Taxi-passenger double-end queuing optimization system and method based on differentiated information level

Abstract

The invention relates to the technical field of intelligent traffic, in particular to a differential information level-based taxi-passenger double-end queuing optimization system and method, and the system comprises a multi-level information issuing module which is used for collecting dynamic and static traffic information, classifying the information, and issuing the classified information to a driver and passengers; the queuing situation deduction module is used for simulating and analyzing queuing decision behaviors of drivers and passengers under different information levels according to the information issued by the multi-level information issuing module and deducing the queuing development situation; and the people and vehicle supply and demand management and control module is used for implementing a social optimal strategy according to the deduced queuing development situation, calculating the optimal queuing adding probability of the taxi and the passengers, and regulating and controlling the arrival rates of the taxis and the passengers. The advantages of real-time interconnection and efficient calculation of a big data platform are fully exerted, and based on differentiated and multi-level information, the double-end queuing time is shortened, the operation order of taxi getting-on points is improved, and the social benefits are greatly improved.

Description

Technical Field

[0001] The present invention relates to the field of intelligent transportation technology, and in particular to a taxi-passenger double-end queuing optimization system and method based on differentiated information levels. Background Art

[0002] Taking a taxi is one of the main modes of transportation for passengers at transportation hubs such as airports and train stations. Usually, these transportation hubs will set up special taxi stations for passengers to take, which forms a typical two-end queue system. In this system, taxis and passengers, as two types of entities, arrive at the two ends of the queue, then match and leave the system. However, many transportation hubs, such as airports and train stations, generally face the problem of long queues.

[0003] In the taxi-passenger two-end queuing system, the opacity of queuing information is an important reason for long queuing time. Since the queuing system is usually not observable in real time, drivers and passengers cannot accurately estimate the benefits of queuing. Therefore, even if the queuing time is too long, drivers and passengers may still choose to join the queue, further exacerbating the congestion of the system. Therefore, queuing information is crucial to the decision-making of drivers and passengers, affecting whether they join the queue and how to optimize their queuing strategy.

[0004] Although there is a research foundation in this field, the existing technology still has shortcomings: (1) In the context of relying on information to optimize the taxi-passenger two-end queue system, most studies only consider information such as arrival rate, queue, and expected waiting time from one side, which makes the queue system only able to obtain extensive management, and its potential for message exchange and instant response has not been fully developed. (2) The existing technology does not fully and dialectically view the role of information in the optimization of the queue system. Experience or subjective belief that the more detailed the information, the more guiding significance it has for whether to join the queue, but the existing methods do not take into account the characteristics of individuals pursuing maximum benefits, and it is very easy to produce situations where refined information brings negative effects.

[0005] Therefore, developing a taxi passenger queue operation management method that can maximize the advantages of multi-level information is of great significance to improving the order of queues at both ends and improving the overall social benefits. Such a system can clarify the optimal release scope of differentiated information, tap the maximum potential of applying information to optimize the queue system, and give clear optimization suggestions to cope with the ever-changing traffic needs and technical challenges. Summary of the invention

[0006] To solve the above problems, the present invention provides a taxi-passenger two-end queuing optimization system and method based on differentiated information levels, which is used to explore the optimization potential of various types of information for the queuing system, promote the efficient operation of taxi-passenger queuing, and improve the overall social benefits.

[0007] To achieve the above object, the present invention provides the following solutions:

[0008] A taxi-passenger dual-end queuing optimization system based on differentiated information levels, including: a multi-level information release module, a queuing situation deduction module, and a passenger and vehicle supply and demand control module;

[0009] The multi-level information release module is used to collect dynamic and static traffic information, classify the information, and release it to drivers and passengers;

[0010] The queue situation deduction module is used to simulate and analyze the queue decision-making behaviors of drivers and passengers at different information levels according to the information released by the multi-level information release module, and deduce the development trend of the queue;

[0011] The passenger and vehicle supply and demand control module is used to implement the social optimal strategy according to the deduced queue development trend, calculate the optimal probability of taxis and passengers joining the queue, and regulate the arrival rate of taxis and passengers.

