A method for dispatching urban rail transit of a certain size during disease control period

Abstract

The application discloses a city rail size inter-route traffic scheduling method suitable for a disease control period, and comprises the following steps: obtaining city rail traffic network information, randomly generating a plurality of small inter-route train operation route schemes which can satisfy the traffic network information; generating, enriching and filtering a small inter-route train departure time solution by using a two-stage discrete water wave optimization algorithm to obtain a final solution population; calculating the fitness of the solution with the highest fitness and the fitness of the adjacent solution based on an operation cost and a train disease infection risk calculation model; performing a migration operation or a mutation operation on the small inter-route train operation route scheme corresponding to the fitness to obtain an evolved small inter-route train operation route scheme; and selecting the solution with the minimum current fitness as an optimal solution to obtain a final small inter-route train operation route scheme and a departure time. The application reduces the disease transmission risk of rail transit in operation in a more economical way.

Description

Technical Field

[0001] This invention belongs to the field of intelligent optimization scheduling, specifically relating to a method for scheduling urban rail transit routes of varying sizes during disease control periods. Background Technology

[0002] A well-developed public transportation system is a crucial foundation for the stable operation of a city. Among these systems, subways offer significant advantages over other modes of transport. An efficient subway passenger network not only enhances residents' travel experience but also promotes urban planning and development. The core basis for subway operation is the timetable, which includes the number of trains, routes, and schedules for different time periods. Therefore, it is a vital aspect of urban rail transit operation and a key research focus for universities and subway companies.

[0003] Traditional train schedule planning relies mainly on human experience, with timetables arranged based on the existing rail infrastructure planning and design. These schedules are periodically optimized and adjusted during actual operation. A large amount of work is done manually, resulting in a large workload and a high error rate. The planning process cannot simulate passenger travel conditions, thus lacking accurate assessment of service quality.

[0004] In the existing technology, Jin Bo et al. proposed a collaborative optimization method for urban rail transit timetables and vehicle operation plans that considers both large and small routes. This method considers two objective functions: passenger service quality and vehicle operating cost, but does not take into account the solution of departure times in disease transmission scenarios, including epidemics.

[0005] Meanwhile, the patent application with patent number CN202211161211 discloses a method for determining the timetable for long and short routes based on passenger flow information. It mainly optimizes departure times with the goal of reducing passenger travel costs. However, the scheme is a single-line route scheduling method, which solves for a balanced departure interval scheme and does not comprehensively consider all traffic routes in the city as a whole. Summary of the Invention

[0006] This invention addresses the problems existing in the prior art by providing a rail transit scheduling method that adopts both short-route and long-route operation schemes during disease prevention and control. Using short-route operation lines and departure times for both short-route and long-route operations as decision variables, a rail transit scheduling model considering disease transmission risk and operating costs is constructed. An improved ecological geography optimization algorithm is used to solve for the short-route operation lines, and a two-stage discrete water wave optimization algorithm is used to solve for the long-route and long-route departure time schemes, generating reasonable short-route operation lines and long-route and long-route departure time schemes, achieving economical travel while reducing the risk of disease transmission.

[0007] This invention is achieved through the following technical solution:

[0008] A method for urban rail transit traffic scheduling with varying numbers of routes during disease control periods includes the following steps:

[0009] S1. Based on the acquired urban rail transit network information, randomly generate multiple short-route train operation route schemes that can meet the requirements of the traffic network information.

[0010] S2. Select a short-route train operation plan;

[0011] S3. Based on the selected small-loop train operation route scheme, the first-stage discrete water wave optimization algorithm is used to generate and enrich the various small-loop train departure time schemes corresponding to the selected line, and the various small-loop train departure time schemes are combined with the small-loop train operation route scheme to form various small-loop train operation schemes.

[0012] S4. Based on the fitness calculation model and the second-stage discrete water wave optimization algorithm, the departure time schemes of large and small loop trains in various large and small loop train operation schemes are evolved to obtain the large and small loop train operation scheme with the minimum fitness value, which is then used as the current optimal large and small loop train operation scheme.

[0013] S5. Repeat steps S2 to S4 to obtain the optimal departure time schemes for both small and large loop trains in the selected small loop train operation route schemes, so as to form the overall operation scheme of small and large loop trains corresponding to the small loop train operation route schemes.

[0014] S6. Based on the overall train operation scheme of the large and small routes, select the overall train operation scheme of the large and small routes with the smallest fitness value as the final overall train operation scheme of the large and small routes.

[0015] Preferably, step S1 further includes the step of:

[0016] The various short-route train operation schemes that are randomly generated and can meet the traffic network information are evolved to obtain the evolved various short-route train operation schemes;

[0017] The short-route train operation plan randomly selected in step S2 is chosen from all the short-route train operation plan before evolution and the short-route train operation plan after evolution.

[0018] Preferably, the evolutionary operation includes the following steps:

[0019] S1.1 Select one of the randomly generated short-route train operation schemes that can satisfy the traffic network information, and calculate its fitness value;

[0020] S1.2 Calculate the immigration rate and emigration rate based on fitness values;

[0021] S1.3 Call the random function to generate random numbers;

[0022] S1.4 Determine whether the random number is less than the immaturity parameter. If so, perform a global migration operation based on the migration rate result to generate a new short-route train operation plan; otherwise, perform a local migration based on the migration rate result.

