An Urban Rail Transit Train Operation Diagram Adjustment Method and Device Based on Genetic Algorithm

Abstract

The present invention provides a method and device for adjusting the operation diagram of urban rail trains based on a genetic algorithm, which relates to the technical field of urban rail train scheduling. First, the present invention analyzes the train punctuality rate, stop time, passenger flow congestion degree, average train delay time, and passenger flow satisfaction rate by obtaining historical data, including train operation plans, station information data, passenger flow data, and train operation data, and generates train operation performance indicators and passenger experience indicators. Then, these two types of indicators are used to construct a riding state model and train it. Taking the historical train operation plan as the initial population, the fitness values of the individuals in the population are calculated and iteratively optimized by the genetic algorithm. Finally, the individual with the highest fitness value is selected as the optimal operation plan. Finally, based on this optimal plan, the current train operation plan is adjusted to formulate a new train scheduling plan to improve the operation efficiency and passenger experience.

Description

Technical Field

[0001] The present invention relates to the technical field of urban rail transit train dispatching, and specifically to an urban rail transit train operation diagram adjustment method and device based on a genetic algorithm. Background Art

[0002] With the acceleration of the urbanization process, urban rail transit has become an important means to solve urban traffic congestion and improve transportation efficiency. However, the matching problem between the operation dispatching of urban rail transit trains and passenger demand has become increasingly prominent, resulting in a series of problems such as low train punctuality rate, long average passenger waiting time, and crowded passenger flow. These problems not only affect the travel experience of passengers, but also reduce the overall operation efficiency and service level of urban rail transit. Therefore, how to effectively adjust the urban rail transit train operation diagram to achieve a rapid response to passenger demand has become an important technical challenge in urban rail transit management.

[0003] In existing dispatching optimization methods, although there are empirical adjustments based on historical data and some traditional optimization algorithms, it is often difficult to comprehensively consider multiple factors. As an adaptive optimization method, the genetic algorithm has strong global search ability and can find the optimal solution in complex multi-objective optimization problems. Therefore, the urban rail transit train operation diagram adjustment method based on the genetic algorithm can more flexibly respond to the dynamically changing passenger flow situation and achieve a more reasonable train dispatching plan.

[0004] In the prior art, the publication number CN113128774B discloses a train operation adjustment optimization method under a fixed train operation line sequence, which configures an initial train operation plan, basic train operation data, and fixed train operation line data; based on the railway line topology structure, the initial train operation plan, the basic train operation data, and the fixed train operation line data, a train operation adjustment optimization model under a fixed train operation line sequence is established based on a spatio-temporal network; an integer programming algorithm is used to solve the train operation adjustment optimization model under a fixed train operation line sequence to obtain a train operation adjustment plan with a fixed train operation line sequence.

[0005] The main problems of the above method are: lack of flexibility. When the actual passenger flow or operation status changes, the plan cannot quickly adapt, resulting in a mismatch between the train operation plan and the actual demand; moreover, when the integer programming algorithm solves the optimization model, it may fall into a local optimal solution, affecting the optimization result; when facing a new line or a new operation plan, it is necessary to re-establish the model and make complex adjustments, with weak adaptability.

[0006] The above information disclosed in the background art section is only used to enhance the understanding of the background of the present disclosure, and thus it may include information that does not constitute the prior art known to those of ordinary skill in the art. Summary of the Invention

[0007] The object of the present invention is to provide a method and device for adjusting the operation diagram of urban rail trains based on a genetic algorithm to solve the problems raised in the above-mentioned background technology.

[0008] To achieve the above object, the present invention provides the following technical solutions:

[0009] A method for adjusting the operation diagram of urban rail trains based on a genetic algorithm, the specific steps include:

[0010] Step 1: Obtain historical data, where the historical data includes the train operation plan and the corresponding state parameters obtained daily. The train operation plan includes the train stop sequence, the train operation time sequence, and the arrival time sequence, and the state parameters include station information data, passenger flow data, and train operation data;

[0011] Step 2: Based on the state parameters, generate the train punctuality rate, the train stop time, the passenger flow congestion degree, the average train delay time, and the passenger flow satisfaction rate. Based on the train punctuality rate and the average train delay time, generate the train operation performance index, and based on the average train stop time, the passenger flow congestion degree, and the passenger flow satisfaction rate, generate the passenger experience index;

