Subway adaptive dispatching method and electronic equipment based on passenger flow classification
By using a passenger flow classification method based on Kmeans++ and Bayesian risk assessment, the subway departure interval is dynamically adjusted, which solves the problems of peak-hour congestion and off-peak waste in subway scheduling, and achieves more efficient capacity utilization and scheduling flexibility.
Patent Information
- Application Number
- CN201911053367.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2019-10-31
- Publication Date
- 2024-11-01
- Estimated Expiration
- 2039-10-31
AI Technical Summary
The existing subway scheduling methods lack flexibility, resulting in congestion during peak hours and wasted capacity during off-peak hours, making it impossible to achieve reasonable capacity allocation and passenger flow management.
Based on data from the subway ground operation and maintenance system, a passenger flow classification system was established using the Kmeans++ algorithm. Combined with the Bayesian risk assessment method, an adaptive scheduling model for the subway was established to dynamically adjust the departure interval.
It improved the utilization rate of subway line capacity, reduced the waste of transport capacity, enhanced the sensitivity of dispatching, and reduced the probability of passenger congestion.
Smart Images

Figure CN112749862B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of artificial intelligence, and in particular to a subway adaptive scheduling method and electronic equipment based on passenger flow classification. Background Art
[0002] The current subway departure intervals are fixed according to the daily peak and normal passenger flow times, which lacks flexibility and causes passenger congestion during peak passenger flow and waste of capacity during low passenger flow times. Summary of the invention
[0003] In view of this, the purpose of the present invention is to provide a subway adaptive scheduling method and electronic equipment for reasonably arranging transportation capacity and guiding passenger flow.
[0004] Based on the above purpose, the present invention provides a subway adaptive scheduling method based on passenger flow classification, which is characterized by comprising:
[0005] According to the subway ground operation and maintenance system data, the data set is obtained;
[0006] Training the data set to obtain a passenger flow classification system, and outputting a passenger flow classification result;
[0007] According to the passenger flow classification results, a subway adaptive scheduling model is established using a risk assessment method;
[0008] Obtaining the departure interval time according to the real-time subway ground operation and maintenance system data, the passenger flow classification system and the subway adaptive scheduling model;
[0009] Subway vehicles are dispatched according to the departure interval.
[0010] In some embodiments, the risk assessment method is a Bayesian risk assessment method.
[0011] In some implementations, the subway ground operation and maintenance system data includes:
[0012] The load L of the subway train and the platform section i where the subway train is located.
[0013] In some implementations, processing subway ground operation and maintenance system data includes:
[0014] The load L is normalized: Among them, w0 and w1 are constants to eliminate acquisition errors, L max is the maximum load of the subway train.
[0015] In some implementations, the w0 and the w1 take values in [0.01, 0.99].
[0016] In some embodiments, the dataset comprises:
[0017] X=[L norm-1 L norm-2 … L norm-n t], where X represents the feature set of the state of the subway train, n represents n platform intervals, and t represents the departure interval.
[0018] In some embodiments, said training said data set comprises:
[0019] The Kmeans++ algorithm is used to train the data set to obtain m cluster centers.
[0020] In some implementations, the using the passenger flow classification system to classify passenger flow includes:
[0021] Obtain the passenger flow characteristics of the past T period;
[0022] Comparing the distance between the passenger flow feature and the cluster center to obtain a comparison result;
[0023] Passenger flow classification is performed according to the comparison result.
[0024] In some implementations, the selection of the T time period is set according to the fluctuation intensity of the line passenger flow.
[0025] On the other hand, the present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method according to any one of claims 1 to 9 when executing the program.
[0026] From the above description, it can be seen that the subway adaptive scheduling method based on passenger flow classification provided by the present invention adopts Kmeans++ to formulate a line passenger flow classification system according to the subway ground operation and maintenance data, and improves the utilization rate of line transport capacity through machine learning methods, reduces transport capacity waste, improves the sensitivity of railway scheduling according to passenger flow, and reduces the probability of subway passenger flow congestion. BRIEF DESCRIPTION OF THE DRAWINGS
[0027] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0028] Figure 1 This is a flow chart of a subway adaptive scheduling method based on passenger flow classification according to an embodiment of the present invention;
[0029] Figure 2 The present invention is a logical framework diagram of a subway adaptive scheduling method based on passenger flow classification according to an embodiment of the present invention. DETAILED DESCRIPTION
[0030] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with specific embodiments and with reference to the accompanying drawings.