[0023] S1.5 Repeat steps S1.1 to S1.4 to complete the evolution of all schemes for various short-route train operation schemes that can be randomly generated and meet the traffic network information.

[0024] Preferably, the immigration rate I m (X), Migration rate E m The formulas for calculating (X) are as follows:

[0025]

[0026]

[0027] f(x) represents the fitness, f min f max ε and E represent the minimum and maximum fitness values ​​calculated in step S1.1, respectively. ε is a positive integer, E is a fixed parameter, and X represents the short-route train operation route scheme.

[0028] Preferably, step S3 specifically includes:

[0029] S3.1 Based on the selected route for small-loop trains, use enumeration to generate several departure time schemes for small-loop and large-loop trains with equal time intervals, so as to generate multiple departure time schemes for small-loop and large-loop trains.

[0030] S3.2. Use propagation operations to search for train departure time schemes for all routes, large and small, to generate new train departure time schemes for all routes, large and small.

[0031] The search is defined as performing a reversal operation on the departure time sub-schemes of each large and small route. Specifically, the reversal operation involves reversing whether a train departs at the corresponding time in the departure time sub-schemes of large and small routes.

[0032] Preferably, step S4 specifically includes the following steps:

[0033] S4.1. Perform a propagation operation on the departure time plans for trains on various routes to generate new departure time plans for trains on various routes.

[0034] S4.2 Based on the fitness calculation model, calculate and compare the fitness values ​​of the current train departure time schemes for each route, and select the train departure time scheme with the smallest fitness as the current optimal scheme.

[0035] S4.3 For the new large and small route train departure time schemes, if the fitness value is less than that of the original large and small route train departure time schemes, the original schemes shall be replaced.

[0036] S4.4. For the newly generated departure time scheme with better fitness than the current optimal train departure time scheme with the smallest fitness, perform a wave-breaking operation to generate adjacent schemes of the same train departure time scheme for this route, and select the scheme with the smallest fitness value among all schemes to replace the optimal scheme.

[0037] All schemes refer to the train departure time schemes for the large and small routes themselves and the adjacent schemes generated after applying wave-breaking operations to them.

[0038] Preferably, step S4 further includes limiting the number of departure time schemes for both large and small route trains in each iteration, and the limitation calculation formula is as follows:

[0039]

[0040] Where NP(t) represents the number of sub-schemes for train departure times of different routes in the t-th iteration, NP max and NP min These represent the maximum and minimum number of preset train departure time sub-schemes for both large and small routes, t max This indicates the maximum number of iterations for the algorithm.

[0041] Preferably, the fitness calculation model is as follows:

[0042] f(x) = α·R + (1-α)·C

[0043] Where f(x) represents fitness, x represents the train operation plan for both large and small routes, α represents the weighting coefficient, C represents the operating cost, and R represents the risk of train disease infection.

[0044] Preferably, the train disease infection risk calculation model is as follows:

[0045]

[0046] Where R represents the risk of disease infection on the train, δ represents a preset constant value, and r represents the city's risk level. R' indicates the risk of infection inside the train, while R' indicates the risk of infection on the platform.

[0047] Preferably, the operating cost calculation model is as follows:

[0048]

[0049] Where C represents operating costs, s represents transportation routes, i represents train numbers, and a s This indicates the number of trains operating on the s-th transportation route, b. s D indicates the number of trains operating on the short-route of transportation line s. s,i Let c' represent the distance traveled by the i-th vehicle on route s, and let c' represent the cost per kilometer traveled.

[0050] The urban rail transit traffic scheduling method for both short and long routes during disease control periods has the following advantages and significant effects compared to existing technologies:

[0051] First, the present invention adopts a combination of large and small routes to accelerate vehicle operation, optimize capacity allocation, avoid resource waste, and thus reduce operating costs;

[0052] Secondly, this invention considers the risk of disease transmission while meeting travel needs, and adjusts the train operation plan according to the risk of disease transmission, thus providing a relatively safe travel environment.

[0053] Third, this invention provides an operational solution that takes into account both disease risk and operational cost pressures, achieving a dual guarantee of economy and safety. Attached Figure Description

[0054] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0055] Figure 1 This is a schematic diagram of the traffic network information of the present invention;

[0056] Figure 2 This is a flowchart of a method for urban rail transit traffic scheduling applicable during disease control periods, according to the present invention. Detailed Implementation

[0057] The present invention will be further described below with reference to specific embodiments and accompanying drawings. These embodiments are provided to enable those skilled in the art to better understand the invention and do not constitute any limitation on the invention.

[0058] Example 1:

[0059] S1. Based on the acquired urban rail transit network information, randomly generate multiple short-route train operation route schemes that can satisfy the traffic network information. Each short-route train operation route scheme includes multiple short-route train operation route schemes corresponding to multiple lines, as detailed below:

[0060] like Figure 1 The urban traffic network information shown includes various routes in the city's traffic system, as well as the stations on each route.

