Vehicle sharing system rebalance and scale optimization method

Abstract

The invention provides a vehicle sharing system rebalance and scale optimization method, and belongs to the field of planning and scheduling of vehicle sharing systems, and the method comprises the steps: creating a migration kinetic equation of passengers, vehicles and dispatchers in each passenger station; according to an M / M / s queuing theory and an M / D / s queuing theory, respectively obtaining a charging queue stability condition, average charging time consumption, a dispatcher rest queue stability condition and dispatcher average rest time consumption; further analyzing to obtain a condition for achieving equalization, a minimum vehicle scale for achieving equalization and a minimum dispatcher scale for achieving equalization; and designing a scheduling strategy of vehicle charging, dispatcher rest and rebalance. According to the method, the rest demand of the dispatcher is considered, and vehicle charging and vehicle and dispatcher coupling rebalance are considered, so that the riding demand of passengers can be met by using fewer vehicles and dispatchers.

Description

Technical Field

[0001] The present invention belongs to the field of planning and scheduling of vehicle sharing systems, and in particular relates to a vehicle sharing system rebalancing and scale optimization method. Background Art

[0002] The vehicle sharing system composed of electric vehicles is widely regarded as the main mode of green travel in the future due to its high efficiency. However, due to the uneven passenger travel demand between regions, a large number of vehicles will accumulate in popular areas, while other areas will have no cars available. Simply increasing the size of the fleet cannot solve this problem, so it is necessary to assign dispatchers to migrate idle vehicles to meet passenger needs. During the operation of the system, in order to avoid service delays caused by the long charging time of electric vehicles, charging scheduling is also essential. In the actual system, the dispatcher frequently rebalances the vehicle and performs charging scheduling operations, and also needs to rest and rebalance himself, so the dispatcher also needs to be rebalanced and rested.

[0003] Currently, many different methods have been proposed to solve this type of problem. Some literature uses fluid models and M / M / s queuing models to describe the customer / vehicle migration and charging process in the system, and proposes a rebalancing strategy with the goal of dispatching the least empty vehicles while meeting passenger demand. On this basis, other literature designs a customer ride pricing mechanism based on the ride demand between different regions, and uses prices to encourage customers to go to high-demand areas near their destinations while actively dispatching empty vehicles for rebalancing. However, the above literature all assumes that the rebalancing operation is completed independently by autonomous vehicles. At present, the commercial operation field of fleets dominated by autonomous vehicles is limited and cannot meet the needs of practical applications. There are also literatures that believe that the vehicle rebalancing and charging process requires the dispatcher to drive the vehicle to operate, so it is necessary to consider the coupled rebalancing of the vehicle and the dispatcher. On this basis, a joint optimization strategy for vehicle charging scheduling and rebalancing is proposed. There are also literatures that use state space-based methods to monitor vehicles as monitoring objects, and need to monitor the vehicle status in real time to obtain the optimization method for rebalancing. However, when the vehicle scale is large, the data processing and calculation amount of monitoring vehicles is very large, which is difficult to meet the real-time requirements. In summary, the current research on rebalancing and scheduling of vehicle sharing systems is mostly focused on autonomous vehicles, and there are few studies on vehicle scheduling performed manually. In addition, most existing studies ignore the charging needs of vehicles and the rest needs of personnel. Although some literature considers the dispatcher to perform vehicle rebalancing and charging scheduling, it ignores the dispatcher's rest needs; while some literature schedules the dispatcher's rebalancing and rest but ignores the charging needs of vehicles, and is not suitable for large-scale systems. Summary of the invention

[0004] In response to the shortcomings of the existing technology, this application proposes a rebalancing and scale optimization method for a vehicle sharing system. For each passenger stop, a rebalancing optimization plan is given in combination with the dispatcher's rest and vehicle charging scheduling. While meeting the needs of passengers, this plan also provides the minimum scale of vehicles and dispatchers required by the system.

[0005] This application proposes a method for rebalancing and optimizing the scale of a vehicle sharing system, including:

[0006] In each passenger station, a migration dynamics equation of passengers, vehicles and dispatchers is created. The migration dynamics equation is created based on the operation mechanism of the vehicle sharing system and takes into account vehicle charging scheduling, dispatcher rest scheduling, and vehicle and dispatcher coupled rebalancing scheduling;

[0007] The M / M / s queuing theory is used for the vehicle charging process to obtain the stability conditions of the charging queue and the average charging time.

