CN103390196B - Bullet train operational plan establishment optimization method and the system of Complex Constraints condition - Google Patents

Abstract

The invention discloses a kind of bullet train operational plan establishment optimization method and system of Complex Constraints condition, method comprises the travel line paving time of drawing is divided into multiple initial time module; Each time module is carried out to service chart structure optimization; Output optimized operation figure. System comprises time module division device, service chart structure optimization device and output device. The present invention, by having realized hourage from the angle of service chart structure optimization and minimize the doulbe-sides' victory between the optimization aim of maximum these two the phase containing of capacity, dynamically realizes the reasonably optimizing of service chart.

Description

Bullet train operational plan establishment optimization method and the system of Complex Constraints condition

Technical field

The present invention relates to train diagram establishment. More specifically, the present invention relates to bullet train operational plan establishment optimization method and the system of Complex Constraints condition.

Background technology

Train operation plan (or being called service chart) is on the basis of the train service frequency specifying in starting scheme, the scheme that stops, according to fixing transport resource (such as road network structure) and mobile transport resource (such as EMUs, crew member) attribute, determine train starting, the passing through order and arriving and leaving moment in detail of terminal station and approach station.

High speed rail train operation planning optimization refers to and is meeting under the constraintss such as all kinds of activity duration standards of high-speed railway, determine that each the train operating on this circuit in starting scheme arrives AT STATION, sets out and pass through the activity duration, ensureing that passenger train starts, in Zhongdao, makes selected service chart optimization aim reach optimum within the reasonable time.

The establishment problem of high speed rail train operation plan is complicated a, multiple constraint, various dimensions, multiobject extensive Solve problems. By the research and practice of in a large number service chart being worked out, can sum up 3 keys: on (1) high-speed railway, the travel line of friction speed train mutually restriction can cause that trip's prompt drop is low; (2) what difference stopped scheme train will have different service chart structures to sending out order, considers that the paving of the mixed race pattern Train of high middling speed is drawn order, and that should take into account bullet train is favourablely dealt into the linking in existing time on the wire of moment and medium trains. (3) train initially layout scheme rationally whether, directly affect Compilation Quality and the actual use value of route map of train, and relevant with wagon flow source with locomotive routing, station equipment ability, station operation progress.

High Speed Railway Train Diagram structure (the namely Rankine-Hugoniot relations between travel line) has four kinds of forms of expression: (1) service chart can represent by form, is timetable; (2) service chart can represent by the form of matrix, mainly contains: the scheme that stops matrix, dwell time matrix, section operation time matrix, overtaking matrix; (3) train travelling process can discretely turn to train AT STATION with interval to outbreak industry, therefore can represent with discrete event dynamic system; (4) service chart can represent with " two-dimension time-space " schematic form, is exactly the service chart of traditional sense. Therefore it is very necessary and important optimizing High Speed Railway Train Diagram structure.

In method of the prior art, external method comprises: (1) PESP model and mutation thereof based on periodic duty figure; (2), on traditional service chart model basis, improve derivation algorithm efficiency; (3) service chart harmony. Domestic method more typically has: (1) service chart is started pattern, as the PESP model based on periodic duty figure; (2) service chart scheduling algorithm, mainly contains sequencing optimized algorithm; (3) service chart optimization and adjustment, has Rolling optimal strategy and hierarchic parallel strategy; (4) service chart stability of equilibrium, as the application of maximum algebra method; (5) the evaluation aspect of service chart, as set up assessment indicator system and the computational methods of High Speed Railway Train Diagram.

To sum up current research both domestic and external mostly solves by setting up mathematical programming model, because China Express Railway has: " many, the line length of point, wide, passenger flow skewness " is difficult to try to achieve optimal solution on the basis of existing research. For service chart optimization problem, be difficult to single-threaded optimizing by mathematic programming methods, what mathematic programming methods was tried to achieve is static optimal solution, due to the dynamic of system, is difficult to real sorcerer. In addition, the factor involving due to system is many, causes variable number, constraints to increase, and makes the iterative process that solves complicated, has greatly reduced solution efficiency, can not meet requirement of real-time, on the basis of existing research, cannot try to achieve optimal solution. Thereby, must seek a kind of new optimization method.

Summary of the invention

The object of the invention is to provide a kind of bullet train operational plan establishment optimization method and system of Complex Constraints condition, dynamically realizes the reasonably optimizing of service chart.

