CN104157150B - Novel isolated intersection traffic Signalized control method - Google Patents

Abstract

A kind of novel isolated intersection traffic Signalized control method, including: (1), the decision model of signal lights is separated with control algolithm: the mathematical model of decision-making is packaged into a function；(2), traffic occur without block up or block up red light waiting time of all vehicles will not be allowed premised on deteriorating the shortest for the purpose of propose Controlling model, this Controlling model feature is characterized as: (2.1), use the shortest Vehicle length model；(2.2), have employed optimum vehicle waiting time model；(2.3), two threshold values are proposed, for phase place Vehicle length threshold value and total Vehicle length threshold value；(2.4), the fusion of above-mentioned three elements is proposed Hybrid Control Model；(3), for the model having built up, a kind of decision making algorithm will be designed, it is achieved the traffic lights in following several traffic lights cycles are planned, obtain the value minimum that above-mentioned model is tried to achieve by a kind of feasible planning, and real-time.The present invention can dispatch common emergency situations very well, controls in real time, controls effect good.

Description

Novel isolated intersection traffic Signalized control method

Technical field

The present invention relates to the Mathematical Models of the control decision of real-time isolated intersection traffic signal lights and solve this model Algorithm.

Background technology

Recently as the economic and development of science and technology, the automobile quantity in city is on the increase, and this also brings city Congested in traffic problem.Nowadays these problems are increasingly subject to people's attention, and each city also gradually builds up some intelligent transportation Project, the information collecting device such as photographic head, sensor, geomagnetic induction coil have collected the transport information of large amount of complex, these numbers How according to acting on, point duty proposes new higher requirement.For these crossing traffic informations, the most effectively control Traffic lights make it effectively to alleviate traffic pressure and become a study hotspot in this year.

Now, Controlling model and the algorithm of most traffic light is all that model is set up and algorithm at the beginning of design Design two pieces of mixing to complete, first establish algorithm model, then design the algorithm solving this model according to this model feature. Therefore there is many problems.Wherein it is primarily present following two problem

1, because model excessively simplifies in terms of modelling, it is impossible to the emergency situations such as vehicle that well scheduling is common is dilute Less, early evening peak etc..

2, for asking effect algorithm excessively complexity not control in real time, or algorithm the most simply controls poor effect.

Summary of the invention

Present invention aim to overcome that prior art is not enough, disclose a kind of novel isolated intersection traffic Signalized control decision-making Model and algorithm, can well dispatch common emergency situations, control in real time, controls effect good.

The inventive method thinking: control decision model and the algorithm of traffic light are separated, designs one more Comprehensively signal lights decision model, redesigns a more effective real time control algorithms.

Principle: control decision model and the algorithm of live signal lamp control problem are separated by the present invention.First set Counted a kind of mixing control decision model, the Signalized control decision-making made occur without traffic conditions deteriorate on the premise of as far as possible Hommization.Secondly, devise dynamic programming algorithm complicated signal lights decision-making is become and can be tried to achieve in polynomial time.? Eventually, in the state of dynamic programming shifts, call Hybrid Control Model, and then reach the combination of model and algorithm so that can be quickly Obtain a Signalized control decision-making with certain hommization.

For solving above technical problem, the present invention adopts the following technical scheme that:

A kind of novel isolated intersection traffic Signalized control decision model and algorithm, it is characterised in that the method include as Lower step:

(1), for the decision model of signal lights being separated with control algolithm, the mathematical model of decision-making is packaged into by the present invention One function, the intersection information in each moment will submit to this mathematical model function, then can obtain a functional value and represent The traffic index at this crossing, the lowest then traffic of numerical value is the best, otherwise the poorest.So, the mathematical model of decision-making is for follow-up calculation Method design is equivalent to black box operation, and i.e. changing model does not affect later algorithm decision-making.

(2), in terms of modelling, the present invention occurs without by traffic to block up or block up and will not allow all cars premised on deteriorating The red light waiting time the shortest for the purpose of propose Controlling model.Model is as follows:

(2.1), in order to prevent from or alleviate blocking up, the present invention uses the shortest Vehicle length model.The a certain moment is calculated Going out the queuing vehicle length at each crossing, suing for peace their length is exactly the functional value of this Controlling model function.

(2.2), in order to make all car waiting time the fewest, present invention employs optimum vehicle waiting time model.To often One car just carries out timing to it when it enters vehicle waiting list, and timing time is the waiting time of this vehicle, correspondence certain In one moment, suing for peace waiting time of all of vehicle is exactly the functional value of this optimum vehicle waiting time pattern function.