[0061] Related explanation: A long route refers to a subway line running the entire route, while a short route refers to a line running through one of the stations on the entire route as a temporary terminus, similar to a "section" on a bus. The subway uses long and short routes to alleviate passenger flow and save energy. This invention uses long and short routes to solve the scheduling problem.

[0062] The city's road network information is described using mathematical models:

[0063] If there are m lines on line s s Each station is numbered sequentially according to the train's direction of travel. The platform set for line s is... This indicates that the collection point of the long-distance bus route is j. s .

[0064] Because the platforms are symmetrical, it is only necessary to know the stops for the short-haul route in the uphill direction. For the uphill direction of the short-haul route, use j. a Indicates the starting platform, j b This indicates the platform for the turnaround; for the downline direction, the starting platform is... The final station is Line S short-haul bus stop platform collection express.

[0065] Because trains need a certain travel time between stations, this manifests as the time difference between the arrival and departure of the same train from different stations.

[0066] To make the model more realistic, different sites in this invention have different research periods. Let the research period for the starting site be (T). a ,T b ], where T a The start time is the departure time of the first bus; T b The end time is the departure time of the last bus, which is a fixed time period depending on the actual situation.

[0067] In this study, the first train is used to determine the study period for each station, so only the arrival and departure times at each platform are calculated. After discretization, the set τ = {1,2,3,…,ω-1,ω} represents the study time node of the starting station. The study start time for other stations has two cases: if the station belongs to a short-route service, the initial study time is the departure time of the first train on the short-route; otherwise, the initial study time is the departure time of the first train on the long-route. The study end time for all stations is the departure time of the last train on the long-route.

[0068] Let the total research period be (T) a ,T c ], where T a Indicates the start time of the study, T c This represents the latest time the last train departs from the terminal station across all routes. After discretization, the set of time nodes for the total study period can be obtained.

[0069] In this mathematical model, train scheduling with both long and short routes needs to be planned for a rail transit system with a regional risk level of r. For the safety of rail transit operations, the maximum time interval between long-route train departures is [missing information]. Minimum time interval is The maximum time interval between departures of short-haul trains is: Minimum time interval is The city's rail transit network has a total of n lines, represented by the set S = {s1, s2, ..., sn}. n} indicates that the maximum number of long-distance trains that can be scheduled on line s is The number of short-haul trains is Line s has a total of m s A site, using This represents the set of platforms on this line, where the distance from platform j-1 to platform j is Δd. s,j The cost of the train traveling one kilometer is c'. There are u transfer stations in the rail transit network, and K represents the set of transfer platforms in the rail transit network.

[0070] The number of passengers arriving at platform j of line s and heading to platform k within a unit time interval is p. s,j,k Passengers should wait for the train in an orderly manner after arriving at platform j of line s. The maximum capacity of platform j is M, the platform ventilation volume is Q', and the train ventilation volume is... If there is I infected individuals among all passengers, in medicine, the risk of airborne transmission of respiratory infectious diseases is predicted using Quanta, which represents the minimum number of pathogens that can infect an infected individual. The Q value for each infected individual is... uanta The production rate is q. Due to prevention and control measures, passengers will wear masks while traveling on rail transit. The passenger's breathing rate is B, and the viral infection rate while wearing a mask is θ.

[0071] Train i on line s arrives at the starting station of the up-line according to the departure timetable and prepares to depart, stopping at each platform. The train departs after a certain time; the travel time between platform j and platform j is Δt. s,j After train i arrives at platform j, passengers board and alight. The train's maximum passenger capacity is D, and the passenger capacity cannot exceed the local occupancy rate requirement β. If platform j is a transfer platform, some disembarking passengers will transfer to another platform, and the proportion of passengers transferring from platform j to platform j on line s' is e. s,s',j,j' Passenger transfer walking time is After the train arrives at the turnaround station, it begins the turnaround process, which takes approximately [time missing]. The turnaround distance is After completing the turnaround operation, the train switches from the up track to the down track and continues its journey. Upon arrival at the terminal station, it enters the waiting queue to await the next departure instruction.

[0072] The problem that needs to be solved is: during the research period (T) a ,T b Plan the routes for short-route trains on n lines and formulate departure schedules for both short and long routes;

[0073] The problem has three decision variables: Z = {Z1, Z2, ..., Zn} n} represents the short-route train lines in the rail transit network, Z s This indicates the route for the upline trains on line s (small route), which is determined by m. s It consists of 0-1 vectors, where This indicates that the direction of travel on line s is selected with the starting station j as the starting point. a Terminal station J b As a short-haul route.

[0074] A vector of dimensions n×ω This indicates the departure schedule for long-distance trains, among which... The departure schedule scheme for the long-haul route s is represented by ω zero-1 variables, as follows: for This means that the train needs to depart at time t; otherwise, it does not need to depart at time t (1≤t≤ω).

[0075] A vector of dimensions n×ω This indicates the departure time plan for short-route trains in the rail transit network, among which... The departure schedule scheme for the short-route line s is represented by ω zero-1 variables. for This means that the train needs to depart at time t; otherwise, it does not need to depart at time t (1≤t≤ω).