[0012] Optionally, the multi-level information publishing module includes: a first data acquisition unit, a first data preprocessing unit, a data analysis unit, a data division unit, a data publishing unit

[0013] The first data collection unit is used to collect dynamic traffic information and static traffic information; wherein the dynamic traffic information includes: vehicle arrival time, vehicle customization, passenger arrival time, passenger customization, and timestamp data; the static traffic information includes: the number of boarding points, taxi queue capacity, and passenger queue capacity;

[0014] The data preprocessing unit is used to perform data cleaning and standardization processing on the dynamic traffic information and the static traffic information;

[0015] The data analysis unit is used to calculate key information indicators for the pre-processed data; wherein the key information indicators include: taxi queue length, passenger queue length, effective arrival rate, taxi waiting time, passenger waiting time, social benefits, driver benefits and passenger benefits;

[0016] The data division unit is used to divide the information into invisible information, partially visible information, and completely visible information based on the key information indicators; wherein the invisible information is no information provided, the partially visible information includes: the expected waiting time of a taxi, the waiting time of a passenger, and the completely visible information includes: the real-time waiting time of a taxi, the real-time waiting time of a passenger;

[0017] The data publishing unit is used to publish information to drivers and passengers based on the information classification results.

[0018] Optionally, the data analysis unit calculates key information indicators for the preprocessed data, including:

[0019]

[0020]

[0021]

[0022]

[0023]

[0024]

[0025]

[0026]

[0027]

[0028]

[0029] in, is the arrival rate of taxis, is the passenger arrival rate, For service intensity, In queue status, is the taxi queue capacity, The queuing capacity for passengers, is the steady-state probability, is the taxi queue length, is the passenger queue length, is the effective arrival rate, Waiting time for taxis, Waiting time for passengers, is the expected value of the indicator, For social benefits, Benefits for drivers, Benefits for passengers.

[0030] Optionally, the queue situation deduction module includes: a second data acquisition unit, a second data preprocessing unit, a benefit calculation unit, a queue decision characterization unit, and a situation deduction unit;

[0031] The second data collection unit is used to collect order fees, passenger waiting costs, taxi driver waiting costs, taxi transportation costs, and passenger income after arriving at the destination;

[0032] The second data preprocessing unit is used to perform data cleaning and standardization processing on the data collected by the second data acquisition unit;

[0033] The benefit calculation unit is used to construct utility functions for drivers and passengers respectively according to the information released by the multi-level information release module and the data preprocessed by the second data preprocessing unit;

[0034] The queuing decision representation unit is used to add a Nash equilibrium strategy to the utility function to solve the probability of different groups joining the queue;

[0035] The situation deduction unit is used to readjust the arrival rates of drivers and passengers according to the probability of joining the queue, perform situation deduction based on the Markov chain, reconstruct the queue system under the influence of information, and recalculate key information indicators;

[0036] Optionally, the utility function constructed by the benefit calculation unit includes:

[0037] The driver's utility function under partial visibility is:

[0038]

[0039] Passenger utility function under partial visibility:

[0040]

[0041] The driver's utility function under full visibility is:

[0042]

[0043] Passenger utility function under full visibility:

[0044]

[0045] in, For order fees, Waiting costs for passengers, is the taxi waiting cost, For transportation costs, For the profit after reaching the destination, The queue length is The probability of joining the queue when .

[0046] Optionally, the Nash equilibrium strategy added to the queuing decision representation unit includes:

[0047] Driver equilibrium strategy under partial visibility:

[0048]

[0049] in, and yes exist Solutions within the range;

[0050] Traveler equilibrium strategy under partial visibility:

[0051]

[0052] in, and yes exist Solutions within the range;

[0053] The driver's equilibrium strategy under full visibility:

[0054]

[0055] in, and yes exist Solutions within the range;

[0056] Traveler equilibrium strategy under partial visibility:

[0057]

[0058] in, and yes exist Solutions within the range.

[0059] Optionally, the social optimal strategy implemented by the passenger and vehicle supply and demand control module includes:

[0060] The socially optimal strategy in the partially visible case after calculating the probability of the driver joining the queue:

[0061]

[0062] The socially optimal strategy in the partially visible case after the probability of passengers joining the queue is calculated:

[0063]

[0064] The socially optimal strategy in the fully visible case after calculating the probability of drivers joining the queue is:

[0065]

[0066] The socially optimal strategy in the fully visible case after the probability of passengers joining the queue is calculated:

[0067]

[0068] in, For order fees, Waiting costs for passengers, Waiting costs for taxi drivers, For taxi transportation costs, The income of passengers after they arrive at their destination.

