A railway train dispatching method based on autonomous collaboration

Abstract

A railway traffic dispatching method based on autonomous collaboration includes the following steps: inputting train operation plans and emergency event information; establishing dispatching subgroups and determining key information involved in the group; establishing a distributed optimization model for the dispatching agent group; designing a distributed collaborative method for agents; and forming a train operation plan within the jurisdiction of the entire dispatching center. In response to the interference of emergency events, the present invention establishes dispatching subgroups corresponding to different stations and an optimization model for the dispatching agent group. Through the distributed collaborative method for agents, the collaborative solution of the receiving and dispatching operations of different stations in the group is achieved, and the entire operation adjustment plan is obtained by merging. This method decomposes a large-scale optimization problem into multiple small-scale sub-problems for decentralized solution, which can effectively improve the solution efficiency.

Description

Technical Field

[0001] The present invention belongs to the technical field of railway traffic dispatching and commanding, and relates to a railway traffic dispatching method based on autonomous collaboration. Background Art

[0002] Train dispatching is the core of high-speed railway operations and a key technology for ensuring safe and efficient train operation. Currently, railway dispatching and command systems primarily utilize a centralized approach, with the dispatching center issuing phased plans to stations. Stations then passively execute these phased plans to generate route control instructions, thus implementing train dispatching and command. When external emergencies (including infrastructure failures, severe wind, rain, and snow), resulting in conflicts in train operating resources and disrupting scheduled operations, dispatchers adjust train operation plans based on their experience. The dispatching and command system lacks the autonomous decision-making capabilities to proactively adjust station reception and dispatch operations based on changes in train status and operating conditions. Therefore, autonomous and coordinated decision-making regarding train operation plans between stations and the dispatching center is needed to further improve train operation efficiency. Summary of the Invention

[0003] The purpose of the present invention is to provide a railway traffic scheduling method based on autonomous collaboration, which can effectively realize the autonomous collaborative scheduling of railway traffic under emergency interference, and its technical solution is as follows;

[0004] A railway train dispatching method based on autonomous collaboration includes the following steps:

[0005] Step 1: Input train operation plan and emergency information, including train delay duration, infrastructure failure and environmental disturbance data;

[0006] Step 2: Based on the train set N and station set J in the current scheduling section, multiple scheduling subgroups are established. Each scheduling subgroup corresponds to a master station agent and its associated neighbor station agents. The scheduling subgroup is used to dynamically obtain train operation status information and determine the train scheduling tasks within the group.

[0007] Step 3: For each scheduling subgroup, a distributed optimization model is established. The model takes minimizing the total train delay time and the number of arrival and departure line adjustments as the optimization goal, and the comprehensive optimization goal is calculated through weighted sum;

[0008] Step 4: Based on the distributed optimization model, design agent coordination rules, including priority determination for receiving and dispatching operations, dynamic allocation of arrival and dispatch lines, and inter-group data interaction mechanisms, and generate local adjustment plans;

[0009] Step 5: Merge the local adjustment plans of all scheduling subgroups, update the global train operation plan, and achieve dynamic optimization through event triggering or timed synchronization mechanism.

[0010] The present invention also discloses a non-volatile storage medium, characterized in that it stores a computer program, which, when executed by a processor, implements the above-mentioned railway traffic scheduling method based on autonomous collaboration, wherein the method generates a train operation plan by dynamically establishing scheduling subgroups and distributed optimization models.

[0011] The present invention also discloses an electronic device, characterized in that it includes a processor and a memory, the memory stores a computer program, and when the processor executes the computer program, it adjusts the railway traffic plan in real time based on the above-mentioned autonomous and collaborative railway traffic scheduling method.

[0012] Beneficial effects

[0013] To address the disruptions caused by emergencies, this paper establishes dispatch subgroups corresponding to different stations and builds an optimization model for these dispatching agents. Through a distributed collaborative approach, this approach achieves a coordinated solution for train pickup and dispatch operations at different stations within the group, merging them into a comprehensive operational adjustment plan. This method decomposes a large-scale optimization problem into multiple smaller subproblems for decentralized solution, effectively improving solution efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 It is a flow chart of a railway train dispatching method based on autonomous collaboration;

[0015] Figure 2 It is a schematic diagram of the train receiving and dispatching operations involved in different station dispatching groups. DETAILED DESCRIPTION

[0016] A railway train dispatching method based on autonomous collaboration, such as Figure 1 As shown, the following steps are included:

[0017] Step 1: Input train operation plan and emergency information.

