Urban rail and urban rail train cooperative operation organization strategy selection method

Abstract

The invention relates to a city region rail and city rail train cooperative operation organization strategy selection method, which comprises the following steps of: constructing a city region rail and city rail train operation organization alternative strategy set from the aspects of route crossing, station stopping and marshalling; on the basis of the alternative strategy set, a feasible strategy suitable for the urban rail and the urban rail train is determined; and based on the feasible strategies, establishing a city region rail and city rail train operation scheme collaborative optimization model, and solving to obtain the cross-road departure frequency, the station stopping scheme and the marshalling number under each feasible strategy. According to the method, a clear solution is made for the strategy of which operation organization level is adopted under the specific condition of the multi-system rail transit, so that the defect that a set of strategy selection method of the operation organization level is lacked for the urban rail and the urban rail to cope with the variable passenger flow is overcome.

Description

Technical Field

[0001] The present invention relates to the technical field of urban rail transit, and particularly to a method for selecting a coordinated operation organization strategy for suburban rail and urban rail trains. Background Art

[0002] With the acceleration of China's urbanization process, regional coordinated development has become a new trend. Within the scope of urban agglomerations and metropolitan areas, a regional rail transit composite network system composed of multiple types of rail transit has become the main force driving the development of urban clusters within the region, with suburban rail and urban rail transit being typical representatives.

[0003] The operation organizations of suburban rail and urban rail present diverse characteristics, and the requirements for operation organizations vary in different scenarios. Existing research on rail transit operation organization strategies has been relatively in-depth, and the advantages, disadvantages, and clear applicable scopes of various operation organization strategies have been given. However, there is little research on suburban rail at present, and few studies have deeply explored the selection of multi-type rail transit operation organization strategies. There is a lack of a clear method for selecting operation organization strategies to cope with the changing passenger flows of suburban rail and urban rail, and it is difficult to clearly answer "which specific coordinated operation organization strategy should be adopted" in specific situations. Summary of the Invention

[0004] The purpose of the present invention is to provide a method for selecting a coordinated operation organization strategy for suburban rail and urban rail trains, which optimizes the rail transit operation network from the perspective of operation organization, realizes the deep integration of suburban and urban rail transit, and thus improves the travel quality of passengers.

[0005] The technical solution for achieving the purpose of the present invention is as follows:

[0006] A method for selecting a coordinated operation organization strategy for suburban rail and urban rail trains, the method comprising:

[0007] S1, constructing a set of alternative strategies for the operation organizations of suburban rail and urban rail trains from the aspects of train operation routes, stopping stations, and carbody formations;

[0008] S2, determining the feasible strategies suitable for suburban rail and urban rail trains based on the set of alternative strategies in S1;

[0009] S3, based on the feasible strategies in S2, taking the minimum passenger travel time cost and the minimum total enterprise operation cost as the objectives, establishing a coordinated optimization model for the train operation plans of suburban rail and urban rail trains, and solving to obtain the train operation route departure frequencies, stopping station plans, and carbody formations under each feasible strategy.

[0010] Furthermore: In S1, the operation strategies of the routes include three strategies: cross-line operation, large-small loop operation, and single loop operation; the stop plans include two strategies: all-stops and non-all-stops; the formation modes include two strategies: fixed formation and mixed operation of multiple formations.

[0011] Furthermore: The steps for determining the feasible strategies for the coordinated operation of urban rail transit and urban rail in S2 include:

[0012] Step1: Determine whether the cross-line conditions are met. If so, go to Step2; otherwise, go to Step4. The cross-line conditions are that the cross-line passenger flow Q 跨 and the cross-line passenger flow ratio λ are both greater than the corresponding set cross-line passenger flow threshold and cross-line passenger flow ratio threshold.

[0013] Step2: If the line section imbalance coefficient β is greater than the set imbalance threshold, it is determined that the passenger flow distribution is uneven and go to Step3; otherwise, go to Step6.

[0014] Step3: If the number of stations where the corresponding passenger flow boarding and alighting coefficient γ is greater than or equal to the set passenger flow boarding and alighting peak value or less than or equal to the set passenger flow boarding and alighting trough value exceeds the set first ratio, it is recommended to operate the non-all-stops strategy; if the number of stations where the corresponding section load factor η is greater than or equal to the set section load factor peak value or less than or equal to the set section load factor trough value exceeds the set second ratio, it is recommended to operate the mixed operation of multiple formations strategy; otherwise, go to Step6.

[0015] Step4: If the line section imbalance coefficient β is greater than the set imbalance threshold, it is determined that the passenger flow distribution is uneven and go to Step5; otherwise, go to Step6.

[0016] Step5: If the number of stations where the corresponding passenger flow boarding and alighting coefficient γ is greater than or equal to the set passenger flow boarding and alighting peak value or less than or equal to the set passenger flow boarding and alighting trough value exceeds the set first ratio, and continue to judge: if such stations are concentrated in the same section formed by two turn-back stations, it is recommended to operate the large-small loop operation strategy; if such stations are scattered in different sections formed by two turn-back stations, it is recommended to operate the non-all-stops strategy; except for the above two situations, it is recommended to operate the large-small loop operation and non-all-stops strategies.

[0017] If the number of stations where the corresponding section load factor η is greater than or equal to the set section load factor peak value or less than or equal to the set section load factor trough value exceeds the set second ratio, it is recommended to operate the mixed operation of multiple formations strategy.

[0018] Otherwise, go to Step6.

[0019] Step6: Maintain the existing operation organization strategy.

