A method for collaborative optimization of multi-line bus corridor timetables based on cloud model
Through a genetic algorithm based on cloud model optimization of the timetables of multiple lines in the bus corridor, the problem of passenger transfer time is solved and more efficient bus operations are achieved.
Patent Information
- Application Number
- CN202411751675.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-02
- Publication Date
- 2025-04-08
AI Technical Summary
The existing technology is difficult to effectively coordinate the timetables of multiple bus routes, resulting in passengers' transfer time in the bus corridor for too long and travel efficiency is low.
A genetic algorithm based on cloud model is adopted to establish a multi-line timetable collaborative optimization model for bus corridors, optimize the cross-and-mutation operations of genetic algorithms, and quickly search for the optimal solution to determine the collaborative departure timetable of bus lines.
By collaboratively optimizing the departure time of bus routes, it significantly reduces the transfer time for passengers in the bus corridor, improves passenger transfer efficiency, and improves the level of bus operation services.
Smart Images
Figure CN119783871A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to a method for collaborative optimization of multi-line timetables in a bus corridor based on a cloud model, and belongs to the technical field of coordinated dispatching of multi-line buses. Background Art
[0002] The bus timetable is an operational plan for bus companies to organize route operations. It comprehensively considers passenger flow demand and company operating costs, and is used to guide the entire process of bus vehicle operations by determining the departure time or frequency of each bus. It is the basic work of urban bus company management.
[0003] The bus corridor is based on medium-capacity buses and includes multiple conventional bus lines that share the same line or transfer with the main bus line, which together form the urban passenger corridor. In order to better meet the travel needs of passengers in the bus corridor, a multi-line timetable collaborative optimization model for the bus corridor is established by comprehensively considering factors such as the traffic capacity of the bus corridor, the service level of the bus station, the service level of the bus operation, the passenger congestion, and the bus operation cost, in order to seek the best configuration of the capacity resources of each line in the bus corridor and minimize the travel time of transfer passengers in the bus corridor.
[0004] The cloud model combines fuzzy mathematics and probability theory, and can be described as a random number set that follows the normal distribution law and has a stable tendency. Its digital characteristics can be expressed by the expected value E. x 、Entropy E n and super entropy H e To express. x is the expected value of the cloud droplet distribution in space, which determines the center position of the cloud model. n An uncertainty measure that reflects a qualitative concept and indicates the horizontal distribution range of cloud droplets. e YesE n The uncertainty measure represents the discrete range of cloud droplets. According to the above three parameters, the corresponding cloud model can be generated by the Y conditional cloud generation algorithm, the forward normal cloud generation algorithm, etc. Among them, the forward normal cloud generation algorithm is the algorithm based on the cloud model, which generates cloud droplets without any prior conditions. The Y conditional cloud generation algorithm is an algorithm for generating cloud droplets when the membership degree (i.e., conditional value) is known, which is suitable for scenes modeled under specific conditions.
[0005] Genetic algorithm is a random search algorithm. When solving complex problems, it is prone to problems such as premature convergence, poor local search ability and slow convergence speed. Summary of the invention
[0006] The purpose of the invention is to solve the problem of coordinated dispatching of multi-line buses.
[0007] In order to achieve the above object, the technical solution of the present invention provides a method for collaborative optimization of multi-line operation plans of bus corridors based on a cloud model, which is characterized by comprising the following steps:
[0008] Establish a coordinated optimization model for multi-line timetables in bus corridors, establish the model objective function by minimizing the overall travel time (including travel time and transfer time) of passengers transferring in the bus corridor, and set constraints based on the roads, platforms, bus vehicle capacity and line scheduling plans in the bus corridor system;
[0009] The solution is solved based on the genetic algorithm of the cloud model, the crossover and mutation operations of the genetic algorithm are optimized, the optimal solution is quickly searched for in the combination scheme of departure intervals of multiple lines in the bus corridor, and the obtained optimal solution is run as the new scheduling scheme.
