CN106652434A - Bus dispatching method based on rail transit coordination - Google Patents

Abstract

The invention discloses a bus dispatching method based on rail transit coordination, comprising the following steps: determining optimized rail transit stations and bus lines; collecting selected bus lines and relevant data of rail traffic; Passengers waiting for the bus at the station are divided into three categories: randomly arriving passengers, transfer passengers and stranded passengers; the waiting time of the three types of passengers is calculated respectively, and a bus scheduling model with the shortest waiting time of passengers is constructed; an improved genetic algorithm is designed and Solve the model to get the departure timetable of the bus; calculate the stop time and travel time of the bus, and then get the departure timetable of the bus. The implementation of the invention can improve the transfer efficiency of passengers, thereby improving the public transport service level of the whole city.

Description

A public transport scheduling method based on rail transit coordination

technical field

The invention belongs to the field of intelligent transportation systems, and in particular relates to a public transportation dispatching method based on rail transit coordination.

Background technique

In recent years, the construction of rail transit in large cities has been significantly accelerated. As a large-capacity, safe and environmentally friendly passenger transportation mode, rail transit has played a very important and key role in solving urban problems with its unique advantages. However, since the overall capacity of the rail transit network will always reach the upper limit and cannot be built endlessly, its service range usually only covers both sides of the line. At this time, the ground bus needs to attract passengers for it to extend its coverage to every corner of the city. Therefore, it is particularly important to coordinate and optimize the timetable of ground buses based on rail transit.

The theory and method of coordinated scheduling of public transport and rail transit have achieved certain results at home and abroad, but there are still some shortcomings, mainly in the following aspects: less consideration is given to the carrying capacity of public transport, and it will There are stranded passengers, which leads to a certain deviation between the theoretical research and the actual situation; the fixed departure interval is used as the assumption or optimization goal, and the departure interval of the ground bus is often uncertain; The stop time of the station, and in practice, the stop time of the vehicle is a key factor affecting whether passengers can transfer successfully.

Therefore, only by fully considering the various influencing factors in the passenger transfer process can a more reasonable bus dispatching plan be formulated.

Contents of the invention

Purpose of the invention: To provide a bus dispatching method based on rail transit coordination to improve passenger transfer efficiency and reduce passenger transit time loss.

Technical solution: a public transport scheduling method based on rail transit coordination, including the following steps:

Step 1: Determine the rail transit station to be optimized, determine the bus line and the research time period;

Step 2: Collect bus line data, rail transit data and other data to be optimized. The bus line data includes the free travel time of the vehicle from the origin to the transfer station, the departure time, the number of research vehicles, the maximum departure interval, and the minimum departure time interval, and the average arrival rate of non-transfer passengers; rail transit data include the arrival time of trains, the number of trains arriving, the number of people arriving at the bus platform for each train, and the transfer time of passengers, as well as the number of trains determined through sampling surveys. The transfer ratio of the bus line; other data include the traffic volume between the origin of the bus and the transfer station;

Step 3: Collect the travel time information of passenger transfers, fit the data, and calculate the mean and variance;

Step 4: Predict the remaining carrying capacity of buses of different shifts when they arrive at the transfer station based on historical data;

Step 5: According to the composition of conventional bus passenger flow, the passengers waiting for the bus at the transfer station are divided into randomly arriving passengers, transfer passengers and stranded passengers, and the transfer passenger flow is divided based on the passenger transfer travel time, and the various types are calculated respectively The waiting time of passengers; taking the departure time of bus vehicles as the optimization variable, a bus scheduling model based on the shortest waiting time of passengers is established;

Step 6: According to the characteristics of the bus dispatching model, a genetic algorithm of ordered integer coding is designed and solved;

Step 7: Based on historical data, predict the number of people who get off at the transfer station of each bus; calculate the number of transfers and boarders of each bus at the station according to the departure time of the bus calculated in step 6 ; Calculate the parking time of the vehicle at the station according to the number of people getting on and off the bus, and then calculate the arrival time of the bus;

Step 8: Based on the BRP function, calculate the travel time of the bus before arriving at the transfer station, and generate the bus departure schedule.

The data collected in the step 2 includes: the free travel time of the bus is t _s , the number of research vehicles is m, the departure time is tbd _j , 1≤j≤m, the maximum departure interval is I _Max , and the minimum departure interval is I _Min , the average arrival rate of non-transfer passengers is λ; the number of arriving trains is n, the arrival time of the train is tr _i , 1≤i≤n, the travel time of passengers is te, and the number of people arriving at the bus platform for each train Q _i , determine the transfer ratio α of the bus line under study through sampling survey, and the traffic volume v between the origin of the bus and the transfer station;

The step 5 is further as follows:

Step 51: Let the arrival time of non-transfer passengers obey the uniform distribution, and the average arrival rate of non-transfer passengers is constant λ, then the waiting time of non-transfer passengers is:

