WO2024174566A1

WO2024174566A1 - Multi-vehicle-type timetable design method and system for intelligent bus system

Info

Publication number: WO2024174566A1
Application number: PCT/CN2023/127740
Authority: WO
Inventors: 王晓伟; 吴松屿; 袁维; 秦兆博; 谢国涛; 秦洪懋; 秦晓辉; 徐彪; 丁荣军
Original assignee: 湖南大学无锡智能控制研究院
Priority date: 2023-02-20
Filing date: 2023-10-30
Publication date: 2024-08-29
Also published as: CN116432386A; CN116432386B

Abstract

Disclosed in the present invention are a multi-vehicle-type timetable design method and system for an intelligent bus system. The method comprises: step 1, setting line operation parameters and multi-vehicle-type bus resource configurations according to the actual operation situation of bus lines; step 2, reading and processing historical operation data of a corresponding line in multiple days, and statistically analyzing the law of passenger flow demands and a travelling duration of a vehicle between stops within a corresponding operation period; step 3, constructing a multi-objective optimization function which is used for evaluating a generated multi-vehicle-type departure timetable scheme; step 4, constructing a constraint condition set for a multi-vehicle-type timetable; step 5, according to the multi-objective optimization function and the constraint condition set for the multi-vehicle-type timetable, performing calculation by using a multi-objective optimization algorithm, so as to obtain a Pareto optimal solution set; and step 6, performing screening on the Pareto optimal solution set, so as to select solutions representing different benefit preferences, and generating a corresponding departure timetable scheme. The present invention can formulate a multi-vehicle-type departure timetable for a single bus line on the basis of set resource configurations and historical passenger flow data.

Description

A multi-type timetable design method and system for intelligent public transportation system

Technical Field

The present invention relates to the field of intelligent transportation technology, and in particular to a multi-vehicle timetable design method and system for an intelligent public transportation system.

Background Art

With the rapid development of urban economy and technology, intelligent connected vehicle technology has been gradually applied to urban public transportation systems, providing new solutions to the problems faced by urban public transportation, such as insufficient resource utilization, imbalance in service quality and cost input. Compared with traditional public transportation systems, autonomous driving public transportation systems have greater potential in reducing energy consumption, improving operational efficiency and public transportation attractiveness, and are considered an important part of smart city construction.

In the existing autonomous bus operation scenarios, the dispatching system can usually assign different types of vehicles to perform shift tasks to meet the time-varying passenger flow demand during the operation period, thereby matching the capacity supply with the service demand. Therefore, under a given line layout and resource allocation, a reasonable bus line departure schedule is the basis for the operation and subsequent dispatch of the autonomous bus system, and it plays a vital role in improving service satisfaction and reliability.

However, traditional bus dispatching systems usually use fixed departure frequencies and a single vehicle model in the design of departure schedules. The few improved solutions only divide the departure time and vehicle model into two-stage problems for calculation. This type of schedule design method has certain shortcomings in terms of adaptability to time-varying passenger flow, rational use of different types of transport resources, and trade-offs between the benefits of different travel participants. It is difficult to give full play to the advantages of autonomous driving bus systems in terms of efficiency, flexibility, and autonomy.

A patent document discloses a bus scheduling method and system based on genetic algorithms, which divides a day into six time periods, calculates and analyzes the line traffic volume, and obtains the initialization schedules of the line in different time periods; then, with the minimum average waiting time and the standard deviation of all waiting times as the goal, the genetic algorithm is used to optimize the schedule offset list under the condition of a certain departure schedule, and the final departure frequency and schedule for different time periods are obtained. However, this patent technology only considers the departure schedule design problem in a single vehicle model operation scenario. This solution cannot be used for bus operation systems with two or more different vehicle models, and the consideration of the optimization target is relatively simple and the calculation is relatively simplified. Therefore, there are great limitations in actual application.

Another patent document discloses a method for optimizing bus departure time intervals based on a genetic algorithm, which takes the maximum vehicle load factor and the minimum passenger waiting time as optimization objectives, and converts the weighted sum of the two into a single objective, and then optimizes the departure time interval through a genetic algorithm, and then formulates a departure schedule plan based on the interval sequence. However, this technology converts the two objectives into a single objective problem by weighted summation. In actual use, it is necessary to balance the two conflicting objectives by adjusting the weight coefficient. For complex and changeable actual scenarios, this adjustment process is cumbersome and time-consuming, which will affect the implementation of the plan. Only the departure interval is optimized, and the heterogeneity of the departure models is not considered.