[0012] Step 3: Construct a riding state model, use the train operation plan in the historical data as the input, and the corresponding train operation performance index and passenger experience index as the labels to train the riding state model;

[0013] Step 4: Use the train operation plan in the historical data as the individuals of the initial population, calculate the fitness value of the individuals based on the train operation performance index and the passenger experience index, take the maximization of the fitness value as the optimization goal, and perform iterative optimization on the individuals based on the genetic algorithm and the riding state model until the predetermined number of iterations is reached, and select the individual with the highest fitness value as the best train operation plan;

[0014] Step 5: Obtain the current train operation plan, and based on the best train operation plan, determine the train dispatching plan for adjusting the current train operation plan to make the current train operation plan close to the best train operation plan.

[0015] Further, the station information data includes the maximum allowable passenger flow of each station; the passenger flow data includes the passenger flow of each station and the number of passengers boarding different trains at this station; the train operation data includes the total number of trains, the number of trains arriving on time, the actual stop time of the trains at each station, the remaining passenger capacity, and the train delay time.

[0016] Further, the principle for generating the train operation performance index is:

[0017] The formula for generating the train punctuality rate is:

[0018]

[0019] Among them, R represents the train punctuality rate, M time represents the number of trains arriving on time, and M represents the total number of trains;

[0020] The formula for generating the average train delay time is as follows:

[0021]

[0022] Among them, Y represents the average train delay time, i represents the index of the train, and i ∈ [1, M], j represents the index of the station, J represents the total number of stations, D i,j represents the train delay time when the i-th train arrives at the j-th station;

[0023] The formula for generating the train operation performance index is as follows:

[0024]

[0025] Among them, A represents the train operation performance index, α1 and α2 respectively represent the weight coefficients of the train punctuality rate and the average train delay time, and α1 + α2 = 1.

[0026] Furthermore, the principle for generating the passenger experience index is as follows:

[0027] The formula for generating the average train dwell time is as follows:

[0028]

[0029] Among them, W represents the average train dwell time, j represents the index of the station, J represents the total number of stations, W i,j represents the actual dwell time of the i-th train at the j-th station;

[0030] The formula for generating the passenger flow congestion degree is as follows:

[0031]

[0032] Among them, C represents the passenger flow congestion degree, C j,cur represents the passenger flow at the j-th station, C j,max represents the maximum allowable passenger flow at the j-th station;

[0033] The formula for generating the passenger flow satisfaction rate is as follows:

[0034]

[0035] Among them, Z represents the passenger flow satisfaction rate, E i,jDenote the remaining passenger capacity of the $i$-th train at the $j$-th station as $U$. i,j Denote the number of passengers boarding the $i$-th train at the $j$-th station.

[0036] The principle for generating the passenger experience index is as follows:

[0037] $B = \beta_1\cdot(1 - W)+\beta_2\cdot(1 - C)+\beta_3\cdot Z$

[0038] Among them, $B$ represents the passenger experience index, $\beta_1$, $\beta_2$, and $\beta_3$ respectively represent the weight coefficients of the average passenger waiting time, the passenger flow congestion degree, and the passenger flow satisfaction rate, $\beta_1+\beta_2+\beta_3 = 1$, and $\beta_2>\beta_3>\beta_1$.

[0039] Furthermore, the principle for generating the individuals of the initial population is as follows:

[0040] Each individual in the population includes the following data elements:

[0041] Train stop sequence, $S$ i $=[s$ i,1 ,$s$ i,2 ,\(\cdots\),$s$ i,j ,\(\cdots\),$s$ i,J ;

[0042] Among them, $S$ i represents the stop sequence of the $i$-th train, $i$ represents the train index, $j$ represents the station index, and $j = 1,2,\cdots,J$, $J$ represents the total number of stations, and $s$ i,j represents the $j$-th stop station of the $i$-th train;

[0043] Train running time sequence, $T$ i $=[t$ i,2 ,$t$ i,3 ,\(\cdots\),$t$ i,j ,\(\cdots\),$t$ i,J ;

[0044] Among them, $T$ i represents the running time sequence of the $i$-th train, and $t$ i,j represents the running time of the $i$-th train from the $j$-th station to the $j - 1$-th station;

[0045] Arrival time sequence, $H$ i $=[h$ i,1 ,$h$ i,2 ,\(\cdots\),$h$ i,j ,\(\cdots\),$h$ i,J ;

[0046] Among them, $H$ i represents the arrival time sequence of the $i$-th train, and $h$ i,jDenote the arrival time of the $i$-th train at the $j$-th station; for each train, the individual formed is:

[0047] I i =[S i ,T i ,H i

[0048] where $I$ i denotes the individual corresponding to the $i$-th train in the population.