[0069] A taxi-passenger double-end queuing optimization method based on differentiated information levels, the method comprising:

[0070] Collect dynamic and static traffic information, classify the information, and release it to drivers and passengers;

[0071] Based on the information released by the multi-level information release module, simulate and analyze the queuing decision-making behaviors of drivers and passengers at different information levels, and deduce the development trend of the queuing;

[0072] According to the deduced queue development trend, the social optimal strategy is implemented, the optimal probability of taxi rentals and passengers joining the queue is calculated, and the arrival rate of taxis and passengers is regulated.

[0073] The beneficial effects of the present invention are:

[0074] This paper proposes a one-dimensional continuous-time Markov model, which expands the research on the customer arrival process in the queuing system by introducing the queuing probability, and can adapt to the real-time dynamic changes in the system state. The model considers the dynamic evolution process of the two-end queuing system under three information conditions: invisible, partially visible, and significantly visible. Through this method, the behavior of the queuing system can be accurately simulated and predicted, especially in complex and uncertain environments, which improves the stability and efficiency of the queuing system and scientifically guides the management department to optimize the modeling of the matching of people and vehicles at the taxi boarding point.

[0075] The present invention constructs a decision utility function and proposes an equilibrium strategy, which quantifies the impact of differentiated information on the optimization of the queuing system under different information environments. By implementing the equilibrium strategy, it can truly reflect the queuing choices of passengers / drivers when faced with differentiated information. Through this method, the optimal release range of different information can be clarified, providing an application reference for the informatization and intelligence of taxi operations.

[0076] This paper proposes a socially optimal strategy, clarifies the optimization direction of the queuing system under differentiated information levels, and deeply explores the optimization space. Through this strategy, the overall social welfare can be maximized, while ensuring system efficiency and avoiding resource waste. The socially optimal strategy provides scientific decision-making support for policy makers and operation managers, especially in high-demand environments, and can achieve optimal resource allocation. BRIEF DESCRIPTION OF THE DRAWINGS

[0077] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0078] Figure 1 A schematic diagram of a taxi-passenger double-end queuing optimization process based on differentiated information levels according to an embodiment of the present invention;

[0079] Figure 2 A state transition diagram of a queuing system without information provision according to an embodiment of the present invention;

[0080] Figure 3 A state transition diagram of a queuing system after providing an expected waiting time to a driver according to an embodiment of the present invention;

[0081] Figure 4 A state transition diagram of a queuing system after providing passengers with an expected waiting time according to an embodiment of the present invention;

[0082] Figure 5 A state transition diagram of a queuing system after providing a real-time waiting time to a driver according to an embodiment of the present invention;

[0083] Figure 6 This is a state transition diagram of a queuing system after providing passengers with real-time waiting time according to an embodiment of the present invention. DETAILED DESCRIPTION

[0084] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0085] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0086] like Figure 1 As shown, this embodiment proposes a taxi-passenger dual-end queuing optimization system based on differentiated information levels, including: a multi-level information release module, a queuing situation deduction module, and a passenger and vehicle supply and demand control module;

[0087] A multi-level information release module is used to collect dynamic and static traffic information, classify the information, and release it to drivers and passengers;

[0088] The queuing situation deduction module is used to simulate and analyze the queuing decision-making behaviors of drivers and passengers at different information levels based on the information released by the multi-level information release module, and deduce the development trend of the queue;

[0089] The passenger and vehicle supply and demand management module is used to implement the social optimal strategy based on the deduced queue development trend, calculate the optimal probability of taxi rentals and passengers joining the queue, and regulate the arrival rate of taxis and passengers.

[0090] Specifically, this embodiment realizes the optimization of the taxi and passenger queuing system on the control logic of multi-level information release → queue situation deduction → supply and demand control of people and vehicles. First, according to the degree of refinement of information and the distribution object, three situations are designed: invisible (no information provided), partially visible (providing expected waiting time), and fully visible (providing real-time waiting time). The key information indicator calculation program of the queuing system is deployed in the MEC server to study the birth and death process of the queuing system; secondly, for the two groups of drivers and passengers, the utility function is constructed, and the Nash equilibrium strategy is deployed to simulate the autonomous selection behavior of the driver and passenger groups, and the development trend of the queuing system under the influence of information is deduced using the Markov chain; finally, considering that information may cause negative effects, the social optimal strategy is deployed to solve the optimal probability of joining the queue at each information level, so as to optimize the supply and demand level when necessary and improve social benefits. This embodiment gives full play to the advantages of real-time interconnection and efficient computing of the big data platform. Based on differentiated and multi-level information, it shortens the queuing time at both ends, improves the operating order of the taxi boarding point, and greatly improves social benefits.

[0091] The implementation of the socially optimal strategy can maximize the overall benefits of the system. In game theory, the arrival distribution of different participants is controlled and strategies are jointly formulated to achieve the optimal solution.

[0092] Further, the multi-level information publishing module includes: a first data acquisition unit, a first data pre-processing unit, a data analysis unit, a data division unit, and a data publishing unit;

[0093] The first data collection unit is used to collect dynamic traffic information and static traffic information; wherein the dynamic traffic information includes: vehicle arrival time, vehicle customization, passenger arrival time, passenger customization, and timestamp data; the static traffic information includes: the number of boarding points, taxi queue capacity, and passenger queue capacity;

[0094] A data preprocessing unit, used for data cleaning and standardization of dynamic traffic information and static traffic information;

[0095] The data analysis unit is used to calculate key information indicators for the pre-processed data; wherein the key information indicators include: taxi queue length, passenger queue length, effective arrival rate, taxi waiting time, passenger waiting time, social benefit, driver benefit and passenger benefit;

[0096] The data division unit is used to divide the information into invisible information, partially visible information and completely visible information based on key information indicators; wherein invisible information means no information is provided, partially visible information includes: expected waiting time of taxis and waiting time of passengers, and completely visible information includes: real-time waiting time of taxis and real-time waiting time of passengers;

[0097] The data publishing unit is used to publish information to drivers and passengers based on the information classification results.

[0098] Specifically, in this embodiment, the multi-level information publishing module mainly senses dynamic and static traffic information (time stamp data , Taxi arrival rate , passenger arrival rate , Taxi queue capacity , passenger queue capacity ), in-depth calculation of key information indicators of the queuing system (steady-state probability , queue length , Expected waiting time etc.), and publish differentiated, multi-level information on variable message boards or user terminals. This module is implemented through the following steps:

[0099] Step 1: Install monitoring equipment at the entrance and exit roads and queuing areas of taxi pick-up points to sense the arrival time and vehicle customization in real time. , passenger arrival time, passenger customization , time stamp data and other dynamic traffic information. Through on-site surveys or data surveys, obtain the number of pick-up points, taxi queue capacity , passenger queue capacity And other static traffic information.

[0100] Step 2: The collected dynamic and static traffic information is uploaded to the mobile edge computing of the taxi operation management department Server. Utilize The server pre-processes the data, including data cleaning and standardization, to prepare for the subsequent calculation of key information indicators of the queuing system.

[0101] Step 3: A queuing system key information indicator calculation program is deployed on the server to analyze the preprocessed data and comprehensively evaluate the performance of the queuing system, as shown below.

[0102]

[0103]

[0104]

[0105]

[0106]

[0107]

[0108]

[0109]

[0110]

[0111]

[0112] in, is the arrival rate of taxis, is the passenger arrival rate, For service intensity, In queue status, is the taxi queue capacity, The queuing capacity for passengers, is the steady-state probability, is the taxi queue length, is the passenger queue length, is the effective arrival rate, Waiting time for taxis, Waiting time for passengers, is the expected value of the indicator, For social benefits, Benefits for drivers, Benefits for passengers.

[0113] The taxi queue length, passenger queue length, effective arrival rate, taxi waiting time, passenger waiting time, social benefits, driver benefits and passenger benefits are key indicators. Arrival rate, service intensity, queue capacity, etc. are all data inputs.

[0114] Step 4: Based on the calculation results of the key information indicators of the queuing system, according to the different publishing objects, The server divides the information into taxi queue related information and passenger queue related information; according to the different levels of refinement, The server divides the information into invisible information (no information provided), partially visible information, and fully visible information. Therefore, the information that can be published includes the expected waiting time for a taxi. , Passenger waiting time , Real-time taxi waiting time , Real-time waiting time for passengers .