There is also a two-stage design method for multi-model departure schedule in the prior art, which takes the departure time and model of the bus route as a two-stage decision-making problem. In the first stage, the departure frequency in each period is determined based on the average value of the maximum cross-sectional passenger flow and the rated passenger capacity of all models, and the departure time sequence of the shift is determined in a uniformly spaced manner; in the second stage, the departure model sequence corresponding to the departure time sequence is optimized based on the meta-heuristic algorithm. Finally, a multi-model departure schedule for the entire operation cycle is obtained. In the treatment of different models, this design method only considers the average passenger capacity of the model to determine the departure frequency, while ignoring the influence of different models on driving costs and the number of available vehicles. Therefore, there are limitations in calculation and the actual performance of the obtained schedule scheme cannot be guaranteed. In addition, this method designs a multi-model schedule in a two-stage manner, calculates the departure time and model in turn, and due to the large differences in the capacity level of vehicles of different models, the capacity of the shift model will directly affect its optimal departure time. Therefore, this technical solution has the problem of lack of optimality of the obtained solution, and there is still a large room for optimization in terms of optimality.

Summary of the invention

In view of the above-mentioned deficiencies in the prior art, the present invention provides a multi-vehicle timetable design method and system for an intelligent public transportation system, which can formulate a multi-vehicle departure timetable for a single bus line based on established resource allocation and historical passenger flow data.

To achieve the above object, the present invention provides a multi-vehicle timetable design method for an intelligent public transportation system, which comprises:

Step 1: Set the route operation parameters and multi-model bus resource configuration according to the actual operation of the bus route;

Step 2: read and process the historical operation data of the corresponding line for multiple days, and statistically analyze the passenger flow demand pattern and the vehicle travel time between stations within the corresponding operation period;

Step 3, construct a multi-objective optimization function for evaluating the generated multi-vehicle departure schedule scheme;

Step 4, construct the constraint set of the multi-vehicle timetable;

Step 5: Based on the multi-objective optimization function and the constraint condition set of the multi-vehicle timetable, a multi-objective optimization is used to The Pareto optimal solution set TT ^ps is obtained by using the algorithm;

Step 6: Select solutions representing different benefit preferences from the Pareto optimal solution set TT ^ps and generate the corresponding departure schedule plan.

Furthermore, the passenger flow demand rules in the operation cycle in step 2 include the passenger flow arrival rate α _kf and the proportion of passengers getting off the bus ρ _kf at each station. Step 2 includes:

Step 21, divide the entire operation cycle [t ₀ ,t _e ] into a number of characteristic time periods f∈F={1,2,…, ^NF } according to the characteristic cycle length τ; where t ₀ and t _e are the start and end times of the operation of a single day, respectively, F is the characteristic time period set, and ^NF is the total number of the divided characteristic time periods f;

Step 22, traverse each departure schedule in the single-day operation data, correspond each arrival time to the corresponding characteristic time period, and accumulate the number of boardings, the proportion of alightings, and the travel time between stations in each characteristic time period;

Step 23, by dividing the accumulated number of boardings in the characteristic time period f by τ, the passenger flow arrival rate of the station in the characteristic time period f on a single day is obtained; by dividing the accumulated proportion of the number of alightings by the number of departures on a single day, the average proportion of the number of alightings at the station in the characteristic time period f on a single day is obtained; by dividing the accumulated travel time between stations by the accumulated number of times, the average travel time between adjacent stations on the line is obtained;

Step 24, traverse all the read daily operation data, repeat steps 22 and 23, average all the daily operation data over the number of days, and obtain the passenger flow arrival rate α _kf , the proportion of passengers getting off the bus ρ _kf and the inter-station travel time T _k at each station in each characteristic time period f.

Furthermore, the multi-objective optimization function in step 3 includes the average waiting time Z _AWT and the total operating cost Z _TOC during the entire operation cycle. Z _AWT and Z _TOC are calculated by the following formulas (1) and (7), respectively:

In the formula, k is the index of the station; N ^K is the number of stations; m is the shift index; N ^M is the total number of departure shifts; The characteristic period obtained from step 2 analysis The passenger flow arrival rate of station k in the feature time period f obtained by step 2 _is the passenger flow arrival rate of station k in the feature time period f. It represents t _mk /τ rounded up; τ is the characteristic cycle length; t _mk is the estimated arrival time of shift m at station k, and t _(m-1)k is the estimated arrival time of shift m-1 at station k;

Where, ^NI is the total number of vehicle types, i is the vehicle type index, and _NTi is the vehicle type index in the departure schedule. is the number of trips for vehicle type i, R is the total mileage of the route; Ci _is the unit driving cost of vehicle type i.

Furthermore, the constraint condition set of step 4 includes the first and last bus departure time constraints, bus departure interval constraints, bus resource utilization constraints, available vehicle resource constraints and vehicle rated passenger capacity constraints.

Furthermore, step 6 specifically includes:

Step 61, sorting the timetable solutions in the Pareto optimal solution set TT ^ps obtained in step 5 in ascending order according to the objective function value Z _AWT , and recording the number N of timetable solutions contained therein.

Step 62, taking the first solution in the Pareto optimal solution set TT ^ps as the timetable scheme with priority on service quality TT-QoS, taking the last solution in the Pareto optimal solution set TT ^ps as the timetable scheme with priority on operating cost TT-Cost, and taking the solution in the middle of TT ^ps as the timetable scheme with balanced service and cost TT-Eq;

Step 63, after marking the average waiting time of passengers, the calculated value of the total operating cost and the required number of vehicles of each type in the timetable solutions TT-QoS, TT-Cost and TT-Eq obtained in step 62, the solutions are output as optional timetable solutions.