[0049] Furthermore, the formula for generating the fitness value is:

[0050] F=w1·e A +w2·e B

[0051] where $F$ represents the fitness value, $A$ and $B$ respectively represent the train operation performance index and the passenger experience index, $w1$ and $w2$ respectively represent the weight coefficients of the train operation performance index and the passenger experience index, $w1>w2$, and $w1 + w2=1$.

[0052] Furthermore, the principle for iteratively optimizing individuals based on the genetic algorithm is:

[0053] Based on the tournament selection method, randomly select two individuals from the initial population, compare the fitness values of the selected individuals, select the individual with higher fitness between the two as the parent, repeat the above process until the required number of parent individuals is selected, which is regarded as completing one round of selection, and randomly perform crossover and mutation on the selected parent individuals to generate new offspring, combine the offspring and the parents to form a new population, input the individuals in the new population into the riding state model, calculate the train operation performance index and the passenger experience index of each individual, calculate the fitness value of the individual based on the train operation performance index and the passenger experience index, record the individual with the highest fitness value as the current optimal solution, the above operations are regarded as one iteration, set the maximum number of iterations to 100, when the number of iterations reaches 100, or the improvement amplitude of the optimal fitness in 10 consecutive iterations is less than 5%, then terminate the iteration, and select the individual of the current optimal solution as the best train operation plan.

[0054] The present invention also provides an urban rail train operation diagram adjustment device based on the genetic algorithm, and the device is used to implement the above-mentioned urban rail train operation diagram adjustment method based on the genetic algorithm, specifically including:

[0055] A data acquisition module, configured to obtain historical data, where the historical data includes the train operation plan and the corresponding state parameters obtained day by day, the train operation plan includes the train stop sequence, the train operation time sequence and the arrival time sequence, and the state parameters include the station information data, the passenger flow data and the train operation data;​

[0056] An index generation module, configured to generate a train punctuality rate, a train stop time, a passenger flow congestion degree, an average train delay time, and a passenger flow satisfaction rate based on status parameters, generate a train operation performance index based on the train punctuality rate and the average train delay time, and generate a passenger experience index based on the average train stop time, the passenger flow congestion degree, and the passenger flow satisfaction rate;

[0057] A model training module, configured to build a riding status model, use the train operation plan in historical data as input, and the corresponding train operation performance index and passenger experience index as labels to train the riding status model;

[0058] A plan optimization module, configured to use the train operation plan in historical data as an individual of an initial population, calculate the fitness value of the individual based on the train operation performance index and the passenger experience index, take maximizing the fitness value as an optimization goal, and perform iterative optimization on the individual based on a genetic algorithm and the riding status model until a predetermined number of iterations is reached, and select the individual with the highest fitness value as the best train operation plan;

[0059] A plan scheduling module, configured to obtain the current train operation plan, and determine a train scheduling plan for adjusting the current train operation plan based on the best train operation plan, so that the current train operation plan is close to the best train operation plan.

[0060] Compared with the prior art, the beneficial effects of the present invention are:

[0061] The present invention utilizes historical train operation plans and status parameter data, which can comprehensively reflect all aspects of train operation. The multi-dimensional evaluation method helps to better evaluate the train operation performance and passenger experience, so as to achieve more targeted optimization. Moreover, the historical data obtained day by day can timely reflect the changes in passenger flow and train operation status, enabling the train operation plan to be dynamically adjusted, improving the operation efficiency and passenger satisfaction; and training the riding status model based on historical data can provide accurate suggestions for current decisions and improve the efficiency of train scheduling and operation decisions.