[0115] Step 5: Based on the classification results of the queuing system information, The deployment information publishing types on the server are invisible, partially visible, and fully visible. In the invisible case, The server does not provide any information for drivers or passengers; in some cases, The server provides expected waiting time information for drivers and passengers; with full visibility, The server provides real-time waiting time information for drivers and passengers.

[0116] Timestamp refers to Greenwich Mean Time Year moon day hour point Seconds (Beijing time Year moon day hour point The total number of seconds from the time of the last digit (seconds) to now.

[0117] Queue capacity refers to the maximum number of taxis or passengers that a boarding point can accommodate at the same time.

[0118] Queue length refers to the number of taxis or passengers waiting in the queue at the same time.

[0119] Waiting time refers to the time a driver or passenger waits in a queue.

[0120] Further, the queue situation deduction module includes: a second data acquisition unit, a second data preprocessing unit, a benefit calculation unit, a queue decision representation unit, and a situation deduction unit;

[0121] The second data collection unit is used to collect order fees, passenger waiting costs, taxi driver waiting costs, taxi transportation costs, and passenger income after arriving at the destination;

[0122] A second data preprocessing unit, used for performing data cleaning and standardization processing on the data collected by the second data acquisition unit;

[0123] A benefit calculation unit, used to construct utility functions for drivers and passengers respectively according to the information released by the multi-level information release module and the data preprocessed by the second data preprocessing unit;

[0124] The queue decision representation unit is used to add the Nash equilibrium strategy to the utility function to solve the probability of different groups joining the queue;

[0125] The situation deduction unit is used to readjust the arrival rates of drivers and passengers according to the probability of joining the queue, reconstruct the queue system under the influence of information, and recalculate key information indicators.

[0126] Specifically, in this embodiment, the queue situation deduction module: in the face of differentiated queue information, the two groups of drivers and passengers will re-estimate the benefits of joining the queue / refusing to queue, and make corresponding decisions, and finally reconstruct the entire queue system. The queue situation deduction module mainly includes three functions: 1. Group benefit calculation; 2. Queuing decision representation; 3. System situation deduction. The group benefit calculation is to construct utility functions for drivers and passengers based on indicators such as waiting time, waiting cost, order cost, fuel consumption, etc., to calculate the benefit estimates of different groups when facing differentiated information. The queue decision representation is to propose a Nash equilibrium strategy to solve the probability of different groups joining the queue on the basis of group benefit calculation, and complete the decision-making link after information perception. The system situation deduction is to use the Markov chain to reconstruct the queue system after group decision-making, and deduce the impact of information on queue evolution and social benefits based on the state transition probability.

[0127] Step 1: Collect order fees through official access, data mining or actual investigation , Passenger waiting costs , Taxi driver waiting cost , Taxi transportation costs , the benefits of passengers after arriving at their destination Similarly, through The server preprocesses the data, including data cleaning and standardization, to prepare for the subsequent construction of utility functions.

[0128] Step 2: The group utility function is deployed on the server, and the utility function is constructed for the driver / passenger group, taking into account the visibility of information. For passengers, the waiting cost will be incurred while waiting in line for a taxi. , and you need to pay the corresponding order fee , and finally get the profit after reaching the destination For taxi drivers, the taxi waiting cost is incurred while waiting in line for passengers. At the same time, transportation costs such as fuel consumption and battery consumption will be incurred during transportation , and finally receive the order fee Under different levels of information visibility, the above groups estimate benefits through waiting time differently. In the case of partial visibility, the driver / passenger group estimates benefits using expected waiting time; in the case of full visibility, the driver / passenger group estimates benefits using real-time waiting time. The utility function is constructed as follows.

[0129] 1. Driver utility function under partial visibility:

[0130]

[0131] 2. Passenger utility function under partial visibility:

[0132]

[0133] 3. Driver utility function under full visibility:

[0134]

[0135] 4. Passenger utility function under partial visibility:

[0136]

[0137] in, For order fees, Waiting costs for passengers, is the taxi waiting cost, For transportation costs, For the profit after reaching the destination, The queue length is The probability of joining the queue when .

[0138] Step 3: To determine the probability of joining the queue, The solver is deployed in the server. Considering that the group decides whether to join the queue based on the estimated benefits, this process generally follows the Nash equilibrium strategy. The equilibrium strategy can be explained as follows: if the benefits in the queue system are obvious, drivers / passengers will join the queue until the waiting time becomes longer and the benefits in the queue system are 0, at which time the number of users joining the queue and the number of users refusing to join the queue reach equilibrium. In solving the probability of joining the queue, the proposed equilibrium strategy is as follows.