Concrete technical scheme is as follows:

A bullet train operational plan establishment optimization method for Complex Constraints condition, comprises the following steps:

The travel line paving time of drawing is divided into multiple initial time module;

Each time module is carried out to service chart structure optimization;

Output optimized operation figure.

Describedly each time module carried out to service chart structure optimization further comprise:

S1, to train dispatch a car combination be optimized;

S2, to current time module dispatch a car sequence be optimized;

S3, obtain the minimum initial launch figure of conflict as alternative by travel line structure optimization;

S4, dissolve by described alternative is conflicted and obtain the optimized operation figure of this time module;

S5, judge whether exist do not optimize time module, as exist enter S6, perform step three output optimized operation figure if do not exist;

S6, using this time module of not optimizing as current time module, repeated execution of steps S2-S5.

It is described that to train, the combination of dispatching a car is optimized further and comprises:

According to train service frequency, passenger flow trip rule and passenger flow fitness, and in conjunction with the consistent principle of the total stops of train, the train combination of each time module is divided.

The described sequence of dispatching a car to current time module is optimized further comprising the steps:

According to Stirling formula, train is arranged and adopted serial method to be optimized.

Described further comprising the steps as alternative by the minimum initial launch figure of travel line structure optimization acquisition conflict:

In the time not considering to conflict, carry out travel line recursion, obtain initial launch figure;

Minimum as alternative using conflicting in described initial launch figure.

The described travel line recursion of carrying out in the time not considering to conflict, acquisition initial launch figure is:

Dispatch a car and stop as constraints with the minimum dwell time using time interval between trains spaced by automatic block signals, obtained by following formula recursion:

$a_{k+1}^i=d_k^i+r_k^i---\left(5\right)$

$a_k^i+{\mathrm{\δ}}_k^i\×s_k^i=d_k^i---\left(6\right)$

Wherein,Represent the stop situation of train i at k station,Represent the time departure of train i at k station;Represent the time of advent of train i at k station;Represent the dwell time of train i at k station;Represent that train i is at the running time in k interval, γ_ikRepresent the minimum dwell time of train i at k station, k represents station index, k=1, and 2 ... n, n is station sum; I represents train index, i=1, and 2 ... m, m is station sum.

Described using this time module of not optimizing as current time module, repeated execution of steps S2-S5 also comprises:

This time module of not optimizing is carried out to module transition, using next time module after transition as current time module, repeated execution of steps S2-S5, described module transition comprises one of travel line using the last item travel line of an optimised upper time module as next time module, makes the initial time of this next time module be adjusted into the time of departure of this last item travel line.

Correspondingly, the present invention also provides a kind of bullet train operational plan establishment optimization system of Complex Constraints condition, and this system comprises:

Time module is divided device, for the travel line paving time of drawing is divided into multiple time module;

Service chart structure optimization device, for carrying out service chart structure optimization to each time module;

Output device, for exporting optimized operation figure.

Described service chart structure optimization device further comprises with lower unit:

The Combinatorial Optimization unit of dispatching a car, for train is dispatched a car, combination is optimized;

Unit is optimized in the sequence of dispatching a car, and is optimized for the sequence of dispatching a car to current time module;

Travel line structure optimization unit, for obtaining the minimum initial launch figure of conflict as alternative by travel line structure optimization;

Unit is dissolved in conflict, for dissolve the optimized operation figure that obtains this time module by described alternative is conflicted;

Judging unit, for judging whether to exist the time module of not optimizing, starts and repeats unit as existed, starts described output device if do not exist;

Repeat unit, for time module that this is not optimized, as current time module, and unit is optimized in the sequence of dispatching a car described in starting.

The described Combinatorial Optimization unit of dispatching a car, is further used for according to train service frequency, passenger flow trip rule and passenger flow fitness, and in conjunction with the consistent principle of the total stops of train, the train combination of each time module is divided;

Unit is optimized in the described sequence of dispatching a car, and is further used for the formula according to Stirling, and train is arranged and adopted serial method to be optimized;

Described travel line structure optimization unit, is further used for carrying out travel line recursion in the time not considering to conflict, and obtains initial launch figure, and minimum as alternative for described initial launch figure is conflicted;

The described unit that repeats further comprises module transition subelement, this module transition subelement is for carrying out module transition by this time module of not optimizing, using next time module after transition as current time module, repeated execution of steps S2-S5, described module transition comprises one of travel line using the last item travel line of an optimised upper time module as next time module, makes the initial time of this next time module be adjusted into the time of departure of this last item travel line.