[0031] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the embodiments of the present invention should be understood by people with ordinary skills in the field to which the present disclosure belongs. The "first", "second" and similar words used in the present disclosure do not represent any order, quantity or importance, but are only used to distinguish different components. "Including" or "comprising" and similar words mean that the elements or objects appearing in front of the word cover the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.
[0032] It should be noted that the method of the embodiment of the present invention can be performed by a single device, such as a computer or a server. The method of this embodiment can also be applied in a distributed scenario and completed by multiple devices cooperating with each other. In the case of such a distributed scenario, one of the multiple devices can only perform one or more steps in the method of the embodiment of the present invention, and the multiple devices will interact with each other to complete the described method.
[0033] As one of the main means of transportation in major cities around the world, subways bear important responsibilities for public transportation in large and medium-sized cities. More than 200 cities around the world have opened subways. In China, there are currently 38 large and medium-sized cities with subways. Among them, Beijing, Shanghai, Guangzhou, and Shenzhen have opened 22, 17, 13, and 8 subway lines respectively. The average daily passenger flow of subway lines in Beijing, Shanghai, and Guangzhou has exceeded 10 million. With the construction of lines and the increase in the number of subway users, the corresponding subway dispatching pressure is also increasing. The current subway dispatch is based on time periods, and trains are dispatched at fixed departure intervals for different time periods. It has low sensitivity to the real-time passenger flow of the line, resulting in passenger congestion during peak passenger flow and waste of capacity during low passenger flow. A reasonable dispatching plan can reduce the waste of subway passenger flow and capacity and reduce the energy consumption of subway operation.
[0034] Since the passenger flow of the subway is divided into peak and off-peak hours according to the time of the day, and factors such as holidays will also cause differences in passenger flow distribution, the scheduling of the subway plays an important role in the diversion of passenger flow. The most reasonable scheduling plan should be based on real-time passenger flow, and the vehicle departure interval should be adjusted according to the passenger flow within a certain departure frequency range. In order to evaluate the passenger flow of the line and to make adaptive adjustments to the vehicle departure interval, the subway adaptive scheduling method and electronic equipment based on predicted passenger flow proposed in the present invention first train the historical data in the subway ground operation and maintenance system, establish a passenger flow classification system based on the Kmeans++ algorithm, and use the Bayes Bayesian risk assessment method to establish a relationship model between different passenger flow classification results and departure intervals, and combine the configuration of line vehicle scheduling to formulate the real-time optimal departure interval through the optimization model. The present invention effectively controls the departure of subway vehicles, on the one hand, improves the resilience of passenger flow peak moments, and on the other hand, reduces the waste of transportation capacity during passenger flow troughs, taking into account the requirements of both transportation capacity and energy saving, and at the same time can guide the vehicle scheduling of the subway line intelligently, and improve the degree of intelligence of scheduling.
[0035] Combine the following Figure 1 A flow chart of a subway adaptive scheduling method based on passenger flow classification according to an embodiment of the present invention; Figure 2 The present invention is further described in the following logical framework diagram of a subway adaptive scheduling method based on passenger flow classification according to an embodiment of the present invention. The present invention provides a subway adaptive scheduling method based on passenger flow classification, which is characterized by comprising:
[0036] S1: Get the data set based on the subway ground operation and maintenance system data;
[0037] The passenger flow analysis and calculation unit in the subway ground operation and maintenance system can realize accurate correspondence and comprehensive calculation of station information, train entry information, and the number of passengers getting on and off each door at each station, and perform real-time calculations on the passengers getting on and off each door of each train, and obtain the total passenger flow carried by the train when it leaves the station, that is, the load L and the platform section i where the subway train is located, and store the different types of data obtained in the database as basic indicators for real-time estimation of different types. Considering the delay and incompleteness of the data upload process, the quality analysis of the various data obtained can be performed, for example, filtering out incomplete information data and obviously abnormal data.