[0020] Furthermore: The specific calculation formula for the cross-line passenger flow ratio λ in Step1 is as follows:

[0021]

[0022] Among them, Q 跨 is the cross-line passenger flow within the statistical time, and Q 本 is the in-line passenger flow within the statistical time;

[0023] The specific calculation formula for the line section imbalance coefficient β is as follows:

[0024]

[0025] Among them, Q max is the maximum section passenger flow, M is the number of one-way line sections, and Q m is the passenger flow of the m-th section;

[0026] The passenger boarding and alighting coefficient γ of station i i The specific calculation formula is as follows:

[0027]

[0028] Among them, is the maximum boarding and alighting volume per hour of station i, Q u is the boarding and alighting volume per hour of station u, and k is the number of stations;

[0029] The specific calculation formula for the section load factor η is as follows:

[0030]

[0031] Among them, is the passenger flow of section m within the statistical time; N is the number of trains passing through section m in a certain direction within the statistical time; C is the train seating capacity.

[0032] Furthermore: The coordinated optimization model for the operation plan of urban rail transit and urban rail trains in S3 includes:

[0033] The objective function is:

[0034] min Z1 = t w ·a 1 ·p w

[0035] min Z2 + Z3

[0036]

[0037]

[0038] The constraint conditions are:

[0039]

[0040] Where: Z1 is the passenger travel cost; t w is the travel time of passenger group w for OD pair; a 1 is the passenger time cost coefficient; p w is the number of passengers in passenger group w for OD pair; Z2 is the train running kilometer cost; is the length of the cross - line route and the short - line route; is the length of the main - line route; is the number of train formations on the main - line route; is the number of train formations on the cross - line route and the short - line route; a 2 is the unit - length operation cost coefficient; is the train operation frequency on the main - line route; is the train operation frequency on the cross - line route and the short - line route; is the minimum and maximum train departure frequencies allowed for the cross - line route and the short - line route; is the minimum and maximum train departure frequencies allowed for the main - line route; is the maximum sectional passenger flow of the cross - line route and the short - line route; is the maximum sectional passenger flow of the main - line route; P is the passenger capacity of the car; η is the train load factor; SE is the set of intervals; is 1 if the main - line route passes through interval se, otherwise 0; is 1 if the cross - line route and the short - line route pass through interval se, otherwise 0; t min is the train - following interval time; is a 0 - 1 variable indicating whether the two terminal reversing stations of the cross - line route stop; C is the set of routes other than the main - line large - route; Z3 is the train stop cost; a 3 is the single - stop cost coefficient.

[0041] Furthermore: The NSGA - II algorithm is used to solve the collaborative optimization model of the urban - rail transit and urban - rail train operation plan, and the route departure frequency, stop plan and formation number under each feasible strategy are obtained.

[0042] Furthermore: The decision variables in the NSGA - II algorithm include the route departure frequency, stop plan and formation number. Each chromosome represents a specific operation plan. The route departure frequency and formation number adopt real - number coding, and the stop plan adopts 0 - 1 coding.

[0043] Furthermore: The Q - Learning algorithm is used to adaptively adjust the crossover and mutation probabilities in the NSGA - II algorithm,

[0044] where the crossover probability P c is adjusted through the reward function r c , and the crossover probability P is adjusted through the reward function rm Adjust the mutation probability P m :

[0045]

[0046] In the formula, is the fitness value of the i-th chromosome in the t-th generation population ; is the fitness value of the i-th chromosome in the (t + 1)-th generation population ; Popsize is the population size; is the best fitness of the t-th generation population;

[0047] is the best fitness of the (t + 1)-th generation population.

[0048] Furthermore: The number of rows of the Q-value table corresponds to the number of environmental states of NSGA-II, and the number of columns corresponds to the number of actions in the action set A; In the initial state, the Q-value table is set as a all-zero matrix.

[0049] Furthermore: The environmental state s of the t-th generation of NSGA-II t is;

[0050] s t = w 1 ·fit * + w 2 ·div * + w 3 ·Best * (w 1 + w 2 + w 3 = 1)

[0051]

[0052]

[0053] Among them, fit * is the average fitness of the population; div * is the diversity of the population; Best * is the best fitness of the population; w 1 , w 2 , w 3 are the weight coefficients of fit * , div * , Best * respectively.

[0054] In addition, the present invention provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the above-mentioned method for selecting the collaborative operation organization strategy of the urban rail transit and urban rail trains are realized.

[0055] The present invention also provides an electronic device, including:

[0056] a memory for storing a computer program;

[0057] a processor for realizing the steps of the above-mentioned method for selecting the collaborative operation organization strategy of the urban rail transit and urban rail trains when executing the computer program.

[0058] Compared with the prior art, the present invention has the following remarkable advantages:

[0059] (1) From three aspects of the selection idea, basis indicators, and specific process, the present invention proposes a method for selecting the collaborative operation organization strategy of the urban rail transit and urban rail, screens out feasible strategies that meet the actual situation of the line, and makes a clear answer to "what kind of strategy at the operation organization level should be adopted" in the specific situation of multi-modal rail transit, thus making up for the deficiency that there is a lack of a set of strategy selection methods at the operation organization level for the urban rail transit and urban rail to cope with the changing passenger flow.

[0060] (2) The invention can be applied to the train operation organization under complex operation modes. During the collaborative optimization process of the train operation plans of the urban rail transit and urban rail trains, all the turn-back frequencies, stop plans, and formation numbers involved in the urban rail transit and urban rail are taken into account; at the same time, the algorithm proposed by the invention has a faster solution efficiency. On the basis of the NSGA-II algorithm framework, the crossover and mutation probabilities of the NSGA-II algorithm are adaptively adjusted by combining reinforcement learning. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] Figure 1 is a schematic diagram of the steps of a method for selecting the collaborative operation organization strategy of an urban rail transit and urban rail trains according to the present invention.