[0062] This statement also calculates the fitness value of an individual through train operation performance indicators and passenger experience indicators, and uses a genetic algorithm for optimization to search for the individual with the maximum fitness value, objectively evaluating the advantages and disadvantages of each scheduling plan from multiple dimensions, generating fitness values to make the optimization process more explicit and targeted; in the face of different operating environments and passenger flow changes, the genetic algorithm can provide flexible scheduling plans, helping the operator quickly adapt to changes and avoid falling into local optimal solutions. The results of each optimization can provide feedback for the next round of scheduling decisions, forming a continuously improving closed loop, enhancing the adaptability of the plan, and because of the fast iterative optimization, multiple possible operation plans can be provided in a short time, helping decision-makers quickly evaluate and select the best plan, and improving the efficiency of decision-making scheduling. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 It is a schematic flowchart of the method of the embodiment of the present invention;

[0064] Figure 2 It is a schematic diagram of the device module of the embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0065] To make the objectives, technical solutions and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below in conjunction with specific embodiments.

[0066] It should be noted that unless otherwise defined, the technical terms or scientific terms used in the present invention should have the ordinary meanings understood by those with ordinary skills in the field to which the present invention belongs. The "first", "second" and similar terms used in the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. The terms such as "including" or "comprising" mean that the elements or objects appearing before this word cover the elements or objects listed after this word and their equivalents, without excluding other elements or objects. The terms such as "connected" or "coupled" are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. The terms such as "upper", "lower", "left", "right" are only used to represent relative positional relationships, and when the absolute position of the object being described changes, the relative positional relationship may also change accordingly.

[0067] Embodiment:

[0068] Please refer to Figure 1 , the present invention provides a technical solution:

[0069] An urban rail train operation diagram adjustment method based on a genetic algorithm, the specific steps include:

[0070] Step 1: Obtain historical data, where the historical data includes the train operation plan and the corresponding status parameters obtained daily. The train operation plan includes the train stop sequence, the train running time sequence, and the arrival time sequence. The status parameters include station information data, passenger flow data, and train operation data;

[0071] In this embodiment, the station information data includes the maximum allowable passenger volume of each station; the passenger flow data includes the passenger volume of each station and the number of passengers boarding different trains at this station; the train operation data includes the total number of trains, the number of trains arriving on time, the actual stop time of the train at each station, the remaining passenger capacity, and the train delay time.

[0072] Step 2: Based on the status parameters, generate the train punctuality rate, the train stop time, the passenger flow congestion degree, the average train delay time, and the passenger flow satisfaction rate. Based on the train punctuality rate and the average train delay time, generate the train operation performance index. Based on the average train stop time, the passenger flow congestion degree, and the passenger flow satisfaction rate, generate the passenger experience index;

[0073] In this embodiment, the principle for generating the train operation performance index is as follows:

[0074] The formula for generating the train punctuality rate is:

[0075]

[0076] Among them, R represents the train punctuality rate, m time represents the number of trains arriving on time, and m represents the total number of trains;

[0077] The train punctuality rate directly reflects the operation efficiency of the train. The stop situation of the train at the station, the running time between stations, and the actual arrival time of the train at the station will all affect the train punctuality rate. The punctuality rate of the train is directly proportional to the number of trains arriving on time;

[0078] The formula for generating the average train delay time is:

[0079]

[0080] Among them, Y represents the average train delay time, i represents the index of the train, and i ∈ [1, M], j represents the index of the station, J represents the total number of stations, D i,j represents the train delay time of the i-th train arriving at the j-th station;

[0081] The average train delay time reflects the average value of the total delay time of each train at all stations among all trains. The larger the average delay time, the more congested the overall train operation;

[0082] The formula for generating train operation performance indicators is as follows:

[0083]

[0084] Among them, A represents the train operation performance indicator, and α1 and α2 represent the weight coefficients of the train punctuality rate and the average train delay time respectively.

[0085] The train operation performance indicator comprehensively considers the train punctuality rate and the delay situation, reflecting the quality of the train operation plan. The higher the train operation performance indicator, the better the train operation plan. The train operation performance indicator is directly proportional to the train punctuality rate and inversely proportional to the average train delay time; and α1 = α2 = 0.5.

[0086] The principle for generating the passenger experience indicator is as follows:

[0087] The formula for generating the average train dwell time is as follows:

[0088]

[0089] Among them, W represents the average train dwell time, j represents the index of the station, J represents the total number of stations, and W i,j represents the actual dwell time of the i-th train at the j-th station;

[0090] The average train dwell time represents the average of the dwell times of each train at all stations among all trains. The shorter the average dwell time, the smoother the train operation and the more convenient it is for passengers to get on and off; the longer the average dwell time, the more congested the train operation.