The present invention also provides a multi-vehicle timetable design system for an intelligent public transportation system, which comprises:

Parameter configuration module, which is used to set line operation parameters and multi-model bus resource configuration according to the actual operation of the bus line;

A data reading module is used to read the historical operation data of the bus line from the database, which includes the arrival and departure times of all departures at each station on the line and the number of passengers getting on and off the bus in multiple single days;

Data analysis module, which is used to statistically analyze the passenger flow demand pattern and vehicle travel time between stations within the corresponding operation cycle;

The timetable generation module is used to construct a multi-objective optimization function for evaluating the generated multi-model departure timetable scheme and a set of constraints for the multi-model timetable, and to use a multi-objective optimization algorithm to solve and obtain a Pareto optimal solution set TT ^ps ;

The solution output module is used to select solutions representing different benefit preferences from the Pareto optimal solution set TT ^ps and generate the corresponding departure timetable solution.

Furthermore, the passenger flow demand rules in the operation cycle in the data analysis module include the passenger flow arrival rate α _kf and the proportion of passengers getting off the bus ρ _kf at each station. The data analysis module includes:

The time period division unit is used to divide the entire operation cycle [t ₀ ,t _e ] into a number of characteristic time periods f∈F＝{1,2,…, ^NF } according to the characteristic cycle length τ; wherein t ₀ and t _e are the operation start time and end time of a single day, respectively, F is the characteristic time period set, and ^NF is the total number of the divided characteristic time periods f;

The calculation unit is used to traverse each departure schedule in the single-day operation data, correspond each arrival time to the corresponding characteristic time period, and accumulate the number of boardings, the proportion of alightings and the number of stops in each characteristic time period. Driving time; by dividing the accumulated number of boardings in the characteristic time period f by τ, the passenger arrival rate of the station in the characteristic time period f on a single day is obtained; by dividing the accumulated proportion of alightings by the number of departures on a single day, the average proportion of alightings at the station in the characteristic time period f on a single day is obtained; by dividing the accumulated travel time between stations by the accumulated number of times, the average travel time between adjacent stations of the line is obtained; finally, all single-day operation data are averaged over the number of days to obtain the passenger arrival rate α _kf , the proportion of alighting passengers ρ _kf and the travel time between stations T _k at each station in each characteristic time period f.

Furthermore, the multi-objective optimization function in the timetable generation module includes the average waiting time Z _AWT and the total operating cost Z _TOC during the entire operation cycle. Z _AWT and Z _TOC are calculated by the following formulas (1) and (7), respectively. The constraint set includes the first and last bus departure time constraints, bus departure interval constraints, bus resource utilization constraints, available vehicle resource constraints, and vehicle rated passenger capacity constraints:

Where, ^NI is the total number of vehicle types, i is the vehicle type index, _NTi is the number of trips of vehicle type i in the departure schedule plan, R is the total mileage of the line; _Ci is the unit driving cost of vehicle type i.

Furthermore, the solution output module includes:

A sorting unit, which is used to sort the timetable solutions in the Pareto optimal solution set TT ^ps obtained by the timetable generation module in ascending order according to the objective function value Z _AWT , update TT ^ps and record the number N of timetable solutions it contains;

a schedule screening unit, configured to select the first solution in the Pareto optimal solution set TT ^ps as the schedule with service quality priority TT-QoS, select the last solution in the Pareto optimal solution set TT ^ps as the schedule with operation cost priority TT-Cost, and select the solution in the middle of TT ^ps as the schedule with service-cost balance TT-Eq;

The departure schedule plan generation unit is used to convert the schedule plan TT-QoS, TT-Cost and TT-Eq After marking the average waiting time of passengers, the calculated value of the total operating cost and the number of vehicles of each type required, they are output as an optional timetable solution.

The present invention has the following advantages due to the adoption of the above technical solution: In particular,

The beneficial effects of the present invention are as follows:

1) Integrating the bus route departure times and vehicle types into the same framework for optimization calculation can better match bus capacity resources with time-varying passenger flow demand, which is superior to the existing two-stage calculation method in terms of resource utilization, balance and service quality;

2) The bus departure schedule design is defined and optimized in the form of multi-objective optimization, which can obtain different schedule plans that balance the benefits of both passengers and bus operators, effectively reduce passenger travel time and bus operating costs, and improve the overall service level of the bus system;

3) Compared with the general bus scheduling method that converts multiple optimization objectives into a single objective problem through scaling and weighted summation, this method can reduce unnecessary parameter settings and can be used by dispatchers in a user-friendly manner. Bus dispatchers or computer systems can select the final implementation plan based on the actual needs and preferences of bus operations.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a flow chart of a method for designing a timetable for a multi-vehicle intelligent public transportation system according to an embodiment of the present invention.

