Optimization Methods and Systems for Arrival Tracks in Marshalling Stations for Virtual Marshalling

Abstract

This invention relates to the field of transportation technology and discloses a method and system for optimizing the use of arrival tracks in marshalling yards for virtual train formations, thereby improving the throughput capacity of arrival throats in marshalling yards within a phased plan. The method includes: constructing a directed graph of the yard throat area; dividing the arrival track bundles according to the directed graph; determining virtual train formations and the number of trains within each virtual train formation; selecting the track bundle with the closest number of tracks and trains among the currently available track bundles for train reception; when the number of trains in the virtual train formation is less than or equal to the maximum capacity of the selected track bundle, the selected track bundle is used directly for train reception; otherwise, after each available track in the selected track bundle is used to receive one train in sequence, the excess trains are de-virtually formed from the preceding train and a new virtual train formation is re-evaluated, and so on, to complete the allocation of receiving tracks and routes for each virtual train formation within the phased plan.

Description

Technical Field

[0001] This invention relates to the field of transportation technology, and in particular to a method and system for optimizing the use of arrival tracks in marshalling yards for virtual marshalling. Background Technology

[0002] Virtual train formation is a train control mode that shortens train tracking time by increasing inter-train communication, enabling multiple trains to operate collaboratively and form a train group. In virtual train formation, the first train in the train group leads the entire group, sending control commands to other trains; this first train is defined as the lead car, and the remaining trains are non-lead cars. Two adjacent trains within the same train group are designated as leading and trailing cars based on their direction of travel. The trailing car shares information with the leading car and maintains a certain tracking interval while following it. Figure 1 As shown, train sets operate using a quasi-moving block system.

[0003] The practical application of virtual train formation will bring about significant changes to railway infrastructure and train operation methods. Currently, trains mostly use fixed block systems to ensure operational safety when running on railway sections. Fixed block systems separate trains into different block sections, controlling the train's position and speed according to signal lights, such as... Figure 2 (a) In virtual train formation mode, a quasi-moving block system is used between train sets, with a minimum headway of 3 minutes. Under the quasi-moving block system, physical block sections are eliminated, and the following distance between trains is based on the absolute braking distance of the following train when the preceding train is stationary, such as... Figure 2 (b) In virtual train formation mode, the following distance between trains within the same trainset is the relative braking distance of the following train considering the speed of the preceding train, not the braking endpoint of the rear of the preceding train. Figure 2 (c) Based on the fact that the train tracking interval within a train group is less than the tracking interval between train groups, conditions for determining virtual train formations can be set. For example, trains located in the same section, running in the same direction, and with a tracking interval of less than 3 minutes belong to the same virtual train formation.

[0004] The growth in railway freight volume and the introduction of national policies have placed higher demands on the throughput capacity and orderly operation of marshalling yards. A well-designed track utilization plan for marshalling yard arrival areas can reduce train dwell time within the station, improve the continuity of various operations, and enhance station throughput capacity. While many scholars have studied track utilization issues in marshalling yards under current transportation organization conditions, no scholars have yet considered the optimization of track utilization under a virtual marshalling operation mode. The virtual marshalling operation mode is expected to significantly reduce train following intervals and greatly improve the throughput capacity of sections and stations. With the development of technologies such as electronics, sensors, and computers, virtual marshalling operation mode will become an important development direction for railway transportation. Therefore, research on the optimization of railway station track utilization for virtual marshalling is essential. Summary of the Invention

[0005] The purpose of this invention is to disclose a method and system for optimizing the use of arrival tracks in marshalling yards for virtual marshalling, so as to improve the throughput capacity of arrival throats in marshalling yards within the phased plan.

[0006] To achieve the above objectives, the present invention discloses an optimization method for the arrival track utilization of marshalling yards for virtual marshalling, comprising:

[0007] Construct a directed graph of the throat area of ​​the parking lot;

[0008] According to the directed graph, the arrival track bundles are divided, and the routes corresponding to any two tracks in any combination within any track bundle can be converted to each other by changing the direction of a turnout.