The present invention has carried out perfect to the optimization of service chart, broken through static mathematical optimization method for solving in the past, has dynamically realized the reasonably optimizing of service chart from the angle of structure.

Brief description of the drawings

Below with reference to accompanying drawings and in conjunction with the embodiments the present invention is specifically described.

Fig. 1 is service chart structure optimization flow chart of steps;

Fig. 2 is that high-speed railway is for drawing friction speed grade train time module schematic diagram;

Fig. 3 is " zero conflict " initial launch figure;

Fig. 4 is " 4 conflicts " initial launch figure;

Fig. 5 is the service chart of optimal case;

Fig. 6 is for optimizing transition schematic diagram.

Detailed description of the invention

With reference to the accompanying drawings and by embodiments of the invention, technical scheme of the present invention is described in detail.

The optimization of bullet train service chart establishment is converted into the optimization to service chart structure by the present invention, from the angle of service chart structure optimization, restriction relation between train is dissolved in the target of model, in feasible, is asked optimum, replaced in the past ask feasible thought with constraint. Propose many decision-makings ground substep dynamic optimization strategy, by optimizing train dispatch a car portfolio ratio and the order of dispatching a car, minimized the restricting relation between travel line; The design of dispatching a car based on global optimization is proposed: with " time interval between trains spaced by automatic block signals I_ZDispatch a car, the minimum dwell time stops, namely orderAs constraints, arrange to high-density travel line, the optimization aim that recessiveness has comprised " minimize hourage " and " capacity maximum " in fact; Using " number of collisions is minimum " as optimization aim, dynamically realize the doulbe-sides' victory between the optimization aim of " minimize hourage " and " capacity maximum " these two phase containing from the angle of structure.

Service chart model of structural optimization used in the present invention is:

(1) parameter-definition:

T: time window;

K: station index, k=1,2 ... n, n is station sum;

I, j: train index (j represents the follow-up train of train i), i=1,2 ... m, m is station sum;

γ_ik: train i is in the minimum dwell time at k station;

λ_i: be worth the loss of time of train i, and train i often delays the value of corresponding loss in a minute;

I_Z: time interval between trains spaced by automatic block signals; Follow the trail of the minimum time interval between running train, be called time interval between trains spaced by automatic block signals;

I_d: by the train consideration that gets to the station, between the adjacent train of different running method, the minimum time interval getting to the station;

I_f: consider from station by train, between the adjacent train of different running method, from the minimum time interval at station

(2) decision variable:

1, binary variable

Train i is in the situation that stops at k station;

2, integer variable

Train i is at the time departure at k station;

Train i is in the time of advent at k station;

Train i is in the dwell time at k station;

Train i, in the running time in k interval, is also the running time between k station and k+1 station, comprises train in the additional time-division of these interval start and stop;

(3) constraints:

1, section operation time-constrain

Suppose that train fixes in interval running time, so according to train from the time of departure at station can be unique time of advent of definite next stop train.

$a_{k+1}^i=d_k^i+r_k^i---\left(8\right)$

2, dwell time constraint

There is the time leaving from station of definite train that time of departure of train and dwell time can be unique

$a_k^i+{\mathrm{\δ}}_k^i\×s_k^i=d_k^i---\left(9\right)$

3, station personal distance retrains, and ensures to have between two continuous events the time interval of minimum safe, occurs without conflict. Comprise station headway I_fWith station interarrival time I_dConstraint

Station headway I_fConstraint:

$d_k^{i+1}\≥d_k^i+I_f---\left(10\right)$

Station interarrival time I_dConstraint:

$a_k^{i+1}\≥a_k^i+I_d---\left(11\right)$

Personal distance constraint of this station is for dissolving module in conflict, and as constraints, by having minimum safety time interval between constraint adjacent train, thereby the conflict that realizes service chart is dissolved.

(4) optimization aim

On high-speed railway, the travel line of friction speed train mutually restricts and causes that trip's prompt drop is low, different trains to send out order will have different service chart structures, it is very necessary and important therefore optimizing High Speed Railway Train Diagram structure.

1, capacity maximum

Within preset time, paving is drawn more train, improves the ability of service chart, is first important goal that service chart is optimized.