(2.3), in order to merge above-mentioned two model, the present invention proposes two threshold values, i.e. phase place Vehicle length threshold value with total Vehicle length threshold value.Phase place Vehicle length threshold value is, the Vehicle length of some phase place at a crossing is proposed threshold value, threshold value Generally the Vehicle length of this phase place is in 2 to 5 times of (it is undetermined that concrete coefficient regards concrete condition) durations of its maximum green perild The most all cars are rolled away from.Total Vehicle length threshold value is that each phase place Vehicle length summation is proposed threshold value, and threshold value is the most right The all Vehicle lengths in this crossing are just enough within 2 to 5 times of times (it is undetermined that concrete coefficient regards concrete condition) in standard traffic lights cycle All cars are rolled away from.

(2.4), the fusion of above-mentioned three elements is proposed the Hybrid Control Model of the present invention.The car of each phase place of current crossing Length less than phase place Vehicle length threshold value and total Vehicle length less than using optimum vehicle to wait during total Vehicle length threshold value Time model, otherwise uses the shortest Vehicle length model.Note: be the Vehicle length of current time as the Vehicle length compared, The intersection information incoming with calling this function is unrelated.(3), for the model having built up, the present invention will design a kind of determining Plan algorithm, algorithm is planned realizing the traffic lights to following several traffic lights cycles, is obtained a kind of feasible planning to above-mentioned mould The value that type is tried to achieve is minimum.And ensure that there is higher real-time while meeting this characteristic.Algorithm steps is as follows:

(3.1), the pretreatment of algorithm.Algorithm pretreatment includes carrying out the existing intersection information data obtained at secondary Reason, the prediction of vehicle that the prediction of future car flow and green light phase are driven away by the data analysis after processing.

(3.2), algorithm model is designed an integer programming model.The present invention this according to the traffic lights display time Subsection (second) is as ultimate unit, and whole model can be with abbreviation just, and the i-th moment jth phase place traffic lights situation is P_ij, P_ijBeing 0/1 variable, 0 represents red light, and 1 represents green light.The most whole problem just can become have various P with abbreviation_ijAsk The planning of a kind of optimum so that the result of the mathematical model set up above is minimum.Certainly P_ijAlso need to meet some traffic lights Distinctive restriction.

(3.3) design dynamic programming method according to the special nature of traffic lights and solve integer programming model.If directly Solve linear programming problem, then state total amount is 2^n*m, n represents the time span of decision-making, and m represents number of phases, it can be seen that This is a np problem, and in actual traffic lights control, it is desirable to he there is the strongest precision simultaneously have to be in short-term In solve out.But this problem itself is not pure barbaric 0/1 planning problem, itself exists Some the strongest constraints, by the research of this 0/1 planning problem and the distinctive character of traffic lights, this problem is to 0/1 rule Draw model to be made that and be revised as the integer programming problem that dynamic programming can be used to solve, finally in polynomial time (in actual motion, time complexity is less than 1 second) solves out.

The innovative point of the present invention is embodied in:

1) control decision model is separated with algorithm.

2) a kind of Hybrid Control Model is designed.

3) a kind of novel decision making algorithm is designed: dynamic programming algorithm based on integer programming.

Accompanying drawing explanation

The general flow chart of Fig. 1 present invention

Fig. 2 Controlling model flow chart

Fig. 3 decision making algorithm flow chart

Formula explanation

Formula 1 the shortest Vehicle length model

The optimum vehicle waiting time model of formula 2

Formula 3 model conversion condition

Formula 4 Hybrid Control Model

Formula 5 algorithm state equation of transfer

Detailed description of the invention

Below in conjunction with accompanying drawing and formula, technical solution of the present invention is described further.

Principle: the present invention proposes Signalized control decision model and the separation of algorithm, and the most respectively to mould Type is set up and algorithm design proposes to optimize.

If Fig. 1 is the general flow chart of the present invention, first pass through and model is separated with algorithm, Controlling model is packaged into one The function of black box, then by decision making algorithm continuous calling model function subsequently, to reach decision-making purpose.

Fig. 2 is the signal lights mathematical model set up of the present invention, his the shortest Vehicle length model of based on formula (1) Optimum vehicle waiting time model with formula (2) adds the decision-making value of formula (3), is combined into a complex decision mould Type, model be mathematically represented as formula (4).

Formula (1): L_mFor total Vehicle length, L_iIt it is the Vehicle length of No. i-th phase place.

Formula (2): T_nFor total vehicle waiting time, t_iIt it is the waiting time of No. i-th vehicle.

Formula (3): L_max1For phase place Vehicle length threshold value, L_max1For total Vehicle length threshold value.

Formula (4): G (t) is pattern function, t correspondence t.