[0008] The M / D / s queuing theory is used for the dispatcher rest process model to obtain the stability conditions of the dispatcher rest queue and the average rest time of the dispatcher.

[0009] According to the migration dynamics equation, the stability condition of the charging queue, the average charging time, the stability condition of the dispatcher's rest queue and the average rest time of the dispatcher, the conditions for reaching equilibrium, the minimum vehicle size to reach equilibrium and the minimum dispatcher size to reach equilibrium are obtained;

[0010] According to the conditions for reaching equilibrium, the minimum vehicle size to achieve equilibrium, and the minimum dispatcher size to achieve equilibrium, a scheduling strategy for vehicle charging, dispatcher rest, and rebalancing is designed, and the scheduling strategy for vehicle charging, dispatcher rest, and rebalancing is used as the result of coupled rebalancing and scale optimization of the vehicle sharing system.

[0011] In each passenger station, migration dynamics equations of customers, vehicles and dispatchers are created, including: migration dynamics equations of customers at passenger stations, migration dynamics equations of vehicles at passenger stations and migration dynamics equations of dispatchers at passenger stations.

[0012] The migration dynamics equation of the passenger station customers is calculated as follows:

[0013] ;

[0014] in, is the customer arrival rate at passenger station i, The rate at which vehicles are assigned to customers at passenger station i, is the idle vehicle existing in passenger station i at the current moment, is the dynamic change rate of customers at the passenger station, is the number of passengers existing in passenger station i at the current moment, and H() is the Heaviside function.

[0015] The migration dynamics equation of vehicles at the passenger station i is composed of the sum of the rate at which passenger vehicles leave the passenger station i, the rate at which passenger vehicles arrive from other passenger stations j, the rate at which the dispatcher drives the low-power vehicle from the passenger station i to the vehicle charging station k, the rate at which the dispatcher drives the vehicle from the vehicle charging station k to the passenger station i, the rate at which the dispatcher drives the rebalancing vehicle to leave the passenger station i, and the rate at which the dispatcher drives the rebalancing vehicle to arrive at the passenger station i.

[0016] The migration dynamics equation of the dispatcher at the passenger station i is composed of the sum of the rate at which the dispatcher drives the low-power vehicle from the passenger point i to the vehicle charging station k, the rate at which the dispatcher drives the vehicle from the vehicle charging station k to the passenger station i, the rate at which the dispatcher drives the rebalancing vehicle to leave the passenger station i, the rate at which the dispatcher drives the rebalancing vehicle to arrive at the passenger station i, the rate at which the dispatcher who needs to rest goes from the passenger station i to the rest station m, the rate at which the dispatcher returns to the passenger station i after resting, the rate at which the rebalancing dispatcher leaves the passenger station i, and the rate at which the rebalancing dispatcher arrives at the passenger station i.

[0017] The charging queue stability condition is that the vehicle arrival rate of the vehicle charging station k is less than the total service rate, and the calculation formula is as follows:

[0018] ;

[0019] in, is the dispatch rate from passenger station i to vehicle charging station k, is the rate at which low-battery vehicles arrive at passenger station i, The probability that a low-battery vehicle at passenger station i chooses to go to vehicle charging station k for charging, is the number of charging piles, For each charging pile service rate, is the total service rate of all charging piles in vehicle charging station k;

[0020] The average charging time is the sum of the vehicle waiting time and the charging time, and is calculated as follows:

[0021] ;

[0022] in, is the sum of the waiting time in queue and the charging time of the vehicle. is the vehicle arrival rate at vehicle charging station k, l ranges from 0 to An integer.

[0023] The dispatcher rest queue stability condition is that the dispatcher arrival rate is less than the total service rate, and the calculation formula is as follows:

[0024] ;

[0025] in, is the rate at which passenger station i sends dispatchers to rest stop m, is the probability that a fatigued dispatcher at passenger station i chooses to go to rest station m for rest, is the proportion of the dispatcher who needs to rest when arriving at passenger station i, and rest station m contains rest space, and each dispatcher has a rest time of , the service rate of rest stop m is ;

[0026] The average rest time of the dispatcher is calculated as follows:

[0027] ;

[0028] in, The busyness of the site, is the dispatcher arrival rate at rest stop m, is the probability that there is no customer queue, is the average rest time of the dispatcher, l represents the time from 0 to An integer.