2, hourage loss reduction

The hourage loss of train is mainly owing to stopping and causing AT STATION, the dwell time part of the dwell time that is above standard, and we are defined as and lose Z hourage, and loss hourage is dropped to minimum second important goal that service chart is optimized of regarding as. Minimum loss hourage represents with following formula:

$\mathrm{MinZ}=\underset{i=1}{\overset{m}{\mathrm{\Σ}}}\underset{k=1}{\overset{n}{\mathrm{\Σ}}}[{\mathrm{\λ}}_i\×(s_k^i-{\mathrm{\γ}}_{\mathrm{ik}})]---\left(12\right)$

Because the target of service chart optimization is minimizing and the maximization of capacity of hourage loss, there is certain contradiction in these both targets, in order to ensure minimizing of overall travel time, will inevitably sacrifice the ability of service chart, and in order to increase service chart ability, the travelling speed of train is affected again. Therefore, under the prerequisite with minimum tracking interval operation, reduce as far as possible because travel line restricts the loss of time causing mutually at starting station train, become the unique method of coordinating the two target.

The present invention adopts many decision-makings ground distribution dynamic optimisation strategy, the optimization of service chart structure is divided into following 4 committed step technology, as shown in Figure 1.

The optimization of one, dispatching a car and combining

According to the trip rule of passenger flow and train service frequency, in order to ensure that the each station of each module can meet the demand stopping, and the fitness of passenger flow, in conjunction with the consistent principle of the total stops of train, the train combination of each time module is divided. We regard train as on the skeleton that forms service chart, because the stop train of scheme of friction speed grade, difference restricts mutually, therefore different train groups is combined in service chart and can shows as different service chart structures, to dispatching a car, combination is optimized, exactly the combination of train is optimized, namely the structure of service chart is carried out to preliminary optimization, thus this step of the Combinatorial Optimization of dispatching a car, can be for providing initial framework to the optimization of service chart structure below.

Two, the optimization to the sequence of dispatching a car.

Rationally travel line is sorted, determine the activity duration of travel line at each station, this is a ultra-large optimization problem. According to Stirling formula,Along with the increase of n, full amount of calculation of arranging is huge, considers herein to set up a kind of one-to-one relationship arranging between a kind of special sequence, is then arranged by sequence generation.

By the sequence (a of n-1 element_n-1,a_n-2,...a₁) set up one-to-one relationship with the arrangement of n element, because satisfy condition 0≤a_i≤ i, the sequence (a of 1≤i≤n-1_n-1,a_n-2,...a₁) total n! Individual, this just and the full arrangement of n element n! Individual integer is corresponding one by one. Consider a_iMathematical characteristics (0≤a_i≤ i, i=1,2 ..., n-1), set up a kind of logical relation herein, establish a_iRepresent to arrange the number of the right of number i+1 position in P number less than i+1, sequence (a_n-1,a_n-2,...a₁) can determine the arrangement P of its unique correspondence. The method of being arranged by sequence generation is as follows:

We can arrange since 1 structure:

1, write 1;

2, consider a₁, it is only desirable 0 or 1 years old. If get 0,2 must be in 1 back; Otherwise it is in 1 front.

3, consider a₂, it is desirable 0,1,2 years old. If get 0,3 back that must be placed on the arrangement of two numbers obtained in the previous step, if a₂=1,3 centres that must be placed on the arrangements of two numbers that second step obtains, if a₂=2,3 fronts that must be placed on two numbers of the arrangement that second step obtains.

K+1, consideration a_k，0≤a_k≤k

(1)a_k=0, k+1 must be placed on this k number that k walks the arrangement obtaining backmost;

(2)a_k=1, k+1 must be placed in the middle of inverse two numbers of k number that k walks the arrangement obtaining;

(j)a_k=j, k+1 must be placed between the j reciprocal and this two number of j+1 of k number that k walks the arrangement obtaining.

Three, travel line structure optimization

Travel line structure optimization comprises travel line recursion and finds the process of the minimum service chart of conflict. About travel line recursion, to each arranging situation basis " time interval between trains spaced by automatic block signals I_ZDispatch a car, the minimum dwell time stops, namely orderRecurrence relation by formula (1)-(3) can obtain initial launch figure, under more all arranging situations, which kind of conflict is minimum, illustrate that under this combination, the mutual restricting relation between travel line is the most weak, for the target that reduces as far as possible mutually to restrict due to travel line the loss of time causing, this arrangement can be considered that optimum is dispatched a car and puts in order.

It should be explained that, with " time interval between trains spaced by automatic block signals I_ZDispatch a car, and minimum dwell time parking (order) " as constraints, arrange to high-density travel line, and using " number of collisions is minimum " as optimization aim, the optimization aim that recessiveness has comprised " minimize hourage " and " capacity maximum " in fact. Taking " number of collisions is minimum " as target, realize the doulbe-sides' victory between the optimization aim of " minimize hourage " and " capacity maximum " these two phase containing from structure, carry out perfect to the optimization of service chart, break through static mathematical optimization method for solving in the past, dynamically realized the reasonably optimizing of service chart from the angle of structure.