$L_m=\underset{i=1}{\overset{m}{\Σ}}{L}_i$

Formula 1 the shortest Vehicle length model

$T_n=\underset{i=1}{\overset{n}{\Σ}}{t}_i$

The optimum vehicle waiting time model of formula 2

L_i≤L_max1(1≤i≤m)∩L_m≤L_max2

Formula 3 model conversion condition

$G\left(t\right)=\left\{\begin{array}{c}\underset{i=1}{\overset{m}{\Σ}}{L}_i(L_i\≤L_{\max 1}\∩L_m\≤L_{\max 2})\\ \underset{i=1}{\overset{n}{\Σ}}{t}_i(L_i>L_{\max 1}\∪L_m>L_{\max 2})\end{array}\right.$

Formula 4 Hybrid Control Model

$F(t,i)=\left\{\begin{array}{c}\underset{g_{\min }\≤g\≤g_{\max }}{\min }\left(F\right(t-g,i-1)+\underset{k=t-g}{\overset{t}{\Σ}}G(k\left)\right)(i\≠m)\\ \underset{g_{\min }\≤g\≤g_{\max }}{\min }\left(F\right(t-g,1)+\underset{k=t-g}{\overset{t}{\Σ}}G(k\left)\right)(i=m)\end{array}\right.$

Formula 5 algorithm state equation of transfer

Such as the decision making algorithm flow process that Fig. 3 is the present invention, algorithm head relative data carries out pretreatment and to future as shown in the figure The prediction of vehicle flowrate, then by signal model feature is set up integer programming model, uses dynamic finally by model feature State planning algorithm solves.

Formula (5) is the state transition equation of dynamic programming algorithm, and wherein t represents that t, i represent current i-th crossing For green light, g is the green time at i crossing.

In sum, the whole flow process that realizes is summarised as:

1) Controlling model of signal lights is separated with decision making algorithm.

2) Controlling model is designed Hybrid Control Model.

3) decision making algorithm is designed an integer programming model and solves by dynamic programming algorithm.

Claims (1)

1. a novel isolated intersection traffic Signalized control method, it is characterised in that the method comprises the steps:

(1), the decision model of signal lights is separated with control algolithm:

The mathematical model of decision-making is packaged into a function, the intersection information in each moment is submitted to this mathematical model function, Then obtaining a functional value and represent the traffic index at this crossing, the lowest then traffic of numerical value is the best, otherwise the poorest；

(2) for the purpose of, traffic occurs without and blocks up or block up red light waiting time of all vehicles will not be allowed premised on deteriorating the shortest Proposing Controlling model, this Controlling model feature is characterized as:

(2.1), the shortest Vehicle length model is used: is calculated the queuing vehicle length at each crossing, to them a certain moment Length summation be exactly the functional value of this pattern function；

(2.2), have employed optimum vehicle waiting time model: it is just entered by each car when it enters vehicle waiting list Row timing, timing time is the waiting time of this vehicle, the corresponding a certain moment, and the waiting time summation to all of vehicle is exactly The functional value of this pattern function；

(2.3), two threshold values are proposed, for phase place Vehicle length threshold value and total Vehicle length threshold value:

Phase place Vehicle length threshold value is, the Vehicle length of some phase place at a crossing is proposed threshold value, and threshold value generally should The Vehicle length of phase place is the most enough in 2 to 5 times of durations of its maximum green perild to be rolled away from all cars；

Total Vehicle length threshold value is that each phase place Vehicle length summation is proposed threshold value, and threshold value is that vehicle all to this crossing is long Degree is the most enough within 2 to 5 times of times in standard traffic lights cycle to be rolled away from all cars；

(2.4), the fusion of above-mentioned three elements is proposed Hybrid Control Model: the Vehicle length of each phase place of current crossing is less than phase Position Vehicle length threshold value and total Vehicle length are less than using optimum vehicle waiting time model during total Vehicle length threshold value, otherwise Use the shortest Vehicle length model；(3), for the model having built up, a kind of decision making algorithm will be designed, it is achieved several to future The traffic lights planning in individual traffic lights cycle, obtains a kind of feasible planning, and the value making above-mentioned Hybrid Control Model try to achieve is minimum, and Having real-time, step is as follows: (3.1), the pretreatment of algorithm: algorithm pretreatment includes the existing intersection information data obtained Carry out after-treatment, the prediction of vehicle that the prediction of future car flow and green light phase are driven away by the data analysis after processing；

(3.2), algorithm model is designed an integer programming model: according to the least unit second conduct of traffic lights display time Ultimate unit, whole model can be with abbreviation just, and the i-th moment jth phase place traffic lights situation is P_ij, P_ijIt is one 0/1 to become Amount, 0 represents red light, and 1 represents green light, and whole problem reduction becomes to have various P_ijAsk the planning of a kind of optimum so that described The result of calculation of integer programming model is minimum；

(3.3) design dynamic programming method and solve integer programming model.