[0076] The number of trains departing on line S is:

[0077]

[0078] If the train numbers are numbered sequentially according to the departure order of the long-distance trains, then the set of long-distance trains operating on line s is:

[0079] The number of short-route trains departing on line S is:

[0080]

[0081] If the train numbers are numbered sequentially according to the departure order of the short-route trains, then the set of short-route trains running on line s is:

[0082] The train set for line s is The departure time corresponding to train i is represented as T. s,i Since there are two different train routes on the line, the following 0-1 variables are defined for ease of description. μ s,i,j The value ρ indicates that train i on line s passes through platform j, and is 1 otherwise. s,i,j The value σ is 1 if platform j on line s is the starting platform of train i on the up line; otherwise, it is 0. s,i,j The value is 1 if platform j on line s is the starting platform of the downline of train i, and 0 otherwise.

[0083] The time when train i arrives at the starting station on the up line is its departure time. During the journey, if platform j is the starting platform of train i on the down line, it means that the train arrives at platform j after a turnaround operation, and the arrival time is the time when the train left the previous platform plus the turnaround time; otherwise, the arrival time is equal to the time when the train left the previous platform plus the inter-station travel time.

[0084]

[0085] The time when train i leaves platform j on line s is equal to the time when the train arrives at the platform plus the stopping time.

[0086]

[0087] When train i arrives at the originating station, the number of passengers on board is 0. During the journey, the number of passengers on board when train i arrives at platform j on line s is equal to the number of passengers on board when the train arrived at the previous platform plus the total number of passengers on board minus the number of passengers getting off.

[0088]

[0089] in This represents the number of passengers disembarking when train i arrives at platform j-1. This represents the number of passengers boarding when train i arrives at platform j-1.

[0090] At time t, the number of people waiting for a train from platform j on line s to platform k is equal to the number of people waiting at the previous time plus the passenger flow per unit time. If platform j is a transfer platform, the number of transfer passengers at the station also needs to be considered. Since different platforms have different study periods, if time t is not within the study period of platform j on line s, the passenger flow is 0. This can be expressed as:

[0091]

[0092] in This represents the number of passengers who arrive at platform j of line s at time t and transfer to platform k.

[0093] The total number of passengers waiting for the train at platform j of station s at time t is:

[0094]

[0095] When train i arrives at platform j on line s, the remaining number of passengers inside the train is the maximum number of passengers the train can accommodate minus the number of passengers inside the train when it arrives at the station plus the number of passengers who get off the train at the station.

[0096]

[0097] Because there are two different operating routes on the line, when calculating the number of passengers boarding, it is necessary to consider both the remaining capacity of the train and whether passengers can board the train to reach the target platform. This invention considers that passengers waiting at the platform before the train departs can board. If the remaining capacity of the train is greater than or equal to the expected number of passengers, that is, the train can carry all the expected passengers. In this case, passengers boarding train i from platform j on line s to platform k can be represented as:

[0098]

[0099] If the remaining capacity of a train upon arrival is less than the expected number of passengers, the train cannot accommodate all expected passengers. Due to the unique nature of rail transit, passengers will always board the first train to arrive when boarding is available. Therefore, assume that the proportion of passengers boarding from platform j to platform k is the same as the proportion of passengers waiting for the train. In this case, the number of passengers boarding train i from platform j to platform k on line s can be represented as:

[0100]

[0101] Based on the above steps, passengers board train i at platform j on line s and proceed to platform k.

[0102]

[0103] The number of passengers boarding train i at platform j on line s is:

[0104]

[0105] When the train leaves the platform, passengers board, so the number of passengers waiting on the platform decreases accordingly, as shown below:

[0106]

[0107]

[0108] The number of passengers disembarking when train i arrives at platform j on line s is equal to the number of passengers boarding at platform j as the destination platform.

[0109]

[0110] If platform j is a transfer platform, some disembarking passengers will transfer. If the transfer destination is platform j on line s', the transferring passengers will... The number of transfer passengers arriving at the target platform at time t can be expressed as:

[0111]

[0112] Since passengers traveling at the same time often have similar travel characteristics, assuming that the proportion of transfer passengers heading to platform k after arriving at the target platform is the same as the proportion of passenger flow within a unit time interval, it can be expressed as:

[0113]

[0114] The mathematical model for this rail transit network needs to satisfy the following constraints:

[0115] (1) Too many routes will seriously increase the difficulty of operation and organization for the operator, so each route should have at most one small route.

[0116]

[0117] (2) The departure time of the last bus is fixed.

[0118]

[0119]

[0120] (3) To ensure the safety of train operation, the interval between trains must be greater than the minimum interval between trains. On the other hand, to avoid passengers waiting for too long, the interval between trains must be less than the maximum interval between trains.

[0121] T min <T s,i -T s,u-1 <T max

[0122] (4) The number of passengers waiting for the train shall not exceed the maximum capacity of the platform. Similar to the maximum capacity of the train, the maximum capacity of the platform shall also meet the load factor requirements.