[0010] Preferably, the model objective function is as follows:
[0011]
[0012] Where F represents the objective function of the model;
[0013] It represents the interval time between trains k and m on lines i and j arriving at the transfer point n; represents the number of passengers at transfer station n when train number k of line i arrives; represents the number of passengers at transfer station n when train number m of route j arrives; is the transfer ratio of train number k on route i at transfer station n, is the transfer ratio of train number m on line j at transfer station n, calculated based on the ratio of the number of transfer passengers to the number of passengers at the station;
[0014] N represents the total number of transfer stations in the bus corridor;
[0015] O represents the total number of bus routes in the bus corridor;
[0016] K represents the last train number of line i in time period T;
[0017] M represents the last train number of route j in time period T;
[0018] S ik represents the departure time of the kth train on line i;
[0019] T in represents the travel time of line i from the starting station to the transfer station n;
[0020] T i ′ nrepresents the travel time of line i from the corridor starting station to the transfer station n; T i ′ represents the total travel time of line i from the corridor starting station to the terminal station;
[0021] S jm represents the departure time of the mth bus on route j;
[0022] T jn It represents the travel time of line j from the starting station to the transfer station n.
[0023] T j ′ n represents the travel time of route j from the corridor starting station to the transfer station n;
[0024] T j ′ represents the total travel time of route j from the corridor starting station to the terminal station;
[0025] Preferably, the constraints include:
[0026] Departure interval constraints include:
[0027] The time from the start of time period T to the first departure time of the line does not exceed the maximum departure interval H of this line in this time period. max ;
[0028] The departure time of the last train on the route within time period T shall not exceed the end time of the time period;
[0029] The departure interval of all trains in time period T is within the maximum departure interval H of the line. max And the line minimum workshop H min within;
[0030] Station capacity constraints, including:
[0031] The number of vehicles arriving at the transfer station at the same time cannot exceed the platform capacity;
[0032] Bus corridor capacity constraints include:
[0033] The buses operating in the bus corridor meet the capacity restrictions of the bus corridor.
[0034] Public transport capacity constraints include:
[0035] The number of passengers on a public bus before and after the transfer cannot exceed the rated passenger capacity of the vehicle.
[0036] Preferably, the solution process includes the following steps:
[0037] Step 1: Concatenate the departure intervals of all buses on all routes in the bus corridor into a chromosome in binary coding form. The departure interval constraints in the model are directly implemented through coding. The overall gene is represented as pop(i) = (x1, x2, x3, ..., x r-1, x r ), x r represents the departure interval between the rth bus and the previous bus, and r is the number of all departures on the bus corridor during the solution period;
[0038] Step 2: Fitness function calculation
[0039] The constraints are converted into fitness penalty functions and added to the objective function. The fitness function is expressed as Among them, F(i) is the model objective function, δ(i) is the fitness penalty function;
[0040] Step 3: Execute the selection algorithm:
[0041] Based on the proportional algorithm, all individuals in the population are sorted according to their relative fitness, and the better individuals are retained in different proportions at the beginning and end of the algorithm;
[0042] Step 4: Perform crossover operation, including the following steps:
[0043] Step 401: randomly generate a certainty μ according to a uniform distribution;
[0044] Step 402: Calculate the crossover operation cloud model parameters and randomly select the parent pop for crossover a and pop b , using the average value of the algorithm of the cross-parent generation as the expected value of the cloud, that is, The search range is based on the difference between the individuals with the highest and lowest fitness in the current population. n = variable search range / m1,H e =E n / m2, where m1 and m2 are control coefficients;
[0045] Step 403: Use the Y condition cloud generation algorithm, input the cloud model parameters, and generate two offspring pop c and pop d ;
[0046] Step 5: Perform mutation operation, including the following steps:
[0047] Step 501: Randomly select a mutant individual pop e ;
[0048] Step 502: Determine the mutation operation cloud model parameters, using the randomly selected mutation individual pop eis the expected value, and the difference between the individuals with the highest and lowest fitness in the current population is used as the base as the search range, E n = variable search range / n1, H e =E n / m2, where n1 and n2 are control coefficients;
[0049] Step 503: Generate offspring pop using the forward normal cloud generation algorithm f , if the certainty μ is less than the mutation probability, the individual is updated;
[0050] Step 6: Determine whether the algorithm is terminated, and use the number of iterations as the termination criterion of the algorithm. Once the algorithm reaches the preset maximum number of iterations, further update operations will be stopped. The algorithm will output the current optimal individual gene, decode this optimal gene according to the encoding rules, and calculate the departure time of each train one by one according to the rules of departure intervals. Through this process, a coordinated departure timetable for multiple lines in the bus corridor is finally constructed.
