A method for determining the number of bus operation status classifications

Abstract

The present invention discloses a method for determining the classification number of a bus operation state, and the steps include: S1. collecting bus travel speed data; S2. setting the value range of the bus operation state classification number; S3. obtaining the bus operation state speed threshold under each classification number; S4. calculating the variance fitting goodness under each classification number; S5. determining the optimal classification number of the bus operation state. Aiming at the problem that there is no basis for the value of the bus operation state classification number, the present invention proposes a method for determining the bus operation state classification number based on the variance fitting goodness and its increment, which can reasonably determine the optimal classification number of the bus operation state.

Description

Technical Field

[0001] The invention relates to the field of urban public transportation systems, and in particular to a method for determining a classification number of public transportation operation states. Background Art

[0002] In recent years, the investment in urban public bus and tram infrastructure (hereinafter referred to as "bus") has continued to increase, but the bus passenger volume has decreased instead of increased. Compared with other motorized transportation modes, the disadvantages of bus such as slow driving speed and long travel time have led to its competitiveness declining year by year. In order to increase the attractiveness of bus, it is crucial to improve its transportation speed. Objectively and impartially evaluating the operation status of bus and accurately identifying the weak points in the operation process of bus are the basis for improving the transportation speed of bus.

[0003] To evaluate the operation status of public transportation, it is necessary to first determine the number of operation status classifications, which is the basis for determining the operation status classification. The existing number of bus operation status classifications mainly refers to the classification method of traffic flow. The early three-parameter flow-density-speed model divides traffic flow into two phases: congested flow and non-congested flow; Kerner proposed a three-phase traffic flow theory, arguing that traffic flow has three phases: free flow, synchronous flow, and wide moving blocking flow; Liu proposed a traffic status evaluation model that combines Gaussian mixture model (GMM) and linear regression (LR). Based on Kerner's traffic flow theory, traffic flow is divided into four phases: free flow, quasi-free flow, synchronous flow, and traffic congestion; Dai Xuezhen et al. divided the traffic operation status into five levels: smooth, basically smooth, mild congestion, moderate congestion, and severe congestion according to the "Urban Traffic Operation Status Evaluation Specification" (GB / T33171-2016). The above traffic flow operation status classification method has a certain reference value for bus flow, but bus flow and traffic flow are different in operation mode, vehicle performance, road right, etc. There is controversy about the direct application of the above traffic flow classification method to bus flow. At present, there is a lack of unified standards and effective determination basis for the classification number of bus operation status. Summary of the invention

[0004] The technical problem to be solved by the present invention is to provide a method for determining the classification number of bus operation status, which can provide an effective basis for determining the optimal classification number of bus operation status and lay a foundation for evaluating the bus operation status.

[0005] The technical solution adopted by the present invention to solve the technical problem is: constructing a method for determining the number of bus operation status classifications, comprising the following steps:

[0006] S1. Collecting bus travel speed data;

[0007] S2. Setting the value range of the bus operation status classification number;

[0008] S3, obtaining the bus running speed threshold under each classification number;

[0009] S4, calculate the variance goodness of fit under each classification number;

[0010] S5. Determine the optimal classification number of bus operation status.

[0011] According to the above scheme, the collected bus travel speed data is the ratio of the bus travel distance to the travel time within a fixed time interval, and the fixed time interval is set to 5 minutes or 10 minutes.

[0012] According to the above scheme, in step S2, referring to relevant national / industry standards and existing research on traffic flow and bus operation status classification numbers, the bus operation status classification number k ranges from 2 to 7.

[0013] According to the above solution, in step S3, the bus running state speed threshold is a critical speed value between different bus running states.

[0014] According to the above scheme, in step S3, the method for obtaining the speed threshold of the bus running state is as follows:

[0015] S301, sorting the bus travel speed data in ascending order to form a bus travel speed sequence statistic;

[0016] S302, using the natural break point method to divide the bus travel speed order statistics into k categories, k = 2, 3, ..., 7;

[0017] S303: Output the critical speed values ​​between the bus running states under each classification number, which are defined as v1, v2, ..., v in ascending order. k-1 .

[0018] According to the above scheme, the critical speed values ​​v1, v2, ..., v k-1 Divide the bus travel speed dataset into k subsets: [v min ,v1),[v1,v2),…,[v k-1 ,v max ],v min is the minimum speed value in the travel speed data set, v max is the maximum speed value in the travel speed data set, and the k speed subsets correspond to k bus operation states.