[0009] Determine a virtual train formation and the number of trains within it, and then allocate the trains within the virtual train formation as follows:

[0010] Select the track bundle with the closest number of tracks to the number of trains among the currently available track bundles to arrange train reception. When the number of trains in the virtual train formation is less than or equal to the maximum capacity of the selected track bundle, the selected track bundle is used directly for train reception. If the number of trains in the virtual train formation exceeds the maximum capacity of the selected track bundle, one train is connected to each available track of the selected track bundle in sequence. Then, the excess trains are removed from the virtual formation mode from the preceding train and a new virtual formation is re-evaluated. This process is repeated to complete the allocation of receiving tracks and routes for each virtual train formation within the phase plan.

[0011] The virtual train formation is designed to enable the lead train to lead the entire train group and send control commands to other trains. Within the same virtual train formation, two adjacent trains in the same direction of travel are divided into a lead train and a follow train. The follow train shares information with the lead train and maintains a certain tracking interval to follow the lead train.

[0012] To achieve the above objectives, the present invention also discloses a marshalling yard arrival track utilization optimization system for virtual marshalling, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the corresponding methods described above.

[0013] The present invention has the following beneficial effects:

[0014] The station's route control system does not fully utilize computer interlocking, allowing for direct switching of turnout directions to create new routes within train intervals. Without virtual marshalling, trains enter the tracks using the traditional entry method: first, the relevant area for the entry route is confirmed to be clear, then the route is arranged and locked until the train has cleared its route, at which point it is unlocked. With virtual marshalling operation, freight trains undergoing refitting can arrive at the marshalling yard with shorter intervals. In this case, the arrival yard will need to arrange receiving routes in two ways depending on the actual situation. First, to avoid conflicts between train entry routes and other operational routes such as locomotive travel within the station, the first car of independently operating trains and virtual formation trains still need to be received using the traditional entry method. Second, for non-first cars waiting to be received in virtual formation trains, the receiving route needs to be arranged according to the degree of overlap between its route and the route of the adjacent train in the preceding direction: when the route of the non-first car waiting to be received only needs to change the direction of one switch on the basis of the route of the adjacent train in the preceding direction, it can be directly received and entered the station by timely switching the switch during the interval between the two trains through the train autonomous route control mode; thereby improving the throughput capacity of the arrival yard throat of the marshalling yard within the phase plan.

[0015] The present invention will now be described in further detail with reference to the accompanying drawings. Attached Figure Description

[0016] The accompanying drawings, which form part of this application, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an undue limitation of the invention. In the drawings:

[0017] Figure 1 This is a schematic diagram of the virtual train formation mode disclosed in this invention.

[0018] Figure 2 These are the different modes of train interval operation diagrams disclosed in this invention.

[0019] Figure 3 These are schematic diagrams of different vehicle reception routes disclosed in this invention.

[0020] Figure 4 This is a schematic diagram of the optimization method for the arrival track of the marshalling yard under the virtual marshalling operation mode disclosed in this invention.

[0021] Figure 5 This is a layout diagram of the downstream arrival yard of Zhuzhou North Marshalling Station disclosed in this invention.

[0022] Figure 6 This is a directed graph corresponding to the arrival site layout diagram of the example disclosed in this invention. Detailed Implementation

[0023] The embodiments of the present invention will be described in detail below with reference to the accompanying drawings, but the present invention can be implemented in many different ways as defined and covered by the claims.

[0024] Example 1

[0025] This embodiment discloses an optimization method for the use of arrival yard tracks in a marshalling yard under a virtual marshalling operation mode. Under the premise of a given throat area and track layout of the marshalling yard, virtual marshalling operation plan of the train section and arrival time, the method considers the impact of the virtual marshalling operation mode on the change of train entry mode, and optimizes the entry route and receiving track of arriving trains with the goal of maximizing the throughput capacity of the throat area of ​​the marshalling yard within the phase plan.