Among them: T ₁ is the waiting time of non-transfer passengers, tbd _j is the departure time of the jth bus, λ is the average arrival rate of non-transfer passengers;

Step 52: Make the passenger transfer travel time te obey the normal distribution, and its probability density function is:

Among them: μ is the mean value of normal distribution, σ is the standard deviation of normal distribution, let te～N(μ,σ ² ), the shortest passenger transfer time temin, the longest passenger transfer time temax and

Organized to get:

Step 53: When the i-th train arrives, everyone who transfers to the bus can take the nearest bus. In this case, the waiting time for transferring passengers is:

Among them: Q _i is the number of people arriving at the bus platform for each train, α is the transfer ratio of the bus line determined through the sample survey, and tr _i is the arrival time of the i-th train;

Step 54: When the i-th train arrives, passengers who walk relatively fast take the nearest bus, but passengers who walk relatively slowly need to wait for the arrival of the next bus. In this case, transfer Passenger waiting times are:

Among them, q _{i, j} is the number of people who transfer from the i-th train to the j-th bus;

Then the total waiting time of transfer passengers is:

Among them: T ₂ is the waiting time of transfer passengers, f _{i, j} is a 0-1 variable, when the transfer passengers of the i-th train can all catch up with the nearest bus, then f _{i, j} takes 1, otherwise takes 0;

Step 55: Assuming that the number of people stranded on the jth bus is d _j , the secondary waiting time for stranded passengers is:

Among them: T3 is the waiting time _of stranded passengers;

Step 56: Taking the departure time of the bus as the optimization variable, the bus scheduling model established is as follows:

The constraints are:

(1) During the research period, the departure interval of the bus should be within the maximum and minimum departure intervals, that is: I _Max ≥ tbd _j+1 -tbd _j ≥ I _Min ;

(2) During the study period, the departure time of the first bus at the transfer station should be within the maximum departure interval of the bus, that is: I _Max ≥ tbd ₁ ≥ 0;

(3) After the last train arrives, it should be ensured that all transfer passengers can transfer, that is, tbd _m ≥ tr _n +temax;

(4) The situation of transferring from rail transit to bus should meet the situation 1 or 2, that is: f _i,j (1-f _i,j )=0, where, f _i,j is a 0-1 variable, when the first If all the passengers on the i train can take the jth bus, it is 1, otherwise it is 0.

The step 6 is further as follows:

Step 61: Set genetic algorithm parameters, including: initial population number, mutation probability, crossover probability and iteration number;

Step 62: Coding, using the method of real number coding, assuming that there are m buses arriving in the research period [0, H], then there are m variables in the optimization model, and the departure times are {x(1), x( 2),x(3),...x(m-2),x(m-1),x(m)}, then each individual in the population uses X={x(1),x(2), The encoding form of x(3),...x(m-2), x(m-1), x(m)};

Step 63: Generation of the initial population. Since the variables of the model are a set of numbers arranged in order and there is no repeated data, the initial population is generated by one of the following methods:

(1) Randomly generate a number x(1) in [0, I _Max ];

(2) Randomly generate a number α between [I _Min , I _Max ], then x(2)=x(1)+α;

(3) Repeat the above operations until x(m) is generated;

Step 64: Calculate the fitness function. The goal of this model is the minimum value problem, so take its reciprocal as the fitness function:

Step 65: Selection operation, using the roulette wheel method, the probability of each individual inheriting to the next generation is equal to the ratio of the fitness of the individual to the fitness of the entire population, then the probability of each individual inheriting to the next generation is:

Among them, Σp _i =1

Randomly generate a number r∈[0,1], if p ₁ +p ₂ +p ₃ …+p _k-1 ≤r≤p ₁ +p ₂ +p ₃ …+p _k , then the kth individual will be selected into the next generation;

Step 66: Cross operation, as follows:

(1) Select two individuals as parents according to the probability of random generation, and randomly select two intersection points, and all variables in the intersection points are used as intersection objects;

(2) Subtract the first variable in the intersection area of parent 2 from each variable in parent 1, and take the absolute value, so that a set of data can be obtained. If the data contains 0 elements, then no Perform crossover, if there is no 0 element, exchange this variable in parent generation 2 with the variable with the smallest difference in parent generation 1;

Step 67: Mutation operation, using the method of uniform mutation, the details are as follows:

(1) A number r∈[0,1] is generated immediately, if r is less than the given mutation probability, the gene at the current position is selected for mutation;

(2) Calculate the new variation value, which can be divided into the following three situations:

① If the position of the mutation is not at the first and last ends, the new variable value after the mutation:

x(i)'=x(i-1)+r×(x(i+1)-x(i))

② If the position of the mutation is at the first point, the new variable value after the mutation:

x(1)'=x(2)+r×x(2)

③If the position of the mutation is at the last point, the new variable value after the mutation:

x(m)'=x(m-1)+r×(R-x(m-1))

Among them, R is the length of research time and represents the maximum value of each variable.