[0019] According to the above scheme, in step S4, the goodness of variance fit, i.e., Goodness of Variance Fit, hereinafter referred to as GVF, is a summary measure of the classification results of the natural break point method. The value of GVF is between 0 and 1, 1 indicates an excellent fit, and 0 indicates an extremely poor fit. The calculation steps are as follows:

[0020] S401. Calculate the sum of squared deviations from the array means of the bus travel speed data set, that is, Sum of squared Deviations from the Array Means, referred to as SDAM. n is the number of all elements in the travel speed data set; v i is the value of the i-th element in the travel speed data set; is the mean of all elements in the travel speed data set;

[0021] S402, calculate the sum of the class total square differences (Sum of Deviationfrom the Class Mean, SDCM for short) of k speed subsets under each classification number—— n j is the number of all elements in the jth subset, and there exists v j,i is the value of the i-th element in the j-th subset; is the mean of all elements in the jth subset;

[0022] S403, calculate the GVF value of the classification result output by the natural break point method under each classification number, GVF k =(SDAM-SDCM k ) / SDAM.

[0023] According to the above scheme, in step S5, when the overall sample is fixed, SDAM is a fixed value, and SDCM decreases as the number of classifications increases. Therefore, the more classifications there are, the larger the GVF value is. In particular, when the number of classifications is equal to the number of elements in the data set, SDCM is 0 and GVF is equal to 1. However, the larger the number of classifications of the bus operation status, the better. Too many classifications have no practical physical meaning. When the GVF increment is very small, the goodness of fit of the classification result does not change much, and continuing to increase the number of classifications will not significantly improve the accuracy of the bus operation status division. Therefore, the present invention selects a classification number with a large GVF value and whose increment with the increase of the number of classifications is no longer significant as the optimal classification number of the bus operation status, that is, the minimum number of classifications that simultaneously meets the following two conditions:

[0024] The first condition is that GVF is greater than 0.8, which ensures that the classification results output by the natural break point method have a good enough fitting effect;

[0025] The second condition: when one more category is added to the current category, the increment of GVF is less than 0.05, and the fitting effect is no longer significantly improved.

[0026] The method for determining the number of bus operation status classifications of the present invention has the following beneficial effects:

[0027] 1. In the present invention, GVF is used as a criterion for evaluating the goodness of fit of the natural break point method, and its value reflects the fitting effect under a specific number of classifications;

[0028] 2. In the present invention, as the number of classifications increases, the GVF value will gradually increase. The GVF increment reflects the degree of improvement in the fitting effect of the classification results as the number of classifications increases. When the increase in the number of classifications no longer significantly improves the fitting effect, the optimal number of classifications for the bus operation status is obtained;

[0029] 3. The present invention determines the optimal classification number of the bus operation status through the GVF values ​​under different classification numbers and the incremental size of GVF as the classification number increases, which provides a reasonable and effective basis for determining the classification number of the bus operation status. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 A flow chart of a method for determining the number of bus operation status classifications of the present invention;

[0031] Figure 2 This is a schematic diagram of GVF values ​​that classify bus operation states into 2 to 7 categories based on bus travel speed data of a certain city on a natural day of November 30, 2019 in an embodiment of the present invention;

[0032] Figure 3 This is a schematic diagram of the GVF average value of the bus operation status classification results of a certain city for 30 natural days based on different classification numbers in an embodiment of the present invention;

[0033] Figure 4 This is a schematic diagram of the increment of the GVF average value based on the bus operation status classification results of a certain city for 30 natural days as the number of categories increases in an embodiment of the present invention. DETAILED DESCRIPTION

[0034] In order to have a clearer understanding of the technical features, purposes and effects of the present invention, specific embodiments of the present invention are now described in detail with reference to the accompanying drawings.

[0035] like Figure 1-4 As shown, the method for determining the number of bus operation status classifications of the present invention comprises the following steps:

[0036] S1. Collecting bus travel speed data; the collected bus travel speed data is the ratio of the bus travel distance to the travel time within a fixed time interval, and the fixed time interval is set to 5 minutes or 10 minutes.

[0037] S2. Setting a value range for the bus operation status classification number; referring to relevant national / industry standards and existing research on traffic flow and bus operation status classification number, the value range of the bus operation status classification number k is set to 2 to 7.

[0038] S3, obtaining the bus running speed threshold under each classification number;

[0039] The bus running state speed threshold is obtained as the critical speed value between different bus running states. The specific method is as follows:

[0040] S301, arranging the bus travel speed data at fixed time intervals from small to large to form a bus travel speed sequence statistic;

[0041] S302, using the natural break point method to divide the bus travel speed order statistics into k categories, k = 2, 3, ..., 7;

[0042] S303 outputs the critical speed values ​​between bus running states under each classification number, which are defined as v1, v2, ..., v in ascending order. k-1 , which is the speed threshold of the bus operation status.

[0043] Critical speed values ​​v1, v2, …, v k-1 Divide the bus travel speed dataset into k subsets: [v min ,v1),[v1,v2),…,[v k-1 ,v max ],v min is the minimum speed value in the travel speed data set, v max is the maximum speed value in the travel speed data set, and the k speed subsets correspond to k bus operation states.

[0044] S4. Calculate the goodness of variance fit for each number of classifications. The goodness of variance fit (GVF) is a summary measure of the classification results of the natural break point method. The value of GVF is between 0 and 1, where 1 indicates an excellent fit and 0 indicates an extremely poor fit. The calculation steps are as follows:

[0045] S401. Calculate the sum of squared deviations from the Array Means (SDAM) of the bus travel speed data set. n is the number of all elements in the travel speed data set; v i is the value of the i-th element in the travel speed data set; is the mean of all elements in the travel speed data set;

[0046] S402, calculate the sum of the class total square differences (Sum of Deviationfrom the Class Mean, SDCM for short) of k speed subsets under each classification number—— n j is the number of all elements in the jth subset, and there exists v j,i is the value of the i-th element in the j-th subset; is the mean of all elements in the jth subset;

[0047] S403, calculate the GVF value of the classification result output by the natural break point method under each classification number, GVF k =(SDAM-SDCM k ) / SDAM.

[0048] S5. Determine the optimal classification number of bus operation status.

[0049] When the overall sample is fixed, SDAM is a constant, and SDCM decreases as the number of classifications increases. Therefore, the more classifications there are, the larger the GVF value is. In particular, when the number of classifications is equal to the number of elements in the data set, SDCM is 0 and GVF is equal to 1. However, the larger the number of classifications of the bus operation status, the better. Too many classifications have no practical physical meaning. When the GVF increment is very small, the goodness of fit of the classification result does not change much, and continuing to increase the number of classifications will not significantly improve the accuracy of the bus operation status division. Therefore, the present invention selects the number of classifications with a large GVF value and whose increment with the increase of the number of classifications is no longer significant as the optimal number of classifications for the bus operation status, that is, the minimum number of classifications that simultaneously meets the following two conditions:

[0050] The first condition is that GVF is greater than 0.8, which ensures that the classification results output by the natural break point method have a good enough fitting effect;

[0051] The second condition: when one more category is added to the current category, the increment of GVF is less than 0.05, and the fitting effect is no longer significantly improved.

[0052] In a preferred embodiment of the present invention, based on the bus travel speed data of a certain city on natural day November 30, 2019, an optimal classification number of the bus operation status of a certain city is determined according to a method for determining the classification number of bus operation status proposed by the present invention.

[0053] The specific implementation steps are as follows:

[0054] S1. Collect and obtain the bus travel speed data of a certain city on November 30, 2019, calculate the bus travel speed at fixed intervals of every 10 minutes, remove some abnormal data with too large values ​​and arrange them from small to large, and obtain the bus travel speed order statistics. Use the natural break point method (implemented with Python software) to divide the bus travel speed order statistics into k categories, k = 2, 3, ..., 7, and output the critical speed values ​​between the bus running states under each classification number, which are defined as v1, v2, ..., v in order from small to large. k-1 , divide the bus travel speed dataset into k subsets: [v min ,v1),[v1,v2),…,[v k-1 ,v max ],v min is the minimum speed value in the travel speed data set, v max is the maximum speed value in the travel speed data set, and the k speed subsets correspond to k bus operation states.