[0091] The formula for generating the passenger flow congestion degree is as follows:

[0092]

[0093] Among them, C represents the passenger flow congestion degree, and C j,cur represents the passenger flow at the j-th station, and C j,max represents the maximum allowable passenger flow at the j-th station;

[0094] The passenger flow congestion degree reflects the relationship between the actual passenger flow at the station and the maximum acceptance capacity of the station. The higher the actual passenger flow, the higher the passenger flow congestion degree and the lower the passenger experience;

[0095] The formula for generating the passenger flow satisfaction rate is as follows:

[0096]

[0097] Among them, Z represents the passenger flow satisfaction rate, and E i,j represents the remaining passenger capacity of the i-th train at the j-th station, and U i,jIndicates the number of passengers taking the i-th train at the j-th station;

[0098] The passenger flow satisfaction rate indicates the satisfaction of the remaining passenger capacity of the train with the passengers at the next station. A high passenger flow satisfaction rate means that the remaining passenger capacity of the train is high, and it is easier to meet the riding needs of the passengers at the next station;

[0099] The principle for generating the passenger experience index is as follows:

[0100] B = β1·(1 - W) + β2·(1 - C) + β3·Z

[0101] Where, B represents the passenger experience index, β1, β2, and β3 respectively represent the weight coefficients of the average train stop time, passenger flow congestion degree, and passenger flow satisfaction rate, β1 + β2 + β3 = 1, and β2 > β3 > β1.

[0102] The passenger experience index comprehensively considers the boarding and alighting time of passengers, the congestion situation at the station, and the passenger-carrying capacity of the train for passengers, reflecting the riding experience of passengers. The higher the passenger experience index, the better the riding experience of passengers, the better the train operation plan. The passenger experience index is inversely proportional to the average train stop time and the passenger flow congestion degree, and directly proportional to the passenger flow satisfaction rate, and β1 = 0.2, β2 = 0.5, β3 = 0.3.

[0103] Step 3: Construct a riding state model, using the train operation plan in the historical data as input, and the corresponding train operation performance index and passenger experience index as labels to train the riding state model.

[0104] In this embodiment, the riding state model is trained based on a deep learning network, using the historical train operation plan as input. The structure of the deep learning network is as follows:

[0105] Input layer: 3 neurons, used to input the train stop sequence, train operation time sequence, and arrival time sequence in the train operation plan;

[0106] First fully connected layer: 128 neurons, using the ReLU activation function;

[0107] Second fully connected layer: 64 neurons, using the ReLU activation function;

[0108] Third fully connected layer: 32 neurons, using the ReLU activation function;

[0109] Output layer: 2 neurons, outputting the train operation performance index and the passenger experience index.

[0110] Step 4: Use the train operation plan in the historical data as an individual in the initial population. Calculate the fitness value of the individual based on the train operation performance index and the passenger experience index. With the maximization of the fitness value as the optimization goal, perform iterative optimization on the individual based on the genetic algorithm and the riding state model until the predetermined number of iterations is reached, and select the individual with the highest fitness value as the best train operation plan;

[0111] In this embodiment, the principle for generating individuals in the initial population is as follows:

[0112] The data elements included in each individual in the population are:

[0113] Train stop sequence, S i =[s i,1 , s i,2 , …, s i,j , …, s i,J ;

[0114] Among them, S i represents the stop sequence of the i-th train, i represents the index of the train, j represents the index of the station, and j = 1, 2, …… J, where J represents the total number of stations, and s i,j represents the j-th stop station of the i-th train;

[0115] The train stop sequence reflects the operation path of the train and is the most basic component of the scheduling. Adjust the stations where the train stops according to the changes and demands of the passenger flow to ensure passing through the main stations during the peak period;

[0116] Train running time sequence, T i =[t i,2 , t i,3 , …, t i,j , …, t i,J ;

[0117] Among them, T i represents the running time sequence of the i-th train, and t i,j represents the running time of the i-th train from the j-th station to the j - 1-th station;

[0118] The train running time sequence reflects the running time of the train between stations and is used to judge whether there is a congestion situation;

[0119] Arrival time sequence, H i =[h i,1 , h i,2 , … h i,j , …, h i,J ;

[0120] Among them, H i represents the arrival time sequence of the i-th train, and h i,jDenote the arrival time of the $i$-th train at the $j$-th station;

[0121] The arrival time series reflects the actual arrival time of the train at the station, which is used to judge whether the train is late or early, so as to judge the congestion situation.