Four, conflict is dissolved

Travel line conflict is dissolved, and is exactly to process train crossing, overtaking scheme after all. Between the train of ad eundem, do not have Avoidance, the overtaking of train only occurs between different brackets train, and we dodge high-grade train by regulation inferior grade train, minimize due to the stand-by period of dodging generation.

The key dissolved of conflict is, by all trains in initial launch figure each station regard node as to the event of sending out, formula (4)-(5) are as constraints, minimize all to being with sequential interval time of sum between an event. The method that route map of train conflict is dissolved is summed up and mainly contains following 6 kinds: (1) translation train path; (2) exchanging operation order; (3) change the scheme that stops; (4) set up train station to be avoided; (5) extend the train dwelling time; (6) adjust the train time-domain that starts.

As scheme system of the present invention comprise time module divide device, service chart structure optimization device and output device. Described time module is divided device for the travel line paving time of drawing is divided into multiple time module; Described service chart structure optimization device, for carrying out service chart structure optimization to each time module; Described output device, for exporting optimized operation figure.

Method of the present invention comprises:

Step1. by system initialization of the present invention. For convenience of description, below, describe as an example of a certain starting scheme of Jing-Hu Railway example.

Step2. read circuit, station, interval, train data and service chart scale. In this example, I_Z=5min，I_d=2min，I_f=3min，γ_ik=2min。

Step3. divide in device and carry out service chart initial time Module Division in described time module, determine and can, for the time section that spreads the line of rowing, afterwards service chart be become to different time module according to train ratio and time shaft initial division.

The operation characteristic of considering high-speed railway, as shown in Figure 2,0:00-4:00 keeps in repair skylight, 6:00-7:00 is the row inspection car time, as shown in Figure 2, it is 7:00-24:00(17 hours that historical facts or anecdotes border can be used for the row time interval of line of paving), can this time interval be divided into such as 3 according to the difference of starting scheme, 4,5,6,7,8 initial time modules, carry out the paving of service chart and draw to each time module.

The collocation that variety classes passenger train (as the G of China, D, Z, K, the prefix trains such as T) is dispatched a car at the starting station is determined by the type of passenger flow and quantity. Therefore, determine that the priority collocation relation at the rational train starting station will analyse in depth the trip rule of variety classes passenger flow. (1) if the whole day passenger flow at certain starting station trip rule has the bimodal feature in the morning, afternoon, so, can arrange intensive dispatching a car of stage at peak period train; (2) the travelling feature on opportunity of statistics different levels demand, determines the precedence relationship that the train of different brackets or kind is dispatched a car quantity and dispatched a car. Passenger flow trip rule is mated with train train density accordingly, and this is by the decision variety classes train sequencing of dispatching a car.

It is as shown in table 1 that the each section of Jing-Hu Railway is started train logarithm. In conjunction with the activity duration of current China Jing-Hu Railway, can there be 17 hours (1020 minutes) for the time section that spreads the line of rowing, 8 initial time modules will be divided between the paving partition of travel line, each time module respectively has 127.5 minutes, 8 time module are respectively: [7:00:00-9:07:30], [9:07:30-11:15:00], [11:15:00-13:22:30], [13:22:30-15:30:00], [15:30:00-17:37:30], [17:37:30-19:45:00], [19:45:00-21:52:30], [21:52:30-24:00:00].

Table 1 Jing-Hu Railway train is started logarithm

Described service chart structure optimization device further comprises the Combinatorial Optimization unit of dispatching a car, unit is optimized in the sequence of dispatching a car, travel line structure optimization unit, conflict are dissolved unit, judging unit and repeat unit. The performed step in each unit will be elaborated below.

Step4. described in, dispatch a car Combinatorial Optimization unit to the train Combinatorial Optimization of dispatching a car, specific as follows:

Step4.1 is according to the consistent principle of train service frequency, to each time module mean allocation train.

Started by Beijing South Station, have 47 trains to Hongqiao direction, wherein 41 row A grade class cars, 6 row B grade trains, each time module on average can spread and draw 6 row trains.