[0051] The present invention introduces a cloud model, fully utilizes the randomness and stable tendency characteristics of the cloud model, optimizes the crossover and mutation operations of the genetic algorithm, and makes the probability value retain both trend and randomness, thereby improving the real-time and effectiveness of the algorithm.
[0052] Compared with the prior art solutions, the beneficial effects of the present invention are:
[0053] The departure timetables of multiple bus lines sharing a bus corridor are optimized collaboratively to reduce the transfer time of passengers and maximize the transfer efficiency of passengers within the bus corridor. In the problem-solving process, the cloud model theory and genetic algorithm are combined to improve the algorithm performance. BRIEF DESCRIPTION OF THE DRAWINGS
[0054] Figure 1 This is a flow chart of the genetic algorithm based on the cloud model in the method of the present invention. DETAILED DESCRIPTION
[0055] The present invention will be further described below in conjunction with specific embodiments. It should be understood that these embodiments are only used to illustrate the present invention and are not intended to limit the scope of the present invention. In addition, it should be understood that after reading the content taught by the present invention, those skilled in the art can make various changes or modifications to the present invention, and these equivalent forms fall within the scope limited by the appended claims of the application equally.
[0056] The problem of coordinated optimization of multi-line timetables in bus corridors can be described as: within a specific time period (generally the morning and evening peaks), the model objective function is established to minimize the overall travel time (including transit time and transfer time) of passengers transferring in the bus corridor, and constraints are set based on the roads, platforms, bus vehicle capacity and line scheduling plans in the bus corridor system.
[0057] 1. Basic assumptions of the problem:
[0058] 1. Each passenger can only take one route at a time;
[0059] 2. Overtaking is not allowed in bus corridors on all routes;
[0060] 3. All vehicles are operated as full-course vehicles;
[0061] 4. Buses on all routes arrive at the station on time within the allowable error range;
[0062] 5. When two lines run on the same line, it is assumed that the transfer passengers transfer at the first station on the same line;
[0063] 6. The time step is calculated in minutes.
[0064] 2. Problem Model:
[0065] 1. Objective function:
[0066] The main goal of the coordinated optimization of multi-line timetables in bus corridors is to minimize the waiting time of passengers at transfer stations in bus corridors, thereby reducing the overall travel time of transfer passengers, improving travel efficiency, and improving the service level of bus operations. Therefore, the objective function of the model is the overall travel time of transfer passengers in the bus corridor (including travel time and transfer time), as shown in the following formula.
[0067]
[0068] Where F represents the objective function of the model;
[0069] It represents the interval time between trains k and m of lines i and j arriving at transfer point n, in minutes; It indicates the number of passengers at the transfer station n when the kth train of line i arrives, in person; It indicates the number of passengers transferring to station n when train number m of route j arrives, in person; is the transfer ratio of train number k on line i at transfer station n, unit: %; It is the transfer ratio of train number m on line j at transfer station n, unit: %.