[0123] N' s,j,t <M·β

[0124] (5) Since the short-route train cannot pass through all platforms, if platform j is the starting station of train i's up-line, it means the train is arriving at the starting station to prepare for departure, so the inter-station travel time does not need to be added; if platform j is the starting station of train i's down-line, it means the train arrives at platform j through a turnaround operation, so the turnaround time should be added instead of the inter-station travel time. Therefore, the total travel time for train i on line s is:

[0125]

[0126] The train schedule plan needs to meet the capacity of the rail transit operator, that is, there must be trains available for dispatch when the train needs to depart at the specified departure time.

[0127]

[0128]

[0129] Based on the above information, NP solutions X for short-interval travel schemes that satisfy the constraints can be randomly generated as the initial population. Here, NP represents the initial parameters defined for the algorithm.

[0130] S2. Select a short-route train operation plan;

[0131] S3. Based on the selected small-loop train operation route scheme, the first-stage discrete water wave optimization algorithm is used to generate and enrich the various small-loop train departure time schemes corresponding to the selected line, and the various small-loop train departure time schemes are combined with the small-loop train operation route scheme to form various small-loop train operation schemes.

[0132] The first stage of the discrete water wave optimization algorithm includes the following steps:

[0133] S3.1. Based on the selected short-route train operation plan, several equal-time interval train departure time plans for both short and long routes are generated using an enumeration method, resulting in a variety of short and long route train departure time plans, as detailed below:

[0134] S3.11. Use the enumeration method to generate a train departure time scheme with equal time intervals for both small and large routes. The train operation scheme consisting of the selected small route train operation route scheme and the small and large route train departure time scheme is denoted as solution x.

[0135] S3.12. If the solution x satisfies the mathematical model constraints of the rail transit network, then pop it into the initial population.

[0136] If the conditions are not met, discard the item.

[0137] S3.13, repeat S3.11 to S3.12 until all equal time interval train departure time schemes for large and small routes that satisfy the maximum time interval constraint and the minimum time interval constraint are generated.

[0138] S3.2. Based on the fitness calculation model, calculate the fitness values ​​of various train operation schemes with different route sizes.

[0139] The fitness calculation model is as follows:

[0140] In the medical field, the Wells-Riley model is commonly used to calculate the probability of transmission and the number of infections. Based on this train disease infection risk calculation model, the risk of infection within the train is:

[0141]

[0142] The risk of infection on the platform can be represented as:

[0143]

[0144] The total risk of infection during train operation can be expressed as:

[0145]

[0146] Where R represents the risk of disease infection on the train, δ represents a preset constant value, and r represents the city's risk level. R represents the risk of infection inside the train, and R' represents the risk of infection on the platform. This formula shows that the risk of epidemic transmission R varies with the regional risk level r and the risk of infection inside the train. The risk of infection (R') on the platform increases exponentially.

[0147] Where δ is a constant greater than 1;

[0148] Since trains on short-route lines cannot stop at all platforms, if platform j is the starting station for train i's up-line journey, indicating that the train is arriving at the starting station and preparing to depart, then the inter-station travel distance Δd does not need to be added. s,j If platform j is the initial station on the down line of train i, it means the train arrived at platform j through a turnaround operation, so the turnaround distance should be added. Instead of the inter-station travel distance Δd s,j Then the travel distance of train i on line s is:

[0149]

[0150] The operating cost calculation model can be expressed as:

[0151]

[0152] Where C represents operating costs, s represents transportation routes, i represents train numbers, and a s This indicates the number of trains operating on the s-th transportation route, b. s D indicates the number of trains operating on the short-route of transportation line s. s,i Let c' represent the distance traveled by the i-th vehicle on route s, and let c' represent the cost per kilometer traveled.

[0153] Different risk levels in different regions will have different emphases on operating costs and epidemic risk. For example, in low-risk areas, more attention will be paid to operating costs than the risk of epidemic transmission. Therefore, the linear weighted (SAW) method is used to construct the objective function:

[0154] min f(χ)=α·R+(1-α)·C

[0155] Where f(x) represents fitness, α represents weighting coefficient, C represents operating cost, and R represents train disease infection risk.

[0156] S3.3. Use propagation operations to search for train departure time schemes for all routes, large and small, to generate new train departure time schemes for all routes, until the maximum number of train departure time schemes for all routes is reached, as detailed below:

[0157] S3.31, Setting parameters: α, NP max ;

[0158] S3.32. Enrich the initial solution set using propagation operations. Perform a propagation operation on each solution x in the initial population pop to generate a new solution x';

[0159] The specific dissemination procedures are as follows:

[0160] Since the problem to be solved is a discrete problem, the propagation operation is defined as performing p searches on the solution x to generate a new solution. To continue the idea in the basic water wave optimization algorithm that the better the fitness of the solution, the smaller the search range of the solution. Therefore, in this invention, the better the fitness of the solution x, the smaller p is; the worse the fitness of the solution x, the larger p is. p is [1, λ]. x A random integer between λ and 1, where λ is a random integer between λ and 1. x The wavelength λ represents the solution x. x The calculation formula is as follows:

[0161]

[0162] Where α is the wavelength attenuation coefficient, f max and f min These are the maximum and minimum fitness of the population, respectively, and ε is a very small positive integer used to prevent division by zero anomalies.