[0038] In a multi-index evaluation system, due to the different properties of each evaluation index, they usually have different dimensions and orders of magnitude. When the levels of each index differ greatly, if the original index value is used directly for analysis, the role of the index with a higher value in the comprehensive analysis will be highlighted, and the role of the index with a lower value will be relatively weakened. Therefore, in order to ensure the reliability of the results, the original index data needs to be standardized. Before data analysis, it is usually necessary to normalize the data first, and use the standardized data for data analysis. Data normalization is the indexation of statistical data. Data normalization mainly includes two aspects: data homogeneity processing and dimensionless processing. Data homogeneity processing mainly solves the problem of data of different properties. Directly summing up indicators of different properties cannot correctly reflect the comprehensive results of different forces. It is necessary to first consider changing the properties of the inverse index data so that the forces of all indicators on the evaluation scheme are homogeneous, and then summing them up to get the correct results. Data dimensionless processing mainly solves the comparability of data. After the above standardization processing, the original data are converted into dimensionless index evaluation values, that is, the values of each indicator are at the same order of magnitude, and comprehensive evaluation and analysis can be carried out. Normalization can make the features of different dimensions comparable in value, which can greatly improve the accuracy of the classifier. On the one hand, normalization can increase the convergence speed of the model; on the other hand, when it comes to distance calculation algorithms, a small range of values will cause a loss of accuracy, and normalization can significantly improve accuracy.
[0039] Therefore, it is known that the maximum load that a bicycle can bear is L max , the load L of the subway train is normalized as follows:
[0040]
[0041] Among them, w0 and w1 are constants with a value range of [0.01, 0.99]. w0 tends to be 0.01, and w1 tends to be 0.99. The specific values can be optimized according to the requirements of the normalization effect, mainly to eliminate errors in the load collection process.
[0042] After the above processing, the feature set of the state of the subway train passing through the platform within the T minute period is defined as X = [L norm-1 L norm-2 … L norm-n t], where the subscripts from 1 to n represent the n platform intervals on the line, and t represents the departure interval of subway trains passing through the platform within the T-minute time period.
[0043] S2: training the data set to obtain a passenger flow classification system, and outputting a passenger flow classification result;
[0044] Clustering algorithm is an algorithm in machine learning that involves grouping data. In a given dataset, it can be divided into different groups through a clustering algorithm. In theory, data within the same group have the same attributes or features, while the attributes or features of data in different groups vary significantly. The clustering algorithm is an unsupervised learning algorithm and is widely used as a common data analysis algorithm in many fields. In the field of data science, by using clustering analysis to group data, data information can be obtained more clearly, and it has advantages such as fast speed and simple calculation.
[0045] As one of the classic clustering algorithms, the K-means clustering algorithm is widely used in industrial-level data processing. Its principle is easy to understand and the implementation code is simple. The original K-means algorithm initially randomly selects K points in the dataset as the clustering centers, while K-means++ selects K clustering centers according to the following idea: Suppose n initial clustering centers have been selected (0 < n < K), then when selecting the (n + 1)-th clustering center: The points that are farther away from the current n clustering centers have a higher probability of being selected as the (n + 1)-th clustering center. When selecting the first clustering center (n = 1), it is also done randomly.
[0046] The selection of the positions of the k initialized centroids has a great impact on the final clustering result and running time. Therefore, it is necessary to select appropriate k centroids. If only randomly selected, it may lead to slow convergence of the algorithm. The K-Means++ algorithm is an optimization of the method of randomly initializing the centroids in K-Means.
[0047] Therefore, the present invention uses the Kmeans++ algorithm to train the feature set X = [L norm-1 L norm-2 …L norm-n t] of the subway train state, generating m clustering centers, denoted as O1, O2, …, O m , and each clustering center corresponds to a passenger flow level.
[0048] S3: According to the passenger flow classification result, use a risk assessment method to establish a subway adaptive scheduling model;
[0049] Bayesian estimation is the use of Bayes' theorem to combine new evidence with previous prior probabilities to obtain new probabilities. It provides a method for calculating the probability of a hypothesis based on the prior probability of the hypothesis, the probability of observing different data given the hypothesis, and the observed data themselves. Bayesian estimation derives the posterior probability (the conditional probability of an event after considering relevant evidence or data) as the result of two priors, prior probabilities (the probability of an event before considering relevant evidence or data) and the likelihood function (derived from the probability model). Bayesian inference calculates the posterior probability based on Bayes' theorem. Bayesian estimation is widely used in artificial intelligence and expert systems. Since the late 1950s, Bayesian estimation techniques have been the basis of computer pattern recognition technology. There are now more and more applications that combine Bayesian estimation with simulation-based Monte Carlo methods, because some complex models cannot be solved analytically using Bayesian analysis, because the graph pattern structure can be combined with some fast simulation methods (such as Gibbs sampling or other Metropolis–Hastings algorithms).