[0029] The condition for reaching equilibrium is calculated as follows:

[0030] ;

[0031] in, , is the dispatch rate of vehicle charging station k to passenger station j, is the proportion of vehicles returning to passenger station j from vehicle charging station k, is the vehicle arrival rate at vehicle charging station k, is the probability that a passenger in passenger station j goes to passenger station i, is the dispatcher rate sent from rest stop m to passenger stop j, is the proportion of dispatchers in rest stop m who choose to go to passenger stop j, is the dispatcher arrival rate at rest stop m, is the rebalancing rate of vehicles from passenger station i to passenger station j, is the rebalancing rate from passenger station j to passenger station i, is the dispatcher's rebalancing rate from passenger station i to passenger station j, is the rebalancing rate of the dispatcher from passenger station j to passenger station i, is the customer arrival rate of station passengers at station i, is the customer arrival rate of station passengers at station j, is the rate at which dispatchers are sent from passenger station i to rest stop m, is the dispatch rate from passenger station i to vehicle charging station k.

[0032] The minimum vehicle scale required to achieve equilibrium is composed of the sum of the number of passenger vehicles and vehicles performing rebalancing tasks, the number of vehicles heading for charging and charging, and the number of vehicles heading to the station after completing charging.

[0033] The minimum dispatcher size required to achieve equilibrium is composed of the sum of the number of dispatchers who perform vehicle rebalancing and self-balancing, the number of dispatchers who perform vehicle charging tasks, the number of dispatchers who go to rest stops and the number of dispatchers who go to stations after resting.

[0034] The scheduling strategy for vehicle charging, dispatcher rest and rebalancing is obtained by solving the following nonlinear programming problem, and the calculation formula is as follows:

[0035] and

[0036] ;

[0037] Constraints:

[0038] ;

[0039] ;

[0040] ;

[0041] in, is the customer arrival rate of station passengers at station i, is the rate at which low-battery vehicles arrive at passenger station i, is the proportion of dispatchers who need to rest when arriving at passenger station i, is the customer arrival rate at passenger station j, is the probability that a passenger in passenger station i goes to passenger station j, is the probability that a passenger in passenger station j goes to passenger station i, is the rebalancing rate of vehicles from passenger station i to passenger station j, is the rebalancing rate from passenger station j to passenger station i, is the dispatch rate from passenger station i to vehicle charging station k, is the dispatch rate of vehicle charging station k to passenger station j, is the dispatcher's rebalancing rate from passenger station i to passenger station j, is the dispatcher's rebalancing rate from passenger station j to passenger station i, is the rate at which the dispatcher at passenger station i sends messages to the rest stop m, is the probability that a fatigued dispatcher at passenger station i chooses to go to rest station m for rest, is the probability that the dispatcher of rest stop m goes to passenger stop j, is the sum of the waiting time in queue and the charging time of the vehicle. is the sum of the dispatcher’s waiting time in queue and rest time. is the number of rest stops m that can provide rest for dispatchers at the same time, is the service rate of rest stop m, is the travel time between passenger station i and passenger station j, is the travel time between passenger station j and passenger station i, is the travel time from passenger station i to vehicle charging station k, is the travel time from vehicle charging station k to passenger station j, is the travel time from passenger station i to rest stop m, is the travel time from rest stop m to passenger stop i, N is the set of passenger stops, M is the set of vehicle charging stations, P is the set of rest stops, is the number of rest spaces, is the service rate for each dispatcher, For each charging pile service rate, is the number of charging piles.

[0042] The scheduling strategy for vehicle charging, dispatcher rest and rebalancing also includes: the number of additionally deployed vehicles and the number of additionally deployed dispatchers, wherein the number of additionally deployed vehicles is calculated as follows:

[0043] ;

[0044] in, is the number of additional vehicles deployed, is the dispatch rate from passenger station i to vehicle charging station k, is the sum of the waiting time in queue and the charging time of the vehicle. is the travel time from passenger station i to vehicle charging station k;

[0045] The number of additional dispatchers deployed is calculated as follows:

[0046] ;

[0047] in, is the number of additional dispatchers deployed, is the rate at which the dispatcher at passenger station i sends messages to the rest stop m, is the travel time from passenger station i to rest stop m, It is the sum of the dispatcher's waiting time in queue and rest time.

[0048] