Rail transit operation scheduling method based on multiple agents

Abstract

The invention relates to the field of rail transit operation scheduling, in particular to a multi-agent-based rail transit operation scheduling method. Comprising the following steps: constructing a multi-agent system; running state information of a train is obtained in real time based on the train intelligent agent, and passenger flow information of a station is collected based on the station intelligent agent; the dispatching center intelligent body generates a dispatching instruction according to the running state information and the passenger flow information and a preset optimization algorithm, and sends the dispatching instruction to the train intelligent body and the station intelligent body; and the train intelligent agent and the station intelligent agent adjust the operation state and the service strategy according to the received scheduling instruction. Through the constructed multi-agent system, the train operation state and the station passenger flow information are collected and transmitted to the dispatching center agent, the dispatching instruction is generated through the preset optimization algorithm, the train and the station agent are guided to adjust the operation state and the service strategy, efficient collaborative dispatching of rail transit is achieved, and the dispatching efficiency is improved. The problem that traditional scheduling is difficult to adapt to a complex operation environment is solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of rail transit operation scheduling, and in particular to a rail transit operation scheduling method based on multi-agent. BACKGROUND

[0002] With the acceleration of urbanization, urban rail transit systems are facing increasing passenger flow pressure and operational complexity. Traditional rail transit operation scheduling methods have many shortcomings.

[0003] Firstly, in the face of complex and variable passenger flow situations, such as sudden large passenger flow during peak hours, local passenger flow surge caused by special events, etc., traditional fixed timetables or scheduling methods based on simple rules are difficult to respond flexibly and efficiently, and cannot adjust the transport capacity in time, which may cause passengers to wait for a long time, overcrowding in carriages, and other problems.

[0004] Secondly, traditional scheduling methods have poor coordination between lines when dealing with complex networks formed by multiple lines. They often only focus on the operation efficiency of a single line, ignoring the mutual influence of train operation between different lines, and are difficult to achieve global optimal scheduling of the entire rail transit network.

[0005] Furthermore, traditional scheduling methods cannot simultaneously consider multiple optimization objectives, such as reducing energy consumption while pursuing train punctuality, or increasing passenger service quality while significantly increasing operating costs. For example, in some large city rail transit systems, the passenger flow at some transfer stations during the morning peak period is huge. Traditional scheduling methods cannot timely increase the train frequency in this area, resulting in long waiting time for passengers at the platform, and even safety hazards. Moreover, the connection time of trains from different lines at transfer stations is unreasonable, causing a large number of passengers to miss the transfer, reducing the efficiency of the entire rail transit system and passenger satisfaction.

[0006] In summary, the main technical problem in the prior art is that traditional rail transit operation scheduling is difficult to adapt to complex and variable operating environments and cannot achieve efficient and coordinated scheduling. SUMMARY

[0007] In order to overcome the technical defects of the traditional rail transit operation scheduling in the prior art, which is difficult to adapt to complex and variable operating environments and cannot achieve efficient and coordinated scheduling, the purpose of the present application is to provide a rail transit operation scheduling method based on multi-agent, by constructing a multi-agent system composed of train agents, station agents and scheduling center agents, collecting train operation status and station passenger flow information and transmitting them to the scheduling center agent, generating scheduling instructions through a preset optimization algorithm, guiding the train and station agents to adjust the operation status and service strategy, and achieving efficient and coordinated scheduling of rail transit, thereby solving the problem that traditional scheduling is difficult to adapt to complex operating environments.

[0008] The application discloses a kind of based on multi-agent track traffic operation scheduling method, comprising the following steps:

[0009] Multi-agent system is constructed, and multi-agent system includes train agent, station agent and dispatching center agent;

[0010] Based on train agent, the running state information of train is acquired in real time, based on station agent, the passenger flow information of station is collected;

[0011] Running state information and passenger flow information are sent to dispatching center agent;

[0012] Dispatching center agent generates scheduling instruction according to running state information and passenger flow information by preset optimization algorithm, and sends scheduling instruction to train agent and station agent;

[0013] Train agent and station agent adjust running state and service strategy according to received scheduling instruction.

[0014] Preferably, the optimization algorithm preset in dispatching center agent is multi-objective optimization algorithm, and the multi-objective in multi-objective optimization algorithm includes train punctuality, passenger waiting time, energy consumption and operating cost, and the objective function of multi-objective optimization algorithm is:

[0015]

[0016] Wherein, F is comprehensive optimization index, P is train punctuality, W is passenger average waiting time, E is the total energy consumption of multi-agent system, C is operating cost, 、 、 、 Train punctuality, passenger average waiting time, energy consumption, operating cost weight coefficient respectively, and + .

[0017] Preferably, the calculation formula of train punctuality P is:

[0018]

[0019] Wherein, The number of trains arriving at destination on time within specified time, Total number of train operation.

[0020] Preferably, the calculation formula of passenger average waiting time W is:

[0021]

[0022] Wherein, n is the total number of passengers entering station in statistical period, The waiting time for the i-th inbound passenger is the time interval between the time when the passenger enters the station and the time when the passenger boards the train.

[0023] Preferably, the multi-agent system comprises a plurality of train agents, and each train agent exchanges information through a wireless network, generates a warning message when a train agent detects that the current train has a fault or an emergency, and sends the warning message to adjacent train agents, and the adjacent train agents adjust their running speed and / or running path according to the warning message.

[0024] Preferably, the calculation formula for the adjacent train agent to adjust its running speed is:

[0025]

[0026] wherein, is the adjusted running speed of the adjacent train agent, is the running speed of the adjacent train agent before adjustment, k is the speed adjustment coefficient when the adjacent train agent adjusts the running speed, and , and the speed adjustment coefficient is determined according to the distance between the train agent generating the warning message and the adjacent train agent, and / or the severity of the fault of the train where the train agent generating the warning message is located.

[0027] Preferably, the station agent predicts the passenger flow trend within the first time range threshold according to the collected passenger flow information, and when the passenger flow within the first time range threshold is greater than the carrying capacity of the corresponding station, the station agent generates a request message and sends it to the dispatch center agent to request an increase in train frequency.

[0028] Preferably, when the station agent predicts the passenger flow trend within the first time range threshold, it calculates based on a time series analysis algorithm to construct a time series model according to the historical passenger flow information sorted by time, and predicts the passenger flow within the first time range threshold according to the time series model.

[0029] Preferably, when the dispatch center agent generates the dispatch instruction, it optimizes the train departure interval and running speed according to the passenger flow information of the transfer station and the running time of each line train for lines with transfer relationship according to the coordination relationship between different lines.

[0030] Preferably, the calculation formula for the dispatch center agent to optimize the departure interval is:

[0031]

[0032] wherein, is the optimized departure interval, the departure interval before optimization, the departure interval adjustment amount, the departure interval adjustment amount determined according to the passenger flow congestion degree of the transfer station and the operation plan of trains of each line, and when the passenger flow congestion of the transfer station is serious, the departure interval is shortened; when the passenger flow of the transfer station is smooth, the departure interval is increased.

[0033] After the above technical scheme is adopted, compared with the prior art, the present application has the beneficial effects that:

[0034] 1. The present application constructs a multi-agent system composed of train agents, station agents and dispatch center agents, collects train operation state and station passenger flow information and transmits them to the dispatch center agent, generates dispatch instructions through a preset optimization algorithm, guides the train and station agents to adjust the operation state and service strategy, realizes efficient collaborative scheduling of rail transit, and solves the problem that the traditional scheduling is difficult to adapt to complex operation environment.

[0035] 2. When generating the dispatch instructions, the dispatch center agent considers the collaborative relationship between different lines, optimizes the train departure interval and running speed according to the transfer station passenger flow information and the train operation time of each line, and effectively improves the transfer efficiency.

[0036] 3. The optimization algorithm adopts a multi-objective optimization algorithm, and the train punctuality rate, passenger average waiting time, energy consumption and operation cost are included in the objective function, and the multi-objective balance is realized by setting different weight coefficients.

[0037] 4. Multiple train agents interact information in real time through a wireless communication network, and when a train agent detects a fault or a sudden situation, it can send a warning message to the adjacent train agent to ensure the safe operation of the line.

[0038] 5. The station agent uses a time series analysis algorithm to predict the passenger flow trend within the first time range threshold, and when it is predicted that the passenger flow will be greater than the carrying capacity of the station, it timely sends a request to increase the number of classes to the dispatch center agent to avoid station congestion and improve passenger travel experience. BRIEF DESCRIPTION OF DRAWINGS

[0039] Figure 1 A step schematic diagram of a rail transit operation scheduling method based on a multi-agent according to the present application. DETAILED DESCRIPTION

[0040] The advantages of the present application are further described below in combination with the drawings and specific embodiments.

[0041] The exemplary embodiments will be described in detail herein with reference to the attached drawings. In the following description, like reference numerals refer to like elements throughout the description. The following exemplary embodiments are described in the context of implementations described herein. These implementations are not intended to represent all implementations consistent with the present disclosure. Instead, they are merely examples consistent with some aspects of the present disclosure as detailed in the appended claims.

[0042] The terminology used in the description of the present disclosure herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the present disclosure. As used in the description of the present disclosure and the appended claims, the singular forms "a", "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It also will be understood that the term "and / or" as used herein refers to and encompasses any and all possible combinations of one or more of the associated listed items.

[0043] It will be understood that, although the terms first, second, third, etc. can be used herein to describe various information, these terms are not intended to denote a particular order or hierarchy. These terms are used merely for the sake of convenience so as to distinguish one from another. For example, a first information can be termed a second information, and, similarly, a second information can be termed a first information, without departing from the scope of the present disclosure. As used herein, the term "if' can be construed to mean "when" or "upon" or "in response to determining" terms denoting the occurrence of an action.

[0044] In the description of the present disclosure, it should be understood that the terms "longitudinal", "lateral", "upper", "lower", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", and the like, indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, and are only used to facilitate the description of the present disclosure and simplify the description, and do not indicate or imply that the device or element referred to must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as a limitation on the present disclosure.

[0045] In the description of the present disclosure, unless otherwise specified and limited, it should be noted that the terms "mounting", "connecting", "connection" should be understood broadly, for example, it can be a mechanical connection or an electrical connection, or a communication between two elements, or a direct connection, or an indirect connection through an intermediate medium, and the specific meaning of the above terms can be understood by those skilled in the art according to the specific circumstances.

[0046] In the following description, the suffixes used to represent elements such as "module", "part", or "unit" are used only for the convenience of the description of the present disclosure, and do not have a specific meaning by themselves. Therefore, "module" and "part" can be used interchangeably.

[0047] Referring to Figure 1 In the present embodiment, a multi-agent-based rail transit operation scheduling method will be described in detail. The multi-agent-based rail transit operation scheduling method comprises the following steps: constructing a multi-agent system, the multi-agent system comprising train agents, station agents and a dispatching center agent; acquiring real-time train operation state information based on the train agents, and collecting passenger flow information of stations based on the station agents; sending the operation state information and the passenger flow information to the dispatching center agent; the dispatching center agent generating a dispatching instruction according to a preset optimization algorithm based on the operation state information and the passenger flow information, and sending the dispatching instruction to the train agents and the station agents; the train agents and the station agents adjusting operation states and service strategies according to the received dispatching instruction.

[0048] In the present embodiment, a multi-agent-based rail transit operation scheduling method comprises the following steps:

[0049] Step S100: Constructing a multi-agent system, the multi-agent system comprising train agents, station agents and a dispatching center agent. The train agents are attached to trains, the station agents are attached to specific stations, and the dispatching center agent is attached to a dispatching center. The train agents, the station agents and the dispatching center agent communicate and transmit information through a wireless communication network.

[0050] Step S200: Acquiring real-time train operation state information based on the train agents, and collecting passenger flow information of stations based on the station agents, and sending the acquired operation state information and passenger flow information to the dispatching center agent by the train agents and the station agents. The operation state information collected by the train agents includes but is not limited to the position, speed and passenger capacity of the train, and the acquisition method includes but is not limited to various sensors, such as a positioning sensor for acquiring the position of the train, a speed sensor for acquiring the speed of the train, and a passenger flow sensor for counting the passenger capacity in the train compartment. The passenger flow information collected by the station agents includes but is not limited to the inbound passenger flow, outbound passenger flow and station-station passenger flow, and the acquisition method includes but is not limited to deploying passenger flow monitoring devices such as infrared induction counters and video monitoring analysis devices at the inbound and outbound entrances of the station and in the key areas of the station to acquire the inbound passenger flow, outbound passenger flow and station-station passenger flow, respectively.

[0051] Step S300: Sending the acquired operation state information and passenger flow information to the dispatching center agent.

[0052] Step S400: the dispatching center agent generates a dispatching instruction according to the received operation state information and passenger flow information and according to a preset optimization algorithm, and sends the generated dispatching instruction to the train agent and the station agent respectively. It should be noted that the dispatching center agent will perform information cleaning after receiving the operation state information and passenger flow information to remove abnormal values and incorrect information, for example, removing unreasonable speed information caused by sensor failure. Unreasonable speed information can be identified by setting a speed range threshold, for example, the speed range threshold is set to 200-250 km / h. If the speed information is 150 km / h, it is not within the set speed range threshold, and it can be removed. That is, when the speed is greater than 250 km / h or less than 200 km / h, it is identified as unreasonable speed information. It should be noted that the speed of the train is low when it departs and arrives at a station, but this is not an "unreasonable speed information" and therefore does not need to be removed. However, through the above conditions, it will be identified as "unreasonable speed information". Therefore, the speed at departure and arrival time can be increased to set the arrival time and departure time, for example, the arrival time is 8:00, and the speed is less than 200 km / h after 7:50, which is not considered "unreasonable speed information" and is not removed. For example, the departure time is 9:00, and the speed is less than 200 km / h before 9:10, which is not considered "unreasonable speed information" and is not removed. In addition, it should be noted that the uniform speed increase or decrease of the speed can also be used to determine whether the train is departing or arriving.

[0053] After cleaning, the information of different lines and different stations will be classified and stored for subsequent processing, such as associating the position information of each train on the same line and the passenger flow information of the corresponding station to analyze as a whole, thereby improving the efficiency of analysis.

[0054] Step S500: The train agent and the station agent adjust the running state and service strategy according to the received scheduling instruction. Specifically, after receiving the scheduling instruction, the train agent and the station agent analyze the scheduling instruction according to their respective functions and operation mechanisms, and perform corresponding operations to realize the coordinated scheduling and efficient operation of rail transit. More specifically, after receiving the scheduling instruction through the wireless communication transmission network, the train agent and the station agent first perform format verification and decoding on the scheduling instruction to ensure the accuracy and integrity of the scheduling instruction. For example, when the train agent receives a scheduling instruction containing speed adjustment, running path change, etc., it will parse the specific adjustment parameters according to the preset data protocol; after the station agent receives a scheduling instruction to adjust the passenger guidance scheme, increase the staff configuration, etc., it will also specify the specific service strategy adjustment content involved in the instruction.

[0055] Further, the optimization algorithm preset in the scheduling center agent is a multi-objective optimization algorithm, and the multiple objectives in the multi-objective optimization algorithm include train punctuality rate, passenger waiting time, energy consumption, and operating cost. The objective function of the multi-objective optimization algorithm is: . Wherein, F is a comprehensive optimization index, P is a train punctuality rate, W is a passenger average waiting time, E is a total energy consumption of a multi-agent system, C is an operating cost, , , , are weight coefficients of the train punctuality rate, the passenger average waiting time, the energy consumption, and the operating cost, respectively, and .

[0056] In this embodiment, the optimization algorithm preset in the scheduling center agent in the above-mentioned embodiment step S400 will be described in detail. The optimization algorithm preset in the scheduling center agent is a multi-objective optimization algorithm, and the multiple objectives in the multi-objective optimization algorithm refer to the train punctuality rate, the passenger waiting time, the energy consumption, and the operating cost, which comprehensively balance multiple key factors in rail transit operation. The train punctuality rate directly affects the passenger travel experience and the reliability of the entire rail transit; the passenger waiting time is related to the passenger satisfaction; the energy consumption and the operating cost are closely related to the economic benefits and sustainable development. The objective function of the multi-objective optimization algorithm is specifically: . Wherein, F is a comprehensive optimization index, P is a train punctuality rate, W is a passenger average waiting time, E is a total energy consumption of a multi-agent system, C is an operating cost, , , , are weight coefficients of the train punctuality rate, the passenger average waiting time, the energy consumption, and the operating cost, respectively, and .​​

[0057] It should be noted that the weight coefficients 、 、 、 can be adjusted according to different operation needs and scenarios. For example, in a specific embodiment, during the peak hours of weekdays, in order to improve the service quality of passengers, the weight coefficient of the average waiting time of passengers can be increased, so that the optimization algorithm pays more attention to reducing the waiting time of passengers when generating the scheduling instructions. If, in a certain time period, according to historical information and real-time monitoring, it is determined that the punctuality rate P of the train is 0.9 (i.e., 90% of the trains can run on time), the average waiting time W of the passengers is 5 minutes, the total energy consumption E of the multi-agent system is 1000 degrees, the operation cost C is 50000, and the weight coefficients 、 、 、 are set, then the comprehensive optimization index F = 0.3 x 0.9 + 0.4 x 5 + 0.1 x 1000 + 0.2 x 50000 = 10102.27 can be obtained by substituting the numerical values into the above objective function. By continuously adjusting the scheduling strategy, the comprehensive optimization index F is gradually optimized, so as to balance multiple goals.

[0058] Further, the calculation formula of the train punctuality rate P is: wherein is the number of trains that arrive at the destination on time within a specified time, is the total number of trains.

[0059] In this embodiment, the punctuality rate P mentioned in the above embodiment will be described in detail. The punctuality rate P is calculated according to the calculation formula wherein is the number of trains that arrive at the destination on time within a specified time, is the total number of trains. For example, in some embodiments, in the rail transit operation of a certain day, the total number of trains is 500, and the number of trains that arrive at the destination on time within a specified time is 450, then the punctuality rate is calculated according to the above calculation formula , i.e., the punctuality rate of the trains on that day is 90%. This punctuality rate can intuitively reflect the on-time degree of train operation, and provide accurate basic data for the multi-objective optimization algorithm, so as to better consider the factor of train punctuality when generating the scheduling instructions.

[0060] Further, the calculation formula of the average waiting time of passengers W is: wherein n is the total number of passengers who enter the station within a statistical time period, is the waiting time of the i-th passenger, and the waiting time is the time interval between the time when the passenger enters the station and the time when the passenger boards the train.

[0061] In the present embodiment, the passenger average waiting time W described in the above embodiment will be described in detail. In the present embodiment, the calculation method of the passenger average waiting time W is clarified, and the overall waiting situation of the passengers at the station can be reflected by the statistics and average calculation of the waiting time of each passenger entering the station, and the calculation formula is , wherein n is the total number of passengers entering the station in the statistics time period, is the waiting time of the i-th passenger, and the waiting time is the time interval between the time when the passenger enters the station and the time when the passenger boards the train. For example, in some embodiments, in a one-hour statistics time period at a certain station, the total number of passengers entering the station n = 1000, and the time when each of the 1000 passengers enters the station and the time when each of the 1000 passengers boards the train are recorded one by one, so as to calculate the waiting time of each passenger . If the 1000 waiting times are added together, then , the passenger average waiting time W can be calculated according to the calculation formula . The passenger average waiting time makes the generated dispatching instructions more in line with the actual needs of the passengers when the dispatching center agent generates the dispatching instructions.

[0062] Further, the multi-agent system includes a plurality of train agents, and each train agent exchanges information through a wireless network. When a train agent detects that the current train has a fault or encounters an emergency, the train agent generates a warning information and sends the warning information to adjacent train agents. The adjacent train agents adjust their running speed and / or running path according to the warning information.

[0063] In the present embodiment, the multi-agent system described above will be described in detail again. In the multi-agent system in the present embodiment, the train agents include a plurality of train agents, the number of which is equal to the number of actually running trains. Each of the actually running trains includes a train agent. Meanwhile, the platform agents in the multi-agent system also include a plurality of platform agents, the number of which is equal to the number of actual platforms. Each of the actual platforms is provided with a platform agent.

[0064] ​The train agents on each of the above actual running trains interact information through wireless network, which can quickly share running state and other information. When a train has a fault, such as equipment failure causing the train to be unable to accelerate normally, or a sudden situation, such as foreign matter on the front track, the train agent on the train will generate a warning information according to the situation, and send the generated warning information to the adjacent train agent through the wireless network. When the adjacent train agent receives the warning information, it will adjust the running speed and running path according to the generated warning information to avoid accidents such as collision and maintain the continuous operation of the entire rail transit. For example, in a specific embodiment, there are three trains A, B and C on a rail transit line in order, and there are train agent a, train agent b and train agent c on train A, train B and train C respectively. When the above three trains are running, train A detects that its braking system is faulty and cannot brake normally, then train agent a on train A generates a warning information according to "cannot brake normally", and train agent a sends the warning information to train agent b through the wireless network. After train agent b receives the warning information sent by train agent a, it adjusts the running speed or adjusts the running path according to the warning information, and train agent b also generates a warning information to send to train agent a and train agent c to inform train agent a and train agent c of the choices they have made, so that train agent a and train agent c can make corresponding responses. If train B detects that its braking system is faulty and cannot brake normally, then train agent b on train B generates a warning information according to this "cannot brake normally", and train agent c needs to send the warning information to train agent a and train agent c. After train agent a and train agent c make choices, they also feedback to train agent b. If train C detects that its braking system is faulty and cannot brake normally, then train agent c on train C generates a warning information according to this "cannot brake normally", and train agent c sends the warning information to train agent b. Train agent makes choices and still feedbacks to c, and sends a warning information to train agent.

[0065] It should be noted that in the above embodiment, when train agent a generates a warning information, train agent a can send a warning information to train agent c after train B drives away from the current rail transit route; similarly, when train agent c generates a warning information, train agent c can send a warning information to train agent a after train B drives away from the current rail transit route.

[0066] It should be noted that in another specific embodiment, if the train agent a generates the early warning information, the early warning information is directly sent to all trains running on the rail transit line.

[0067] Further, the calculation formula for the adjacent train agent to adjust its running speed is: wherein, is the adjusted running speed of the adjacent train agent, is the running speed of the adjacent train agent before adjustment, k is the speed adjustment coefficient when the adjacent train agent adjusts the running speed, and , and the speed adjustment coefficient is determined according to the distance between the train agent generating the early warning information and the adjacent train agent, and / or the severity of the fault of the train where the train agent generating the early warning information is located.

[0068] In this embodiment, the adjustment of the running speed of the adjacent train agent according to the early warning information in the above-mentioned embodiment will be described in detail. For example, in the above-mentioned embodiment, after the train agent b on train B sends the early warning information to the train agent a and the train agent c, since train A is in front of train B, the running speed of train A cannot be lower than that of train B for safety, and therefore when the train agent a calculates the required running speed of train A, it is calculated based on the formula . For example, k is 0.5. Assuming that the running speed of train B is 60 km / h, then according to the formula , it is calculated that the running speed of train A should be 60 km / h x (1+0.5) = 90 km / h to maintain a safe distance from train B. Conversely, since C is behind train B, train C needs to maintain a safe distance from train B and can only slow down, and therefore the train agent c calculates the running speed of train C by the formula . For example, k is still 0.5. Assuming that train B still runs at a speed of 60 km / h, then according to the calculation of the running speed of train C, the train agent c needs to be reduced to 60 km / h x (1-0.5) = 30 km / h to maintain a safe distance from train C.

[0069] Further, the station agent predicts the passenger flow trend within the first time range threshold according to the collected passenger flow information, and when the passenger flow within the first time range threshold is greater than the carrying capacity of the corresponding station, the station agent generates a request information and sends it to the dispatch center agent to request an increase in train frequency.

[0070] In this embodiment, the station agent in the above embodiment will be described in detail again. When the station agent predicts the passenger flow trend in the first time range threshold in the future according to the collected passenger flow information, and predicts that the carrying capacity of the station will be exceeded, the station agent will generate a request information and send it to the dispatch center agent, and the dispatch center agent will reasonably increase the train frequency according to the actual prediction in the request information. For example, in a specific embodiment, a station agent d is deployed in station D. By analyzing the passenger flow data of the same time period every day in the past week, combined with the real-time information of the boarding passenger flow, the alighting passenger flow and the in-station passenger flow, etc. on the same day, the station agent d predicts that the passenger flow of the station D will reach the saturation state and exceed the carrying capacity in the next half hour, and then the station agent a generates a request information.

[0071] Further, when the station agent predicts the passenger flow trend in the first time range threshold, the calculation is based on the time series analysis algorithm, so as to construct a time series model according to the historical passenger flow information sorted by time, and predict the passenger flow in the first time range threshold according to the time series model.

[0072] In this embodiment, the specific analysis method of the station agent in the above embodiment to predict the passenger flow trend will be described in detail. The station agent calculates based on the time series analysis algorithm to obtain the passenger flow trend. The time series analysis algorithm is a data analysis algorithm suitable for processing data sorted by time sequence. The station agent sorts and arranges the historical passenger flow information collected by the mobile phone, such as the boarding passenger flow and the alighting passenger flow in different time periods every day in the past week, according to the time sequence, and constructs a time series model. For example, but not limited to, ARIMA (autoregressive integrated moving average model) is used to construct a time series model. Through fitting and parameter estimation of historical passenger flow information, the autoregressive integrated moving average model can achieve the change rule of passenger flow, including trend, seasonality and other characteristics. Then, by inputting the future time information into the constructed time series model, the passenger flow data in the future period can be predicted. In some specific embodiments, assuming that the time series analysis model predicts that the boarding passenger flow of a station between 9:00 and 10:00 the next day will increase by 30%, the station agent will generate a request information according to this prediction result, and the dispatch center will select to increase the appropriate frequency according to the request information and the increase of 30%.

[0073] Further, when the dispatch center agent generates the dispatch instruction, according to the coordination relationship between different lines, for the lines with transfer relationship, the dispatch center agent optimizes the train departure interval and running speed according to the passenger flow information of the transfer station and the operation time of each line train.

[0074] In this embodiment, the dispatch center agent in the above embodiment will be described in detail again. In actual rail transit, the transfer efficiency between different lines directly affects the efficiency of the entire traffic operation and the passenger experience. The dispatch center agent collects passenger flow information at the transfer station, including transfer passenger flow, transfer time distribution, etc., as well as train operation time of each line, etc., and adjusts the train departure interval and running speed by using the transfer optimization algorithm. For example, in a certain urban rail transit, line A and line B have a transfer relationship at a transfer station. The dispatch center agent analyzes the passenger flow information of the transfer station and finds that when a train of line A arrives at the transfer station, the corresponding transfer train of line B still has a long time to arrive, resulting in a large number of passengers waiting at the transfer station. Therefore, the dispatch center agent appropriately shortens the departure interval of the subsequent train on line A according to this situation, and at the same time increases the running speed of the corresponding train on line B, so that the trains of the two lines can better connect at the transfer station, improving the transfer efficiency and reducing the waiting time of passengers.

[0075] Further, the calculation formula of the dispatch center agent optimizing the departure interval is: . Wherein, is the optimized departure interval, is the departure interval before optimization, is the departure interval adjustment amount, and the departure interval adjustment amount is determined according to the passenger flow congestion degree of the transfer station and the running plan of each line train, and when the passenger flow congestion of the transfer station is serious, , the departure interval is shortened; when the passenger flow of the transfer station is smooth, , the departure interval is increased.

[0076] In this example, the transfer optimization algorithm in the above dispatch center will be described in detail. In this embodiment, the calculation formula optimizes the departure interval time, wherein is the optimized departure interval, is the departure interval before optimization, is the departure interval adjustment amount, and the departure interval adjustment amount is determined according to the passenger flow congestion degree of the transfer station and the running plan of each line train, and when the passenger flow congestion of the transfer station is serious, , the departure interval is shortened; when the passenger flow of the transfer station is smooth, , the departure interval is increased. This calculation formula is based on the original departure interval as the basis, by introducing the departure interval adjustment amount , the dynamic optimization of the train departure interval is realized. Wherein, the original departure interval is the departure interval preset according to the daily train operation plan, line capacity and other factors; As a core adjustment parameter, its value is determined by the degree of passenger congestion at transfer stations and the train operation plans of each line. For example, during normal operating hours, the original departure interval of a certain line... The timeframe is set to 8 minutes. As passenger flow changes, the dispatch center's intelligent agent calculates... The optimized departure interval is then obtained through calculation. For example, when passenger congestion at transfer stations is severe, it means a large number of passengers are gathered at the transfer station, requiring more trains to disperse the crowd. At this time... This means shortening the departure interval. For example, during the morning rush hour on weekdays (5 PM to 7 PM), a certain transfer station experiences a surge in passenger flow due to the concentration of large commercial and office areas nearby, leading to platform congestion. The dispatch center's intelligent agent, through passenger flow information uploaded by the station's intelligent agent, determines that the transfer station is in a state of verified congestion and calculates... =-2min, if If set to 8 minutes, then =8 + (-2) = 6min, so the optimized departure interval is 6min. Conversely, if the transfer station has low passenger flow, such as at night (10 PM to 11 PM), the calculated interval is... =3, then =8+3=11min, so the optimized departure interval is 11min to reduce unnecessary train operations.

[0077] It should be noted that the embodiments of the present invention have better implementability and are not intended to limit the present invention in any way. Any person skilled in the art may use the above-disclosed technical content to change or modify it into equivalent effective embodiments. However, any modifications or equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall still fall within the scope of the technical solution of the present invention.

Claims

1. A multi-agent-based rail transit operation scheduling method, characterized in that, Includes the following steps: Construct a multi-agent system, which includes a train agent, a station agent, and a dispatch center agent; The train intelligent agent acquires real-time train operation status information, and the station intelligent agent collects station passenger flow information. The operational status information and the passenger flow information are sent to the dispatch center intelligent agent; The dispatch center intelligent agent generates dispatch instructions based on the operating status information and the passenger flow information using a preset optimization algorithm, and sends the dispatch instructions to the train intelligent agent and the station intelligent agent; The train agent and the station agent adjust their operating status and service strategies according to the received scheduling instructions.

2. The rail transit operation scheduling method based on multi-agent technology according to claim 1, characterized in that, The optimization algorithm preset in the dispatch center agent is a multi-objective optimization algorithm. The multiple objectives in the multi-objective optimization algorithm include train punctuality rate, passenger waiting time, energy consumption, and operating cost. The objective function of the multi-objective optimization algorithm is: ， Where F is the comprehensive optimization index, P is the train punctuality rate, W is the average passenger waiting time, E is the total energy consumption of the multi-agent system, and C is the operating cost. , , , These are the weighting coefficients for the train punctuality rate, average passenger waiting time, energy consumption, and operating costs, respectively. + .

3. The rail transit operation scheduling method based on multi-agent technology according to claim 2, characterized in that, The formula for calculating the train punctuality rate P is: ， in, The number of trains that arrive at their destination on time within the specified time. This represents the total number of trains in operation.

4. The rail transit operation scheduling method based on multi-agent technology according to claim 2, characterized in that, The formula for calculating the average passenger waiting time W is: ， Where n is the total number of passengers entering the station during the statistical period. Let be the waiting time for the i-th passenger entering the station, where the waiting time is the time interval between the passenger's entry time and the time they board the train.

5. The rail transit operation scheduling method based on multi-agent technology according to claim 1, characterized in that, The multi-agent system includes multiple train agents, and each train agent interacts with other train agents through a wireless network. When a train agent detects a malfunction or an emergency in the current train, it generates an early warning message and sends the message to neighboring train agents. The neighboring train agents adjust their speed and / or path according to the warning message.

6. The rail transit operation scheduling method based on multi-agent technology according to claim 5, characterized in that, The formula for calculating the speed adjustment of adjacent train agents is as follows: ， in, The adjusted operating speed for the adjacent train agents. Let k be the speed of the adjacent train agent before adjustment, and k be the speed adjustment coefficient when the adjacent train agent adjusts its speed. The speed adjustment coefficient is determined based on the distance between the train agent that generated the warning information and the adjacent train agents, and / or the severity of the fault of the train to which the train agent that generated the warning information is located.

7. The rail transit operation scheduling method based on multi-agent technology according to claim 1, characterized in that, The station agent predicts the passenger flow trend within the first time range threshold based on the collected passenger flow information. When the passenger flow within the first time range threshold is greater than the carrying capacity of the corresponding station, the station agent generates a request message and sends it to the dispatch center agent to request an increase in train services.

8. The rail transit operation scheduling method based on multi-agent technology according to claim 7, characterized in that, When the station agent predicts the passenger flow trend within the first time range threshold, it performs calculations based on a time series analysis algorithm to construct a time series model after sorting historical passenger flow information by time, and then predicts the passenger flow within the first time range threshold based on the time series model.

9. The rail transit operation scheduling method based on multi-agent technology according to claim 1, characterized in that, When generating the scheduling instructions, the intelligent agent in the scheduling center optimizes the train departure interval and running speed based on the coordination relationship between different lines. For lines with transfer relationships, it optimizes the train departure interval and running speed based on the passenger flow information of the transfer stations and the running times of trains on each line.

10. The rail transit operation scheduling method based on multi-agent technology according to claim 9, characterized in that, The formula for the dispatch center agent to optimize the departure interval is: ， in, For the optimized departure interval, The departure interval before optimization. The departure interval adjustment amount, the departure interval adjustment amount The timing is determined based on the level of passenger congestion at transfer stations and the train operation plans for each line. Furthermore, when passenger congestion at transfer stations is severe... To shorten the departure interval; when passenger flow at transfer stations is smooth, Increase departure intervals.

Images