[0070] N represents the total number of transfer stations in the bus corridor;
[0071] O represents the total number of bus routes in the bus corridor;
[0072] K represents the last train number of line i in time period T;
[0073] M represents the last train number of route j in time period T;
[0074] S ik Indicates the departure time of the kth train on line i, unit: minute;
[0075] T in It represents the travel time of line i from the starting station to the transfer station n, in minutes;
[0076] T′ in It represents the travel time of line i from the corridor starting station to the transfer station n, in minutes;
[0077] T′ i It represents the total travel time of line i from the starting station to the terminal station of the corridor, in minutes;
[0078] S jm Indicates the departure time of the mth bus on route j, in minutes;
[0079] T jn It represents the travel time of line j from the starting station to the transfer station n, in minutes.
[0080] T′ jn It represents the travel time of route j from the corridor starting station to the transfer station n, in minutes;
[0081] T′ j It represents the total travel time of line j from the corridor starting station to the terminal station, in minutes;
[0082] 2. Constraints
[0083] 1) Departure interval constraints:
[0084] The time from the start of time period T to the first departure time of the line does not exceed the maximum departure interval H of this line in this time period. max :
[0085]
[0086]
[0087] The departure time of the last train on a route within time period T does not exceed the end time of the time period:
[0088] S iK ≤T,i∈O
[0089] S jM ≤T,j∈O
[0090] The departure interval of all trains in time period T is within the maximum departure interval H of the line. max And the line minimum workshop H min Within:
[0091]
[0092] Wherein, T represents the time period for departure time coordination, and the peak period can be selected; Indicates the maximum departure interval of line i, j in time period T, unit: minute; It indicates the minimum departure interval of lines i and j in time period T, in minutes.
[0093] 2) Platform capacity constraints:
[0094] According to the platform design specifications, based on comprehensive considerations of factors such as the distance passengers travel and the length of buses, the number of berths at a bus stop is usually no more than 4. In order to avoid buses queuing outside the platform, causing passengers to wait for additional transfers, it is necessary to ensure that the number of vehicles arriving at the transfer station at the same time does not exceed the platform capacity. Therefore, the platform capacity should meet the following constraints:
[0095]
[0096] Among them, C n represents the capacity of transfer station n; It is a 0-1 variable, which is 1 when the vehicles of lines i and j arrive at the transfer point n at the same time, otherwise it is 0.
[0097] 3) Bus corridor capacity constraints:
[0098] In order to ensure the traffic efficiency of private vehicles in the bus corridor, the buses operating in the bus corridor need to meet the traffic capacity restrictions of the bus lanes, as shown in the following formula.
[0099]
[0100]
[0101] Where: C lp represents the maximum capacity of buses between stations in the bus corridor, C cp is the maximum capacity of buses at the intersection between stations p and p+1; N′ represents the total number of stations in the bus corridor; It means that there are vehicles on route i running between bus corridor stations p and p+1.
[0102] 4) Bus capacity constraints:
[0103] The number of passengers on a public bus before and after the transfer cannot exceed the rated passenger capacity of the vehicle.
[0104]
[0105] Where: C vi Indicates the rated passenger capacity of the vehicle on route i, in persons; C vj Represents the rated passenger capacity of the vehicle on route j, unit: person.
[0106] 3. Problem Solving Methods:
[0107] The above model is solved by using a genetic algorithm based on the cloud model. x 、Entropy E n and super entropy H e By controlling the three parameters, the algorithm search range can be dynamically adjusted in the crossover and mutation operations through the Y condition cloud generation algorithm and the forward normal cloud generation algorithm to improve the performance of the algorithm.
[0108] like Figure 1 As shown, the key steps in the genetic algorithm of the method of the present invention are as follows:
[0109] Step 1: Determine the gene code. Since the bus departure interval only needs to be accurate to the minute, the code is in binary code. The departure intervals of all buses on all routes in the bus corridor are concatenated into a chromosome. The departure interval constraints in the model can also be directly implemented through coding. The overall gene can be expressed as pop(i) = (x1, x2, x3, ..., x r-1, x r ), x r It represents the departure interval between the rth bus and the previous bus, and r is the total number of departures of the bus lines in the bus corridor during the solution period.