[0026] This embodiment studies the train entry method using virtual formation operation mode, such as... Figure 1 As shown, in this mode, the train uses a new train control system. After arrival at the station, the onboard equipment exchanges information with ground and trackside equipment to obtain the receiving route and track. The station route control method does not fully adopt computer interlocking, allowing direct switching of turnout direction to form a new route within the train interval. When the virtual marshalling mode is not used, the train enters the track according to the traditional station entry mode, that is, first confirming that the relevant area of ​​the entry route is empty, then arranging the route and locking it until the train has cleared the track and unlocking it. When the virtual marshalling operation mode is adopted, freight trains undergoing refitting can arrive at the marshalling yard with a shorter running interval. At this time, the arrival yard will need to arrange the receiving route in two ways according to the actual situation. First, to avoid conflicts between train entry routes and other operational routes such as locomotive travel within the station, the first car of independently operating trains and virtual formation trains must still use the traditional entry method. Second, for non-first cars waiting to be received in virtual formation trains, the receiving route needs to be arranged according to the degree of overlap with the route of the adjacent preceding train: when the route of the non-first car waiting to be received only needs to change the direction of one switch on the basis of the route of the adjacent preceding train, it can be directly received and entered the station through the train autonomous route control mode during the running interval between the two trains; when the route of the non-first car waiting to be received needs to change the direction of multiple switches on the basis of the route of the adjacent preceding train, considering the short following interval between the preceding and following trains, the long time required for multiple switches to be changed sequentially, and the uncertainty of whether the new route will conflict with other operational routes within the station, the traditional entry method must be used to ensure train operation safety.

[0027] like Figure 3 As shown, when two trains in a virtual train formation enter the receiving track using route 1, if the following train enters another track using route 2, since route 2 differs from the route of the preceding train only in the direction of turnout a, the following train can enter the station using autonomous route control without waiting for the route to be locked. If the following train enters the station using route 3, since it passes through several different turnouts compared to the route of the preceding train, it is necessary to rearrange its route and lock it before opening the entry signal to enter the station.

[0028] This embodiment considers the use of tracks in the marshalling yard under the virtual train formation operation mode. It fully considers track availability constraints, track occupancy uniqueness constraints, track occupancy safety constraints, train autonomous route control constraints, and traditional entry route interlocking constraints. It constructs an optimization model with the objective of minimizing the average time that trains occupy the throat area within the stage plan, and designs a heuristic algorithm to solve for the optimal track utilization scheme.

[0029] Key symbols: D represents the set of arrival yard receiving tracks, D = {1, 2, ..., j, ..., m}; D1, D2, ..., D h For each track bundle, D = D1∪D2∪,...,∪D h L is the set of arriving trains within the phase plan, L = {1, 2, ..., i, ..., n}; i is the index of the arriving train, i ∈ L; j is the index of the arrival yard track, j ∈ D; m is the total number of tracks used for receiving trains in the arrival yard; n is the total number of trains arriving within the phase plan; G is the directed graph of the station throat area, G = (N, A); s1, s2, ..., s p These are the directed source points of the graph, i.e., the starting points of all work routes in the throat area; v1, v2, ..., v r Let t1, t2, ..., t be the midpoint of the directed graph, i.e., the center of all turnouts in the throat area; q S is the destination point of a directed graph, representing the endpoint of all operational routes in the throat region; i S is the station entry route for train i to access the track. i ={v a ,v b ,...,v c S is the set of the first k shortest paths to each track; The path that enters lane j and is located at the p-th position in set S; Let i be the time when train i arrives at the station; The time for arranging the route and unlocking the route for train i; t b b is the time required for one switch change; i This represents the number of times train i needs to switch tracks to enter the station; The time that train i travels in the throat area of ​​the station; This refers to the time when train i begins to occupy the track. The time for train i to clear the track; T gj T is the earliest time when a train can enter track j; T is the safe interval between two trains entering the same track.

[0030] Relevant decision variables illustration: α i β is a 0-1 variable; it is 1 if train i needs an interlocking route, and 0 otherwise. iThe variable is 0-1; it is 1 if train i needs to switch points on its route, and 0 otherwise. ij It is a 0-1 variable; if train i occupies track j, it is 1, otherwise it is 0.

[0031] The model optimization objective is to minimize the average time that trains occupy the throat area within the planned phase. The arrival yard receiving capacity is mainly affected by the throat throughput capacity at the arrival yard entrance and the track receiving capacity. In the case of coordinated track receiving and pushing / unloading operations, the throat throughput capacity becomes the bottleneck. Therefore, this research model takes minimizing the average time Z that trains occupy the throat area within the planned phase as its optimization objective. Specifically...