Step 68: optimal individual preservation strategy; its specific operation process is as follows:

(1) By calculating the fitness value, find out the maximum fitness value in the population in the current iteration;

(2) Compare the individual with the highest current fitness with the best individual so far, if the fitness of the individual with the best current fitness is greater than that of the best individual so far, then take the individual with the best current fitness as the so far The individual with the best fitness so far;

(3) After selection, crossover, mutation and other operations, find out the individual with the worst fitness, and replace it with the individual with the highest fitness so far;

(4) Repeat the above operations;

Step 69: output the result, the individual with the best fitness so far after the last iteration is the optimal solution of the problem, that is, the optimized bus departure timetable, and output the result.

The step 7 is further as follows:

Step 71: For the number of passengers getting off the bus, use the gray prediction model to predict the number of passengers getting off the bus at the transfer station for different shifts on weekdays or holidays;

Step 72: According to the departure time of the bus calculated in step 6, calculate the number of boarders of the bus as:

Step 73: The stop time of the bus is:

Among them: C _j is the remaining carrying capacity of the bus, t _d is the parking time of the bus, t _oc is the opening and closing time of the bus door, T _a is the time for getting off passengers at the back door of the bus, T _b is the time for boarding passengers at the front door of the bus, t _b is the average boarding time of a single passenger, t _a is the average time of getting off the bus of a single passenger, P _ja is the number of passengers getting off, and P _jb is the number of passengers getting on board;

Step 74: The arrival time of the bus is:

tba _j ＝tbd _j -t _d ＝tbd _j -t _oc -max(T _a ,T _b )＝tbd _j -t _oc -max(t _a P _ja ,t _b P _ja )

Where: tba _j is the arrival time of the jth bus;

The step 8 is further as follows:

Step 81: The travel time of the bus from the departure point to the transfer station is:

Among them, t _s is the actual running time of the bus from the departure station to the transfer station, t _f is the free travel time of the road section, v is the traffic volume passing through the road section at that time, c is the actual traffic capacity of the road section, β is an undetermined parameter of the model, and its values are 0.15 and 0.4 respectively;

Step 82: Calculate the departure time of the bus: t _0j =tbd _j -t _s ;

Among them, t _0j is the departure time of the jth bus.

Beneficial effects: the invention comprehensively considers factors such as passenger transfer time, bus passenger capacity and vehicle stop time, divides conventional bus passenger flow, and proposes a new scheduling method. It is of great significance to give full play to the respective advantages of rail transit and ground public transportation, improve transfer efficiency, and meet passengers' transfer needs.

Description of drawings

Figure 1 is a flow chart of the method of the present invention.

Fig. 2 is an analysis diagram of the transfer behavior of the present invention.

Fig. 3 is a result diagram of the genetic algorithm of the present invention.

detailed description

Describe technical scheme of the present invention in detail below: a kind of public transport scheduling method based on rail traffic coordination mainly comprises the steps:

Step 1: Determine the rail transit station to be optimized, determine the bus line and the research time period;

Step 2: Collect bus line data, rail transit data and other data to be optimized. The bus line data includes the free travel time of the vehicle from the origin to the transfer station, the departure time, the number of research vehicles, the maximum departure interval, and the minimum departure time interval, and the average arrival rate of non-transfer passengers; rail transit data include the arrival time of trains, the number of trains arriving, the number of people arriving at the bus platform for each train, and the transfer time of passengers, as well as the number of trains determined through sampling surveys. The transfer ratio of the bus line; other data include the traffic volume between the origin of the bus and the transfer station;

Step 3: Collect the travel time information of passenger transfers, fit the data, and calculate the mean and variance;

Step 4: Predict the remaining carrying capacity of buses of different shifts when they arrive at the transfer station based on historical data;

Step 5: According to the composition of conventional bus passenger flow, the passengers waiting for the bus at the transfer station are divided into randomly arriving passengers, transfer passengers and stranded passengers, and the transfer passenger flow is divided based on the passenger transfer travel time, and the various types are calculated respectively The waiting time of passengers; taking the departure time of bus vehicles as the optimization variable, a bus scheduling model based on the shortest waiting time of passengers is established;

Step 6: According to the characteristics of the bus dispatching model, a genetic algorithm of ordered integer coding is designed and solved;

Step 7: Based on historical data, predict the number of people who get off at the transfer station of each bus; calculate the number of transfers and boarders of each bus at the station according to the departure time of the bus calculated in step 6 ; Calculate the parking time of the vehicle at the station according to the number of people getting on and off the bus, and then calculate the arrival time of the bus;

Step 8: Based on the BRP function, calculate the travel time of the bus before arriving at the transfer station, and generate the bus departure schedule.