[0050] Therefore, the present invention adopts the Bayeux tree risk assessment method to evaluate the subway train status feature set X=[L norm-1 L norm-2 … L norm-n t], taking t as the target variable, a Bayesian evaluation model based on the relationship between cluster centers and departure intervals is established, denoted as BM;
[0051] S4: obtaining the departure interval according to the real-time subway ground operation and maintenance system data, the passenger flow classification system and the subway adaptive scheduling model;
[0052] According to the real-time data of the subway ground operation and maintenance system, the characteristics of the line passenger flow in the T-minute period are obtained. The selection of T can be set according to the fluctuation intensity of the line passenger flow. T =[L norm-1 ,L norm-2 ,…,L norm-n ]; compare the distance between X and the cluster center in S2 to determine the passenger flow level of the line; import the passenger flow level into BM to obtain the corresponding departure interval T N .
[0053] S5: Dispatching subway vehicles according to the departure interval.
[0054] According to the departure interval T obtained by S4 N , to dispatch subway trains.
[0055] As a more specific embodiment of the present invention, the present invention is mainly aimed at the problem of passenger flow control in urban rail transit, and combines driving monitoring, passenger flow simulation, passenger flow classification, time planning, and vehicle scheduling control, and combines the limited rationality of passengers to propose a subway adaptive scheduling method based on passenger flow classification. Taking any subway line in Beijing as an example, some historical data of the morning peak is captured from the subway ground operation and maintenance system data, and the maximum and minimum values corresponding to the morning peak are selected; the interval between the station corresponding to the maximum value of the morning peak and the station corresponding to the minimum value of the morning peak is used as the morning peak downward congestion interval. Similarly, some historical data of the evening peak is captured from the subway ground operation and maintenance system data, and the maximum and minimum values corresponding to the evening peak are selected to obtain the total passenger flow carried by the train when it leaves the station in any time period during these two peak periods, that is, the load L and the platform section i where the subway train is located, and the load of the subway train is normalized. It is known that a single vehicle can withstand a maximum load L max :
[0056]
[0057] In order to eliminate the error in the load collection process, w0 and w1 are constants with a value range of [0.01, 0.99]. w0 tends to 0.01 and w1 tends to 0.99. The specific values can be optimized according to the requirements of the normalization effect.
[0058] After the above processing, the feature set of the state of the subway train passing through the platform within the T minute period is defined as X = [L norm-1 L orm-2 … L norm-n t], where the subscripts from 1 to n represent the n platform intervals on the line, and t represents the departure interval of subway trains passing through the platform within the T-minute time period.
[0059] Use Kmeans++ algorithm to analyze the feature set X=[L norm-1 L orm-2 … L norm-n t] to train and generate m cluster centers, which are denoted as O1, O2, ..., O m , each cluster center corresponds to a passenger flow level.
[0060] The Bayeux tree risk assessment method is used to evaluate the subway train status feature set X = [L norm-1 L orm-2 … L norm-n t], taking t as the target variable, a Bayesian evaluation model based on the relationship between cluster centers and departure intervals is established, denoted as BM;
[0061] According to the real-time data of the subway ground operation and maintenance system, the characteristics of the line passenger flow in the T-minute period are obtained. The selection of T can be set according to the fluctuation intensity of the line passenger flow. T =[L norm-1 ,L norm-2 ,…,L norm-n ]; X and cluster centers O1, O2, ..., O m Compare the distances to determine the passenger flow level of the route; import the passenger flow level into the Bayesian evaluation model to obtain the corresponding departure interval time T N According to the obtained departure interval time T N , combined with the relevant regulations of Beijing Urban Rail Transit Operation Management, this paper proposes a plan for the division of the operating time periods of the target line on weekdays. Because the first departure time of the target line in the upward direction is 06:00, the last departure time is 23:00, and the whole operation time is as long as 58 minutes, the actual operation time of the line is 06:00-23:58, so the actual operation time division of the target line includes the period of 6:00-24:00. In this operation time division plan, the whole day of the operation day is divided into 8 operation time periods with different time intervals. The first category has a time interval of 40 minutes, the second and third categories are both 60 minutes, the fourth category is 400 minutes, the fifth and sixth categories are both 160 minutes, the seventh category is 120 minutes, and the eighth category is 80 minutes. It can be seen that the passenger flow distribution law in the operation period of 8:40-15:20 lasts for a long time, and has the characteristics of relatively dense travel distribution. The passenger flow distribution transfer law in each operation period is relatively uniform, while there are large differences between different operation periods. Based on this, the rail transit operation management department can select reasonable departure intervals within each operating period according to the actual passenger flow in each operating period, construct a full-day departure plan with equal intervals in different time periods for the target line, and formulate corresponding train operation plans. It should be noted that the effective departure time period is 06:00-23:00 during the operating period, which is about one hour less than the actual operating time. Therefore, the effective departure time of the last operating period divided by this plan is only 22:40-23:00, which does not completely overlap with the operating period divided by the plan.