[0032] The constraints on the model are as follows:

[0033] (1) [Stockway Availability Constraints]

[0034] When selecting a receiving track for an arriving train, a track that is currently idle must be selected. This constraint is expressed as:

[0035]

[0036] (2) [Uniqueness constraint of lane occupancy]

[0037] An arriving train can only connect to one track and continue until it leaves that track, with no intermediate track switching required. This constraint is expressed as follows:

[0038]

[0039] (3) [Safety Constraints on Stock Lane Occupancy]

[0040] When two consecutive trains enter the same track, the interval between their entry into the track must meet the minimum safety interval requirement. This constraint is expressed as:

[0041]

[0042] (4) Constraints on Autonomous Train Route Control

[0043] When a non-leading car in a virtual trainset meets the autonomous route control conditions, it can track and enter the throat area without waiting for the preceding car's route to unlock. It then enters the designated receiving track by changing the direction of a single turnout. This constraint is expressed as 0 ≤ (α... i -α i+1 )|S i ∪S i+1 |≤max{|S i |,|S i+1 |}+1;

[0044] (5) [Traditional Station Entry Route Interlocking Constraints]

[0045] If the arrival interval between two adjacent trains is greater than the time required to arrange a route and interlocking for the following train, then the following train's entry route must be interlocked using the traditional entry method. This constraint is expressed as:

[0046] The specific idea behind the heuristic algorithm is as follows:

[0047] This algorithm treats virtual train sets as a whole and allocates receiving tracks accordingly. It comprises three parts: directed graph construction of the depot throat area, track bundle partitioning sub-algorithm, and receiving track allocation algorithm. The algorithm flowchart is shown below. Figure 4 As shown.

[0048] (1) Construct a directed graph of the throat area of ​​the parking lot

[0049] Based on the layout diagram of the throat area of ​​the marshalling yard, a directed graph G = (N, A) is constructed, where N is the set of vertices in the directed graph, and A is the set of arcs with directions. The vertex set N includes three subsets: the source vertex set S, the sink vertex set T, and the intermediate vertex set V; the relationship between the vertex sets is: N = S ∪ T ∪ V, and... Source set S = {s1, s2, ..., s} p} and sink set T = {t1, t2, ..., t q} represent the starting and ending points of all work routes in the throat area, respectively; the intermediate point set V = {v1, v2, ..., v...} r} represents the center of all turns in the throat area; arc set A = { i ,v j >,..., <v k ,v K >,..., <v k ,t j The arc '>' represents a directed edge between adjacent connected points, with the arc pointing in the same direction as the work path.

[0050] (2) Divide the arrival track bundle

[0051] If a route entering two tracks passes through only one different turnout, then these two tracks can be used to arrange for the connection of two adjacent trains in virtual train formation operation. Therefore, it is necessary to compare all routes in the throat area, and then determine the relationship between all receiving tracks in the arrival yard to obtain the track bundle. The track bundle partitioning sub-algorithm 1 is constructed as follows:

[0052] Sub-Algorithm 1:

[0053] Step 1: Input the directed graph G of the parking lot throat area constructed in (1) into the computer, specifically including the source vertex set S = {s1, s2, ..., s} of the directed graph.​p The sink set T = {t1, t2, ..., t} q The intermediate point set V = {v1, v2, ..., v} r} and arc set A = { i ,v j >,..., <v k ,v K >,..., <v k ,t j >}.

[0054] Step 2: Use the K-shortest path algorithm to find the top 5 shortest paths from each source to the sink, forming the set of all paths. in, j∈D, p is a path The position number in set S is indicated by j, which is the number of the path leading to the track.

[0055] Step 3: Compare all paths in the path set pairwise. When a path can be transformed into another path by changing the direction of a turnout, record the track numbers reached by these two paths. Specifically, set the initial values ​​p = 1 and q = p + 1.

[0056] Step 4: Select the p-th path from the complete path set S. With the q-th path

[0057] Step 5: Verify the path With path satisfy If so, the track numbers j1 and j2 entered by the two paths are recorded in list R in the form of list [j1,j2].

[0058] Step 6: After the comparison is completed, the p-th path remains unchanged, and the q-th path is handled in two cases: if the q-th path is not the last path in the set, it becomes the next path, i.e., q = q + 1, and returns to Step 3 to continue; if the q-th path is already the last path in the set, it directly proceeds to Step 7.

[0059] Step 7: Perform operations based on the p-th path. If the p-th path is not the last path in the set, then p = p + 1, set the q-th path as the path after the p-th path, that is, set q = p + 1 and return to Step 4. If the p-th path is the last path in the set, then proceed to Step 8.

[0060] Step 8: The algorithm ends.

[0061] ​After the algorithm finishes, multiple pairs of track numbers are obtained, indicating that there are at least two paths leading to these two tracks respectively. These paths can be switched by changing the direction of a turnout, i.e., list R = [[j1,j2],[j1,j3],...,[ja,jb],...]. When [j1,j2], [j1,j3], and [j2,j3] exist simultaneously, the three track number pairs are merged into a single list [j1,j2,j3].

[0062] Check all track number pairs [ja,jb] ([ja,jb]∈R) in list R. If any of these pairs exist, merge them. Otherwise, do nothing and retain the original track number pairs. The final result is:

[0063] Each element in the list R = [[j1,j2,...,ja],[j1,jb],...,[jc,...,jd]] is a track bundle, namely D1 = [j1,j2,...,ja], D2 = [j1,jb],...,Dh = [jc,...,jd], and the total number of track bundles is set D, D = {D1,D2,...,Dh}.

[0064] (3) Allocate receiving tracks

[0065] To optimize the use of receiving tracks within the phase plan, it is necessary to uniformly allocate receiving tracks for multiple trains in virtual formation operation. Following the overall algorithm design, the Sub-Algorithm 2 for receiving track allocation is constructed as follows:

[0066] Sub-Algorithm 2:

[0067] Step 1: Input the set of trains arriving within the planned phase, L = {1,2,...,i,...,n}.

[0068] Step 2: Identify the arriving train sets in the virtual formation operation.

[0069] Step 2.1: Set i = 1, L1 = [i].

[0070] Step 2.2: When the arrival time of the following train differs from that of the preceding train by less than 3 minutes, the following train is considered virtually coupled to the preceding train and belongs to the same virtual train group. That is, when... If i+1 is added to L1, proceed to Step 2.3; otherwise, proceed to Step 3.

[0071] Step 2.3: i = i + 1, proceed to Step 2.2.

[0072] Step 3: Determine the currently available lane.

[0073] Step 3.1: Check if the j-th track is free. Set j = 1 and temp = []. The temp set is used to store the numbers of the free tracks.

[0074] Step 3.2: When the time for the track to be cleared is less than the time for the train to arrive, the track is considered to be free and can be used. Add j to temp.

[0075] Step 3.3: Check the availability of the next track. Let j = j + 1, and proceed to Step 3.2. After the last track has been checked, proceed to Step 4.

[0076] Step 4: Update the track bundle set, removing non-idle tracks from the initial track bundle, D = {D1, D2, ..., Dh}.

[0077] Step 4.1: Set i = 1.

[0078] Step 4.2: Remove the non-empty tracks from the initial track bundle to form the current track bundle Di = Di∩temp.

[0079] Step 4.3: i = i + 1. If i ≤ h, proceed to Step 4.2; otherwise, proceed to Step 5.

[0080] Step 5: Determine the number of trains in the current trainset arriving at the station, and select the track bundle with the closest number of tracks to the number of trains to arrange train reception. That is, select Du such that ||L1|-|Du||=min{|L1-D1|,|L1-D2|,...,|L1-Dh|}.

[0081] Step 6: When the number of trains arriving in the train set is less than the maximum number that the track bundle can accommodate, arrange for the trains to be received directly and then proceed to Step 10. If the number of trains arriving in the train set exceeds the maximum number of tracks in the track bundle, then proceed to Step 7.

[0082] Step 7: Determine the number of trains exceeding the limit, which is the difference between the number of trains and the number of tracks within the maximum track bundle. Let diff = |L1| - |Du|.

[0083] Step 8: If the number of excess trains is too large, exceeding the sum of the number of empty tracks in the remaining track bundles, then after each empty track receives one train, the remaining trains wait outside the station and proceed to Step 10; otherwise, proceed to Step 9.

[0084] Step 9: When the number of excess trains is less than or equal to the sum of the number of empty tracks in the remaining track bundles, the excess trains and the preceding train shall be released from the virtual formation mode in advance in the section, and the trains shall be reassigned to receiving tracks in accordance with Step 5 as a new virtual formation train group.

[0085] Step 10: Algorithm ends.

[0086] The optimal stock trading strategy is finally obtained through a solution.

[0087] Example 2

[0088] This embodiment uses the downhill arrival yard of Zhuzhou North marshalling yard as a reference for case study. Its layout diagram and corresponding directed graph are shown below. Figure 5 and Figure 6 As shown.