[0062] Figure 2 is a schematic diagram of the idea for selecting feasible strategies for collaborative operation organization in the present invention.

[0063] Figure 3 is a schematic diagram of the specific process for selecting feasible strategies for collaborative operation organization in the present invention.

[0064] Figure 4 is a flow chart of algorithm solution.

[0065] Figure 5 is a schematic diagram of the fusion of genetic algorithm and reinforcement learning.

[0066] Figure 6It is the graph of the boarding and alighting coefficients of each station on Line S7 - S1 - 3 from 7:00 to 8:00 in the morning.

[0067] Figure 7 It is the graph of the full load rates of each section on Line S7 - S1 - 3 from 7:00 to 8:00 in the morning. Specific implementation manners

[0068] The following will describe the implementation manners of the present invention in detail with reference to the accompanying drawings.

[0069] The present invention provides a method for selecting a collaborative operation organization strategy for suburban rail and urban rail trains. Based on the analysis of common operation organization strategies, an alternative set of collaborative operation organization strategies for suburban rail and urban rail is constructed; further, a method for selecting feasible strategies for the collaborative operation organization of suburban rail and urban rail is proposed from three aspects: selection ideas, basis indicators, and specific processes, and feasible strategies that conform to the actual situation of the line are screened out; on the basis of obtaining the feasible strategies, a collaborative optimization model for the train operation plan of suburban rail and urban rail is established to determine the specific departure frequencies, stopping stations, and formation numbers under each feasible strategy. The invention aims to solve the problem of selecting the collaborative operation organization strategy for suburban rail and urban rail, and has a wide application prospect in the regional rail transit intermodal transport mode.

[0070] To achieve the above object, as Figure 1 shown, the present invention adopts the following technical solutions:

[0071] S1. Based on the existing common operation organization modes, construct an alternative strategy set for the operation organization of suburban rail and urban rail trains from the aspects of routes, stops, and formations.

[0072] S2. Based on the alternative strategy set of S1, propose a method for selecting feasible strategies for the collaborative operation organization from three aspects: selection ideas, basis indicators, and specific processes, and determine the feasible strategies suitable for suburban rail and urban rail.

[0073] S3. Based on the feasible strategies determined in S2, establish a collaborative optimization model for the train operation plan of suburban rail and urban rail, and obtain in detail the specific departure frequencies, stopping stations, and formation numbers under each feasible strategy. With the goal of minimizing the passenger travel time cost and the enterprise operation cost, considering constraints such as the train full load rate and the line passing capacity, establish a collaborative optimization model for the train operation plan of suburban rail and urban rail.

[0074] S4. Design a solving algorithm. On the basis of the NSGA - Ⅱ algorithm framework, adaptively adjust the crossover and mutation probabilities of the NSGA - Ⅱ algorithm by combining reinforcement learning, and solve the model established in S3 to determine the specific departure frequencies, stopping plans, and formation numbers under each feasible strategy.

[0075] Furthermore, the specific content of S1 includes:

[0076] S101. Collect the existing common operation organization strategies of urban rail transit and urban rail. In terms of train operation patterns, there are three strategies including cross-line operation pattern, short and long operation patterns, and single operation pattern. In terms of stop plans, there are two strategies including all-stops and non-all-stops. In terms of formation modes, there are two strategies including fixed formation and mixed operation of multiple formations. Note that the cross-line operation pattern and the short and long operation patterns do not appear on the same line. There are 3×2×2 = 12 combination methods, and an alternative set of operation organization strategies is constructed. The set includes Strategies 1 - 12, and the strategy set is shown in Table 1:

[0077] Table 1 Alternative set of operation organization strategies

[0078]

[0079] Furthermore, the S2 specifically includes:

[0080] S201. Put forward the selection ideas for feasible strategies of coordinated operation organization of urban rail transit and urban rail, as Figure 2 shown.

[0081] Adopt a hierarchical analysis method from large to small. First, consider the connection between different types of tracks at the network level, and judge whether cross-line operation is needed according to the cross-line passenger flow and its proportion.

[0082] When cross-line operation is adopted for a line, further calculate the imbalance coefficient to judge the balance of passenger flow distribution. When the passenger flow distribution is relatively balanced, the existing operation organization strategy can be maintained and optimized by changing the train departure frequency; when the passenger flow distribution is unbalanced, combined with the passenger flow coefficients of each station, a stop strategy is formulated. In addition, according to the load factors of different sections, the formation strategy of trains is determined.

[0083] When single-line operation is adopted for a line, calculate the imbalance coefficient to judge the balance of passenger flow distribution in the same way. When the passenger flow distribution is relatively balanced, the existing operation organization strategy can be maintained and optimized by changing the train departure frequency; when the passenger flow distribution is unbalanced, calculate whether the distribution of boarding and alighting volumes at each station is concentrated. If the station distribution is concentrated, consider whether it is necessary to add short and long operation patterns to the single-line operation line; if the station distribution is scattered, combined with the passenger flow coefficients of each station, a stop strategy is formulated; if the station distribution has no obvious pattern, the two can be combined. In addition, when the passenger flow distribution is unbalanced, according to the load factors of different sections, the formation strategy of trains is determined.

[0084] S202. Put forward the basis indicators for the selection of feasible strategies for the coordinated operation organization of urban rail transit and urban rail.