[0122] For each train, the individual formed is:

[0123] $I$ i $=$ [$S$ i , $T$ i , $H$ i

[0124] Where $I$ i represents the individual corresponding to the $i$-th train in the population;

[0125] The formula for calculating the fitness value of an individual based on train operation performance indicators and passenger experience indicators is:

[0126] $F = w_1\cdot e$ A $+ w_2\cdot e$ B

[0127] Where $F$ represents the fitness value, $A$ and $B$ respectively represent the train operation performance indicators and passenger experience indicators, $w_1$ and $w_2$ respectively represent the weight coefficients of the train operation performance indicators and passenger experience indicators, $w_1 > w_2$, and $w_1 + w_2 = 1$.

[0128] The fitness value comprehensively considers the train operation situation and the passenger's riding and waiting experience. The higher the fitness value, the smoother the train operation and the better the passenger's riding experience, and the better the corresponding train operation plan. The fitness value is proportional to the train operation performance indicators and passenger experience indicators. In actual scheduling, the operation quality of the train line is given priority, so the weight coefficient of the train operation performance indicator is larger, $w_1 = 0.7$, $w_2 = 0.3$.

[0129] Based on the tournament selection method, randomly select two individuals from the initial population, compare the fitness values of the selected individuals, and select the individual with higher fitness between the two as the parent. Repeat the above process until the required number of parent individuals is selected, which is regarded as completing one round of selection. Then, randomly perform crossover and mutation on the selected parent individuals to generate new offspring. Combine the offspring and the parents to form a new population. Input the individuals in the new population into the riding state model, and calculate the train operation performance indicators and passenger experience indicators of each individual. Calculate the fitness value of each individual based on the train operation performance indicators and passenger experience indicators, and record the individual with the highest fitness value as the current optimal solution. The above operations are regarded as one iteration. Set the maximum number of iterations to 100. When the number of iterations reaches 100, or the improvement amplitude of the optimal fitness in 10 consecutive iterations is less than 5%, terminate the iteration, and select the individual of the current optimal solution as the best train operation plan.

[0130] Step 5: Obtain the current train operation plan, and based on the best train operation plan, determine the train scheduling plan for adjusting the current train operation plan.

[0131] In this embodiment, the scheduling strategies include:

[0132] Train catching up: Adjust the train running time and stopping time to shorten the overall delay;

[0133] Train skipping: Allow some trains to run past stations with less passenger flow;

[0134] Train adding: Add trains on key lines according to the passenger flow prediction data;

[0135] Train empty running: Reduce the train operation frequency on some lines according to the passenger flow prediction data.

[0136] Please refer to Figure 2 , the present invention also provides an urban rail train operation diagram adjustment device based on the genetic algorithm. The device is used to implement the above-mentioned urban rail train operation diagram adjustment method based on the genetic algorithm, and specifically includes:

[0137] A data acquisition module, which is used to obtain historical data. The historical data includes the train operation plan and the corresponding state parameters obtained day by day. The train operation plan includes the train stop sequence, running time sequence, and arrival time sequence. The state parameters include station information data, passenger flow data, and train operation data;

[0138] An index generation module, which is used to generate the train punctuality rate, train stop time, passenger flow congestion degree, average train delay time, and passenger flow satisfaction rate based on the state parameters. Generate train operation performance indicators based on the train punctuality rate and average train delay time, and generate passenger experience indicators based on the average train stop time, passenger flow congestion degree, and passenger flow satisfaction rate;

[0139] A model training module, configured to build a riding state model, use the train operation plan in historical data as input, and the corresponding train operation performance indicators and passenger experience indicators as labels to train the riding state model;

[0140] A plan optimization module, configured to use the train operation plan in historical data as an individual of the initial population, calculate the fitness value of the individual based on the train operation performance indicators and passenger experience indicators, take maximizing the fitness value as the optimization goal, and perform iterative optimization on the individual based on the genetic algorithm and the riding state model until the predetermined number of iterations is reached, and select the individual with the highest fitness value as the best train operation plan;

[0141] A plan scheduling module, configured to obtain the current train operation plan, and based on the best train operation plan, determine a train scheduling plan for adjusting the current train operation plan, so that the current train operation plan is close to the best train operation plan.