[0089] Depend on Figure 5 It is evident that the Zhuzhou North marshalling yard has 15 tracks in the downhill arrival area and 21 turnouts in the throat area. Applying the aforementioned model and algorithm to allocate receiving tracks and routes for trains arriving within the planned timeframe at this station, the objective function value under the conventional operation mode is significantly higher than that under the virtual marshalling operation mode. Compared to the autonomous route control method, the traditional entry method results in a longer time spent in the throat area. When trains are not arriving in concentrated numbers, route arrangement can be done before train arrival, saving time. However, under the virtual marshalling mode, with concentrated train arrivals, route arrangement will affect the reception of subsequent trains. After using the model and algorithm presented in this paper to arrange receiving tracks, the average time each train occupies the throat area is 213.6 seconds. Currently, according to the receiving tracks arranged by the station duty officer, the average time each train occupies the throat area is 300 seconds.

[0090] Refer to above Figure 3 The arrival times and receiving tracks of the freight trains that were actually reorganized during a certain period are shown in Table 1. Since the virtual formation operation mode has not yet been applied, the train arrival times under the virtual formation operation mode are simulated in Table 2 by shortening the tracking interval of freight trains and increasing the number of freight trains, without affecting the scheduled operation of passenger trains in the train operation schedule.

[0091] Table 1 Actual Train Arrival Timetable

[0092]

[0093]

[0094] Table 2 Train arrival timetable under simulated virtual train formation operation mode

[0095] serial number Train number time serial number Train number time 1 31016 18:38 14 19182 21:25 2 virtual 1 18:40 15 virtual 5 21:27 3 virtual 2 18:42 16 25049 21:38 4 40036 18:52 17 41027 21:53 5 virtual 3 18:54 18 virtual 6 21:55 6 41075 18:58 19 25053 22:23 7 25041 19:19 20 virtual 7 22:25 8 virtual 4 19:21 21 virtual 8 22:27 9 40070 19:36 22 25059 22:58 10 25057 19:38 23 40090 23:26 11 31018 20:34 24 virtual 9 23:28 12 25047 20:57 25 virtual 10 23:30 13 25009 21:09

[0096] Directed graph constructed based on arrival fields of computational examples Figure 6 Based on the solution algorithm of this embodiment, the following solution steps are performed:

[0097] Step 1: Read the directed graph. Read the source vertex set S = {s1} and sink vertex set T = {t1, t2, ..., t3} in graph G. 13 The intermediate point set V = {v1, v2, ..., v} 21} and arc set A = {<s1,v1> ,...,<v2,v6> ,..., <v 13 ,t 13 >}.

[0098] Step 2: Search for the first 5 shortest paths. Use the K-shortest path algorithm to find the first 5 shortest paths from each source to the sink, forming the complete set of paths. in, j∈D, p is a path The position of the path in set S is labeled, where the sink point of the path is numbered j.

[0099] Step 3: Set initial values ​​p = 1, q = p + 1.

[0100] Step 4: Select the p-th path from the complete path set S. With the q-th path

[0101] Step 5: Verify the path With path satisfy If so, then the list [j1,j2] will be nested and recorded in list R.

[0102] Step 6: q = p + 1. If q ≤ |S|, return to Step 3. If q > |S|, proceed to Step 7.

[0103] Step 7: p = p + 1. If p ≤ |S| - 1, q = p + 1 and return to Step 4. If p > |S| - 1, proceed to Step 8.

[0104] Step 8: The algorithm ends.

[0105] After the algorithm completes, we obtain a list R = [[t1,t2],[t2,t3],...,[t12,t13]]. We iterate through the sublists in R, checking if any other lists contain the same elements. If they do, we perform the following operation; otherwise, we replace the current sublist with the next one and continue the search. We check if another list in R contains the other elements from both of the above lists. If it does, we merge the three lists in R into a single list; otherwise, we retain the original list. Until all sublists within R are merged or retained, the final list R = [[t1,t2,t3],[t3,t4,t5,t6],[t6,t7,t8,t9,t10,t11],[t12,t13]] is obtained. Each element in the list is a track bundle, namely D1 = [t1,t2,t3], D2 = [t3,t4,t5,t6], D3 = [t6,t7,t8,t9,t10,t11], D4 = [t12,t13]. The total number of track bundles is set D, D = {D1,D2,D3,D4}.

[0106] After obtaining the initial track bundle, the receiving tracks are assigned to the trains scheduled to arrive in the second phase according to the following algorithm.