In a further embodiment, the data collected in step 2 includes: the free travel time of the bus is t _s , the number of research vehicles is m, the departure time is tbd _j , 1≤j≤m, and the maximum departure interval is I _Max , the minimum departure interval is I _Min , the average arrival rate of non-transfer passengers is λ; the number of arriving trains is n, the arrival time of the train is tr _i , 1≤i≤n, and the travel time of passengers is te, every The number of people arriving at the bus platform by a train is Q _i , and the transfer ratio α of the bus line under study is determined by sampling survey, and the traffic volume v of the road section between the origin of the bus and the transfer station;

In a further embodiment, the step 5 is further as follows:

Step 51: Let the arrival time of non-transfer passengers obey the uniform distribution, and the average arrival rate of non-transfer passengers is constant λ, then the waiting time of non-transfer passengers is:

Among them: T ₁ is the waiting time of non-transfer passengers, tbd _j is the departure time of the jth bus, λ is the average arrival rate of non-transfer passengers;

Step 52: Make the passenger transfer travel time te obey the normal distribution, and its probability density function is:

Among them: μ is the mean value of normal distribution, σ is the standard deviation of normal distribution, let te～N(μ,σ ² ), the shortest passenger transfer time temin, the longest passenger transfer time temax and

Organized to get:

Step 53: When the i-th train arrives, everyone who transfers to the bus can take the nearest bus. In this case, the waiting time for transferring passengers is:

Among them: Q _i is the number of people arriving at the bus platform for each train, α is the transfer ratio of the bus line determined through the sample survey, and tr _i is the arrival time of the i-th train;

Step 54: When the i-th train arrives, passengers who walk relatively fast take the nearest bus, but passengers who walk relatively slowly need to wait for the arrival of the next bus. In this case, transfer Passenger waiting times are:

Among them, q _{i, j} is the number of people who transfer from the i-th train to the j-th bus;

Then the total waiting time of transfer passengers is:

Among them: T ₂ is the waiting time of transfer passengers, f _{i, j} is a 0-1 variable, when the transfer passengers of the i-th train can all catch up with the nearest bus, then f _{i, j} takes 1, otherwise takes 0;

Step 55: Assuming that the number of people stranded on the jth bus is d _j , the secondary waiting time for stranded passengers is:

Among them: T3 is the waiting time _of stranded passengers;

Step 56: Taking the departure time of the bus as the optimization variable, the bus scheduling model established is as follows:

The constraints are:

(1) During the research period, the departure interval of the bus should be within the maximum and minimum departure intervals, that is: I _Max ≥ tbd _j+1 -tbd _j ≥ I _Min ;

(2) During the study period, the departure time of the first bus at the transfer station should be within the maximum departure interval of the bus, that is: I _Max ≥ tbd ₁ ≥ 0;

(3) After the last train arrives, it should be ensured that all transfer passengers can transfer, that is, tbd _m ≥ tr _n +temax;

(4) The situation of transferring from rail transit to bus should meet the situation 1 or 2, that is: f _i,j (1-f _i,j )=0, where, f _i,j is a 0-1 variable, when the first If all the passengers on the i train can take the jth bus, it is 1, otherwise it is 0.

In a further embodiment, the step 6 is further as follows:

Step 61: Set genetic algorithm parameters, including: initial population number, mutation probability, crossover probability and iteration times; Step 62: Coding, using real number coding, assuming that there are m buses in the research period [0,H] When the car arrives, there are m variables in the optimization model, and the departure time is respectively {x(1), x(2), x(3),...x(m-2), x(m-1), x(m )}, then each individual in the population uses X={x(1),x(2),x(3),…x(m-2),x(m-1),x(m)} coded form;

Step 63: Generation of the initial population. Since the variables of the model are a set of numbers arranged in order and there is no repeated data, the initial population is generated by one of the following methods:

(1) Randomly generate a number x(1) in [0, I _Max ];

(2) Randomly generate a number α between [I _Min , I _Max ], then x(2)=x(1)+α;

(3) Repeat the above operations until x(m) is generated;

Step 64: Calculate the fitness function. The goal of this model is the minimum value problem, so take its reciprocal as the fitness function:

Step 65: Selection operation, using the roulette wheel method, the probability of each individual inheriting to the next generation is equal to the ratio of the fitness of the individual to the fitness of the entire population, then the probability of each individual inheriting to the next generation is:

Among them, Σp _i =1

Randomly generate a number r∈[0,1], if p ₁ +p ₂ +p ₃ …+p _k-1 ≤r≤p ₁ +p ₂ +p ₃ …+p _k , then the kth individual will be selected into the next generation;

Step 66: Cross operation, as follows:

(1) Select two individuals as parents according to the probability of random generation, and randomly select two intersection points, and all variables in the intersection points are used as intersection objects;

(2) Subtract the first variable in the intersection area of parent 2 from each variable in parent 1, and take the absolute value, so that a set of data can be obtained. If the data contains 0 elements, then no Perform crossover, if there is no 0 element, exchange this variable in parent generation 2 with the variable with the smallest difference in parent generation 1;

Step 67: Mutation operation, using the method of uniform mutation, the details are as follows:

(1) A number r∈[0,1] is generated immediately, if r is less than the given mutation probability, the gene at the current position is selected for mutation;

(2) Calculate the new variation value, which can be divided into the following three situations:

① If the position of the mutation is not at the first and last ends, the new variable value after the mutation:

x(i)'=x(i-1)+r×(x(i+1)-x(i))

② If the position of the mutation is at the first point, the new variable value after the mutation:

x(1)'=x(2)+r×x(2)

③If the position of the mutation is at the last point, the new variable value after the mutation:

x(m)'=x(m-1)+r×(R-x(m-1))

Among them, R is the length of research time and represents the maximum value of each variable.