[0139] 1. Driver equilibrium strategy under partial visibility:

[0140]

[0141] in, and yes exist Solutions within the range.

[0142] 2. Passenger equilibrium strategy under partial visibility:

[0143]

[0144] in, and yes Solutions are in the range [0,1].

[0145] 3. Driver equilibrium strategy under full visibility:

[0146]

[0147] in, and yes exist Solutions within the range.

[0148] 4. Passenger equilibrium strategy under partial visibility:

[0149]

[0150] in, and yes exist Solutions within the range.

[0151] Step 4: Pass The equilibrium strategy in the server solves the probability of drivers / passengers joining the queue and readjusts the arrival rate of drivers / passengers according to the probability of joining the queue, thereby reconstructing the queue system under the influence of information. The reconstructed queue system is shown in the attached Figure 2-Figure 6 .

[0152] Step 5: For the reconstructed queuing system, Based on the calculation program of key information indicators of queuing system, the formula , re-deduce the changing trends of queue length, queue time and social benefits in the queuing system.

[0153] Nash equilibrium means that if the strategy chosen by any participant is optimal when the strategies of all other participants are determined, the combination is defined as Nash equilibrium.

[0154] A Markov chain is a discrete-time random process in mathematics that has the Markov property.

[0155] Specifically, in this embodiment, the supply and demand control module for people and vehicles: in the face of differentiated queuing information, the equilibrium strategy can only reflect the spontaneous selection behavior of drivers / passengers, and cannot give full play to the maximum advantage of differentiated information. The supply and demand control module for people and vehicles mainly includes two functions: supply and demand regulation guidance and expected benefit visualization. The supply and demand regulation guidance is to calculate the optimal probability of both ends joining the queue by implementing the social optimal strategy, and provide effective guidance for relevant management departments to regulate the traffic supply and demand of taxi pick-up points. The expected benefit visualization is to use the optimal probability of both ends joining the queue to readjust the dual-end arrival rate, and then calculate the social benefits under the corresponding circumstances, and visualize it in the management center.

[0156] Step 1: For the optimal value of the probability of joining the queue, a solution program needs to be deployed on the MEC server. Considering that drivers / passengers will adjust their own probability of joining the queue when facing different information levels, it will affect the waiting time of the queue system and further affect the social benefits of the entire queue system. In solving the probability of joining the queue, the proposed social optimal strategy is as follows.

[0157] 1. Socially optimal strategy under partial visibility (calculation of driver joining queue probability):

[0158]

[0159] 2. Socially optimal strategy under partial visibility (calculation of passenger joining queue probability):

[0160]

[0161] 3. Socially optimal strategy under full visibility (calculation of driver joining queue probability):

[0162]

[0163] 4. Socially optimal strategy under full visibility (calculation of the probability of passengers joining the queue):

[0164]

[0165] Step 2: Pass The server publishes the calculated optimal queue joining probability to the relevant management department, and adjusts the arrival rate of taxis and passengers respectively by changing the charging rules or queuing rules; The server downloads the calculated optimal social benefits to the relevant management departments and visualizes them in the form of line graphs or heat maps.

[0166] Step 3: The relevant departments compare the results of the invisible, partially visible, and fully visible situations. The server chooses to publish the expected waiting time or the real-time waiting time to the variable information board or user terminal at the taxi pick-up point to guide drivers and passengers to make better decisions.

[0167] This embodiment proposes a one-dimensional continuous-time Markov model, which expands the research on the customer arrival process in the queuing system by introducing the queuing probability, and can adapt to the real-time dynamic changes in the system state. The model considers the dynamic evolution of the two-end queuing system under three information conditions: invisible, partially visible, and significantly visible. Through this method, the behavior of the queuing system can be accurately simulated and predicted, especially in complex and uncertain environments, which improves the stability and efficiency of the queuing system and scientifically guides the management department to optimize the modeling of the matching of people and vehicles at the taxi boarding point.