[0018] The train operation plan is key information for train dispatching. It defines the train schedule, including arrival and departure times at different stations, station arrival and departure line utilization, the number of trains, the number of stations, the minimum safe interval between stops, the minimum operating time between sections, and the minimum tracking interval between trains. Unexpected events can disrupt train operation according to the planned schedule, leading to deviations in train operation status and delays. Specific emergency information includes delayed trains and their duration (including arrival and departure delays).

[0019] Step 2: Establish a dispatching subgroup and determine the key information involved in the group, including train operation status information, train dispatching tasks, station intelligent entities, etc.

[0020] Autonomous and collaborative railway train dispatching uses stations as individuals (also known as station intelligent entities). Each station intelligent entity completes functions such as route execution and active adjustment, and interacts with other station intelligent entities in the group to complete collaborative decision-making within the group.

[0021] Step 2.1: For a certain dispatching section, N and J are the train set and station set respectively. Each station corresponds to a scheduling subgroup. This group contains multiple station agents, including the current master station agent and neighboring station agents. Station agents in different groups may overlap.

[0022] Step 2.2: Obtain train operation status information, including the time the train runs in the section and the time it stops at stations.

[0023] Step 2.3: Determine the train dispatching task and dispatching group range. Set the station dispatching time range T, for station j, according to the first unfinished train receiving and dispatching operation time The time period corresponding to the scheduling time range T The trains involved in the station determine the number of trains and route information involved in each station agent. At the same time, each dispatching group will The other stations that the trains involved in the master station agent pass through are determined to belong to the neighbor station agents in the group. For group j (master station agent j), the set of trains involved is , the station collection is . Figure 2 This is a schematic diagram of the train receiving and dispatching operations involved in four station dispatching groups. The corresponding solid-line boxes are the stations, trains and station operations involved in the groups.

[0024] Step 2.4: The dispatch center can set the time trigger mode or event trigger mode. After the trigger, the group train scheduling task and the scheduling group range will be updated. For example, the station time range If a new pickup or dispatch operation is added, the group is updated and distributed collaborative optimization is performed. If a new sudden interference scenario occurs, the involved groups are updated and distributed collaborative optimization is performed.

[0025] Step 3: Establish a distributed optimization model for the scheduling agent group.

[0026] The overall decision variable is the group's train operation plan within a given timeframe. Each agent shares local optimization decision variables to jointly optimize the group's overall objective function. The agent's local decision variables are the route control information state for train reception and dispatch operations at the corresponding station, including train arrival and departure times, sequence, and arrival and departure track allocations.

[0027] The model is constructed by establishing an individual utility value evaluation method (such as the total train delay time, the number of arrival and departure line adjustments, etc.), calculating the local optimization objective function, and considering the relevant constraints of the station arrival and departure lines.

[0028] Step 3.1: Design the optimization objective of a single station agent.

[0029] (1) Minimize the total train delay time. For station j (master station agent j) in group j, the total train delay time involved is Minimum.

[0030] (1)

[0031] Where, , is the actual arrival time of train i at station j, is the time when train i actually leaves station j, The planned arrival time of train i at station j is, Schedule train i to leave station j.

[0032] (2) Minimize the number of arrival and departure line adjustments. For station j, the total number of arrival and departure lines involved in the train adjustment is Minimum.

[0033] (2)

[0034] Where, , Is train i actually occupying the departure line at station j? , which is 1 when occupied, otherwise 0. Whether train i is scheduled to occupy the departure line at station j , which is 1 when occupied, otherwise 0. Gather at the departure line for station j.

[0035] For the above two optimization objectives, the comprehensive optimization objective of a single station agent (station j) is calculated by weighted sum :

[0036] (3)

[0037] Where, , w1, w2 are weights.

[0038] Step 3.2: Construct an optimization model for decision-making within the group. For group j (master station agent j), the corresponding optimization goal is The sum of the comprehensive optimization targets for the corresponding stations in the group. Station weights or train weights can be set specifically for this optimization indicator.

[0039] (4)

[0040] Optimization decisions within a group do not involve the time range within the group Train operation plans outside the group and train operation plans for corresponding stations outside the group. Therefore, other decision options are treated as constraints for the current group. Subsequent operations outside the station time range are postponed according to the planned order. For example, for downlink trains, the train operation plan before entering the stations involved in group j, including parameters such as arrival order, arrival and departure line occupancy, and arrival time, constrains the corresponding decision variables in group j. The plan after group j completes the optimization decision is used as a parameter for subsequent groups.

[0041] Step 3.3: Build a global optimization model. This model no longer performs local optimization on stations and trains within a group, but instead targets all stations and trains within the entire section. This can be used to verify the approximation of distributed computing results.

[0042] Step 4: Design a distributed collaborative method for intelligent agents. Each intelligent agent is based on the current situation information and the local optimization indicators of the decision-making within the group. Through heuristic methods, it realizes collaborative decision-making among different groups.

[0043] For train reception and dispatch operations that are not affected by interference, they are executed as planned without adjustment. Therefore, a distributed collaborative solution is performed for train reception and dispatch operations that are affected by the presence of section operation interference before the execution of train reception operations for arriving trains or station operation interference before the execution of train dispatch operations for departing trains.

[0044] Step 4.1: Consider the train in the down direction, taking station j as an example, there is a delay caused by arrival or departure interference, station The corresponding train reception and dispatching operation sequence and times are already determined and serve as known parameters. The disrupted train reception and dispatching operations at station j are considered the decision variables to be optimized. Since the order of train reception operations at station j is known, the primary concern is determining the order of dispatching operations within the overall station operations. A set of pending operations is established, and a constructive heuristic algorithm is constructed using the following rules to sequentially determine the current station's train reception and dispatching operations:

[0045] (1) The receiving operation for any train must be earlier than the departure operation;

[0046] (2) When there is no unexecuted train receiving operation at the station, the current train receiving operation is executed;

[0047] (3) If the number of train receiving operations at the current station is the same as the number of train arrival and departure lines, only the departure operation will be performed;

[0048] (4) The more the objective function (4) corresponding to a station's train receiving and dispatching operation decreases, the higher the priority of executing the train receiving and dispatching operation;

[0049] (5) For vehicle pickup operations, the planned arrival and departure lines can be used as much as possible. If the planned arrival and departure line is occupied, other unoccupied arrival and departure lines should be selected.

[0050] Step 4.2: Station reception and dispatching operations allocation method.

[0051] Step 4.2.1: When determining each operation at the station, first determine the currently executable operation type (train receiving operation or train dispatching operation) based on the set of operations to be executed.

[0052] Step 4.2.2: If only a pick-up operation is included, then this operation is directly selected as the operation assigned to the current station. This complies with the second rule in step 4.1.

[0053] Step 4.2.3: If there are currently two operations, namely, pick-up and dispatch, then according to the fourth rule in step 4.1, calculate the objective function (4) under the pick-up operation (Scheme 1) and the dispatch operation (Scheme 2 to the maximum scheme) respectively. , with a maximum of operations to be dispatched The difference between the objective function (4) under the first-come-first-served operation and the objective function (4) under the first-come-first-served operation is calculated. The larger the difference, the better the effect of the operation scheme on train delay recovery, and the scheme is selected as the operation assigned to the current station.

[0054] It should be noted that the objective function (4) also involves the station set of group j The arrival and departure times of all trains at the station. For down trains, the arrival and departure times of the train before station j have been determined. Based on the partial station operations determined in step 4.2, the station operation plan is determined by executing the current train receiving operation or train dispatching operation. Subsequent operations are calculated according to the first-come-first-served rule. The arrival and departure times of station j and subsequent stations are determined by running on time according to the train tracking interval constraints, arrival and departure line occupancy safety constraints, and minimum running time of the section, with the above-mentioned station operation order determined. The arrival and departure times of the trains are determined; the arrival and departure lines are allocated according to the fifth rule in step 4.1. The objective function calculations of all possible station operation plans are completed respectively.

[0055] Step 4.2.4: If only dispatch operations are included, calculate the dispatch operations (Scheme 1 to Scheme 2) according to Rule 3 in Step 4.1. , waiting for dispatch The difference between the objective function (4) under the first-come, first-served operation and the objective function (4) under the first-come, first-served operation is calculated. The larger the difference, the better the effect of the operation scheme on train delay recovery. The scheme is selected as the operation assigned to the current station. The specific method is similar to the departure operation part in step 4.2.3.

[0056] Step 4.2.5: Complete the station pick-up and dispatch operation allocation in station j, and then continue to calculate the station The train arrival and departure times and the arrival and departure line allocation plan for station j are used as model parameters. This is done until all stations in the entire section have completed the train arrival and departure allocation, achieving collaborative decision-making within the group.

[0057] Step 5: After completing autonomous coordination within the group, the operation plans of all station agents are combined and submitted to the dispatch center, forming a train operation plan for the entire dispatch center's jurisdiction. This global train operation plan is updated after the station agents within any group complete distributed collaborative optimization. If any group has not completed autonomous coordination within the group, the previous stage plan is used as the global train operation plan.

[0058] The above shows and describes the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions merely illustrate the principles of the present invention. Various changes and modifications may be made to the present invention without departing from the spirit and scope of the present invention. Such changes and modifications are intended to fall within the scope of the present invention. The scope of protection claimed by the present invention is defined by the appended claims and their equivalents.

Claims

1. A railway train dispatching method based on autonomous collaboration, characterized by: The steps include: Step 1: Input train operation plan and emergency information, including train delay duration, infrastructure failure and environmental disturbance data; Step 2: Based on the train set N and station set J in the current scheduling section, multiple scheduling subgroups are established. Each scheduling subgroup corresponds to a master station agent and its associated neighbor station agents. The scheduling subgroup is used to dynamically obtain train operation status information and determine the train scheduling tasks within the group. Step 2.1: For a certain dispatching section, N and J are the train set and station set respectively; set There are scheduling subgroups, that is, each station corresponds to a scheduling group; the group contains multiple station agents, including the current main station agent and neighboring station agents; station agents in different groups are repeated; Step 2.2: Obtain train operation status information: including the time the train runs in the section and the time it stops at the station; Step 2.3: Determine the train dispatching task and the dispatching group range; set the station dispatching time range T, and for station j in the group, according to the first unfinished train receiving and dispatching operation time The time period corresponding to the scheduling time range T The trains involved in the internal stations determine the number of trains and route information involved in each station agent; Each scheduling group is based on the time range The other stations that the trains involved in the main station agent pass through are determined to belong to the neighbor station agents in the group; the neighbor station agents of the scheduling subgroup are dynamically associated through the train running path; Step 2.4: The dispatch center sets the timer trigger mode or event trigger mode; Step 3: For each scheduling subgroup, a distributed optimization model is established. The model takes minimizing the total train delay time and the number of arrival and departure line adjustments as the optimization goal, and calculates the comprehensive optimization goal through weighted sum to obtain the local optimization objective function; Step 4: Based on the distributed optimization model, design the agent coordination rules, including priority determination for train pickup and dispatch operations, dynamic allocation of arrival and departure lines, and a data exchange mechanism between groups. Based on the set of pending operations, determine the currently executable operation categories. Calculate the difference between the objective function for train pickup and / or dispatch operations and the objective function for first-come, first-served operations. Select the solution with the largest difference as the current station assignment operation and generate a local adjustment solution. Step 5: Merge the local adjustment plans of all scheduling subgroups, update the global train operation plan, and achieve dynamic optimization through event triggering or timed synchronization mechanism.

2. The railway traffic dispatching method based on autonomous collaboration according to claim 1, wherein step 3 further comprises the following: The model is constructed by establishing an individual utility value evaluation method, calculating the local optimization objective function, and considering the constraints related to the station arrival and departure lines; Step 3.1: Design the optimization objective of a single station agent; (1) Minimize the total train delay time: For station j in the group, the total delay time of the train involved is smallest; （1） Where, , is the actual arrival time of train i at station j, is the time when train i actually leaves station j, The planned arrival time of train i at station j is, The train i is scheduled to leave station j at time; the train set is , G is the group set of stations; (2) Minimize the number of arrival and departure line adjustments: For station j, the total number of arrival and departure lines involved in the train adjustment Minimum, （2） Where, , Is train i actually occupying the departure line at station j? , 1 when occupied, otherwise 0; Whether train i is scheduled to occupy the departure line at station j , 1 when occupied, otherwise 0; Gather at the departure line for station j; For the above two optimization objectives, the comprehensive optimization objective of a single station agent is calculated by weighted sum : （3） Where, , w1, w2 are weights; Step 3.2: Construct an optimization model for decision-making within the group. For station j within the group, the corresponding optimization goal is is the sum of the comprehensive optimization objectives of the corresponding stations in the group; for this optimization objective, the station weight or train weight is set specifically to obtain the local optimization objective function: （4） Optimization decisions within a group do not involve the time range within the group Train operation plans outside the group and train operation plans of corresponding stations outside the group; Step 3.3: Build a global optimization model. In the global optimization model, local optimization is no longer performed based on the stations and trains within the group. Instead, the optimization is performed for all stations and trains in the entire section to verify the degree of distributed computing results.

3. The railway traffic dispatching method based on autonomous collaboration according to claim 1 is characterized by: The step 4 includes the following contents: Step 4.1: Consider the train in the down direction. Station j has a delay caused by arrival or departure interference. The corresponding train reception and dispatching operation sequence and time have been determined as known parameters. The disturbed train reception and dispatching operations at station j are used as decision variables to be optimized. Since the order of train reception operations at station j is known, the main thing to do is to determine the order of dispatching operations within the entire station operation. A set of pending operations is established, and a constructive heuristic algorithm is established using the following rules to sequentially determine the train reception and dispatching operations at the current station. (1) The receiving operation for any train must be earlier than the departure operation; (2) When there is no unexecuted train receiving operation at the station, the current train receiving operation is executed; (3) If the number of train receiving operations at the current station is the same as the number of train arrival and departure lines, only the departure operation will be performed; (4) The more the corresponding objective function of a station's train receiving and dispatching operation decreases, the higher the priority of executing the train receiving and dispatching operation; (5) For vehicle pickup operations, use the planned arrival and departure lines as much as possible. If the planned arrival and departure line is occupied, select other unoccupied arrival and departure lines; Step 4.2: Station reception and dispatching operations allocation method.

4. The method for railway traffic dispatching based on autonomous collaboration according to claim 3, wherein step 4.2 further comprises the following: Step 4.2.1: When determining each operation at the station, first determine the currently executable operation type based on the set of operations to be executed; Step 4.2.2: If it only includes the pick-up operation, directly select this operation as the operation assigned to the current station; Step 4.2.3: If both a train pick-up operation and a train departure operation are currently available, then according to the fourth rule in step 4.1, calculate the difference between the objective function under the train pick-up operation and the objective function under the train departure operation and the objective function under the first-come, first-served operation. The larger the difference, the better the effect of the operation plan on train delay recovery. The plan is then selected as the operation assigned to the current station. Step 4.2.4: If only the departure operation is included, then according to the third rule in step 4.1, calculate the difference between the objective function under the departure operation and the objective function under the first-come, first-served operation. The larger the difference, the better the effect of the operation plan on train delay recovery. Select this plan as the operation assigned to the current station. Step 4.2.5: Complete the station pick-up and dispatch operation allocation in station j, and then continue to calculate the station The train arrival and departure times and arrival and departure line allocation schemes of station j are used as model parameters; until the train arrival and departure operation allocation of all stations in the entire section is completed, collaborative decision-making within the group is achieved.

5. A non-volatile storage medium, characterized in that: A computer program is stored, and when the computer program is executed by a processor, it implements the railway traffic scheduling method based on autonomous collaboration as described in any one of claims 1 to 4, wherein the method generates a train operation plan by dynamically establishing scheduling subgroups and distributed optimization models.

6. An electronic device, characterized in that: The method comprises a processor and a memory, wherein the memory stores a computer program, and when the processor executes the computer program, the method adjusts the railway traffic plan in real time based on the railway traffic scheduling method based on autonomous collaboration according to any one of claims 1 to 4.

Images