Step 68: optimal individual preservation strategy; its specific operation process is as follows:

(1) By calculating the fitness value, find out the maximum fitness value in the population in the current iteration;

(2) Compare the individual with the highest current fitness with the best individual so far, if the fitness of the individual with the best current fitness is greater than that of the best individual so far, then take the individual with the best current fitness as the so far The individual with the best fitness so far;

(3) After selection, crossover, mutation and other operations, find out the individual with the worst fitness, and replace it with the individual with the highest fitness so far;

(4) Repeat the above operations;

Step 69: output the result, the individual with the best fitness so far after the last iteration is the optimal solution of the problem, that is, the optimized bus departure timetable, and output the result.

In a further embodiment, the step 7 is further as follows:

Step 71: For the number of passengers getting off the bus, use the gray prediction model to predict the number of passengers getting off the bus at the transfer station for different shifts on weekdays or holidays;

Step 72: According to the departure time of the bus calculated in step 6, calculate the number of boarders of the bus as:

Step 73: The stop time of the bus is:

Among them: C _j is the remaining carrying capacity of the bus, t _d is the parking time of the bus, t _oc is the opening and closing time of the bus door, T _a is the time for getting off passengers at the back door of the bus, T _b is the time for boarding passengers at the front door of the bus, t _b is the average boarding time of a single passenger, t _a is the average time of getting off the bus of a single passenger, P _ja is the number of passengers getting off, and P _jb is the number of passengers getting on board;

Step 74: The arrival time of the bus is:

tba _j =tbd _j -t _d =tbd _j -t _oc -max(T _a ,T _b )

＝tbd _j -t _oc -max(t _a P _ja ,t _b P _ja )

Where: tba _j is the arrival time of the jth bus;

In a further embodiment, the step 8 is further:

Step 81: The travel time of the bus from the departure point to the transfer station is:

Among them, t _s is the actual running time of the bus from the departure station to the transfer station, t _f is the free travel time of the road section, v is the traffic volume passing through the road section at that time, c is the actual traffic capacity of the road section, β is an undetermined parameter of the model, and its values are 0.15 and 0.4 respectively;

Step 82: Calculate the departure time of the bus: t _0j =tbd _j -t _s ;

Among them, t _0j is the departure time of the jth bus.

An implementation example is described below.

In conjunction with Fig. 1, describe the public transport scheduling method based on coordination with rail traffic of the present invention, comprise the following steps:

Step 1: Select the City Sports Park Station of Metro Line 2 in a certain city as the transfer station. According to the survey, there are more people transferring to bus No. 329 at this station, so it is selected as the research object and its departure timetable is optimized;

Step 2: According to the survey, bus No. 329, the uplink route is from Jingwei Clear Ecological Peninsula to Xi’an North Railway Station, and the downlink route is from Xi’an North Railway Station to Jingwei Clear Ecological Peninsula. The total length is 25400m, the driving speed is about 20km/h, and the whole operation time is about 1.5h ;During the research period, the number of vehicles is 9, the maximum departure interval is 21 minutes, and the minimum departure interval is 9 minutes; collect the arrival timetable of Xi'an Metro Line 2, and count the number of people who transfer to buses for each train;

Step 3: To collect the transfer travel time of passengers, the sample size should be determined first, and the calculation method is as follows:

Among them: n is the size of the sample size; z is the confidence level of the standard error, if the confidence level is 0.95, then z is 1.96; σ is the overall standard deviation. In the case of unknowns, samples can be used to estimate first; E is the allowable error of transfer travel time, and the allowable error can be set to 10%;

First select some samples and make a preliminary estimate. If the variance of passenger transfer travel time is 0.55, the minimum sample size is:

Obtain sample data through follow-up investigation, and use normal distribution to fit the data, and then conduct k-s test. The test results are shown in Table 1:

Table 1 K-S test results

It can be seen from the test results that the mean is 240s (about 4min), the standard deviation is 33s (about 0.55min), and the significance level is 0.849, which is greater than 0.05, so it can be considered that the data is subject to a normal distribution.