FIG. 2 is a block diagram of a system for designing a timetable for a multi-vehicle intelligent public transportation system according to an embodiment of the present invention.

FIG. 3 is a schematic diagram of a coding method for a timetable solution of a multi-vehicle intelligent public transportation system in an embodiment of the present invention.

DETAILED DESCRIPTION

The present invention is described in detail below with reference to the accompanying drawings and embodiments.

As shown in FIG1 , the method for designing a departure schedule for a multi-type intelligent public transportation system provided by an embodiment of the present invention includes:

Step 1: Set the route operation parameters and multi-model bus resource configuration according to the actual operation of the bus route.

The line operation parameters include: the start and end time _t0 and _te of the line operation period, the type of operation day (weekday or holiday), the total mileage of the line R, the rest time between single trains ^Ts , the maximum departure interval _Hmax , the minimum departure interval _Hmin and the lower limit of the train utilization ε.

The multi-model bus resource configuration includes: all models i, i∈I＝{1,2,…, ^NI } that can be dispatched for line operation, as well as the rated passenger capacity _Si , unit travel cost _Ci and available number of vehicles _Vi corresponding to each model.

Step 2: Read and process the historical operation data of the corresponding line for multiple days, and statistically analyze the passenger flow demand pattern and the vehicle travel time between stations during the corresponding operation cycle.

The historical operation data of the line includes: the arrival and departure times and the number of passengers getting on and off at each station of the line for all departure trips on multiple single days under the same operation day type. Where k represents the index of the station, k∈K＝{1,2,…, ^NK }.

The structure of a single-day historical operation data is shown in the following table (example):

In one embodiment, step 2 includes:

Step 21, divide the entire operation cycle [t ₀ , t _e ] into several characteristic time periods f∈F={1,2,…,N ^F } according to the characteristic cycle length τ. Wherein, t ₀ and t _e are the operation start time and end time of a single day, F is the characteristic time period set, and N ^F is the total number of the divided characteristic time periods f.

Step 22, traverse each departure schedule in the single-day operation data, correspond each arrival time to the corresponding characteristic period, and accumulate the number of boardings, the proportion of alightings, and the travel time between stations in each characteristic period. For example: the schedule set is represented by m∈M＝{1,2,…,N ^M }, j _m represents the departure time of schedule m, and i _m represents the departure vehicle model index of schedule m.

Step 23, by dividing the accumulated number of boardings in the characteristic time period f by τ, the passenger arrival rate at the station in the characteristic time period f on a single day is obtained; by dividing the accumulated proportion of alightings by the number of departures on a single day, the average proportion of alightings at the station in the characteristic time period f on a single day is obtained; by dividing the accumulated travel time between stations by the accumulated number of times, the average travel time between adjacent stations of the line is obtained.

Step 24, traverse all the read daily operation data, repeat steps 22 and 23, average all the daily data over the number of days, and obtain the passenger flow arrival rate α _kf , the proportion of passengers getting off the bus ρ _kf and the inter-station travel time T _k at each station in each characteristic time period f.

Here, “average all single-day operation data over the number of days to obtain the passenger flow arrival rate α _kf , the proportion of passengers getting off the bus ρ _kf and the travel time between stations T _k at each characteristic time period f” is used as an example: there are past The operation data of 10 working days can be used to calculate the three data (α _kf , ρ _kf , T _k ) calculated from each single-day operation data through steps 22 and 23. Step 24 averages the 10 single-day data to obtain the final data required by the method.

The above embodiment not only takes into account the arrival pattern of passengers, but also the proportion of passengers getting off at the station. Of course, the operation cycle can also be divided into multiple time periods, and only the passenger flow in each time period is counted to implement step 2.

Step 3: Construct a multi-objective optimization function for evaluating the generated multi-vehicle departure schedule.

In the target selection of the multi-objective optimization function, since the average waiting time of passengers in the entire operation cycle represents the benefit of traveling passengers, and the total operating cost in the entire operation cycle represents the benefit of the bus operator, there is a conflict between the two. Therefore, this embodiment uses the two as the multi-objective optimization function in step 3, thereby effectively weighing the benefits of both passengers and bus operators. Of course, other factors can also be considered on this basis, such as the degree of congestion of the bus. The following is an example of the average waiting time of passengers and the total operating cost in the entire operation cycle selected by the multi-objective optimization function.

