A method for calculating retention volume and average retention time

Abstract

The present invention relates to a method for calculating the retention quantity and the average retention duration. The method comprises the following steps: historical data modeling: processing the historical data in the database to form a mathematical model for calculation; model storage: storing the prediction data obtained from the mathematical model; calculating the real-time retention quantity; calculating the real-time retention quantity according to the mathematical model and retrieving the prediction data, and performing noise reduction processing on the real-time flow rate through a regression algorithm; calculating the real-time retention duration: obtaining the real-time retention duration according to the real-time retention quantity through a calculus formula. By using the method of the present invention, the real-time retention quantity and the average retention duration can be calculated accurately under the condition of only unidirectional accurate people flow. Through historical data modeling, the data is preprocessed by means of function fitting during the modeling process, and after preprocessing, the data is stored in the mathematical model. Relevant preprocessed data is retrieved from the mathematical model, and the retention quantity and the average retention duration are calculated accurately and reasonably.

Description

Technical Field

[0001] The present invention relates to the field of video analysis and monitoring, and particularly to a method for calculating the residence quantity and the average residence duration. Background Art

[0002] At present, with the rapid development of social economy and science and technology, industries such as large shopping mall complexes, catering, transportation, and medical care have risen one after another, and the flow of people in various public places and infrastructure has become more frequent. How to reasonably and accurately analyze the number of people has become a major problem to be solved urgently. In this context, as one of the key analysis indicators of the number of people, how to accurately and real-time count the residence quantity and residence duration is very important.

[0003] In traditional methods, the real-time residence quantity and residence duration are both obtained by calculating the real-time in-and-out passenger flow. The real-time in-and-out passenger flow is generally obtained by installing cameras at the entrances and exits of various public places, and then counting the number of people through the AI intelligent algorithm chip inside the camera for video image analysis. However, there are still problems with the above single-camera video image analysis algorithm as follows:

[0004] There are multiple entrances and exits in some public places or infrastructure, so accurate statistics of two-way passenger flow cannot be achieved, while the accuracy rate of one-way passenger flow is within an acceptable range. Therefore, on the premise of only ensuring the accuracy of one-way passenger flow, it is impossible to calculate and analyze the real-time residence quantity and residence duration using traditional calculation methods. Summary of the Invention

[0005] Aiming at the above-mentioned shortcomings of the prior art, the purpose of the present invention is to provide a method for calculating the residence quantity and the average residence duration to solve one or more problems in the prior art.

[0006] To achieve the above purpose, the technical solution of the present invention is as follows:

[0007] A method for calculating the residence quantity and the average residence duration includes the following steps:

[0008] Historical data modeling: processing the historical data in the database to form a mathematical model for calculation;

[0009] Model storage: storing the prediction data obtained from the mathematical model;

[0010] Calculating the real-time residence quantity; calculating the real-time residence quantity according to the mathematical model and retrieving the prediction data, and performing noise reduction processing on the real-time flow quantity through a regression algorithm;

[0011] Calculating the average residence duration: obtaining the average residence duration according to the real-time residence quantity through a calculus formula.

[0012] Furthermore, the historical data modeling includes the following steps:

[0013] Visual analysis;

[0014] Statistical operations;

[0015] Function fitting.

[0016] Furthermore, the visual analysis includes the following steps:

[0017] Obtain basic data: Obtain the basic data of the monitored area through a video input device in combination with a human recognition algorithm; the basic data includes area number data, time data, inbound passenger flow data, and outbound passenger flow data;

[0018] Draw a relationship chart based on the basic data: The relationship chart includes a two-dimensional scatter plot of the daily retention volume, a three-dimensional scatter plot of time, retention volume, and inbound passenger flow, a two-dimensional scatter plot of inbound passenger flow and retention volume, and a two-dimensional scatter plot of inbound passenger flow and retention volume at a specified time;

[0019] Obtain the first functional relationship between the retention volume s at unit time k, the total inbound passenger flow at unit time k, and the total outbound passenger flow at unit time k at the same time every day according to the relationship chart. Furthermore, the statistical operations are used to preprocess the data with a positive correlation between the retention volume and the inbound passenger flow, and it includes the following steps: k Equal proportion scaling, obtain the first retention volume s at unit time k according to the first functional relationship;

[0020] Supplementation, if the first retention volume s at unit time k 1k ;

[0021] is not equal to zero, then return to perform equal proportion scaling, and obtain the second retention volume s at unit time K according to the second functional relationship; 1k ; Then judge whether the second retention volume s at unit time K 2K is equal to zero and satisfies the logical constraint. 2K

[0022] Furthermore, the conditions of the logical constraint include that the retention volume S at unit time k k is greater than or equal to 0, the retention volume S at unit time k k is less than or equal to the total inbound passenger flow at unit time k, the total inbound passenger flow at unit time k is greater than or equal to the total outbound passenger flow at unit time k, and the first old total outbound passenger flow before unit k is greater than or equal to the first old total outbound passenger flow before unit k - 1.

[0023] Furthermore, the function fitting is used to make the second retention volume s at unit time K 2K ​The weights are calculated according to the third functional formula, and the weights are clustered to obtain unique weights.

[0024] Furthermore, the calculation of the real-time retention volume includes the following steps:

[0025] Model usage: The mathematical model is used to obtain the retention volume S at time k k ;

[0026] Noise reduction fitting: The regression algorithm is used to perform noise reduction processing on the obtained retention volume S at time k k Noise reduction processing.

[0027] Furthermore, the first functional formula is:

[0028]

[0029] where s 1k represents the first retention volume at unit time k, represents the first total incoming passenger flow at unit time k, represents the first total outgoing passenger flow at unit time k.

[0030] Furthermore, the second functional formula is:

[0031]

[0032] where the second retention volume at unit time K is s 2K , represents the new total outgoing passenger flow before time K, represents the new total outgoing passenger flow before time K-1.

[0033] Furthermore, the third functional formula is:

[0034]

[0035] where the weight is represented as w k , s 2K represents the second retention volume at unit time K, and a k represents the incoming passenger flow at unit time k.

[0036] Furthermore, the calculus formula is:

[0037]

[0038] where f(k) is the function obtained by fitting the retention volume in this area on this day, and the total incoming passenger flow at unit time k is

[0039] Compared with the prior art, the beneficial technical effects of the present invention are as follows

[0040] Using the method of the present invention, the real-time retention volume and the average retention duration can be calculated accurately when there is only one-way passenger flow. It models through historical data and preprocesses the data by function fitting during the modeling process. By means of function fitting, the error between the data to be obtained and the actual data can be reduced. After preprocessing, the data is stored in the mathematical model, and then relevant preprocessed data is retrieved from the mathematical model later to achieve fast, accurate and reasonable calculation of the retention volume and the average retention duration. Brief Description of the Drawings

[0041] Figure 1 Fig. shows a schematic flowchart of a method for calculating the retention volume and the average retention duration according to an embodiment of the present invention.

[0042] Figure 2 Fig. shows a two-dimensional scatter plot of the daily retention volume in a method for calculating the retention volume and the average retention duration according to an embodiment of the present invention.

[0043] Figure 3 Fig. shows a three-dimensional scatter plot of the analysis of the three relationships of time, retention volume and incoming passenger flow in a method for calculating the retention volume and the average retention duration according to an embodiment of the present invention.

[0044] Figure 4 Fig. shows a two-dimensional scatter plot of the analysis of the two relationships between the incoming passenger flow and the retention volume in a method for calculating the retention volume and the average retention duration according to an embodiment of the present invention.

[0045] Figure 5 Fig. shows a two-dimensional scatter plot of the analysis of the relationship between the incoming passenger flow and the retention volume at a specified moment in a method for calculating the retention volume and the average retention duration according to an embodiment of the present invention.

[0046] Figure 6 Fig. shows a line graph comparing the original prediction data with the data after noise reduction and fitting in a method for calculating the retention volume and the average retention duration according to an embodiment of the present invention.

[0047] Figure 7 Fig. shows a line graph of the area of the integral calculus value for calculating the real-time retention duration in a method for calculating the retention volume and the average retention duration according to an embodiment of the present invention. Detailed Embodiments

[0048] In order to make the objectives, technical solutions and advantages of the present invention clearer and more understandable, the following further elaborates in detail on a method for calculating the retention amount and average retention duration proposed by the present invention in combination with the accompanying drawings and specific embodiments. According to the following description, the advantages and features of the present invention will be clearer. It should be noted that the accompanying drawings are in a very simplified form and all use non-precise scales, only for conveniently and clearly assisting in explaining the objectives of the embodiments of the present invention. In order to make the objectives, features and advantages of the present invention more obvious and understandable, please refer to the accompanying drawings. It should be noted that the structures, scales, sizes, etc. shown in the drawings of this specification are only used to cooperate with the content disclosed in the specification for those familiar with this technology to understand and read, and are not used to limit the limiting conditions for the implementation of the present invention. Therefore, they do not have technical substantive significance. Any modification of the structure, change of the proportional relationship or adjustment of the size, without affecting the effects that the present invention can produce and the objectives that can be achieved, should still fall within the scope covered by the technical content disclosed by the present invention.

[0049] A method for calculating the retention amount and average retention duration includes the following steps:

[0050] S1: Historical data modeling, that is, processing the historical data in the database to form a mathematical model for calculation. The historical data modeling is used to analyze the relationships between the required data to facilitate the use of the formed mathematical model.

[0051] Specifically, the historical data modeling includes the following steps:

[0052] S1.1: Visual analysis.

[0053] S1.2: Statistical operations.

[0054] S1.3: Function fitting.

[0055] Specifically, the visual analysis includes the following specific steps:

[0056] S1.1.1: Obtain basic data, and obtain the basic data of the monitoring area through a video input device in combination with a human recognition algorithm. The above human recognition algorithm is a well-known algorithm in the art. The video input device is preferably a camera. The basic data includes any one or more of area number data, time data, incoming passenger flow data, and outgoing passenger flow data. The area number represents the area corresponding to this record, and the time represents the time corresponding to this record.

[0057] S1.1.2: Draw a relationship chart based on the above basic data:

[0058] Specifically, the relationship chart includes a two-dimensional scatter plot of the daily retention volume, a three-dimensional scatter plot of time, retention volume, and incoming passenger flow, a two-dimensional scatter plot of incoming passenger flow and retention volume, and a two-dimensional scatter plot of incoming passenger flow and retention volume at a specified time.

[0059] S1.1.3: Obtain the data relationship according to the relationship chart. Specifically, please refer to Figure 2 , in the two-dimensional scatter plot of the daily retention volume, the abscissa represents the time point of the day, and the ordinate represents the daily retention volume.

[0060] Please refer to Figure 3 , by plotting the three-dimensional scatter plot of time, retention volume, and incoming passenger flow, it can be seen that the change of incoming passenger flow within a day has certain rules, and the specific rules are the four constraint conditions in the following logical constraints.

[0061] Please refer to Figure 4 , by plotting the two-dimensional scatter plot of incoming passenger flow and retention volume, it can be analyzed that there is a positive correlation between the retention volume and the incoming passenger flow at the same time every day. Specifically, according to the above-mentioned drawings, it can be obtained that the retention volume on weekends is greater than that on weekdays. Therefore, the change trend within a day can be divided into weekdays and weekends, so modeling is carried out separately according to the week.

[0062] Furthermore, combined with the positive correlation between the retention volume and the incoming passenger flow obtained from the visual analysis, statistical operations are carried out. The purpose of this statistical operation is to preprocess the correlation in the positive correlation between the retention volume and the incoming passenger flow obtained below.

[0063] Before starting the statistical operation, first record the incoming passenger flow at the unit time k (0 ≤ k ≤ n, where n is the time corresponding to the end of a day) as a k , the outgoing passenger flow as b k , and the retention volume at the unit time k as s k . Therefore, the total incoming passenger flow at the unit time k can be expressed by the following functional formula:

[0064]

[0065] The total outgoing passenger flow at the unit time k can be expressed by the following functional formula:

[0066]

[0067] From this, it can be obtained that the retention volume S at the unit time k k is equal to the first functional formula of the total incoming passenger flow at the unit time k minus the total outgoing passenger flow at the unit time k:

[0068]

[0069] According to the above first functional equation, when the one-way passenger flow is accurate, the total incoming passenger flow and the total outgoing passenger flow at the end of the day are not equal. Therefore, the following processing is required:

[0070] B1: Scale proportionally to obtain the first outgoing passenger flow b at unit time k 1 k. In this embodiment, since the accuracy of the incoming passenger flow is ensured, scaling the outgoing passenger flow proportionally is selected. Of course, in other embodiments of the present invention, scaling the incoming passenger flow proportionally can also be selected when ensuring the accuracy of the outgoing passenger flow.

[0071] Specifically, the functional relationship for proportional scaling is as follows:

[0072]

[0073] Where in relation (1) represents the first new total outgoing passenger flow before unit time k,

[0074] represents the first old total outgoing passenger flow before unit time k, represents the first total incoming passenger flow at the end of the day, represents the first total outgoing passenger flow at the end of the day.

[0075] In relation (2), the first outgoing passenger flow at unit time k is b 1 k, represents the first new total outgoing passenger flow before unit time k, represents the first new total outgoing passenger flow before unit time k - 1.

[0076] Therefore, the first retention amount s at unit time k 1k can be obtained according to the following second functional equation:

[0077]

[0078] Where in relation (3), s 1k represents the first retention amount at unit time k, represents the first total incoming passenger flow at unit time k, represents the first total outgoing passenger flow at unit time k.

[0079] B2: Supplement the data relationship. If the first total incoming passenger flow at unit time k minus the first total outgoing passenger flow at unit time k, that is, the first retention amount s 1k is not equal to zero, return to step B1 to continue proportional scaling. If the first total incoming passenger flow at unit time k minus the first total outgoing passenger flow at unit time k, that is, the first retention amount s 1kIf it is equal to zero, the function fitting step is executed. Here, taking the first retention amount s 1k not equal to zero, specifically, for s 1k <0 data is supplemented, and the specific functional relationship is as follows:

[0080]

[0081] In this relationship (4), K represents the moment when the retention is less than 0, and the first retention amount at the unit moment K is s 1K , represents the new total outgoing passenger flow before the moment K represents the old total outgoing passenger flow before the moment K.

[0082] At this time, the second retention amount at the unit moment K is s 2K can be obtained through the following third functional formula:

[0083]

[0084] Among them, the second retention amount at the unit moment K is s 2K , represents the new total outgoing passenger flow before the moment K, represents the new total outgoing passenger flow before the moment K - 1.

[0085] Furthermore, in the above relationship (5), if the new total outgoing passenger flow before the moment K minus the new total outgoing passenger flow before the moment K - 1 is not equal to zero, then it continues to return to perform equal - ratio scaling, which can be expressed by the following relationship:

[0086]

[0087] In relationship (6) represents the second new total outgoing passenger flow before the unit k, represents the second old total outgoing passenger flow before the unit k, represents the second total incoming passenger flow at the end of the day, represents the second total outgoing passenger flow at the end of the day.

[0088] In relationship (7), the second outgoing passenger flow at the unit moment k is b 2 k, represents the second new total outgoing passenger flow before the unit k moment, represents the second new total outgoing passenger flow before the unit k - 1 moment.

[0089] Furthermore, the second retention amount at the unit moment K is s 2KIf it is equal to zero and each parameter relationship in the augmentation and equal proportion scaling satisfies the logical constraints, the function fitting step can be entered. The logical constraints include the following function relations:

[0090] (1) The retention amount S at unit time k k is greater than or equal to 0, and its function relation is expressed as follows:

[0091] s k ≥0

[0092] (2) The retention amount S at unit time k k is less than or equal to the total incoming passenger flow at unit time k and its function relation is expressed as follows:

[0093]

[0094] (3) The total incoming passenger flow at unit time k is greater than or equal to the total outgoing passenger flow at unit time k and its function relation is expressed as follows:

[0095]

[0096] (4) The first old total outgoing passenger flow before unit k is greater than or equal to the first old total outgoing passenger flow before unit k - 1 and its function relation is expressed as follows:

[0097]

[0098] The specific process of function fitting is described as follows:

[0099] The function fitting refers to obtaining the weight value according to the fourth function formula for the second retention amount s at unit time K, that is, substituting the second retention amount into the relation formula with unknown parameters between the retention amount and the incoming passenger flow at the same time every day, so as to obtain the function relation between the retention amount and the incoming passenger flow. 2K At present, common fitting methods include the least squares curve fitting method, polynomial fitting method, custom function fitting method, Gaussian curve fitting method, etc. The present invention preferably adopts the least squares curve fitting method. The solving process of the least squares curve fitting method is the same as the existing well-known method, and the present invention will not elaborate in detail. Please refer to