Step4.2 is according to the trip rule of passenger flow, the train of friction speed grade is carried out to demand assignment: the trip rule of passenger flow refers to passenger's preference, namely passenger is in these 8 time module, the requirement to train class at different time according to passenger, distributes the train of corresponding speed grade to different time module. Generally, we can make the train class distributed uniform of different time module, can all high-grade or inferior grades of a time module. Here, each time module on average has 5 row category-A cars, 1 row category-B car.

Step4.3 passenger flow fitness: ensure that the each station of each time module can meet the demand stopping, and in conjunction with the consistent principle of the total stops of train, the train combination of each time module is distributed. On conversational implication, say to be exactly, the station of passenger under any one time period of one day arrives him, he can realize the object of travelling. The scheme that the stops table 2(of south, Beijing starting train is shown in description 21-22 page) shown in.

The dispatch a car optimum results of combination of 8 time module trains is as follows:

Very first time module: [A2601 (3), A2602 (3), A201 (4), A1 (9), A11 (8), B3 (14)] stops for 41 times totally

The second time module: [A2603 (3), A2604 (3), A202 (4), A2 (9), A4 (8), A12 (9)] stops for 36 times totally

The 3rd time module: [A2605 (3), A2606 (3), A203 (4), A3 (9), A10 (9), A7 (9)] stops for 37 times totally

The 4th time module: [A2607 (3), A2608 (3), A204 (4), A5 (9), A13 (8), B1 (10)] stops for 37 times totally

The 5th time module: [A2609 (3), A2610 (3), A205 (4), A6 (8), A15 (8), B2 (14)] stops for 40 times totally

The 6th time module: [A2611 (3), A2612 (3), A206 (4), A8 (8), A17 (8), B4 (17)] stops for 43 times totally

The 7th time module: [A2613 (3), A2614 (3), A14 (7), A16 (7), A19 (8), B5 (15)] stops for 43 times totally

The 8th time module: [A2615 (3), A2616 (3), A9 (9), A18 (8), B6 (17)] stops for 40 times totally

Wherein: the digitized representation stops in round parentheses, A, B represents category of trains. Below select train groups [A2601 (3), A2602 (3), A201 (4), A1 (9), A11 (8), B3 (14)] carry out instance analysis, the scheme that stops of this train groups is shown in 22 pages, description in Table 3().

Step5. read the scheme that stops, section operation time and dwell time in current time module.

Step6. to current time module, carry out the initial optimizing of service chart structure optimal case

Step6.1 optimizes unit by the described sequence of dispatching a car the sequence of dispatching a car of current time module is optimized:

For the train combination [A2601 (3), A2602 (3), A201 (4), A1 (9), A11 (8), B3 (14)] of very first time module, its number of permutations is total full arrangement alwaysKind, adopt serial method and optimize putting in order herein, can get rid of a hemiidentic scheme, finally have 360 kinds of arranging situations. In the present invention, for convenience of description, by train groups set [A2601 (3), A2602 (3), A201 (4), A1 (9), A11 (8), B3 (14)] use S set={ 1,2,3,4,5,6} represents, and generate whole arrangements of S with serial method, each arrangements P of the corresponding train generating of each sequence, is shown in description 23-25 page as shown 4() as shown in.

Described in Step6.2, travel line structure optimization unit is not being considered under the prerequisite of conflict, according to service chart scale and travel line recursion, with " time interval between trains spaced by automatic block signals I_ZDispatch a car, it (is exactly order that the minimum dwell time stops) " as constraints, by the recurrence relation of formula (1)-(3), paving is drawn initial launch figure, this initial launch figure comprises the service chart corresponding to each train is arranged in S set.

This travel line structure optimization unit of Step6.3 is further found all dispatching a car and is put in order in corresponding initial launch figure, and the minimum initial launch figure of number of collisions is as alternative.

For example, for the train combination [A2601 (3) of very first time module, A2602 (3), A201 (4), A1 (9), A11 (8), B3 (14)], Fig. 3 is " zero conflict " initial launch figure, and we can regard one of alternative of service chart structure optimum as the train combination of this initial launch figure. As seen from Figure 3, although " zero conflict " initial launch figure does not conflict, but its time that takies service chart is that 140min is longer. Then find suboptimum service chart, we find only have " 4 conflicts " service chart as shown in Figure 4. It is 100min that train in this figure combination arrangement mode takies the service chart time, and the time that takies service chart is shorter, and we select this assembled arrangement mode as two of the alternative of service chart structure optimum equally.

Step7. described conflict is dissolved unit described alternative is conflicted and dissolved the optimized operation figure that tries to achieve this time module. Particularly, conflict and dissolve the minimum initial launch figure of described number of collisions in employing formula (4)-(5), finds the minimum scheme of service chart time that takies.