[0062] On the other hand, the present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements a method as described in any one of embodiments 1 to 9 above when executing the program.
[0063] The apparatus of the above-mentioned embodiment is used to implement the corresponding method in the above-mentioned embodiment, and has the beneficial effects of the corresponding method embodiment, which will not be described in detail here.
[0064] Those skilled in the art should understand that the discussion of any of the above embodiments is merely illustrative and is not intended to imply that the scope of the present disclosure (including the claims) is limited to these examples. Within the scope of the present invention, the above embodiments or technical features in different embodiments may be combined, the steps may be implemented in any order, and there are many other variations of the different aspects of the present invention as described above, which are not provided in detail for the sake of simplicity.
[0065] In addition, to simplify the description and discussion, and in order not to obscure the present invention, known power / ground connections to integrated circuit (IC) chips and other components may or may not be shown in the provided figures. In addition, the devices may be shown in the form of block diagrams to avoid obscuring the present invention, and this also takes into account the fact that the details of the implementation of these block diagram devices are highly dependent on the platform on which the present invention will be implemented (i.e., these details should be fully within the scope of understanding of those skilled in the art). Where specific details (e.g., circuits) are set forth to describe exemplary embodiments of the present invention, it will be apparent to those skilled in the art that the present invention may be implemented without these specific details or with variations in these specific details. Therefore, these descriptions should be considered illustrative rather than restrictive.
[0066] Although the invention has been described in conjunction with specific embodiments of the invention, many replacements, modifications and variations of these embodiments will be apparent to those skilled in the art from the foregoing description. For example, other memory architectures (e.g., dynamic RAM (DRAM)) may use the embodiments discussed.
[0067] The embodiments of the present invention are intended to cover all such substitutions, modifications and variations that fall within the broad scope of the appended claims. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.
Claims
1. A subway adaptive scheduling method based on passenger flow classification, characterized in that: include: Preprocessing the subway ground operation and maintenance system data, and obtaining a data set according to the preprocessed subway ground operation and maintenance system data; wherein the subway ground operation and maintenance system data includes: a load L of a subway train and a platform section i where the subway train is located, the load L of the subway train is the passenger flow carried by the subway train when it leaves the station, and the preprocessing of the subway ground operation and maintenance system data is to normalize the load L of the subway train; The load L is normalized by the following formula: Wherein, w0 and w1 are constants for eliminating acquisition errors, Lmax is the maximum load of the subway train, and w0 and w1 are values in [0.01, 0.99]; The data set is trained to obtain a passenger flow classification system, and the passenger flow classification result is output; wherein the data set is trained by a K-means++ algorithm to generate m cluster centers, each cluster center corresponding to a passenger flow level; According to the data set, a Bayesian risk assessment method is used to establish a Bayesian assessment model; According to the real-time subway ground operation and maintenance system data, the passenger flow classification system and the Bayesian evaluation model, the departure interval is obtained; wherein, it includes: according to the real-time subway ground operation and maintenance system data, the characteristic X of the line passenger flow within the T minute period is obtained. T , X T =[L norm-1 , L norm-2 , …L norm-n ], T is determined according to the fluctuation intensity of the line passenger flow, L norm is the normalized load L, n represents n station intervals; The feature X T Compare the distance with the cluster center to determine the passenger flow level of the line; import the passenger flow level into the pre-built Bayesian evaluation model to determine the departure interval time T N ; Subway vehicles are dispatched according to the departure interval.
2. The subway adaptive scheduling method based on passenger flow classification according to claim 1 is characterized in that: The dataset includes: X=[L norm-1 , L norm-2 …L norm-n t], where X represents the feature set of the state of the subway train, n represents n platform intervals, and t represents the departure interval.
3. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the method according to any one of claims 1 to 2 is implemented.
Citation Information
Patent Citations
Self-adaption dispatching method used for public traffic system
CN103198648A
Urban railway transit timetable optimizing technology application framework
CN104200286A
Bus dispatch optimization method for predicting passenger flow based on ARIMA model
CN104573849A
Subway station passenger flow micro statistics and scheduling method
CN105488751A
Rail transfer station passenger flow jam risk assessment method based on AFC (automatic fare collection) card swiping data
CN106779429A