[0163] Each search is defined as randomly selecting a path s for a solution x, and then searching for x... s The [d1,d2] dimension is reversed to generate a new solution x'. Here, d1 and d2 are two randomly generated positive integers (0, 1, 2, ...). <d1<ω,d1<d2<ω)。

[0164] S3.33. If the new solution x' satisfies the constraints of the problem model, then add it to the population pop; otherwise, discard it.

[0165] S3.34, repeat S3.32 to S3.33 until the population pop count reaches NP. max ;

[0166] S4. Based on the fitness calculation model and the second-stage discrete water wave optimization algorithm, the departure time sub-schemes of large and small loop trains in various large and small loop train operation schemes are evolved to obtain the large and small loop train operation scheme with the smallest fitness value, which is used as the optimal large and small loop train operation scheme corresponding to the currently selected line.

[0167] The second-stage discrete water wave optimization algorithm includes the following steps:

[0168] S4.1 Setting parameters: K N ;

[0169] S4.2. Propagate the departure time schemes for each route (both large and small) to generate new departure time schemes. If the fitness of the newly generated departure time scheme is better than the optimal departure time scheme, trigger the wave-breaking operation to generate adjacent schemes for the same route departure time scheme, and select the route with the smallest fitness value from all schemes. Specifically:

[0170] S4.21, The optimal solution in the pop population is x. * ;

[0171] S4.22. Select a solution x from the pop population and perform a propagation operation to generate a new solution x'. If the fitness of the new solution x' is better than x, replace x; otherwise, keep x.

[0172] S4.23. If the fitness of x' is better than that of x * This will trigger a wave-breaking operation. K independent waves are generated around x' for a local search, and k is determined using an adaptive method:

[0173]

[0174] Among them, K N Indicates the control parameter; f(x) * Let f(x') represent the fitness of the old optimal solution, and f(x') represent the fitness of the newly generated optimal solution; ε is a very small positive integer used to prevent division by zero anomalies. This method of updating k means that the more the new optimal solution improves upon the old optimal solution, the more neighbor searches will be performed in the vicinity of x';

[0175] Each independent wave will randomly select one path s from the solution x', and then process x. s The two dimensions d1 and d2 are swapped. If the generated independent wave has a fitness better than x... * Then it replaces x * Conversely, retain x. * ;

[0176] S4.24. Perform steps S4.21 to S4.23 once for each solution in the population;

[0177] S4.3. Each iteration also includes limiting the number of departure time schemes for both long and short route trains, as detailed below:

[0178] S4.31, Setting parameters: NP min ;

[0179] S4.32. By employing a population reduction strategy to remove stagnant or low-quality solutions, convergence is accelerated. Population size update method:

[0180]

[0181] Where NP(t) represents the population size, NP max and NP min These represent the maximum and minimum population sizes, respectively, and t is the current iteration number. maxThis represents the maximum number of iterations for the algorithm. Assuming the current population size is NP, if NP(t) is less than NP, then before proceeding to the next iteration, the individual with the worst fitness, NP-NP(t), is removed from the population.

[0182] S4.4, Iterative steps S4.1 to S4.3. When the algorithm reaches the maximum number of iterations, the train departure time scheme with the smallest fitness value is selected as the final scheme.

[0183] S5. Repeat steps S2 to S4 to obtain the train operation plans corresponding to all small-loop train operation plans generated in S1.

[0184] S6. Select a train operation plan with both long and short routes;

[0185] S7. Based on the selected large and small route train operation schemes, the small route train operation route schemes are further refined. Based on the newly generated small route train operation route schemes, the newly generated small route train operation route schemes and their corresponding large and small route train departure time schemes are combined to form a large and small route train operation scheme;

[0186] S7.1 Select a short-route train operation plan;

[0187] S7.2 Calculate the immigration rate and emigration rate, as follows:

[0188] S7.21. Based on the solution x in the population, obtain the solution X for the short-route train operation.

[0189] S7.22, Setting parameters: I, E;

[0190] S7.23. An ecogeographical algorithm based on local topological connectivity is adopted, mainly using two migration operators (local migration and global migration) to evolve the population. Each solution X is assigned an I. m (X) Immigration rate and emigration rate E m (X) is calculated as follows:

[0191]

[0192]

[0193] Here, I and E are two control parameters, which are generally set to 1, and f max and f min Let I be the maximum and minimum fitness values ​​in the current population, respectively. For each solution X, let I be the maximum and minimum fitness values. m The probability of (X) is used for migration, which allows solutions with better (poor) fitness to have smaller (larger) migration probabilities.

[0194] S7.3. Determine if the random number is less than the immaturity parameter. If so, perform a global migration operation based on the migration rate result to generate a new short-route train operation plan; otherwise, perform a local migration based on the migration rate result. Details are as follows:

[0195] S7.31, Setting parameters: η max η min ;

[0196] S7.32. Select one of the paths to solve X and generate a random number r = rand(0,1);

[0197] S7.33, if r m If (X), then a migration operation is required; otherwise, proceed to step S7.38.