[0100] , by plotting the two-dimensional scatter plot of the relationship analysis between the incoming passenger flow and the retention amount at the specified time, it can be found that the fitting function approaches a linear function of the first order by using the least squares curve fitting method. Therefore, the second retention amount s at unit time K Figure 5 and the incoming passenger flow a at unit time k 2K ​k The ratio, i.e., the slope, is recorded as the weight w k , which can be represented by the following fourth functional expression:

[0101]

[0102] After obtaining the above weights, cluster the weights at the same moment k for the historical days of the same region and the same week. The final unique value obtained. In this embodiment, it is found through statistical calculation that the value of the median of the weights is close to the clustering center. Therefore, directly select the median of the weights as the unique weight W 1k , that is, W 1k ≈ median(w 0 , w 1 , w 2 ,…, w k ). Of course, in other embodiments of the present invention, it may also be other values other than the median of the weights. In this regard, the present invention will not be further described. The above is the completion of the entire process of historical data modeling.

[0103] S2: Model storage. The model storage step is used to store the above required data in a specified data structure to achieve structured storage of the data, and further achieve rapid addition, deletion, modification, retrieval, and operation of the data.

[0104] Furthermore, in the embodiment of the present invention, the stored data includes first data and second data. The first data is all the classified weights, and the second data includes the unique weight extracted after clustering all the weights. The storage of the above first data facilitates subsequent addition, deletion, and modification operations. In the specific use process, it mainly involves the addition of the latest data every day, making the mathematical model have growth. And the second data is used to complete the operation of the mathematical model to support subsequent real-time prediction using the mathematical model.

[0105] S3: Calculate the real-time retention amount. The steps for calculating the real-time retention amount are as follows:

[0106] S3.1: Model usage: According to the current incoming passenger flow a obtained at the unit time k k , with the help of using the above mathematical model, retrieve the weight W corresponding to the corresponding week and the corresponding moment k , to obtain the real-time retention amount S k at the current moment k.

[0107] Among them, the above calculation method can be expressed by the following functional expression:

[0108] S k = a k × W k

[0109] S3.2: Use a regression algorithm to process the obtained retention amount S at time k k for noise reduction and to satisfy logical constraints; in this embodiment, the regression algorithm uses locally weighted linear regression. The specific calculation process of this regression algorithm is a well-known technology in the art. Of course, in other embodiments of the present invention, other regression methods such as simple linear regression, multiple linear regression, and polynomial regression, except for locally weighted linear regression, can also be used. The present invention does not make further limitations in this regard.

[0110] Locally weighted scatterplot smoothing (LOWESS or LOESS) is a powerful tool for examining the relationship between two-dimensional variables. Its main idea is to take a certain proportion of local data and fit a polynomial regression curve in this subset of data, so as to observe the laws and trends shown by the data locally.

[0111] While ordinary regression analysis often models based on all data, which can describe the overall trend. By advancing the local range from left to right in sequence, a continuous curve can ultimately be calculated. The smoothness of the curve is related to the proportion of data we select: the smaller the proportion, the less smooth the fitting, that is, too much emphasis is placed on local properties, and vice versa, the smoother it is.

[0112] Further, please refer to Figure 6 , Figure 6 to visually see the retention amount a at time k as described above k W k the noise reduction comparison between the noise reduction fitting data obtained after noise reduction processing and the original prediction data (i.e., the real-time retention amount in the above S3.1 step). After noise reduction fitting, check the retention amount S after the above noise reduction fitting k to see if it satisfies the logical constraints. The functional relationship of the logical constraints is the same as that of the logical constraints in the above addition and equal ratio scaling. The present invention does not make further descriptions in this regard.

[0113] S4: Calculate the average retention duration. Specifically, please refer to Figure 7 , in the method for calculating the retention amount and average retention duration according to the embodiment of the present invention, the average retention duration t is realized through the following calculus formula:

[0114]

[0115] where f(k) is the function obtained by fitting the retention amount in this area on this day, and the total incoming passenger flow at unit time k is

[0116] where the above retention amount is the retention amount a at time k obtained after noise reduction fitting k Wk , the total incoming passenger flow at unit time k is known, while the total outgoing passenger flow at unit time k is obtained during statistical operations.