[0168] This embodiment constructs a decision utility function and proposes a balance strategy, which quantifies the impact of differentiated information on the optimization of the queuing system under different information environments. By implementing the balance strategy, it can truly reflect the queuing choices of passengers / drivers when faced with differentiated information. Through this method, the optimal release range of different information can be clarified, providing an application reference for the informatization and intelligence of taxi operations.

[0169] This example proposes a socially optimal strategy, clarifies the direction of queue system optimization under differentiated information levels, and deeply explores the optimization space. Through this strategy, the overall social welfare can be maximized, while ensuring system efficiency and avoiding resource waste. The socially optimal strategy provides scientific decision-making support for policy makers and operation managers, especially in high-demand environments, and can achieve optimal resource allocation.

[0170] This embodiment also proposes a taxi-passenger double-end queuing optimization method based on differentiated information levels, including:

[0171] Collect dynamic and static traffic information, classify the information, and release it to drivers and passengers;

[0172] Based on the information released, simulate and analyze the queuing decision-making behaviors of drivers and passengers at different information levels, and deduce the development trend of the queue;

[0173] According to the deduced queue development trend, the social optimal strategy is implemented, the optimal probability of taxi rentals and passengers joining the queue is calculated, and the arrival rate of taxis and passengers is regulated.

[0174] Furthermore, dynamic and static traffic information is collected and classified and released to drivers and passengers, including:

[0175] Collect dynamic traffic information and static traffic information; dynamic traffic information includes: vehicle arrival time, vehicle customization, passenger arrival time, passenger customization, and timestamp data; static traffic information includes: number of boarding points, taxi queue capacity, and passenger queue capacity;

[0176] Carry out data cleaning and standardization of dynamic and static traffic information;

[0177] Calculate key information indicators for the preprocessed data, including steady-state probability, queue length, and expected waiting time.

[0178] Based on key information indicators, information is divided into invisible information, partially visible information, and fully visible information; invisible information means no information is provided, partially visible information includes: expected taxi waiting time, passenger waiting time, and fully visible information includes: real-time taxi waiting time, passenger waiting time;

[0179] According to the information classification results, information is released to drivers and passengers.

[0180] Based on the information released, the queuing decision-making behaviors of drivers and passengers under different information levels are simulated and analyzed, and the development trend of the queuing is deduced, including:

[0181] Collect order fees, passenger waiting costs, taxi driver waiting costs, taxi transportation costs, and passenger income after arriving at the destination;

[0182] Clean and standardize the collected data;

[0183] Based on the published information and preprocessed data, utility functions are constructed for drivers and passengers respectively;

[0184] For the utility function, the Nash equilibrium strategy is added to solve the probability of different groups joining the queue;

[0185] The arrival rates of drivers and passengers are readjusted according to the probability of joining the queue, the queue system under the influence of information is reconstructed, and key information indicators are recalculated.

[0186] The embodiments described above are only descriptions of the preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Without departing from the design spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by ordinary technicians in this field should all fall within the protection scope determined by the claims of the present invention.

Claims

1. A taxi-passenger double-end queuing optimization system based on differentiated information levels, characterized in that: include: Multi-level information release module, queue situation deduction module, and people and vehicle supply and demand control module; The multi-level information release module is used to collect dynamic and static traffic information, classify the information, and release it to drivers and passengers; The queue situation deduction module is used to simulate and analyze the queue decision-making behaviors of drivers and passengers at different information levels according to the information released by the multi-level information release module, and deduce the development trend of the queue; The passenger and vehicle supply and demand control module is used to implement the social optimal strategy according to the deduced queue development trend, calculate the optimal probability of taxis and passengers joining the queue, and regulate the arrival rate of taxis and passengers.

2. The taxi-passenger double-end queuing optimization system based on differentiated information levels according to claim 1 is characterized in that: The multi-level information publishing module includes: a first data acquisition unit, a first data preprocessing unit, a data analysis unit, a data division unit, and a data publishing unit. The first data collection unit is used to collect dynamic traffic information and static traffic information; wherein the dynamic traffic information includes: vehicle arrival time, vehicle customization, passenger arrival time, passenger customization, and timestamp data; the static traffic information includes: the number of boarding points, taxi queue capacity, and passenger queue capacity; The data preprocessing unit is used to perform data cleaning and standardization processing on the dynamic traffic information and the static traffic information; The data analysis unit is used to calculate key information indicators for the pre-processed data; wherein the key information indicators include: taxi queue length, passenger queue length, effective arrival rate, taxi waiting time, passenger waiting time, social benefits, driver benefits and passenger benefits; The data division unit is used to divide the information into invisible information, partially visible information, and completely visible information based on the key information indicators; wherein the invisible information is no information provided, the partially visible information includes: the expected waiting time of a taxi, the waiting time of a passenger, and the completely visible information includes: the real-time waiting time of a taxi, the real-time waiting time of a passenger; The data publishing unit is used to publish information to drivers and passengers based on the information classification results.