[0198] S7.34, Based on the emigration rate E m (X) Select the neighboring location for migration. The specific selection method is as follows: Generate a random number r1 = rand(0,1), and iterate through the neighboring solutions connected to solution X. Each solution is denoted as X. nb If r1 <E m (X nb If so, then choose X. nb As the place of origin;

[0199] S7.35. Calculating the immaturity parameter. To control the probability of global and local migrations during iteration, an immaturity parameter η is designed to ensure a balance between global and local search by maximizing global migrations in the early stages of the algorithm and maximizing local migrations in the later stages. Each migration operation has a probability of η for global migration and a probability of (1-η) for local migration. η decreases linearly with the number of iterations, as shown below:

[0200]

[0201] Where t represents the current iteration number, t max η represents the maximum number of iterations. max and η min These are the control parameters, representing the maximum and minimum values ​​of η;

[0202] S7.36. Generate a random number r2 = rand(0,1). If r2 > η, perform a local migration. This invention designs the migration operation for all stations j on line s, setting X(s,j) = X nb (s,j), that is, the solution X nb The short-trip operation plan for line s in the middle section will be simultaneously provided to X;

[0203] ​S7.37, if r2≤η, perform global migration. In addition to the neighbor X nb , a non-neighbor solution X that is not connected to solution X is selected according to the emigration rate E m (X). If the fitness of X far is better than that of X nb , perform migration operation on X far ; otherwise, perform migration operation on X nb ; otherwise, perform migration operation on X far ;

[0204] S7.38, loop through steps S7.31 to S7.36 until all routes of the selected solution X are traversed, and a new solution X' is generated;

[0205] S7.38, use steps S3 to S4 to solve the departure time scheme of full-length and short-turn trains corresponding to the new solution X', and form a new full-length and short-turn train operation scheme with the solution and the corresponding departure time scheme of full-length and short-turn trains, recorded as x'. If f(x')<f(x), replace x with x'; otherwise, do not replace;

[0206] S7.4, perform mutation operation on the short-turn operation route scheme to generate a new short-turn operation route scheme. For the new short-turn operation route scheme, use steps S3 to S4 to obtain the departure time of full-length and short-turn trains corresponding to the short-turn operation route scheme, and form a full-length and short-turn train operation scheme. The details are as follows:

[0207] S7.41, set parameters: π max , π min ;

[0208] S7.41, according to different evolution processes, an adaptive mutation factor method is proposed to balance the relationship between global search and local search, thereby improving the efficiency and convergence of problem solving. Each solution X has a probability Π to mutate, and the specific calculation formula of mutation rate π is as follows:

[0209]

[0210] wherein, t represents the current number of iterations, t max represents the maximum number of iterations, π max and π min are control parameters, representing the maximum and minimum mutation rates respectively.

[0211] S7.42, generate a random number r3=rand(0,1), if r3<π, mutation operation is required; otherwise, jump to step S7.5. The mutation operation is defined as randomly selecting a route s of solution X and exchanging its two dimensions d1 and d2 to generate a new short-turn operation route scheme X';

[0212] S7.43, the departure time scheme of full- and partial-route trains corresponding to the new solution X' is solved by using steps S3 to S4, and the new full- and partial-route train operation scheme is formed by combining the solution with the corresponding departure time scheme of full- and partial-route trains, which is recorded as x'. If f(x')<f(x), replace x with x'; otherwise, do not replace;

[0213] S7.5, repeat steps S7.1 to S7.4 until all full- and partial-route train operation schemes are traversed;

[0214] S8, repeat steps S6 to S7. When the algorithm reaches the maximum number of iterations, the full- and partial-route train operation scheme with the minimum fitness value is selected as the final full- and partial-route train operation scheme.

[0215] In the above embodiment, according to preferred parameter settings, in said step S3.31, α=1.0026, NP max =50. In said step S4.1, K N =12. In said step S4.31, NP min =6. In said step S7.22, I=1, E=1. In said step S7.31, η max =0.6, η min =0.3. In said step S7.41, π max =0.3, π min =0.05.

[0216] Beneficial effects of the present invention are:

[0217] First, the present invention designs a rail transit scheduling method based on a hybrid heuristic algorithm to solve this problem, adopts the full- and partial-route operation mode for rail transit, accelerates vehicle turnover, improves vehicle utilization efficiency, and enables the transportation capacity of sections with large passenger flow to meet the demand of passenger volume;

[0218] Second, when ensuring scheduling, the present invention adds a disease risk calculation model and an operation cost model to optimize capacity allocation, so as to reduce the epidemic transmission risk caused by passenger travel, and achieve the reduction of disease transmission risk including epidemic caused by people taking rail transit in a more economical way.

[0219] The above described embodiment only describes the preferred embodiment of the present invention, and does not limit the scope of the present invention. Without departing from the design spirit of the present invention, various variations and improvements made by those of ordinary skill in the art to the technical solution of the present invention shall all fall within the protection scope of the present invention.