Such as, the service chart that has 4 conflicts of Fig. 4 to be conflicted and dissolved, the time that the scheme after optimization takies service chart is 120min, as shown in Figure 5. Fig. 5 service chart after dissolving compares with " zero conflicts " service chart of Fig. 3, its time that takies service chart is minimum, therefore two of this alternative is the scheme of service chart structure optimum, thereby can obtain the scheme of service chart structure optimum, (seeing description 26-27 page) as shown in table 5, and the arrangement of dispatching a car: [B3, A2601, A2602, A11, A201, A1].

Step8. described judging unit judges whether to exist the time module of not optimizing, and enters Step9. if exist, and repeats unit described in startup, if do not exist, enters Step10., starts output device.

Step9. usually, described in repeat unit using this time module of not optimizing directly as current time module, thereby described in startup, dispatch a car sequence optimize unit repeated execution of steps Step5 to Step8. Preferably, described in the present embodiment, repeat unit and also comprise module transition subelement, this module transition subelement carries out module transition to this time module of not optimizing, using next time module after transition as current time module, thus and then the sequence optimization unit repeated execution of steps Step5 to Step8 that dispatches a car described in starting. Described module transition comprises Step9.1 and Step9.2:

The boundary adjustment of Step9.1 time module: be in course of adjustment, according to the border of next time module of location positioning of the last item travel line in the time module of having optimized, this position is also this end time border of having optimized time module simultaneously. As shown in Figure 6, the time span of establishing each initial time module is T, has two initial time module [t in figure₀,t₀+ T] and [t₀+T,t₀+ 2T], establish initial time module [t₀,t₀+ T] optimised, can determine so the border of time module according to the position of the last item travel line (shown in thick line) of this time module, this travel line is at the t at the time of departure at starting station a station₁Be the new time boundary of next time module, that is to say that the initial time of next time module is by the time t of initial setting up₀+ T is adjusted into t₁, the beginning and ending time of the current time module of not optimizing is [t₁,t₀+ 2T]. Wherein, the reference that in the time module of having optimized, the last item travel line is only optimized as next time module, do not participate in the arrangement optimization of dispatching a car of next module, that is to say the parameter that can not change this last item travel line in the time that the service chart of next time module is optimized.

The transition of Step9.2 module, proceeds to Step5.

The last item travel line (shown in thick line) in last time module is reentered into the current time module of not optimizing, thereby composition has next optimization module of 7 travel lines. Afterwards, for this next time module [t₁,t₀+ 2T] service chart structure, carry out Step5 to Step8 be optimized. But, because the position of this last item travel line is determined, therefore in fact still to the 6 row trains [A2603, A2604, A202, A2, A4, A12] in this next time module, carry out dispatch a car arrangement optimization, travel line structure optimization and conflict and dissolve.

This technology of module transition, mainly contains 2 effects: (1) carries out the boundary alignment of adjacent time module, avoids service chart time waste. It should be noted that, optimization of the present invention is not the optimization of rectangular window form, but be close to flexibly the end time of having optimized time module, start the optimization of new module, such as 7 beginnings of very first time module, 8 are just through with, and the so current time module of not optimizing, just since 8 optimizations, is avoided time waste. (2) the collision detection effect between adjacent two time module, because Step7 only dissolves for the conflict of time module, adopt module transitional technology, the last item travel line of a upper time module is put into current not optimization time module, in the process that the current time module of not optimizing is optimized to this, will realize cross-module conflict and dissolve, avoid the conflict of 8 time module of simple optimization to omit.

Step10. service chart optimization completes, and exports complete optimized operation figure.

Should be appreciated that the above detailed description of technical scheme of the present invention being carried out by preferred embodiment is illustrative and not restrictive. Those of ordinary skill in the art modifies reading the technical scheme that can record each embodiment on the basis of description of the present invention, or part technical characterictic is wherein equal to replacement; And these amendments or replacement do not make the essence of appropriate technical solution depart from the spirit and scope of various embodiments of the present invention technical scheme. Protection scope of the present invention is only limited by the claims of enclosing.