3. The taxi-passenger double-end queuing optimization system based on differentiated information levels according to claim 2 is characterized in that: The data analysis unit calculates key information indicators for the preprocessed data, including: in, is the arrival rate of taxis, is the passenger arrival rate, For service intensity, In queue status, is the taxi queue capacity, The queuing capacity for passengers, is the steady-state probability, is the taxi queue length, is the passenger queue length, is the effective arrival rate, Waiting time for taxis, Waiting time for passengers, is the expected value of the indicator, For social benefits, Benefits for drivers, Benefits for passengers.

4. The taxi-passenger double-end queuing optimization system based on differentiated information levels according to claim 3 is characterized in that: The queue situation deduction module includes: a second data acquisition unit, a second data preprocessing unit, a benefit calculation unit, a queue decision representation unit, and a situation deduction unit; The second data collection unit is used to collect order fees, passenger waiting costs, taxi driver waiting costs, taxi transportation costs, and passenger income after arriving at the destination; The second data preprocessing unit is used to perform data cleaning and standardization processing on the data collected by the second data acquisition unit; The benefit calculation unit is used to construct utility functions for drivers and passengers respectively according to the information released by the multi-level information release module and the data preprocessed by the second data preprocessing unit; The queuing decision representation unit is used to add a Nash equilibrium strategy to the utility function to solve the probability of different groups joining the queue; The situation deduction unit is used to readjust the arrival rates of drivers and passengers according to the probability of joining the queue, perform situation deduction based on the Markov chain, reconstruct the queue system under the influence of information, and recalculate key information indicators.

5. The taxi-passenger double-end queuing optimization system based on differentiated information levels according to claim 4 is characterized in that: The utility function constructed by the benefit calculation unit includes: The driver's utility function under partial visibility is: Passenger utility function under partial visibility: The driver's utility function under full visibility is: Passenger utility function under full visibility: in, For order fees, Waiting costs for passengers, is the taxi waiting cost, For transportation costs, For the income after reaching the destination, The queue length is The probability of joining the queue when .

6. The taxi-passenger double-end queuing optimization system based on differentiated information levels according to claim 5 is characterized in that: The Nash equilibrium strategy added to the queuing decision representation unit includes: Driver equilibrium strategy under partial visibility: in, and yes exist Solutions within the range; Traveler equilibrium strategy under partial visibility: in, and yes exist Solutions within the range; The driver's equilibrium strategy under full visibility: in, and yes exist Solutions within the range; Traveler equilibrium strategy under partial visibility: in, and yes exist Solutions within the range.

7. The taxi-passenger double-end queuing optimization system based on differentiated information levels according to claim 6 is characterized in that: The social optimal strategy implemented by the vehicle supply and demand control module includes: The socially optimal strategy in the partially visible case after calculating the probability of the driver joining the queue: The socially optimal strategy in the partially visible case after the probability of passengers joining the queue is calculated: The socially optimal strategy in the fully visible case after calculating the probability of drivers joining the queue is: The socially optimal strategy in the fully visible case after the probability of passengers joining the queue is calculated: in, For order fees, Waiting costs for passengers, Waiting costs for taxi drivers, For taxi transportation costs, The income of passengers after they arrive at their destination.

8. A taxi-passenger double-end queuing optimization method based on differentiated information levels, characterized in that: Using the system as described in any one of claims 1 to 7, the method comprises: Collect dynamic and static traffic information, classify the information, and release it to drivers and passengers; Based on the information released by the multi-level information release module, simulate and analyze the queuing decision-making behaviors of drivers and passengers at different information levels, and deduce the development trend of the queuing; According to the deduced queue development trend, the social optimal strategy is implemented, the optimal probability of taxi rentals and passengers joining the queue is calculated, and the arrival rate of taxis and passengers is regulated.

Images