[0107] Step 1: Input the set of trains arriving within the phase plan L = {1,2,...,i,...,25}.

[0108] Step 2: Identify the arriving train sets in the virtual formation operation.

[0109] Step 2.1: Set i = 1, L1 = [i].

[0110] Step 2.2: If If i+1 is added to L1, proceed to Step 2.3; otherwise, proceed to Step 3.

[0111] Step 2.3: i = i + 1, proceed to Step 2.2.

[0112] Step 3: Determine the currently available lane.

[0113] Step 3.1: Set j = 1, temp = [].

[0114] Step 3.2: If Then add j to temp.

[0115] Step 3.3: j = j + 1. If j ≤ m, then proceed to Step 3.2; otherwise, proceed to Step 4.

[0116] Step 4: Update the set of stock bundles, D = {D1, D2, ..., Dh}.

[0117] Step 4.1: Set i = 1.

[0118] Step 4.2: Di = Di ∩ temp.

[0119] Step 4.3: i = i + 1. If i ≤ h, proceed to Step 4.2; otherwise, proceed to Step 5.

[0120] Step 5: Find Du such that ||L1|-|Du||=min{|L1-D1|,|L1-D2|,...,|L1-Dh|}.

[0121] Step 6: If |L1|≤|Du|, then allocate the first |L1| tracks in Du to the trains arriving at L1 and proceed to Step 10; otherwise, proceed to Step 7.

[0122] Step 7: Let diff = |L1| - |Du|.

[0123] Step 8: If diff > |D1| + |D2| + ... + |Dh| (excluding Du), then remove the last diff - (|D1| + |D2| + ... + |Dh|) trains in L1, and assign the remaining trains to the tracks in D1, D2, ..., Dh (excluding Du) in sequence, and proceed to Step 10; otherwise, proceed to Step 9.

[0124] Step 9: If diff ≤ max{|D1|,|D2|,...,|Dh|} (excluding Du), then let L1 = diff, D = D - Du, and return to Step 5.

[0125] Step 10: Algorithm ends.

[0126] The algorithm determined that a total of 25 trains would arrive in the planned stage of this embodiment, including 8 virtual train formations. For 5 virtual train formations, only the lead car was assigned an entry route; subsequent cars simply changed direction at a switch. The remaining virtual train formations were deactivated early because there were no tracks with similar routes for receiving them.

[0127] Table 3 shows the optimal train receiving track and route allocation results obtained from the solution (the virtual train numbers in the table below represent trains added in the virtual formation mode):

[0128] Table 3 shows the results of the allocation of receiving tracks and routes in the calculation example.

[0129]

[0130]

[0131] Therefore, after analyzing the route arrangement for trains to access the arrival yard under the virtual train formation operation mode, the track utilization scheme for the arrival yard suitable for virtual train formation is more effective than the conventional track utilization scheme in improving the station's throughput capacity.

[0132] Example 3

[0133] This embodiment discloses a marshalling yard arrival track utilization optimization system for virtual marshalling, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the series of steps corresponding to the above method embodiment, which includes at least:

[0134] Step S1: Construct a directed graph of the throat area of ​​the parking lot.

[0135] Step S2: Divide the arrival track bundles according to the directed graph. The routes corresponding to any two tracks in any combination within any track bundle can be converted to each other by changing the direction of a turnout.

[0136] Step S3: Determine a virtual train formation and the number of trains within it. Then, allocate the trains within the virtual train formation as described below: Select the track bundle with the closest number of tracks and trains among the currently available track bundles to arrange for train reception. When the number of trains in the arriving train's virtual train formation is less than or equal to the maximum capacity of the selected track bundle, directly arrange for train reception using that selected track bundle. If the number of trains in the arriving train's virtual train formation exceeds the maximum capacity of the selected track bundle, connect one train to each available track of the selected track bundle in sequence. Then, remove the excess trains from the virtual train formation mode and re-determine whether a new virtual train formation can be formed. Repeat this process to complete the allocation of receiving tracks and routes for each virtual train formation within the phase plan.

[0137] The specific implementation of each of the above steps can be referred to in Embodiments 1 and 2, and will not be repeated here. Among them, the virtual formation is to enable the lead car to lead the entire train group and send control commands to other trains. In the same virtual formation, two adjacent trains in the same direction of operation are divided into a leading car and a trailing car. The trailing car shares information with the leading car and maintains a certain tracking interval to follow the leading car.