[0085] 1) Calculate the cross-line passenger flow and its proportion. The cross-line passenger flow Q 跨The passenger flow with the origin and destination located on two different lines can be obtained from the passenger flow OD table. The cross-line passenger flow ratio λ is the ratio of the cross-line passenger flow volume on the line to the total passenger flow volume, and the total passenger flow volume consists of the cross-line passenger flow volume and the in-line passenger flow volume. The specific calculation formula is as follows:

[0086]

[0087] Among them, Q 跨 is the cross-line passenger flow volume within the statistical time, and Q 本 is the in-line passenger flow volume within the statistical time.

[0088] In the present invention, for the lines carrying out cross-line operation, the required cross-line passenger flow volume Q 跨 ≥ 3240 persons / h, and the proportion of cross-line passenger flow λ reaches more than 50%. If either of the two conditions is not met, cross-line operation cannot be carried out.

[0089] 2) Calculate the line section imbalance coefficient β. The line section imbalance coefficient β is used to judge whether the passenger flow distribution is balanced, and the specific calculation formula is as follows:

[0090]

[0091] Among them, Q max is the maximum section passenger flow volume, M is the number of one-way line sections, and Q m is the passenger flow volume of the m-th section.

[0092] In the present invention, when the coefficient β ≤ 1.5, it indicates that there is no imbalance problem in the passenger flow distribution, and the existing strategy can be maintained. On the contrary, adjustments need to be made in terms of train operation patterns, stopping stations, and car body formations.

[0093] 3) Calculate the passenger boarding and alighting coefficient γ of each station. The passenger boarding and alighting coefficient of the station is used to describe the passenger boarding and alighting levels of different stations, and the specific calculation formula is as follows:

[0094]

[0095] Among them, is the maximum hourly boarding and alighting volume of the i-th station, Q u is the maximum hourly boarding and alighting volume of the u-th station, and k is the number of stations.

[0096] In the present invention, when the boarding and alighting coefficient γ ≤ 0.5, it indicates that the number of people boarding and alighting at this station is too small; when the boarding and alighting coefficient γ ≥ 1.5, it indicates that the number of people boarding and alighting at this station is too large. Observe the distribution of stations with the boarding and alighting coefficient γ ≥ 1.5 or γ ≤ 0.5 to judge whether to adopt the large and small train operation patterns and the non-stop-at-all-stations strategy.

[0097] 4) Calculate the full-load rate η of each section. The full-load rate of a section represents the ratio of the actual passenger volume carried by the section to the actual operating capacity of the section, reflecting the utilization efficiency of the transport capacity within a section.

[0098]

[0099] Among them, is the passenger flow volume of section i within the statistical time; N is the number of trains passing through a certain direction of this section within the statistical time; C is the passenger capacity of the train.

[0100] In the present invention, when the average full-load rate η of the sections in multiple sections of the line is η≥120% or η≤20%, it indicates that the full-load rates of the sections in the section vary greatly, and a multi-formation strategy can be selected to achieve a faster evacuation effect.

[0101] S203. Propose the specific steps of the feasible strategy for the coordinated operation organization of urban rail transit and urban rail, such as Figure 3 shown.

[0102] Step1: Judge whether the cross-line condition is satisfied, and calculate the cross-line passenger flow volume Q 跨 and the cross-line passenger flow ratio λ. If both are satisfied, go to Step2; if any condition is not satisfied, go to Step4.

[0103] Step2: Under the cross-line operation condition, calculate the line section imbalance coefficient β, judge whether the passenger flow distribution is balanced, and judge whether the imbalance coefficient satisfies β>1.5. If satisfied, go to Step3; if not satisfied, go to Step6.

[0104] Step3: Calculate the passenger flow boarding and alighting coefficient γ and the full-load rate η of each station. If the number of stations where γ≥1.5 or γ≤0.5 exceeds the set threshold, a non-stop-at-all-stations strategy is recommended; if the number of stations where η≥120% or η≤20% exceeds the set threshold, a multi-formation mixed operation strategy is recommended; otherwise, go to Step6.

[0105] Step4: Under the single-line operation condition, judge whether the passenger flow distribution is balanced, calculate the line section imbalance coefficient β, and judge whether the imbalance coefficient satisfies β>1.5. If satisfied, go to Step5; if not satisfied, go to Step6.

[0106] Step5: Calculate the passenger flow boarding and alighting coefficient γ and the full-load rate η of each station;

[0107] If the number of stations satisfying γ≥1.5 or γ≤0.5 exceeds the set threshold, and: if these stations are concentrated in the same section formed by two reversing stations, it is recommended to implement the strategy of running trains with different stopping patterns; if these stations are scattered in different sections formed by two reversing stations, it is recommended to implement the strategy of non-stop-at-all-stations; except for the above two cases, if the distribution of stations is chaotic and there is no obvious regularity, it is recommended to implement both the strategy of running trains with different stopping patterns and the strategy of non-stop-at-all-stations;

[0108] If the number of stations satisfying η≥120% or η≤20% exceeds the set threshold, it is recommended to implement the strategy of mixed operation of multiple formations;

[0109] Otherwise, go to Step6.

[0110] Step6: Maintain the existing operation organization strategy and design the train departure frequency.

[0111] Furthermore, the specific content of S3 includes:

[0112] S301, calculate the time cost of passenger travel.

[0113] The time cost of passenger travel is equal to the product of the passenger travel time and the time cost coefficient. The passenger travel time includes the waiting time after entering the station, the time on the train, and the time required for transfer.

[0114]

[0115] Among them, p w is the number of passengers in the passenger group of OD pair w; Z1 is the passenger travel cost; a 1 is the passenger time cost coefficient; t w is the travel time of the passenger group of OD pair w; is the average time on the train of the passenger group of OD pair w; is the average transfer time of the passenger group of OD pair w; is the average waiting time of the passenger group of OD pair w.