For example, the average passenger waiting time Z _AWT is calculated by the following formula:

In the formula, k and k′ are the indexes of the two stations respectively; N ^K is the number of stations; m is the trip index; N ^M is the total number of departure trips; The characteristic period obtained from step 2 analysis The passenger arrival rate at inner station k; The characteristic period obtained from step 2 analysis The proportion of passengers getting off at inner station k; = represents the rounding up of t _mk /τ; τ is the characteristic cycle length; t _mk is the estimated arrival time of bus m at station k, t _(m-1)k is the estimated arrival time of bus m-1 at station k, both of which are calculated by the following formula (2); j _m represents the departure time of bus m; T _k′ is the travel time from station (k′-1) to station k′; L _mk is the number of passengers stranded at station k due to capacity limitation of bus m, L _(m-1)k is the number of passengers stranded at station k due to capacity limitation of bus m-1, L _(m-1)k , is the number of passengers stranded at station k′ due to capacity limitation of bus m-1, all three are calculated by the following formula (3); B _mk is the number of passengers boarding bus m at station k, calculated by the following formula (4); i _m is the vehicle model index of bus m; is the rated passenger capacity of vehicle type i _m ; A _mk is the number of passengers getting off at station k of shift m, which is calculated by the following formula (5); Q _mk is the number of passengers in the vehicle when shift m leaves station k, and Q _m(k-1) is the number of passengers in the vehicle when shift m leaves station k-1, which are calculated by the following formula (6) is calculated; K is the site set;

In particular, equations (3) to (6) recursively model passenger travel behavior from two dimensions, m and station k, and then calculate the number of passengers left at the station under the passenger capacity limit. Considering the waiting time of such passengers can further guarantee the quality of bus service.

For example, the total operating cost Z _TOC is related to the number of departures of each type of vehicle and can be calculated using the following formula (7):

Where i is the vehicle type index, ^NI is the total number of vehicle types, such as a bus operator has three types of vehicles: 5, 12, and 25 seats, then ^NI = 3; _NTi is the number of trips of vehicle type i in the departure schedule plan, R is the total mileage of the route; _Ci is the unit driving cost of vehicle type i.

It should be noted that the calculation of operating costs is not limited to the vehicle's driving costs, but also includes other vehicle-related unit distance maintenance, supporting facilities construction, average vehicle purchase costs and other related costs.

Compared with the prior art, the embodiment of the present invention takes into account the design problem of bus departure times for various types of vehicles through steps 2 and 3. Therefore, the present invention is adaptable to different vehicle composition forms of the bus system, and by jointly optimizing the departure times and vehicle types of bus lines, the obtained timetable can be made closer to the global optimal solution.

Step 4: Construct a set of constraints for the multi-vehicle timetable to ensure the feasibility and rationality of the resulting timetable solution under the given settings and resource allocation. This includes but is not limited to the following constraints:

First and last bus departure time constraints: To ensure that the departure times in the timetable can cover the entire operating period, the first and last bus departure times are specified at the start and end times of the operating period, that is, j ₁ = t ₀ ,

Train departure interval constraint: To avoid the distance between adjacent trains being too short or too long, which may affect the stability of line operation, the difference between the departure times of adjacent trains should be within the range set in step 1, that is, H _min ≤ j _m -j _(m-1) ≤ H _max .

Constraint on shift resource utilization: To avoid the phenomenon of oversupply of shift vehicles with a large number of empty seats during their entire service process, the maximum passenger capacity of each shift should be greater than the lower limit of the shift utilization set in step 1, that is,

Available vehicle resource constraints: To avoid the situation where the departure times of a single vehicle type are too concentrated in the timetable, so that all available vehicles are on duty when a certain vehicle type is required to depart, the number of departure times of a single vehicle type within any line cycle time r ^rt in the entire operation cycle should not exceed the number of available vehicles _Vi set in step 1.

Vehicle rated passenger capacity limit: For driving safety reasons, it should be ensured that the number of passengers on a bus at any time does not exceed the rated passenger capacity determined by the vehicle model. This may result in all arriving passengers being unable to get on the bus and having to wait for subsequent buses.

Step 5: Use a multi-objective optimization algorithm to obtain the Pareto optimal solution set.

Pareto optimal solution set: There is no solution that can achieve the optimal solution for all objectives at the same time in a multi-objective optimization problem, so only the Pareto optimal solution set PS can be obtained, which contains multiple feasible solutions that cannot be compared with each other. That is, for any solution s∈PS, there is no solution s′ in the other solutions {PS\s} in the solution set that can make each of its objectives absolutely better than s.

Taking the non-dominated sorting genetic algorithm as an example, the specific implementation of step 5 is described, including the following steps:

Step 51: The computer randomly generates N _s feasible timetable solutions TT _p as initial parent solutions, satisfying the constraints described in step 4, and performs real number encoding on the initial parent solutions, as shown in FIG3 .

Step 52: Calculate two objective function values of each solution according to the method described in step 3, and perform non-dominant sorting of each solution according to the objective function values.

Step 53: Use the league selection method to select N _s winning solutions, and change the timetable plan by inserting new shifts, canceling existing shifts, or reordering the vehicle model sequence within a certain period to obtain a new solution TT _s .

Step 54: Combine TT _p and TT _s to obtain TT _c , calculate two objective function values of each solution according to the method described in step 3, and perform non-dominant sorting on each solution according to the objective function values.

Step 55: According to the ranking level in step 54, the first _Ns timetable solutions are selected as the next generation TT _p , and the solutions with level 1 are recorded in TT ^ps .

Step 56: Repeat steps 53 to 55 until a final Pareto optimal solution set TT ^ps is output after a predetermined number of iterations.