[0138] Similarly, the corresponding methods executed by the system in this embodiment may also include any one or any combination of the following:

[0139] For trains not in virtual formations, track allocation is performed in the traditional manner.

[0140] When assigning trains in a virtual formation, if the route of the non-leading car in the virtual formation needs to be changed by at least two turnout directions based on the route of the adjacent train, the traditional station entry method shall be used to receive the train.

[0141] When allocating trains in virtual train formations, the objective optimization model aims to minimize the average time that trains occupy the throat area within the phase plan. Furthermore, the objective optimization model includes constraints on track availability, track occupancy uniqueness, track occupancy safety, autonomous train route control, and traditional station entry route interlocking. Specifically, the autonomous train route control constraint states that when a non-first train in a virtual train formation meets the autonomous route control conditions, it can enter the throat area without waiting for the preceding train's route to unlock, and then enter the designated receiving track by changing the direction of a single turnout.

[0142] In summary, the optimization methods and systems for the arrival yard track utilization of marshalling yards for virtual marshalling, as disclosed in the above embodiments of the present invention, improve the throughput capacity of the arrival yard throat of marshalling yards within the phased plan compared with conventional track utilization schemes, that is, they can better improve the throughput capacity of the station.

[0143] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A marshalling yard arrival field track assignment optimization method oriented to virtual marshalling, characterized in that, The method comprises the following steps: constructing a throat area directed graph; dividing arriving yard track beams according to the directed graph, and converting the routes of any two tracks in any combination in a track beam to each other by changing the direction of a turnout; determining a virtual marshalling and the number of trains in the virtual marshalling, and performing the following allocation on the trains in the virtual marshalling: selecting a track beam with the closest number of idle tracks to the number of trains in the current idle track beams to arrange the arrival of trains, and when the number of trains in the virtual marshalling is less than or equal to the maximum number of trains that can be accommodated in the selected track beam, directly arranging the arrival of trains in the selected track beam; if the number of trains in the virtual marshalling exceeds the maximum number of trains that can be accommodated in the selected track beam, arranging one train in each idle track in the selected track beam in the order of priority, and then re-determining whether a new virtual marshalling can be formed after the excess trains and the preceding train are released from the virtual marshalling mode, and so on, to complete the allocation of the arrival tracks and routes of the trains in each virtual marshalling in the stage plan; wherein the virtual marshalling is a marshalling that meets the requirement that the first train leads the entire train group and sends control instructions to other trains, and two adjacent trains in the same virtual marshalling are divided into a preceding train and a following train in the same running direction, the following train shares information with the preceding train and maintains a certain tracking interval to follow the preceding train; when performing the allocation of the virtual marshalling trains, the target optimization model is optimized to minimize the average time of the arriving trains in the stage plan occupying the throat area; the target optimization model is specifically: ； wherein, is a 0-1 variable, 1 if train i needs to be scheduled an interlocked route, otherwise 0, is the time for train i to be scheduled a route and to be released from the route, is the time for train i to run through the throat area of the station, L is the set of trains arriving at the station in the stage plan, L = {1, 2,..., i,..., n}, Z is the average time for trains arriving at the station in the stage plan to occupy the throat area, n is the number of trains arriving at the station in the stage plan; in the target optimization model, there are track availability constraints, track occupation uniqueness constraints, track occupation safety constraints, train autonomous route control constraints, and traditional arriving route interlocking constraints; wherein the train autonomous route control constraint is specifically: when the non-first train to be received in the virtual marshalling train group meets the autonomous route control condition, the non-first train can enter the throat area without waiting for the preceding train route to be unlocked, and enter the designated arrival track by switching the direction of a single turnout.

2. The method of claim 1, wherein, The method further comprises: allocating tracks to non-virtual marshalling trains in a traditional manner.

3. The method of claim 1, wherein, The method further comprises: when performing the allocation of the virtual marshalling trains, if the route of the non-first train in the virtual marshalling needs to switch the direction of at least two turnouts based on the route of the immediately preceding train, the train is received in a traditional arriving manner.

4. A marshalling yard arrival field track assignment optimization system oriented to virtual marshalling, comprising a memory, a processor and a computer program stored in the memory and capable of running on the processor, characterized in that, The processor executes the computer program to implement the method of any one of claims 1 to 3.

Images