Step 4: The specific process of using the gray system model to predict the remaining carrying capacity of buses of different shifts at the transfer station is as follows:

The first step is to calculate the level ratio of historical data X ⁽⁰⁾ = (X ⁽⁰⁾ (1), X ⁽⁰⁾ (2),...,X ⁽⁰⁾ (n))

Among them, X ⁽⁰⁾ (m) represents the mth historical data, and k (m) is the level ratio;

If all the levels of the sequence k(m) fall in , the gray system model can be used to predict the data. Otherwise, the original data needs to be processed, and an appropriate constant K can be taken for translation transformation, so that the level ratio falls into the range that can be covered Inside;

In the second step, the historical data sequence is accumulated once to generate a sequence, and the formula is as follows:

The third step is to establish the gray differential equation, the formula is as follows:

X ⁽¹⁾ (t)＝(X ⁽¹⁾ (0)-u/a)e- ^at +u/a

In the formula, the undetermined coefficient A composed of coefficients a and u can be obtained by the least square method, and the formula is:

A=(a,u) ^T

Can be set Yn＝[X ⁽⁰⁾ (2),X ⁽⁰⁾ (3),…X ⁽⁰⁾ (n)] ^T , then A＝(B ^T B)- ¹ B ^T Yn, the response time equation can be established, Then the predicted value can be obtained;

The fourth step is to check the residual. The calculation formula of the residual is as follows:

In general, if ε(m)<0.2, the inspection can be considered qualified.

Similarly, the GM(1,1) prediction process mentioned in step 71 of the present invention is the same as above.

Step 5: the parameter calibration required in the modeling process of the present invention is as shown in Table 2:

Table 2 Calibration table of some model parameters

Use matlab to write the model program, and then bring the above parameters into the model.

Step 6: The parameter values of the algorithm are shown in Table 3:

Table 3 Genetic algorithm parameter value table

According to the genetic algorithm steps described in step 6, the algorithm program is written, as shown in Figure 3, after 1000 iterations, the optimized bus departure timetable can be obtained, as shown in Table 4.

Step 7: Using the gray system model prediction method mentioned above, the number of people getting off at the transfer station can be predicted for each trip, and the results are shown in Table 4.

According to the departure time of the bus calculated in step 6, calculate the number of passengers to be connected by each bus, and then compare it with the remaining passenger capacity of the bus predicted in step 4, and take the minimum value as the number of passengers.

Calculate the bus stop time according to the number of passengers getting on and off. Among them, through a large number of data surveys, the door opening and closing t _oc of the bus is 3s, the boarding time t _b of a unit passenger is 1.2s, and the disembarking time t _a is 0.9s. Then the calculation results of the stop time of each bus are shown in the table below:

Table 4 Calculation results of bus stop time

Then subtract the stop time from the departure time of the bus to obtain the arrival timetable of the bus.

Step 8: According to the survey, the free-flow travel time of the bus from the origin to the transfer site is about 1200s, the actual traffic volume v is 1400pcu/h, the traffic capacity of the lane is 1500pcu/h, and the number of lanes is 3, then the vehicle The actual travel time for is:

The calculated bus departure schedule is shown in the table below:

Table 5 bus schedule

The optimization result of the present invention is compared with the present situation, and the comparison result is as shown in the table, as can be seen, the total waiting time of passengers after optimization is 1.041 * 105s, compared with before optimization, the total waiting time has reduced by 0.409 * 105s, The optimization range is 28.29%, and the improvement effect is relatively obvious. Therefore, the dispatching model and algorithm involved in the present invention have a relatively significant optimization effect and have certain practical value.

Table 6 Comparison table before and after optimization

In summary, the present invention proposes a bus dispatching method based on coordination with rail transit. The invention divides the passenger flow of the bus platform on the basis of considering factors such as transfer travel time, vehicle stop time and bus capacity limitation, establishes a bus dispatching model with the shortest waiting time for passengers, and designs an algorithm to solve it. In case analysis, compared with the current situation, the total waiting time of passengers obtained by applying the present invention can effectively reduce the travel time of passengers and improve transfer efficiency.

The preferred embodiments of the present invention have been described in detail above, but the present invention is not limited to the specific details in the above embodiments. Within the scope of the technical concept of the present invention, various equivalent transformations can be carried out to the technical solutions of the present invention. These equivalent transformations All belong to the protection scope of the present invention.

In addition, it should be noted that the various specific technical features described in the above specific implementation manners may be combined in any suitable manner if there is no contradiction. In order to avoid unnecessary repetition, various possible combinations are not further described in the present invention.

In addition, various combinations of different embodiments of the present invention can also be combined arbitrarily, as long as they do not violate the idea of the present invention, they should also be regarded as the disclosed content of the present invention.