In the above embodiments, the multi-objective optimization algorithm in step 5 is not limited to using the multi-objective genetic Algorithm, other evolution-based multi-objective optimization algorithms can also solve the model, such as the multi-objective particle swarm optimization algorithm.

The embodiment of the present invention balances the two types of objectives involving operators and passengers through step 5, so that solutions representing different benefit preferences can be obtained for the dispatcher to choose.

Step 6: Filter solutions representing different benefit preferences from the optimal solution set and generate corresponding departure schedule plans for selection by dispatchers or computer systems.

Taking the three benefit preferences of service quality priority, operating cost priority and service-cost balance as an example, the following steps are included:

Step 61, sort the timetable solutions in the Pareto optimal solution set TT ^ps obtained in step 5 in ascending order according to the objective function value Z _AWT , update TT ^ps and record the number N of timetable solutions it contains.

Step 62, take the first solution in TT ^ps as the timetable scheme TT-QoS (Chinese full name "Timetable-Quality of Service First") with priority on service quality. Under the timetable scheme TT-QoS, the average waiting time of passengers is minimized. Take the last solution in TT ^ps as the timetable scheme TT-Cost (Chinese full name "Timetable-Operating Cost First") with priority on operating cost. Under the timetable scheme TT-Cost, the total operating cost is minimized. Take the solution in the middle of TT ^ps , for example, the N/2th solution, and take (N-1)/2 when N is an even number, as the timetable scheme TT-Eq (Chinese full name "Timetable-Equilibrium Model") with service-cost balance.

Step 63, after marking the average waiting time of passengers, the calculated value of the total operating cost and the required number of vehicles of each type of vehicle of the timetable solutions TT-QoS, TT-Cost and TT-Eq obtained in step 62, the corresponding departure timetable solution is output to facilitate the dispatcher or computer system to make subsequent decisions.

The output optional timetable method information structure is shown in the following table (taking three types of vehicles as an example):

The above steps 1 to 6 can be represented by FIG. 1 .

The departure times of bus routes and their corresponding bus models are integrated into the same framework for optimization calculation, and solved using a multi-objective optimization method to obtain a timetable solution representing different target preferences for bus dispatchers to choose from, thereby better balancing the benefits of both bus operators and passengers.

As shown in FIG2 , an embodiment of the present invention further provides a multi-vehicle timetable design system for an intelligent public transportation system. For example, the system can execute the above method. The system includes a parameter configuration module, a data reading module, a data analysis module, a timetable generation module, and a solution output module, wherein:

The parameter configuration module is used to set line operation parameters and multi-model bus resource configuration according to the actual operation of the bus line.

The data reading module is used to read the historical operation data of the bus line from the database. The data includes the arrival and departure times of all departure trips at each station on the line and the number of passengers getting on and off the bus in multiple single days.

The data analysis module is used to statistically analyze the passenger flow demand patterns and vehicle travel time between stations within the corresponding operating cycle.

The timetable generation module is used to construct a multi-objective optimization function for evaluating the generated multi-model departure timetable scheme and a set of constraints for the multi-model timetable, and uses a multi-objective optimization algorithm to solve and obtain the Pareto optimal solution set TT ^ps .

In one embodiment, the passenger flow demand rules in the operation cycle in the data analysis module include the passenger flow arrival rate α _kf and the proportion of passengers getting off the bus ρ _kf at each station. The data analysis module includes a time period division unit and a calculation unit, wherein:

The time period division unit is used to divide the entire operation cycle [t ₀ ,t _e ] into several time periods according to the characteristic period length τ. Feature time period f∈F＝{1,2,…, ^NF }; where _t0 and _te are the start and end times of the operation on a single day, respectively, F is the feature time period set, and ^NF is the total number of feature time periods f divided.

The calculation unit is used to traverse each departure trip in the single-day operation data, correspond each arrival time to the corresponding characteristic time period, and accumulate the number of boardings, the proportion of alightings and the travel time between stations in each characteristic time period; by dividing the accumulated number of boardings in the characteristic time period f by τ, the passenger flow arrival rate of the station in the characteristic time period f of a single day is obtained; by dividing the accumulated proportion of alightings by the number of departure trips on a single day, the average proportion of alightings at the station in the characteristic time period f of a single day is obtained; by dividing the accumulated travel time between stations by the accumulated number of times, the average travel time between adjacent stations of the line is obtained; finally, all the single-day operation data are averaged over the number of days to obtain the passenger flow arrival rate α _kf , the proportion of alighting passengers ρ _kf and the travel time between stations T _k of each station in each characteristic time period f.

In one embodiment, the multi-objective optimization function in the timetable generation module includes the average waiting time Z _AWT of passengers and the total operating cost Z _TOC in the entire operation cycle. Z _AWT and Z _TOC are calculated by formula (1) and formula (7) respectively. The constraint condition set includes the first and last bus departure time constraint, bus departure interval constraint, bus resource utilization constraint, available vehicle resource constraint and vehicle rated passenger capacity limit.