Claims

1. A method for urban rail transit traffic scheduling with varying routes during disease control periods, characterized in that, Including the following steps: S1. Based on the acquired urban rail transit network information, randomly generate multiple short-route train operation route schemes that can meet the requirements of the traffic network information. S2. Select a short-route train operation plan; S3. Based on the selected small-loop train operation route scheme, the first-stage discrete water wave optimization algorithm is used to generate and enrich the various small-loop train departure time schemes corresponding to the selected line, and the various small-loop train departure time schemes are combined with the small-loop train operation route scheme to form various small-loop train operation schemes. S4. Based on the fitness calculation model and the second-stage discrete water wave optimization algorithm, the departure time schemes of large and small loop trains in various large and small loop train operation schemes are evolved to obtain the large and small loop train operation scheme with the minimum fitness value, which is then used as the current optimal large and small loop train operation scheme. S5. Repeat steps S2 to S4 to obtain the optimal departure time schemes for both small and large loop trains in the selected small loop train operation route schemes, so as to form the overall operation scheme of small and large loop trains corresponding to the small loop train operation route schemes. S6. Based on the overall train operation plan for both long and short routes, select the overall train operation plan for both long and short routes with the smallest fitness value as the final overall train operation plan for both long and short routes. Step S1 also includes the following steps: The various short-route train operation schemes that are randomly generated and can meet the traffic network information are evolved to obtain the evolved various short-route train operation schemes; The short-route train operation plan selected in step S2 is chosen from all the short-route train operation plan before evolution and the short-route train operation plan after evolution. The evolutionary operation includes the following steps: S1.1 Select one of the randomly generated short-route train operation schemes that can satisfy the traffic network information, and calculate its fitness value; S1.2 Calculate the immigration rate and emigration rate based on fitness values; S1.3 Call the random function to generate random numbers; S1.4 Determine whether the random number is less than the immaturity parameter. If so, perform a global migration operation based on the migration rate result to generate a new short-route train operation plan; otherwise, perform a local migration based on the migration rate result. S1.5 Repeat steps S1.1 to S1.4 to complete the evolution of all schemes for various short-route train operation routes that can be randomly generated and meet the traffic network information. The fitness calculation model is as follows: in, Indicates fitness. This indicates the train operation plan for both long and short routes. Indicates the weighting coefficient. R represents operating costs, and R represents the risk of disease infection on the train.

2. The method for urban rail transit scheduling during disease control as described in claim 1, characterized in that, The migration rate emigration rate The calculation formulas are as follows: ， ， Indicates fitness. , These represent the minimum and maximum fitness values ​​calculated in step S1.1, respectively. It is a positive integer. As a fixed parameter, X represents the route plan for short-distance trains.

3. The method for urban rail transit scheduling during disease control as described in claim 1, characterized in that, Step S3 specifically includes: S3.1 Based on the selected short-route train operation route scheme, use the enumeration method to generate several long-route train departure time schemes with equal time intervals, so as to generate multiple long-route train departure time schemes. S3.

2. Use propagation operations to search for train departure time schemes for all routes, large and small, to generate new train departure time schemes for all routes, large and small. The search is defined as performing a reversal operation on the departure time schemes of trains on various routes, specifically reversing whether a train departs at a certain time in the departure time schemes of various routes.

4. A method for urban rail transit scheduling during disease control as described in claim 1, characterized in that, Step S4 specifically includes the following steps: S4.

1. Perform a propagation operation on the departure time plans for trains on various routes to generate new departure time plans for trains on various routes. S4.2 Based on the fitness calculation model, calculate and compare the fitness values ​​of the current train departure time schemes for each route, and select the train departure time scheme with the smallest fitness as the current optimal scheme. S4.3 For the new large and small route train departure time schemes, if the fitness value is less than that of the original large and small route train departure time schemes, the original schemes shall be replaced. S4.

4. For the newly generated departure time scheme with better fitness than the current optimal train departure time scheme with the smallest fitness, perform a wave-breaking operation to generate adjacent schemes of the same route departure time scheme, and select the scheme with the smallest fitness value among all schemes to replace the optimal scheme.

5. A method for urban rail transit scheduling during disease control as described in claim 1, characterized in that, In step S4, each iteration also includes limiting the number of departure time schemes for both large and small route trains. The limitation calculation formula is as follows: in, This represents the number of train departure time schemes for both large and small routes in the t-th iteration. and These represent the maximum and minimum number of pre-set train departure time schemes for both long and short routes. This indicates the maximum number of iterations for the algorithm.

6. A method for urban rail transit scheduling during disease control as described in claim 1, characterized in that, The train disease infection risk calculation model is as follows: in, Indicates the risk of disease infection on the train. This represents a preset constant value. Indicates the city's risk level. Indicates the risk of infection on the train. This indicates the risk of infection on the platform.

7. A method for urban rail transit scheduling during disease control as described in claim 1, characterized in that, The operating cost calculation model is as follows: in, Indicates operating costs, Indicates transportation routes. Indicates the train number. Indicates the first The number of trains on the No. 1 transportation line is a long-distance route. Indicates the first The short-route service on the No. 1 transportation line will operate [number] times. Indicates the first The No. 1 transportation route Vehicle travel distance, This represents the cost per kilometer traveled.

Images