[0116] The time on the train refers to the sum of the running time required for each section of the travel path of a certain OD passenger group and the stopping time at all intermediate stops except the starting point and the ending point. The running time of the train in each section and the stopping time at each intermediate station can be obtained through operation data.

[0117]

[0118] Among them, t se is the interval running time; γ is the train stop time; is a 0-1 variable. If the passenger group of OD pair w takes the train running on the route c n and l m and travels in the se interval, it is 1; otherwise, it is 0.

[0119] The waiting time refers to the total waiting time of passengers during the entire journey on various different routes.

[0120]

[0121] Among them, is the train operation frequency on the main line route; is the train operation frequency on the cross-line route and the small route.

[0122] The transfer time refers to the total time required for passengers to transfer from one route to another during the journey. The transfer time is equal to the number of transfers multiplied by the walking time for a single transfer, where the number of transfers is equal to the number of route segments passed by the passengers during the journey minus 1.

[0123]

[0124] Among them, t walk is the walking time for transfer. a 1 is the time cost coefficient for passengers.

[0125] S301, calculate the total operating cost of the enterprise.

[0126] The total operating cost of the enterprise includes the train running kilometer cost and the stop cost. The train running kilometer cost is expressed as the product of the total running kilometers of all operating trains and the unit kilometer running cost of the train.

[0127]

[0128] Among them, Z2 is the train running kilometer cost; is the length of the cross-line route and the small route; is the length of the main line route; is the number of carriages of the train on the main line route; is the number of carriages of the train on the cross-line route and the small route; a 2 is the unit length operation cost coefficient.

[0129] The train stop cost refers to the cost incurred by the train due to stopping at stations during operation, which is closely related to the total number of stops of the train. The train stop cost is equal to the number of stops multiplied by the single stop cost coefficient.

[0130]

[0131] Among them, Z3 is the train stop cost; a 3 is the single stop cost coefficient.

[0132] S303, set the model constraint conditions.

[0133] 1) Constraints on the upper and lower limits of the departure frequencies of trains on each route

[0134] When formulating the train operation plan, it is necessary to comprehensively consider the line passing capacity and the turning-back capacity of the turning-back stations, and upper and lower limit constraints should be set for the departure frequencies of trains on each route.

[0135]

[0136] Among them, The minimum and maximum train departure frequencies allowed for cross-line routes and short-turn routes; The minimum and maximum train departure frequencies allowed for local routes;

[0137] 2) Constraints on the train load factor

[0138] When formulating the train operation plan, ensure that sufficient operation capacity is provided for passengers to avoid passenger congestion. The maximum sectional passenger flow should not exceed the transport capacity provided by the train.

[0139]

[0140] Among them, The maximum sectional passenger flow of cross-line routes and short-turn routes; The maximum sectional passenger flow of local routes; The seating capacity of the carriages; η The train load factor.

[0141] 3) Constraints on the line passing capacity

[0142] In the line network, cross-line routes, short-turn routes and local long-turn trains operate in parallel. In the sections where multiple routes operate together, it is necessary to ensure that the sum of the departure frequencies of all routes passing through this section does not exceed the line passing capacity.

[0143]

[0144] Among them, If the local route passes through the section se, it is 1, otherwise it is 0; If the cross-line route and the short-turn route pass through se, it is 1, otherwise it is 0; t min The tracking interval time between trains.

[0145] 4) Stopping constraints

[0146] According to the operation requirements, the train must stop at the turning-back stations of the route. For cross-line routes, in addition to the first and last stations where it must stop, it can be arbitrarily selected whether to stop at intermediate stations.

[0147]

[0148] Among them, A 0-1 variable indicating whether to stop at the two turning-back stations of the cross-line route.

[0149] 5) Formation number constraint

[0150]

[0151] That is, the formation number can only be four, six, or eight.

[0152] Furthermore, the specific steps of S4 are as follows:

[0153] S401, set up the NSGA-II algorithm framework.

[0154] The NSGA-II algorithm starts by initializing the population and calculating the objective function values. Subsequently, it performs fast non-dominated sorting and crowding degree calculation to select the parent population. During the iteration process, it dynamically adjusts the crossover and mutation probabilities by combining reinforcement learning. It generates offspring through selection, crossover, and mutation operations, and then merges with the parent population and performs non-dominated sorting and crowding degree calculation again to form a new parent population until the preset number of iterations is reached, and finally outputs a set of Pareto optimal solutions. The algorithm framework is as Figure 4 shown.

[0155] 1) Chromosome encoding. The decision variables include the departure frequency of the train path, the stop plan, and the formation number. Each chromosome represents a specific train operation plan. Combining the real number encoding and 0-1 encoding methods, the departure frequency of the train path and the formation number use real number encoding, while the stop plan uses 0-1 encoding.

[0156] 2) Selection of fitness function. Using fast non-dominated sorting, all individuals in the population are stratified according to their quality. The individuals in the first layer are the best, and so on. Using the crowding degree operator to compare the quality of individuals in the same layer, the larger the crowding distance, the better the individual.

[0157] 3) Genetic operations. The tournament selection method is a commonly used selection strategy, which preferentially selects individuals with a smaller non-dominated level. When the non-dominated levels of two individuals are the same, it selects the individual with a larger crowding distance. The probabilities of crossover and mutation are obtained through training and optimization by the reinforcement learning model, and the probabilities of crossover and mutation are dynamically adjusted.

[0158] 4) Algorithm termination condition. Completing the preset maximum number of iterations is used as the stopping condition of the algorithm.