[0117] Based on the above total incoming passenger flow and total outgoing passenger flow, a coordinate system is established with the x-axis representing time and the y-axis representing the number of people as shown in Figure 7 . The area enclosed by the function where the total incoming passenger flow is located and the function where the total outgoing passenger flow is located is equivalent to the area enclosed by the function where the retention volume is located and the x-axis. Therefore, this enclosed area can also be further interpreted as the total retention duration, which is . After dividing it by the total incoming passenger flow, the average retention duration of this area is obtained (i.e., according to the above calculus formula).

[0118] The technical features of the above-described embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope described in this specification.

[0119] The above-described embodiments only represent several implementation manners of the present invention. Their descriptions are relatively specific and detailed, but they 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 invention, several modifications and improvements can still be made, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the invention patent should be subject to the appended claims.

Claims

1. A method for calculating the retention quantity and the average retention duration, characterized in that it includes the following steps: Historical data modeling: Process the historical data in the database to form a mathematical model; Model storage: Store the prediction data obtained from the mathematical model; Calculate the real-time retention quantity; Calculate the real-time retention quantity according to the mathematical model and retrieve the prediction data, Perform noise reduction processing on the real-time retention quantity through a regression algorithm; Calculate the average retention duration: Obtain the average retention duration according to the real-time retention quantity through the calculus formula; The historical data modeling includes the following steps: Obtain basic data, and obtain the basic data of the monitoring area through a video input device in combination with a human body recognition algorithm; Draw a relationship chart based on the basic data. From the relationship chart, it can be obtained that at the same time every day, the retention amount at the unit time k is S k , , the incoming passenger flow at the unit time k is a k , the outgoing passenger flow is b k , unit time, 0 ≤ k ≤ n, where n is the time corresponding to the end of a day; Perform preprocessing in combination with visual analysis, including B1 and B2: B1: Scale proportionally to obtain the first outbound passenger flow b1 at unit time k k , represents the first total incoming passenger flow at unit time k, represents the first total outgoing passenger flow at unit time k; B2: Supplement the data of S 1k Supplement the data of <0, where K represents the moment of residence less than 0, and the first residence volume at moment K is S 1k , represents the new total outgoing passenger flow before moment K, and the second residence volume at moment K is s 2k , ; If the first retention amount s 1k is not equal to zero, return to step B1 and continue to perform equal proportion scaling; If the first retention amount s at unit time k 1k is equal to zero, then perform the function fitting step; The second retention amount S at time K 2k If it is not equal to zero, continue to return for equal scaling; If it is equal to zero, and each parameter relationship in supplementation and equal proportion scaling satisfies the logical constraint, then enter the function fitting step; Function fitting: the second retention volume S at K 2k According to Calculate the weight W K ; Based on the current incoming passenger flow a obtained at unit time k k , retrieve the corresponding weight W K , and obtain the real-time retention quantity S at the current k moment k , S k =a k ×W k ; The average retention duration t is realized through the following calculus formula: where f(k) is the retention amount a at time k obtained after noise reduction fitting k W k , per unit time The total incoming passenger flow at time k is t represents the average residence time.

2. The method for calculating the retention quantity and the average retention duration according to claim 1, characterized in that: The visual analysis includes the following steps: Obtain basic data: Obtain the basic data of the monitoring area through a video input device in combination with a human body recognition algorithm; The basic data includes area number data, time data, incoming passenger flow data, and outgoing passenger flow data; Draw a relationship chart according to the basic data: The relationship chart includes a two-dimensional scatter plot of the daily retention quantity, a three-dimensional scatter plot of time, retention quantity, and incoming passenger flow, a two-dimensional scatter plot of incoming passenger flow and retention quantity, and a two-dimensional scatter plot of incoming passenger flow and retention quantity at a specified time; According to the relationship chart, the retention amount at the same time every day at unit time k is S k The functional formula between the total incoming passenger flow at unit time k and the total outgoing passenger flow at unit time k 3. The method for calculating the retention quantity and the average retention duration according to claim 1, characterized in that: The conditions of the logical constraint include the retention volume S at time k k greater than or equal to 0, the retention volume S at time k k less than or equal to the total incoming passenger flow at unit time k The total incoming passenger flow at time k is greater than or equal to the total outgoing passenger flow at time K, The total outgoing passenger flow before unit k is greater than or equal to the total outgoing passenger flow before unit k - 1

Images