[0110] Step 2: Calculate the fitness function. Since the model objective function is to minimize and is not 0, the individual fitness value is set to the inverse of the model objective function. To avoid premature convergence of the model, the constraint condition is converted into a fitness penalty function and added to the objective function. The fitness function can be expressed as Among them, F(i) is the model objective function, and δ(i) is the fitness penalty function.
[0111] Step 3: Execute the selection algorithm. Taking into account the characteristics of the algorithm's early and late optimization, comprehensively consider ensuring the diversity of the population in the early stage and ensuring the global optimality of the algorithm in the later stage. Based on the proportional algorithm, all individuals in the population are sorted according to their relative fitness, and better individuals are retained in different proportions in the early and late stages of the algorithm.
[0112] Step 4: Perform crossover operation, including the following steps:
[0113] 1) Randomly generate the certainty μ according to uniform distribution;
[0114] 2) Calculate the crossover operation cloud model parameters and randomly select the parent pop for crossover a and pop b , using the average value of the algorithm of the cross-parent generation as the expected value of the cloud, that is, The search range is based on the difference between the individuals with the highest and lowest fitness in the current population. n = variable search range / m1,H e =E n / m2, where m1 and m2 are control coefficients;
[0115] 3) Use the Y condition cloud generation algorithm, input the cloud model parameters, and generate two offspring pop c and pop d .
[0116] Step 5: Perform mutation operation, including the following steps:
[0117] 1) Randomly selected mutant individuals pop e ;
[0118] 2) Determine the mutation operation cloud model parameters to randomly select mutation individuals pop e is the expected value, that is, E x =pop e , taking the difference between the individuals with the highest and lowest fitness of the current population as the base as the search range, E n = variable search range / n1, H e =E n / m2, where n1, n2 are control coefficients;
[0119] 3) Generate offspring pop using the forward normal cloud generation algorithm f , if the certainty μ is less than the mutation probability, the individual is updated.
[0120] Step 6: Determine whether the algorithm is terminated, and use the number of iterations as the termination criterion of the algorithm. Once the algorithm reaches the preset maximum number of iterations, further update operations will be stopped. The algorithm will output the current optimal individual gene, decode this optimal gene according to the encoding rules, and calculate the departure time of each train one by one according to the rules of departure intervals. Through this process, a coordinated departure timetable for multiple lines in the bus corridor is finally constructed.
[0121] The above-mentioned embodiments only express several implementation methods of the present invention, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.
Claims
1. A method for collaborative optimization of multi-line timetables in bus corridors based on a cloud model, characterized in that: The following steps are involved: A collaborative optimization model for multi-line timetables in bus corridors is established, and the model objective function is established to minimize the overall travel time (including transit time and transfer time) of transfer passengers in the bus corridor. Constraints are set based on the roads, platforms, bus capacity and line scheduling plans in the bus corridor system. The solution is solved based on the genetic algorithm of the cloud model, the crossover and mutation operations of the genetic algorithm are optimized, the optimal solution is quickly searched for in the combination scheme of departure intervals of multiple lines in the bus corridor, and the obtained optimal solution is run as the new scheduling scheme.