[0159] S402, adaptively adjust the crossover and mutation probabilities.

[0160] By combining Q-learning and the NSGA-II algorithm, a method for adaptively adjusting the crossover and mutation probabilities is provided, as Figure 5 shown. The currently obtained state s tInput into the Q - learning algorithm, which uses its internal Q - value table to estimate the value of taking different actions (adjusting the crossover and mutation probabilities) in different states; update the Q - value table according to the output of the Q - learning algorithm; select the optimal action a according to the updated Q - value table t , that is, the best solution for adjusting the crossover and mutation probabilities, which causes a change in the environmental state and generates a new state s t+1 , and so on in a cycle.

[0161] 1) Parameter initialization

[0162] The parameters include: the NSGA - Ⅱ environmental state set S, the action set A, and the Q - value table;

[0163] In the initial state, the Q - value table is set to a all - zero matrix, the number of rows of which corresponds to the number of states of NSGA - Ⅱ, and the number of columns corresponds to the number of actions in the action set A.

[0164] 2) Calculate the environmental state s of the t - th generation of the NSGA - Ⅱ algorithm t ;

[0165] s t = w 1 ·fit * + w 2 ·div * + w 3 ·Best * (w 1 + w 2 + w 3 = 1)

[0166]

[0167] Among them, is the i - th chromosome of the t - th generation population; fit * is the average fitness of the population; div * is the diversity of the population; Best * is the best fitness of the population; is the fitness value of chromosome ; is the best fitness of the t - th generation population; w 1 、w 2 、w 3 are the weight coefficients of fit * 、div * 、Best * respectively; Popsize is the number of populations.

[0168] 3) Calculate the reward function r t+1, the reward function values r corresponding to adjusting the crossover probability and adjusting the mutation probability t+1 are different. When the executed action is to adjust the mutation probability, the corresponding reward function r t+1 is When the executed action is to adjust the crossover probability, the corresponding reward function r t+1 is

[0169]

[0170] where is the i-th chromosome of the (t + 1)-th generation population; is the best fitness of the (t + 1)-th generation population.

[0171] 4) Select an action a according to the ε-greedy strategy t+1 ;

[0172]

[0173] where a is the action variable, a ∈ A; a random represents randomly selecting an action in A; r 0-1 represents generating a random number in [0, 1]; ε is the strategy greed rate.

[0174] 5) Execute the action a t+1 , observe the reward function value r t+1 and the new state s t+1 ;

[0175] 6) Calculate and update the Q(s t , a t ) value;

[0176] Q(s t , a t ) = (1 - α)Q(s t , a t ) + α(r t+1 + γmaxQ(s t+1 , a t+1 ))

[0177] where α is the learning rate; γ is the learning discount rate; r t+1 is the reward function value, which varies depending on the executed action.

[0178] 7) According to the current action a t+1 , change the crossover and mutation probabilities P c , P m .

[0179] The embodiments of the present invention will be described in detail below with reference to the accompanying drawings. To more intuitively illustrate the method for selecting the collaborative operation organization strategy of the urban rail transit and the urban rail train, the present invention is demonstrated through an application example.

[0180] Taking the regional rail transit network composed of Nanjing Metro Line S7 - S1 - Line 3 as an example for verification, in the inbound direction, the time period from 7:00 to 8:00 in the morning is selected to obtain the specific operation organization strategy that meets the line during this period. The boarding and alighting coefficients and the section load factor of each station from 7:00 to 8:00 in the morning are as Figure 6 、 Figure 7 shown.

[0181] Calculate the passenger flow characteristic coefficients of Lines S1 and S7 from 7:00 to 8:00 in the morning, which are used as the basis for selecting the feasible operation organization strategy. The results are shown in Table 2.

[0182] Table 2 Calculation based on indicators for Lines S1 and S7 from 7:00 to 8:00 in the morning

[0183]

[0184]

[0185] First, judge whether the cross - line condition is met. The cross - line passenger flow is 3423 people / h, and the cross - line passenger flow ratio is 82.2%, which meets the condition. It is recommended to operate a cross - line train formation; then judge the section imbalance coefficient. The section imbalance coefficient is 2.24, which is greater than 1.5; judge the boarding and alighting coefficients of the stations, and for stations with overly large or small boarding and alighting coefficients, it is recommended not to stop at every station; judge the section load factor of the intervals. Since the section load factors vary greatly, it is recommended to adopt a multi - formation mixed - running strategy; according to the strategy selection method, three feasible strategies are generated for Lines S1 and S7:

[0186] ① Cross - line + non - stop - at - every - station; ② Cross - line + multi - formation mixed - running; ③ Cross - line + non - stop - at - every - station + multi - formation mixed - running.

[0187] Calculate the passenger flow characteristic coefficients of Line 3 from 7:00 to 8:00 in the morning, which are used as the basis for selecting the operation organization strategy. The results are shown in Table 3.

[0188] Table 3 Calculation based on indicators for Line 3 from 7:00 to 8:00 in the morning

[0189]

[0190] First, it is judged whether the cross-line condition is met. It is not met. Then, the cross-section imbalance coefficient is judged. The cross-section imbalance coefficient is 1.68, which is greater than 1.5. Next, the boarding and alighting coefficient of stations is judged. Stations with too large boarding and alighting coefficients are concentrated in the section from Forest Farm to Shengtaixi Road. It is recommended to operate trains with different numbers of cars on short and long routes. Then, the full-load rate of the cross-section of the section is judged. The difference in the full-load rate of the cross-section is too large, and the sections with a full-load rate greater than 1.2 are too concentrated. It is recommended to implement the strategy of mixed operation with multiple formations. According to the strategy selection method, three feasible strategies are generated for Line 3:

[0191] ① Trains with different numbers of cars on short and long routes; ② Mixed operation with multiple formations; ③ Trains with different numbers of cars on short and long routes + mixed operation with multiple formations.