In one embodiment, the plan output module includes a sorting unit, a schedule plan generating unit and a departure schedule plan generating unit, wherein:

The sorting unit is used to sort the timetable solutions in the Pareto optimal solution set TT ^ps obtained by the timetable generation module in ascending order according to the objective function value Z _AWT , update TT ^ps and record the number N of timetable solutions it contains.

The schedule screening unit is used to use the first solution in the Pareto optimal solution set TT ^ps as the schedule plan with service quality priority TT-QoS, use the last solution in the Pareto optimal solution set TT ^ps as the schedule plan with operation cost priority TT-Cost, and use the solution in the middle of TT ^ps as the schedule plan with service-cost balance TT-Eq.

The departure schedule plan generating unit is used to mark the average waiting time of passengers, the calculated value of the total operating cost and the required number of vehicles of each type in the schedule plans TT-QoS, TT-Cost and TT-Eq, and output them as optional schedule plans.

Finally, it should be pointed out that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them. Those skilled in the art should understand that the technical solutions described in the above embodiments may be modified, or some of the technical features thereof may be replaced by equivalents; these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

A multi-vehicle timetable design method for an intelligent public transportation system, characterized by comprising:

Step 1: Set the route operation parameters and multi-model bus resource configuration according to the actual operation of the bus route;

Step 2: read and process the historical operation data of the corresponding line for multiple days, and statistically analyze the passenger flow demand pattern and the vehicle travel time between stations within the corresponding operation period;

Step 3, construct a multi-objective optimization function for evaluating the generated multi-vehicle departure schedule scheme;

Step 4, construct the constraint set of the multi-vehicle timetable;

Step 5: Based on the multi-objective optimization function and the constraint condition set of the multi-vehicle timetable, a multi-objective optimization algorithm is used to obtain the Pareto optimal solution set TT ps ;

Step 6: Select solutions representing different benefit preferences from the Pareto optimal solution set TT ps and generate the corresponding departure schedule plan.
The multi-type timetable design method for an intelligent public transportation system according to claim 1, characterized in that the passenger flow demand law within the operation cycle in step 2 includes the passenger flow arrival rate α kf and the proportion of passengers getting off the bus ρ kf at each station, and step 2 includes:

Step 21, divide the entire operation cycle [t 0 ,t e ] into a number of characteristic time periods f∈F={1,2,…, NF } according to the characteristic cycle length τ; where t 0 and t e are the start and end times of the operation of a single day, respectively, F is the characteristic time period set, and NF is the total number of the divided characteristic time periods f;

Step 22, traverse each departure schedule in the single-day operation data, correspond each arrival time to the corresponding characteristic time period, and accumulate the number of boardings, the proportion of alightings, and the travel time between stations in each characteristic time period;

Step 23, by dividing the accumulated number of boardings in the characteristic time period f by τ, the passenger flow arrival rate of the station in the characteristic time period f on a single day is obtained; by dividing the accumulated proportion of the number of alightings by the number of departures on a single day, the average proportion of the number of alightings at the station in the characteristic time period f on a single day is obtained; by dividing the accumulated travel time between stations by the accumulated number of times, the average travel time between adjacent stations on the line is obtained;

Step 24, traverse all the read daily operation data, repeat steps 22 and 23, average all the daily operation data over the number of days, and obtain the passenger flow arrival rate α kf , the proportion of passengers getting off the bus ρ kf and the inter-station travel time T k at each station in each characteristic time period f.
The multi-type timetable design method for an intelligent public transportation system as claimed in claim 2, wherein The characteristic is that the multi-objective optimization function in step 3 includes the average waiting time Z AWT and the total operating cost Z TOC of passengers in the entire operation cycle, and Z AWT and Z TOC are calculated by the following formulas (1) and (7) respectively:

In the formula, k is the index of the station; N K is the number of stations; m is the shift index; N M is the total number of departure shifts; The characteristic period obtained from step 2 analysis The passenger flow arrival rate of station k in the feature time period f obtained by step 2 is the passenger flow arrival rate of station k in the feature time period f. It means t mk /τ is rounded up; τ is the characteristic cycle length; t mk is the estimated arrival time of shift m at station k, and t (m-1)k is the estimated arrival time of shift m-1 at station k;

Where, NI is the total number of vehicle types, i is the vehicle type index, NTi is the number of trips of vehicle type i in the departure schedule plan, R is the total mileage of the line; Ci is the unit driving cost of vehicle type i.
The multi-type timetable design method for an intelligent public transportation system as described in claim 2 is characterized in that the constraint condition set of step 4 includes the first and last bus departure time constraints, the bus departure interval constraints, the bus resource utilization constraints, the available vehicle resource constraints and the vehicle rated passenger capacity restrictions.
The method for designing a multi-type timetable for an intelligent public transportation system according to any one of claims 1 to 4, characterized in that step 6 specifically comprises:

Step 61, sorting the timetable solutions in the Pareto optimal solution set TT ps obtained in step 5 in ascending order according to the objective function value Z AWT , and recording the number N of timetable solutions contained therein.

