}{\underset{j=1}{\overset{9}{\mathrm{\Σ}}}\mathrm{\μ}\left(T_j\right)}$

T ^*Value can't directly take back usefulness, need passing ratio factor conversion can be applied in the optimal control of back and go, in fact the T value is:

T＝40+K ₃*T ^*。

4. crossing as claimed in claim 1 group traffic optimization control method, it is characterized in that, in the described step (3), optimizing signal lamp control traffic flow process is, from certain a moment, write down in each control cycle wagon flow through the quantity of crossing group, calculate the changing value of preceding two cycle vehicle flowrates since the 3rd cycle, of the input of the wagon flow variable quantity in current vehicle flowrate and preceding two cycles as controller;

Be input in the model of mind of fuzzy reasoning gathering the vehicle flowrate that comes, obtain the numerical value in next cycle, adopt genetic algorithm optimization on average to prolong the function that minimum is a target, obtain optimum split, period allocated is gone down to realize control according to this split with vehicle.

5. crossing as claimed in claim 4 group traffic optimization control method is characterized in that, calculates optimum split method to be, and be that target is set up Optimization Model to the crossing group with vehicle mean delay minimum, the split of choosing " crossing group " is a variable; For the simplification problem, with g ₃, g ₄, g ₅Be independent variable, (wherein, $g_1=1-\underset{m=2}{\overset{5}{\mathrm{\Σ}}}{g}_m,$ g ₂For known) be ternary minimizing problem with four variable minimizing problem reductions, the expression formula (3) of " crossing group " average vehicle delay can be expressed as:

d(t)＝P(g ₃，g ₄，g ₅) (3)

Then objective function and constraint condition are:

min?d(t)＝P(g ₃，g ₄，g ₅)

$s.t.\left\{\begin{array}{c}\frac{10s}{T}\≤g_3\≤\frac{T-t_2-30s}{T}\\ \frac{10s}{T}\≤g_4\≤\frac{T-t_2-30s}{T}\\ \frac{10s}{T}\≤g_5\≤\frac{T-t_2-30s}{T}\end{array}\right\}---\left(4\right)$

(4) formula is the minimizing problem, adopts genetic algorithm to be optimized and need convert maximum problem to, thereby exist objective function to the fitness function mapping problems, gets following conversion:

f＝c-αd(t) (5)

Wherein, f is a fitness function, and c makes f get a constant of positive number, and α is a conversion coefficient, and its value increases gradually along with the convergence of desired value;

Through the conversion of formula (5), ask objective function minimum problem to be converted into and ask fitness function f maximum problem, adopt genetic algorithm optimization to obtain optimal control parameter;

The genetic algorithm optimization step:

The genetic algorithms use real coding is with string [c ₁, b ₁, a ₁] chromosome of expression, [a ₁], [b ₁], [c ₁] represent the timing time t of phase place 1, phase place 2, phase place 3 respectively ₁, t ₂, t ₃, and generating initial population and must consider to satisfy bound for objective function through intersection, the new part kind group time of mutation operator generation.

6. crossing as claimed in claim 5 group traffic optimization control method is characterized in that described genetic algorithm is as follows to the algorithm of this problem optimizing:

1) initialization, setting cycle value, the vehicle fleet size of this cycle and last one-period, population number, chromosome length, the total algebraically of iteration, the probability that duplicates, intersects, makes a variation;

2) according to the data in nearest two cycles, use discreet value and variable quantity fuzzy reasoning next cycle value; Adopt real number coding method, produce predetermined population number purpose chromosome at random;

3) randomly draw in candidate solution colony according to predetermined crossing-over rate and several candidate solution is carried out interlace operation;

4) extract according to predetermined hybridization, aberration rate that candidate solution intersects, mutation operation;

5) calculate the target function value of each candidate solution, eliminate the poorest person of target function value in the candidate solution according to selected mortality, generation from objective function is separated is preferably put in institute's omission;

6) judge whether predetermined iterations,, otherwise change (3) if then continue next step;

7) calculate each phase place timing by the optimal-adaptive value;

8) estimate vehicle flowrate of following one-period, change (2) and enter next loop cycle.

Images