The scheme (timetable) of the service chart structure optimum of table 5 very first time module

Claims (8)

1. the bullet train operational plan of a Complex Constraints condition establishment optimization method, is characterized in that, bagDraw together following steps:

The travel line paving time of drawing is divided into multiple initial time module;

Each time module is carried out to service chart structure optimization; This service chart structure optimization step further comprises:

S1, to train dispatch a car combination be optimized;

S2, to current time module dispatch a car sequence be optimized;

S3, obtain the minimum initial launch figure of conflict as alternative by travel line structure optimization;

S4, dissolve by described alternative is conflicted and obtain the optimized operation figure of this time module;

S5, judge whether exist do not optimize time module, as exist enter next step, if do not existCarry out output optimized operation figure;

S6, using this time module of not optimizing as current time module, repeated execution of steps S2-S5;

Output optimized operation figure.

2. method according to claim 1, is characterized in that, described to train dispatch a car combination carry out excellentChange and further comprise:

According to train service frequency, passenger flow trip rule and passenger flow fitness, and in conjunction with the total stops of trainConsistent principle, divides the train combination of each time module.

3. method according to claim 2, is characterized in that, described dispatching a car to current time moduleSequence is optimized further comprising the steps:

According to Stirling formula, train is arranged and adopted serial method to be optimized.

4. method according to claim 3, is characterized in that, describedly obtains by travel line structure optimizationThe initial launch figure that must conflict minimum is further comprising the steps as alternative:

In the time not considering to conflict, carry out travel line recursion, obtain initial launch figure;

Minimum as alternative using conflicting in described initial launch figure.

5. method according to claim 4, is characterized in that, describedly in the time not considering to conflict, transportsLine recursion, acquisition initial launch figure is:

Dispatch a car and stop as constraints with the minimum dwell time using time interval between trains spaced by automatic block signals, by following formulaRecursion obtains:

$a_{k+l}^i=d_k^i+r_k^i---\left(2\right)$

$a_k^i+{\mathrm{\δ}}_k^i\×s_k^i=d_k^i---\left(3\right)$

Wherein,Represent the stop situation of train i at k station,Represent the time departure of train at station;RepresentTrain i is in the time of advent at k station;Represent the dwell time of train i at k station;Represent that train i is in k intervalRunning time, k represents station index, k=1, and 2 ... n, n is station sum; I represents train index, i=1,2 ... m, m is station sum.

6. method according to claim 1, is characterized in that, described by this time module of not optimizingAs current time module, repeated execution of steps S2-S5 also comprises:

This time module of not optimizing is carried out to module transition, using next time module after transition as currentTime module, repeated execution of steps S2-S5, described module transition comprises an optimised upper time moduleThe last item travel line as one of travel line of next time module, make the initial of this next time moduleTime is adjusted into the time of departure of this last item travel line.

7. the bullet train operational plan of a Complex Constraints condition establishment optimization system, is characterized in that, shouldSystem comprises:

Time module is divided device, for the travel line paving time of drawing is divided into multiple time module;

Service chart structure optimization device, for carrying out service chart structure optimization to each time module; Described fortuneRow graph structure optimization device further comprises with lower unit:

The Combinatorial Optimization unit of dispatching a car, for train is dispatched a car, combination is optimized;

Unit is optimized in the sequence of dispatching a car, and is optimized for the sequence of dispatching a car to current time module;

Travel line structure optimization unit, for obtaining the minimum initial launch of conflict by travel line structure optimizationFigure is as alternative;

Unit is dissolved in conflict, for dissolve acquisition this time module by described alternative conflictOptimized operation figure;

Judging unit, for judging whether to exist the time module of not optimizing, starts and repeats as existedUnit, starts output device if do not exist;

Repeat unit, as current time module, and start institute for time module that this is not optimizedState the sequence of dispatching a car and optimize unit;

Output device, for exporting optimized operation figure.

8. system according to claim 7, is characterized in that,

The described Combinatorial Optimization unit of dispatching a car, be further used for according to train service frequency, passenger flow trip rule andPassenger flow fitness, and in conjunction with the consistent principle of the total stops of train, to the train combination of each time moduleDivide;

Unit is optimized in the described sequence of dispatching a car, and is further used for the formula according to Stirling, and train is arranged and adopted orderRow method is optimized;

Described travel line structure optimization unit, is further used for carrying out travel line recursion in the time not considering to conflict,Obtain initial launch figure, and minimum as alternative for described initial launch figure is conflicted;

The described unit that repeats further comprises module transition subelement, and this module transition subelement is used for willThis time module of not optimizing is carried out module transition, using next time module after transition as current time mouldPiece, repeated execution of steps S2-S5, described module transition comprises last by an optimised upper time moduleArticle one, travel line, as one of travel line of next time module, adjusts the initial time of this next time moduleWhole is the time of departure of this last item travel line.