[0192] Considering that the change in the full-load rate within each area of Line 3 is small, while the change between areas is large, in the case of only operating one long-route train, the mixed operation with multiple formations is not applicable to Line 3; for S1 - S7, only using cross-line + mixed operation with multiple formations and stopping at every station will still cause waste of transport capacity due to the small number of passengers boarding at stations. Therefore, the cross-line + mixed operation with multiple formations is not applicable.

[0193] Therefore, for the suburban rail and urban rail S7 - S1 - Line 3, the operation organization strategies obtained according to the strategy selection are shown in Table 4.

[0194] Table 4 Feasible operation organization strategies obtained from strategy selection

[0195]

[0196] For the above four feasible strategies, the train operation plan is optimized collaboratively. Taking the time cost and operation cost generated by the current actual plan as the yardstick, a set of optimal solutions with lower costs is found. Under each strategy, 3 Pareto optimal solutions are selected. It should be noted that all the selected plans are regarded as the optimal train operation plans, and the optimal plans are summarized in Table 5.

[0197] Table 5 Optimization results of the four strategies

[0198]

[0199] Under different feasible strategies, the optimization effects vary. Among them, the third plan of Strategy 1 is the most prominent in terms of operating costs. That is, during the morning peak period from 7:00 to 8:00, on Line 3, 1 six-car formation train runs on the short turn-back section; 12 six-car formation trains run on the long turn-back section; on Line S1, 6 four-car formation trains run on the line's own turn-back section; on Line S7, 4 four-car formation trains run on the line's own turn-back section; 1 four-car formation train runs on the cross-line operation turn-back section between S1 and S7, and stops at Nanjing South Station, Cuiping Mountain, Focheng West Road, Zhengfang Middle Road, Xiangyu North Road, Airport Lishui Station, Wolong Lake, Zhongshan Lake, Xingzhuang, and Wuxiang Mountain. This specific operation organization strategy reduces the operating cost by 10.99%. This significant cost savings is mainly due to the fine-tuning of train formation, stop arrangements, and departure frequencies, making resource utilization more efficient and reducing unnecessary energy consumption and labor costs.

[0200] On the other hand, the sixth plan of Strategy 2 has the largest reduction in time cost. That is, during the morning peak period from 7:00 to 8:00, on Line 3, 2 six-car formation trains run on the short turn-back section; 12 four-car formation trains run on the long turn-back section; on Line S1, 10 four-car formation trains run on the line's own turn-back section; on Line S7, 6 four-car formation trains run on the line's own turn-back section; 1 four-car formation train runs on the cross-line operation turn-back section between S1 and S7, and stops at Nanjing South Station, Cuiping Mountain, Focheng West Road, Xiangyu South Road, Lukou Airport, Qunli, Lishui, Zhongshan Lake, Xingzhuang, and Wuxiang Mountain. This specific operation organization strategy reduces the time cost by 8.26%.

[0201] The operator can select the most suitable strategy according to the actual situation to achieve the best balance between time and operating costs, thereby improving the overall operation efficiency and service quality.

[0202] The above-described embodiments only represent several implementation manners of the present application, and their descriptions are relatively specific and detailed, but should not be construed as limiting the scope of the invention patent. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several deformations and improvements can still be made, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application should be subject to the appended claims.

[0203] The present application also provides an electronic device, including: a memory and a processor; a computer program is stored on the memory, and when the computer program is executed by the processor, the above-mentioned method for selecting a coordinated operation organization strategy for urban rail transit and urban rail trains is implemented.

[0204] The present application further provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the above-mentioned method for selecting a cooperative operation organization strategy for urban rail transit and urban rail trains are implemented. The computer-readable storage medium may include: various media such as USB flash drives, mobile hard disks, read-only memories (ROM), random access memories (RAM), magnetic disks, or optical discs that can store program codes.

[0205] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the above-mentioned method embodiments. Among them, any reference to a memory, storage, database, or other medium used in the various embodiments provided in the present application can include non-volatile and / or volatile memories. Non-volatile memories can include read-only memories (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memories. Volatile memories can include random access memories (RAM) or external cache memories. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and Rambus dynamic RAM (RDRAM), etc.

Claims

1. A method for selecting a strategy for the coordinated operation of urban rail and urban rail trains, characterized in that: The method includes: S1, construct a set of alternative strategies for organizing the operation of urban rail and urban rail trains from the perspectives of route, station and marshaling; S2, based on the set of alternative strategies in S1, determine feasible strategies suitable for urban rail and urban rail trains; S3, based on the feasible strategy of S2, with the goal of minimizing the passenger travel time cost and the total operating cost of the enterprise, establishes a collaborative optimization model for the operation plan of urban rail and urban rail trains, and solves the route departure frequency, stop plan and number of train formations under each feasible strategy.

2. The method according to claim 1, characterized in that: The S1 routes include three strategies: cross-line routes, large and small routes, and single routes; the stop plans include two strategies: stop at every station and stop at no station; the marshaling modes include two strategies: fixed marshaling and multi-marshaling mixed running.