[0055] S2, Goodness of Variance Fit (GVF) is a summary measure of the classification results of the natural break point method. The value of GVF is between 0 and 1, 1 means an excellent fit and 0 means a very poor fit. The calculation steps are as follows:

[0056] S201. Calculate the sum of squared deviations from the Array Means (SDAM) of the bus travel speed data set. n is the number of all elements in the travel speed data set; v i is the value of the i-th element in the travel speed data set; is the mean of all elements in the travel speed data set;

[0057] S202. Calculate the sum of the class mean (SDCM) of the k speed subsets under each classification number. n j is the number of all elements in the jth subset, and there exists v j,i is the value of the i-th element in the j-th subset; is the mean of all elements in the jth subset;

[0058] S203, calculate the GVF value of the classification result output by the natural break point method under each classification number, GVF k =(SDAM-SDCMk ) / SDAM.

[0059] The GVF values ​​of a city for 30 natural days under different classification numbers are as follows: Figure 2 As shown in the figure, the GVF values ​​of 30 natural days under the same number of classifications are basically the same, and they continue to increase with the increase in the number of classifications. When divided into two categories, the GVF is distributed around 0.6, when divided into three categories, the GVF is about 0.8, when divided into four categories, the GVF exceeds 0.8 and is close to 0.9, and when divided into five categories or more, the GVF exceeds 0.9.

[0060] S3, calculate the GVF average value of the results of the 30 natural days of bus operation status in a city divided into 2 to 7 categories (see Figure 3 ) and the increment of the mean value of GVF as the number of categories increases (see Figure 4 ). The GVF increment decreases with the increase of the number of categories. The GVF of category three increases by nearly 0.2 compared with that of category two, the GVF of category four increases by nearly 0.1 compared with that of category three, and the GVF of category five increases by about 0.04 compared with that of category four, which is lower than 0.05. Therefore, the optimal number of categories for the city’s public transportation operation status should be four.

[0061] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the enlightenment of the present invention, ordinary technicians in this field can also make many forms without departing from the scope of protection of the purpose of the present invention and the claims, which all fall within the protection of the present invention.

Claims

1. A method for determining the number of bus operation status classifications, characterized in that: The following steps are involved: S1. Collecting bus travel speed data; S2. Setting the value range of the bus operation status classification number; S3, obtaining the bus running speed threshold under each classification number; The bus running state speed threshold is the critical speed value between different bus running states; The method for obtaining the speed threshold of the bus running status is as follows: S301, sorting the bus travel speed data in ascending order to form a bus travel speed sequence statistic; S302, using the natural break point method to divide the bus travel speed order statistics into k kind, k =2, 3,…,7; S303: Output the critical speed value between the bus running states under each classification number, which are defined in ascending order as v 1, v 2,…, v k-1 ; The critical speed value v 1, v 2,…, v k-1 The bus trip speed dataset is divided into k Subsets: [ v min , v 1), [ v 1, v 2),…, [ v k-1 , v max ], v min is the minimum speed value in the travel speed data set, v max is the maximum speed value in the travel speed data set, k The speed subsets correspond to k bus operation status; S4, calculate the variance goodness of fit under each classification number; Goodness of Variance Fit, also known as Goodness of Variance Fit, referred to as GVF is the summary measure of the classification results using the natural breaks method, GVF The value of is between 0 and 1, 1 means an excellent fit, and 0 means an extremely poor fit. The calculation steps are as follows: S401. Calculate the sum of squared deviations from the Array Means of the bus travel speed data set. SDAM —— , n is the number of all elements in the travel speed dataset; v i The travel speed data set i The value of an element; is the mean of all elements in the travel speed data set; S402, calculate the number of each classification k The sum of the total squared differences of the speed subsets (Sum of Deviation from the Class Mean, referred to as SDCM )—— , n j For the j The number of all elements in the subsets, and there exists ; v j,i For the j In the subset i The value of an element; For the j The mean of all elements in the subsets; S403, calculate the natural break point method output classification results under each classification number GVF value, ; S5, determining the optimal classification number of the bus operation status; The optimal number of categories for bus operation status is the minimum number of categories that satisfies the following two conditions at the same time: First condition: GVF Greater than 0.8, ensuring that the classification results output by the natural break point method have a sufficiently good fitting effect; Second condition: When adding a category number based on the current category number, GVF When the increment is less than 0.05, the fitting effect is no longer significantly improved.

2. The method for determining the number of bus operation status classifications according to claim 1, characterized in that: In step S1, the collected bus travel speed data is the ratio of the bus travel distance to the travel time within a fixed time interval, and the fixed time interval is set to 5 minutes or 10 minutes.

3. The method for determining the number of bus operation status classifications according to claim 1, characterized in that: In step S2, the bus operation status classification number k The value range is 2 to 7.

Images