[0142] The above formulas are all dimensionless and take their numerical values for calculation. The formulas are obtained by collecting a large amount of data for software simulation to obtain a formula closest to the actual situation. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0143] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. Those skilled in the art can realize that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are executed by hardware or software methods depends on the specific application and design constraints of the technical solution.

[0144] The units described as separate components may or may not be physically separated. The components shown as units may or may not be physical units. They may be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0145] The above is only the specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed in the present application, and all of them should be covered within the protection scope of the present application.

Claims

1. An adjustment method for the operation diagram of urban rail trains based on genetic algorithm, characterized in that, The specific steps include: Step 1: Obtain historical data, which includes the train operation plan and corresponding status parameters obtained daily. The train operation plan includes the train stop sequence, train operation time sequence, and arrival time sequence, and the status parameters include station information data, passenger flow data, and train operation data. Step 2: Based on the status parameters, generate the train punctuality rate, train stop time, passenger flow congestion degree, average train delay time, and passenger flow satisfaction rate. Generate a train operation performance index based on the train punctuality rate and the average train delay time, and generate a passenger experience index based on the average train stop time, passenger flow congestion degree, and passenger flow satisfaction rate. Step 3: Construct a riding state model. Use the train operation plan in the historical data as the input, and the corresponding train operation performance index and passenger experience index as labels to train the riding state model. Step 4: Use the train operation plan in the historical data as an individual in the initial population. Calculate the fitness value of the individual based on the train operation performance index and the passenger experience index. With the maximization of the fitness value as the optimization goal, perform iterative optimization on the individual based on the genetic algorithm and the riding state model until the predetermined number of iterations is reached, and select the individual with the highest fitness value as the best train operation plan. Step 5: Obtain the current train operation plan, and based on the best train operation plan, determine a train scheduling plan for adjusting the current train operation plan to make the current train operation plan close to the best train operation plan.

2. The method for adjusting the operation diagram of urban rail trains based on the genetic algorithm according to claim 1, wherein: In step 1, the station information data includes the maximum allowable passenger flow of each station; the passenger flow data includes the passenger flow of each station and the number of passengers boarding different trains at that station. The train operation data includes the total number of trains, the number of trains arriving on time, the actual stop time of the trains at each station, the remaining passenger capacity, and the train delay time.

3. The method for adjusting the operation diagram of urban rail trains based on the genetic algorithm according to claim 1, characterized in that: The principle for generating the train operation performance index in step 2 is: The formula for generating the train punctuality rate is: Among them, R represents the train punctuality rate, M time represents the number of trains arriving on time, and M represents the total number of trains; The formula for generating the average train delay time is: Among them, Y represents the average train delay time, i represents the index of the train, and i ∈ [1, M], j represents the index of the station, J represents the total number of stations, and D i,j represents the train delay time when the i-th train arrives at the j-th station; The formula for generating the train operation performance index is: Where A represents the train operation performance index, α1 and α2 respectively represent the weight coefficients of the train punctuality rate and the average train delay time, and α1 + α2 = 1.

4. A method for adjusting the operation diagram of urban rail trains based on the genetic algorithm according to claim 1, characterized in that: The principle for generating the passenger experience index in step 2 is: The formula for generating the average train stop time is: Among them, \(W\) represents the average train stopping time, \(j\) represents the index of the station, \(J\) represents the total number of stations, and \(W_{ij}\) i,j represents the actual stopping time of the \(i\)-th train at the \(j\)-th station; The formula for generating the passenger flow congestion degree is: Among them, C represents the passenger flow congestion degree, C j,cur represents the passenger flow volume of the j-th station, C j,max represents the maximum allowable passenger flow volume of the j-th station; The formula for generating the passenger flow satisfaction rate is: Among them, Z represents the passenger flow satisfaction rate, and E i,j represents the remaining passenger capacity of the i-th train at the j-th station, and U i,j represents the number of passengers taking the i-th train at the j-th station; The principle for generating the passenger experience index is: B = β1·(1 - W) + β2·(1 - C) + β3·Z Where B represents the passenger experience index, β1, β2, and β3 respectively represent the weight coefficients of the average passenger waiting time, passenger flow congestion degree, and passenger flow satisfaction rate, β1 + β2 + β3 = 1, and β2 > β3 > β1.