3. The method according to claim 1, characterized in that: The steps for determining the feasible strategy for organizing the coordinated operation of urban rail and urban rail in S2 include: Step 1: Determine whether the cross-line condition is met. If so, go to Step 2, otherwise go to Step 4; the cross-line condition is the cross-line passenger flow Q 跨 and the cross-line passenger flow ratio λ are both greater than the corresponding set cross-line passenger flow threshold and cross-line passenger flow ratio threshold; Step 2: If the line section imbalance coefficient β is greater than the set imbalance threshold, the passenger flow distribution is judged to be unbalanced and go to Step 3, otherwise go to Step 6; Step 3: If the number of stations whose corresponding passenger flow coefficient γ is greater than or equal to the set passenger flow peak value or less than or equal to the set passenger flow valley value exceeds the set first proportion, it is recommended to implement the non-stop strategy; if the number of stations whose corresponding section load rate η is greater than or equal to the set section load rate peak value or less than or equal to the set section load rate valley value exceeds the set second proportion, it is recommended to implement the multi-marshaling mixed running strategy; otherwise, go to Step 6; Step 4: If the line section imbalance coefficient β is greater than the set imbalance threshold, the passenger flow distribution is judged to be unbalanced and go to Step 5, otherwise go to Step 6; Step 5: If the number of stations whose corresponding passenger flow coefficient γ is greater than or equal to the set passenger flow peak value or less than or equal to the set passenger flow valley value exceeds the set first proportion, and continue to judge: if such stations are concentrated in the same section formed by two turn-around stations, it is recommended to implement the strategy of large and small routes; if such stations are scattered in different sections formed by two turn-around stations, it is recommended to implement the strategy of non-stopping at every station; except for the above two situations, it is recommended to implement the strategy of large and small routes and non-stopping at every station; If the number of stations whose corresponding section full load rate η is greater than or equal to the set section full load rate peak value or less than or equal to the set section full load rate valley value exceeds the set second proportion, it is recommended to implement a multi-marshaling mixed running strategy; Otherwise go to Step 6; Step 6: Maintain the existing operational organizational strategy.

4. The method according to claim 3, characterized in that: The specific calculation formula of the cross-line passenger flow ratio θ in Step 1 is as follows: Among them, Q 跨 is the cross-line passenger flow during the statistical period, Q 本 The passenger flow of this line during the statistical period; The specific calculation formula of the line section imbalance coefficient β is as follows: Among them, Q max is the maximum cross-sectional passenger flow, M is the number of one-way line cross-sections, Q m is the passenger flow of the mth section; Passenger flow pick-up and drop-off coefficient γ at station i i The specific calculation formula is as follows: in, is the maximum number of passengers and passengers per hour at station i, Q u is the hourly passenger and passenger volume of station u, k is the number of stations; The specific calculation formula of section full load rate η is as follows: in, is the passenger flow of section m during the statistical period; N is the number of trains passing through section m in a certain direction during the statistical period; C is the capacity of the train.

5. The method according to claim 1, characterized in that: The collaborative optimization model of the urban rail and urban rail train operation plan in S3 includes: The objective function is: min Z1=t w ·a1·p w min Z2+Z3 The constraints are: Where: Z1 is the passenger travel cost; t w is the travel time of OD for group w passengers; a1 is the passenger time cost coefficient; p w is the number of passengers in the w passenger group for OD; Z2 is the train running kilometer cost; The length of cross-line intersections and small intersections; is the length of the line intersection; is the number of train formations on this line; is the number of train formations on cross-line routes and small routes; a2 is the unit length operation cost coefficient; The frequency of trains running on this line; The frequency of trains running on cross-line routes and small routes; The minimum and maximum train departure frequencies allowed for cross-line routes and small routes; The minimum and maximum train departure frequencies allowed on this line; The maximum cross-sectional passenger flow of cross-line intersections and small intersections; is the maximum cross-sectional passenger flow of this line; P is the capacity of the carriage; η is the train load factor; SE is the interval set; If the line crosses the interval se, it is 1, otherwise it is 0; If the cross-line intersection and small intersection pass through the interval se, it is 1, otherwise it is 0; t min To track the train intervals; A 0-1 variable indicating whether the train stops at the turnaround stations at both ends of the cross-line intersection; C is the intersection set except the main intersection of the line; Z3 is the train stop cost; a 3 is the cost coefficient of a single stop.

6. The method according to claim 1, characterized in that: The NSGA-II algorithm is used to solve the collaborative optimization model of urban rail and urban rail train operation plans, and the route departure frequency, stop plan and number of train formations under each feasible strategy are obtained.

7. The method according to claim 6, characterized in that: The decision variables in the NSGA-II algorithm include route departure frequency, stop plan and number of train formations. Each chromosome represents a specific operation plan. The route departure frequency and number of train formations are coded with real numbers, and the stop plan is coded with 0-1.

8. The method according to claim 6, characterized in that: The Q-Learning algorithm is used to adaptively adjust the crossover and mutation probabilities in the NSGA-II algorithm. The reward function r c Adjust the crossover probability P c , through the reward function r m Adjust the mutation probability P m : In the formula, is the i-th chromosome of the t-th generation population The fitness value of is the i-th chromosome of the t+1 generation population The fitness value of ; Popsize is the population size; is the optimal fitness of the population in generation t; is the optimal fitness of the t+1th generation population.

9. The method according to claim 8, characterized in that: The number of rows in the Q-value table corresponds to the number of environmental states of NSGA-Ⅱ, and the number of columns corresponds to the number of actions in the action set A; in the initial state, the Q-value table is set to an all-zero matrix.

10. The method according to claim 8, characterized in that: The tth generation environmental state s of NSGA-Ⅱ t for; s t =w1·fit * +w2·div * +w3·Best * (w1+w2+w3=1) Among them, fit * is the average fitness of the population; div * For population diversity; Best * is the optimal fitness of the population; w1, w2, and w3 are fit * , div * 、Best * The weight coefficient of .

Images