Claims (6)

1. A public transport dispatching method based on rail traffic coordination, is characterized in that, comprises the steps: Step 1: Determine the rail transit station to be optimized, determine the bus line and the research time period; Step 2: Collect bus line data, rail transit data and other data to be optimized. The bus line data includes the free travel time of the vehicle from the origin to the transfer station, the departure time, the number of research vehicles, the maximum departure interval, and the minimum departure time interval, and the average arrival rate of non-transfer passengers; rail transit data include the arrival time of trains, the number of trains arriving, the number of people arriving at the bus platform for each train, the transfer travel time of passengers, and the number of trains determined by sampling survey. The transfer ratio of the bus line; other data include the traffic volume between the origin of the bus and the transfer station; Step 3: Collect the travel time information of passenger transfers, fit the data, and calculate the mean and variance; Step 4: Predict the remaining carrying capacity of buses of different shifts when they arrive at the transfer station based on historical data; Step 5: According to the composition of the conventional bus passenger flow, the passengers waiting for the bus at the transfer station are divided into randomly arriving passengers, transfer passengers and stranded passengers, and the transfer passenger flow is divided based on the passenger transfer travel time, and the various types are calculated respectively The waiting time of passengers; taking the departure time of bus vehicles as the optimization variable, a bus scheduling model based on the shortest waiting time of passengers is established; Step 6: According to the characteristics of the bus dispatching model, a genetic algorithm of ordered integer coding is designed and solved; Step 7: Based on historical data, predict the number of people who get off at the transfer station of each bus; calculate the number of transfers and boarders of each bus at the station according to the departure time of the bus calculated in step 6 ; Calculate the parking time of the vehicle at the station according to the number of people getting on and off the bus, and then calculate the arrival time of the bus; Step 8: Based on the BRP function, calculate the travel time of the bus before arriving at the transfer station, and generate the bus departure schedule. 2. The public transport scheduling method based on rail transit coordination as claimed in claim 1, wherein the data collected in the step 2 include: the free travel time of public transport vehicles is t _s , the number of research vehicles is m, and The time is tbd _j , 1≤j≤m, the maximum departure interval is I _Max , the minimum departure interval is I _Min , the average arrival rate of non-transfer passengers is λ; the number of arriving trains is n, and the arrival time of the train is tr _i , 1≤i≤n, the travel time for passengers to transfer is te, the number of people arriving at the bus platform for each train is Q _i , the transfer ratio α of the bus line under study is determined by sampling survey, the origin of the bus to the transfer station The traffic volume of the link between v. 3. the public transport scheduling method based on rail traffic coordination as claimed in claim 2, is characterized in that, described step 5 is further: Step 51: Let the arrival time of non-transfer passengers obey the uniform distribution, and the average arrival rate of non-transfer passengers is constant λ, then the waiting time of non-transfer passengers is: Among them: T ₁ is the waiting time of non-transfer passengers, tbd _j is the departure time of the jth bus, λ is the average arrival rate of non-transfer passengers; Step 52: Make the passenger transfer travel time te obey the normal distribution, and its probability density function is: Among them: μ is the mean value of normal distribution, σ is the standard deviation of normal distribution, let te～N(μ,σ ² ), the shortest passenger transfer travel time te min, the longest passenger transfer travel time te max and Organized to get: Step 53: When the i-th train arrives, everyone who transfers to the bus can take the nearest bus. In this case, the waiting time for transferring passengers is: Among them: Q _i is the number of people arriving at the bus platform for each train, α is the transfer ratio of the bus line determined through the sample survey, and tr _i is the arrival time of the i-th train; Step 54: When the i-th train arrives, passengers who walk relatively fast take the nearest bus, but passengers who walk relatively slowly need to wait for the arrival of the next bus. In this case, transfer Passenger waiting times are: Among them, q _{i, j} is the number of people who transfer from the i-th train to the j-th bus; Then the total waiting time of transfer passengers is: Among them: T ₂ is the waiting time of transfer passengers, f _{i, j} is a 0-1 variable, when the transfer passengers of the i-th train can all catch up with the nearest bus, then f _{i, j} takes 1, otherwise takes 0; Step 55: Assuming that the number of people stranded on the jth bus is d _j , the secondary waiting time for stranded passengers is: Among them: T3 is the waiting time _of stranded passengers; Step 56: Taking the departure time of the bus as the optimization variable, the bus scheduling model established is as follows: The constraints are: (1) During the research period, the departure interval of the bus should be within the maximum and minimum departure intervals, that is: I _Max ≥ tbd _j+1 -tbd _j ≥ I _Min ; (2) During the study period, the departure time of the first bus at the transfer station should be within the maximum departure interval of the bus, that is: I _Max ≥ tbd ₁ ≥ 0; (3) After the last train arrives, it should be ensured that all transfer passengers can transfer, that is, tbd _m ≥ tr _n +temax; (4) The situation of transferring from rail transit to bus should meet the situation 1 or 2, that is: f _i,j (1-f _i,j )=0, where, f _i,j is a 0-1 variable, when the first If all the passengers on the i train can take the jth bus, it is 1, otherwise it is 0. 