5. The method for adjusting the operation diagram of urban rail trains based on the genetic algorithm according to claim 1, characterized in that: The principle for generating an individual in the initial population in step 4 is: The data elements included in each individual in the population are: Train stopping sequence, S i = [s i,1 , s i,2 , …, s i,j , …, s i,J ; Among them, S i represents the stop sequence of the i-th train, where i represents the index of the train, j represents the index of the station, and j = 1, 2,..., J, where J represents the total number of stations, and s i,j represents the j-th stop station of the i-th train; Travel time series, T i = [t i,2 , t i,3 , …, t i,j , …, t i,J ; Among them, T i represents the train operation time series of the i-th train, and t i,j represents the running time of the i-th train from the j-th station to the (j - 1)-th station; Arrival time series, H i = [h i,1 , h i,2 , … h i,j …, h i,J ; Among them, H i represents the arrival time sequence of the i-th train, and h i,j represents the arrival time of the i-th train at the j-th station; For each train shift, the individual formed is: I i = [S i , T i , H i ​ Among them, I i represents the individual corresponding to the i-th train in the population.

6. The method for adjusting the operation diagram of urban rail trains based on genetic algorithm according to claim 1 is characterized in that: The formula for generating the fitness value in step 4 is: F = w1·e A + w2·e B Among them, F represents the fitness value, A and B respectively represent the train operation performance index and the passenger experience index, w1 and w2 respectively represent the weight coefficients of the train operation performance index and the passenger experience index, w1 > w2, and w1 + w2 = 1.

7. A method for adjusting the operation diagram of urban rail trains based on the genetic algorithm according to claim 1, characterized in that: The principle for iteratively optimizing individuals based on the genetic algorithm in step 4 is as follows: Based on the tournament selection method, randomly select two individuals from the initial population, compare the fitness values of the selected individuals, and select the individual with a higher fitness between the two as the parent. Repeat the above process until the required number of parent individuals is selected, which is regarded as completing one round of selection. Then, randomly perform crossover and mutation on the selected parent individuals to generate new offspring. Combine the offspring and the parents to form a new population. Input the individuals in the new population into the riding state model, and calculate the train operation performance index and the passenger experience index of each individual. Calculate the fitness value of the individual based on the train operation performance index and the passenger experience index, and record the individual with the highest fitness value as the current optimal solution. The above operations are regarded as one iteration. Set the maximum number of iterations to 100. When the number of iterations reaches 100, or the improvement amplitude of the optimal fitness in 10 consecutive iterations is less than 5%, terminate the iteration, and select the individual with the current optimal solution as the best train operation plan.

8. An urban rail transit train operation diagram adjustment device based on a genetic algorithm, characterized in that: The device is used to implement the urban rail train operation diagram adjustment method based on the genetic algorithm according to any one of claims 1 - 7, and specifically includes: A data acquisition module, which is used to obtain historical data. The historical data includes the train operation plan and the corresponding state parameters obtained daily. The train operation plan includes the train stop sequence, the train operation time sequence, and the arrival time sequence. The state parameters include station information data, passenger flow data, and train operation data; An index generation module, which is used to generate the train punctuality rate, the train stop time, the passenger flow congestion degree, the average train delay time, and the passenger flow satisfaction rate based on the state parameters. Generate the train operation performance index based on the train punctuality rate and the average train delay time, and generate the passenger experience index based on the average train stop time, the passenger flow congestion degree, and the passenger flow satisfaction rate; A model training module, which is used to construct a riding state model, use the train operation plan in the historical data as the input, and the corresponding train operation performance index and passenger experience index as the labels to train the riding state model; A plan optimization module, which is used to use the train operation plan in the historical data as the individuals of the initial population, calculate the fitness value of the individuals based on the train operation performance index and the passenger experience index, take maximizing the fitness value as the optimization goal, and iteratively optimize the individuals based on the genetic algorithm and the riding state model until the predetermined number of iterations is reached, and select the individual with the highest fitness value as the best train operation plan; A plan scheduling module, which is used to obtain the current train operation plan, and determine the train scheduling plan for adjusting the current train operation plan based on the best train operation plan, so that the current train operation plan is close to the best train operation plan.

Images