2. The method for collaborative optimization of multi-line timetables in bus corridors based on a cloud model as claimed in claim 1, characterized in that: The model objective function is shown as follows: i, j∈O 1≤k≤K-1 1≤m≤M-1 Where F represents the objective function of the model; It represents the interval time between trains k and m on lines i and j arriving at the transfer point n; represents the number of passengers at transfer station n when train number k of line i arrives; represents the number of passengers at transfer station n when train number m of route j arrives; is the transfer ratio of train number k on route i at transfer station n; is the transfer ratio of train number m on line j at transfer station n. N represents the total number of transfer stations in the bus corridor; O represents the total number of bus routes in the bus corridor; K represents the last train number of line i in time period T; M represents the last train number of route j in time period T; S ik represents the departure time of the kth train on line i; T in represents the travel time of line i from the starting station to the transfer station n; T′ in It represents the travel time of line i from the corridor starting station to the transfer station n, in minutes; T′ i It represents the total travel time of line i from the starting station to the terminal station of the corridor, in minutes; S jm represents the departure time of the mth bus on route j; T jn It represents the travel time of line j from the starting station to the transfer station n. T′ jn It represents the travel time of route j from the corridor starting station to the transfer station n, in minutes; T′ j It represents the total travel time of line j from the starting station to the terminal station of the corridor, in minutes.
3. The method for collaborative optimization of multi-line timetables in bus corridors based on a cloud model as claimed in claim 1, characterized in that: The constraints include: Departure interval constraints include: The time from the start of time period T to the first departure time of the line does not exceed the maximum departure interval H of this line in this time period. max : The departure time of the last train on the route within time period T shall not exceed the end time of the time period; The departure interval of all trains in time period T is within the maximum departure interval H of the line. max And the line minimum workshop H min within; Station capacity constraints, including: The number of vehicles arriving at the transfer station at the same time cannot exceed the platform capacity; Bus corridor capacity constraints include: The buses operating in the bus corridor meet the capacity restrictions of the bus lanes. Public transport capacity constraints include: The number of passengers on a public bus before and after the transfer cannot exceed the rated passenger capacity of the vehicle.
4. The method for collaborative optimization of multi-line timetables in bus corridors based on a cloud model as claimed in claim 1, characterized in that: The solution process includes the following steps: Step 1: Concatenate the departure intervals of all buses on all routes in the bus corridor into a chromosome in binary coding form. The departure interval constraints in the model are directly implemented through coding. The overall gene is represented as pop(i) = (x1, x2, x3, ..., x r-1 , x r ), x r represents the departure interval between the rth bus and the previous bus, and r is the number of all departures on the bus corridor during the solution period; Step 2: Fitness function calculation The constraints are converted into fitness penalty functions and added to the objective function. The fitness function is expressed as Among them, F(i) is the model objective function, δ(i) is the fitness penalty function; Step 3: Execute the selection algorithm: Based on the proportional algorithm, all individuals in the population are sorted according to their relative fitness, and the better individuals are retained in different proportions at the beginning and end of the algorithm; Step 4: Perform crossover operation, including the following steps: Step 401: randomly generate a certainty μ according to a uniform distribution; Step 402: Calculate the crossover operation cloud model parameters and randomly select the parent pop for crossover a and pop b , using the average value of the algorithm of the cross-parent generation as the expected value of the cloud, that is, The search range is based on the difference between the individuals with the highest and lowest fitness in the current population. n = variable search range / m1,H e =E n / m2, where m1 and m2 are control coefficients; Step 403: Use the Y condition cloud generation algorithm, input the cloud model parameters, and generate two offspring pop c and pop d ; Step 5: Perform mutation operation, including the following steps: Step 501: Randomly select a mutant individual pop e ; Step 502: Determine the mutation operation cloud model parameters, using the randomly selected mutation individual pop e is the expected value, and the difference between the individuals with the highest and lowest fitness in the current population is used as the base as the search range, E n = variable search range / n1, H e =E n / m2, where n1 and n2 are control coefficients; Step 503: Generate offspring pop using the forward normal cloud generation algorithm f , if the certainty μ is less than the mutation probability, the individual is updated; Step 6: Determine whether the algorithm is terminated, and use the number of iterations as the termination criterion of the algorithm. Once the algorithm reaches the preset maximum number of iterations, further update operations will be stopped. The algorithm will output the current optimal individual gene, decode this optimal gene according to the encoding rules, and calculate the departure time of each train one by one according to the rules of departure intervals. Through this process, a coordinated departure timetable for multiple lines in the bus corridor is finally constructed.