4. the public transport scheduling method based on rail traffic coordination as claimed in claim 3, is characterized in that, described step 6 is further: Step 61: Set genetic algorithm parameters, including: initial population number, mutation probability, crossover probability and iteration number; Step 62: Coding, using the method of real number coding, assuming that there are m buses arriving in the research period [0, H], then there are m variables in the optimization model, and the departure times are {x(1), x( 2),x(3),...x(m-2),x(m-1),x(m)}, then each individual in the population uses X={x(1),x(2), The encoding form of x(3),...x(m-2), x(m-1), x(m)}; Step 63: Generation of the initial population. Since the variables of the model are a set of numbers arranged in order and there is no repeated data, the initial population is generated according to one of the following methods: (1) Randomly generate a number x(1) in [0, I _Max ]; (2) Randomly generate a number α between [I _Min , I _Max ], then x(2)=x(1)+α; (3) Repeat the above operations until x(m) is generated; Step 64: Calculate the fitness function. The goal of this model is the minimum value problem, so take its reciprocal as the fitness function: Step 65: Selection operation, using the roulette wheel method, the probability of each individual inheriting to the next generation is equal to the ratio of the fitness of the individual to the fitness of the entire population, then the probability of each individual inheriting to the next generation is: Among them, Σp _i =1 Randomly generate a number r∈[0,1], if p ₁ +p ₂ +p ₃ …+p _k-1 ≤r≤p ₁ +p ₂ +p ₃ …+p _k , then the kth individual will be selected into the next generation; Step 66: Crossover operation, the details are as follows: (1) Select two individuals as parents according to the probability of random generation, and randomly select two intersection points, and all variables in the intersection points are used as intersection objects; (2) Subtract the first variable in the intersection area of parent 2 from each variable in parent 1, and take the absolute value, so that a set of data can be obtained. If the data contains 0 elements, then no Perform crossover, if there is no 0 element, exchange this variable in parent generation 2 with the variable with the smallest difference in parent generation 1; Step 67: Mutation operation, using the method of uniform mutation, the details are as follows: (1) A number r∈[0,1] is generated immediately, if r is less than the given mutation probability, the gene at the current position is selected for mutation; (2) Calculate the new variation value, which can be divided into the following three situations: ① If the position of the mutation is not at the first and last ends, the new variable value after the mutation: x(i)'=x(i-1)+r×(x(i+1)-x(i)) ② If the position of the mutation is at the first point, the new variable value after the mutation: x(1)'=x(2)+r×x(2) ③If the position of the mutation is at the last point, the new variable value after the mutation: x(m)'=x(m-1)+r×(R-x(m-1)) Among them, R is the length of research time and represents the maximum value of each variable. Step 68: optimal individual preservation strategy; its specific operation process is as follows: (1) By calculating the fitness value, find out the maximum fitness value in the population in the current iteration; (2) Compare the individual with the highest current fitness with the best individual so far, if the fitness of the individual with the best current fitness is greater than that of the best individual so far, then take the individual with the best current fitness as the so far The individual with the best fitness so far; (3) After selection, crossover, mutation and other operations, find out the individual with the worst fitness, and replace it with the individual with the highest fitness so far; (4) Repeat the above operations; Step 69: output the result, the individual with the best fitness so far after the last iteration is the optimal solution of the problem, that is, the optimized bus departure timetable, and output the result. 5. the public transport scheduling method based on rail traffic coordination as claimed in claim 4, is characterized in that, described step 7 is further: Step 71: For the number of passengers getting off the bus, use the gray prediction model to predict the number of passengers getting off the bus at the transfer station for different shifts on weekdays or holidays; Step 72: According to the departure time of the bus calculated in step 6, calculate the number of boarders of the bus as: Step 73: The stop time of the bus is: Among them: C _j is the remaining carrying capacity of the bus, t _d is the parking time of the bus, t _oc is the opening and closing time of the bus door, T _a is the time for getting off passengers at the back door of the bus, T _b is the time for boarding passengers at the front door of the bus, t _b is the average boarding time of a single passenger, t _a is the average time of getting off the bus of a single passenger, P _ja is the number of passengers getting off, and P _jb is the number of passengers getting on board; Step 74: The arrival time of the bus is: tba _j =tbd _j -t _d =tbd _j -t _oc -max(T _a ,T _b ) ＝tbd _j -t _oc -max(t _a P _ja ,t _b P _ja ) Among them: tba _j is the arrival time of the jth bus. 6. the public transport scheduling method based on rail traffic coordination as claimed in claim 5, is characterized in that, described step 8 is further: Step 81: The travel time of the bus from the departure point to the transfer station is: Among them, t _s is the actual running time of the bus from the departure station to the transfer station, t _f is the free travel time of the road section, v is the traffic volume passing through the road section at that time, c is the actual traffic capacity of the road section, β is an undetermined parameter of the model, and its values are 0.15 and 0.4 respectively; Step 82: Calculate the departure time of the bus: t _0j =tbd _j -t _s ; Among them, t _0j is the departure time of the jth bus.