Step 62, taking the first solution in the Pareto optimal solution set TT ps as the timetable scheme with priority on service quality TT-QoS, taking the last solution in the Pareto optimal solution set TT ps as the timetable scheme with priority on operating cost TT-Cost, and taking the solution in the middle of TT ps as the timetable scheme with balanced service and cost TT-Eq;

Step 63, after marking the average waiting time of passengers, the calculated value of the total operating cost and the required number of vehicles of each type in the timetable solutions TT-QoS, TT-Cost and TT-Eq obtained in step 62, the solutions are output as optional timetable solutions.
A system for designing a multi-type timetable for an intelligent public transportation system, characterized by comprising:

The parameter configuration module is used to set the line operation parameters and Multi-model bus resource allocation;

A data reading module is used to read the historical operation data of the bus line from the database, which includes the arrival and departure times of all departures at each station on the line and the number of passengers getting on and off the bus in multiple single days;

Data analysis module, which is used to statistically analyze the passenger flow demand pattern and vehicle travel time between stations within the corresponding operation cycle;

The timetable generation module is used to construct a multi-objective optimization function for evaluating the generated multi-model departure timetable scheme and a set of constraints for the multi-model timetable, and to use a multi-objective optimization algorithm to solve and obtain a Pareto optimal solution set TT ps ;

The solution output module is used to select solutions representing different benefit preferences from the Pareto optimal solution set TT ps and generate the corresponding departure timetable solution.
The multi-type timetable design system for an intelligent public transportation system according to claim 6, characterized in that the passenger flow demand law within the operation cycle in the data analysis module includes the passenger flow arrival rate α kf and the proportion of passengers getting off the bus ρ kf at each station, and the data analysis module includes:

The time period division unit is used to divide the entire operation cycle [t 0 ,t e ] into a number of characteristic time periods f∈F＝{1,2,…, NF } according to the characteristic cycle length τ; wherein t 0 and t e are the operation start time and end time of a single day, respectively, F is the characteristic time period set, and NF is the total number of the divided characteristic time periods f;

A calculation unit is used to traverse each departure trip in the single-day operation data, correspond each arrival time to the corresponding characteristic time period, and accumulate the number of boardings, the proportion of alightings and the travel time between stations in each characteristic time period; by dividing the accumulated number of boardings in the characteristic time period f by τ, the passenger flow arrival rate of the station in the characteristic time period f of a single day is obtained; by dividing the accumulated proportion of alightings by the number of departure trips on a single day, the average proportion of alightings at the station in the characteristic time period f of a single day is obtained; by dividing the accumulated travel time between stations by the accumulated number of times, the average travel time between adjacent stations of the line is obtained; finally, all the single-day operation data are averaged over the number of days to obtain the passenger flow arrival rate α kf , the proportion of alighting passengers ρ kf and the travel time between stations T k at each station in each characteristic time period f.
The multi-type timetable design system for an intelligent public transportation system as claimed in claim 6 is characterized in that the multi-objective optimization function in the timetable generation module includes the average waiting time Z AWT of passengers in the entire operation cycle and the total operation cost Z TOC , Z AWT and Z TOC are calculated by the following formula (1) and formula (7) respectively, and the constraint condition set includes the first and last bus departure time constraints, the bus departure interval constraints, the bus resource utilization constraint, the available vehicle resource constraint and the vehicle rated passenger capacity limit:

In the formula, k is the index of the station; N K is the number of stations; m is the shift index; N M is the total number of departure shifts; The characteristic period obtained from step 2 analysis The passenger flow arrival rate of station k in the feature time period f obtained by step 2 is the passenger flow arrival rate of station k in the feature time period f. It represents t mk /τ rounded up; τ is the characteristic cycle length; t mk is the estimated arrival time of shift m at station k, and t (m-1)k is the estimated arrival time of shift m-1 at station k;

Where, NI is the total number of vehicle types, i is the vehicle type index, NTi is the number of trips of vehicle type i in the departure schedule plan, R is the total mileage of the line; Ci is the unit driving cost of vehicle type i.
The multi-type timetable design system for an intelligent public transportation system according to any one of claims 6 to 8, characterized in that the solution output module comprises:

A sorting unit, which is used to sort the timetable solutions in the Pareto optimal solution set TT ps obtained by the timetable generation module in ascending order according to the objective function value Z AWT , update TT ps and record the number N of timetable solutions it contains;

a schedule screening unit, configured to select the first solution in the Pareto optimal solution set TT ps as the schedule with service quality priority TT-QoS, select the last solution in the Pareto optimal solution set TT ps as the schedule with operation cost priority TT-Cost, and select the solution in the middle of TT ps as the schedule with service-cost balance TT-Eq;

The departure schedule plan generating unit is used to mark the average waiting time of passengers, the calculated value of the total operating cost and the required number of vehicles of each type in the schedule plans TT-QoS, TT-Cost and TT-Eq, and output them as optional schedule plans.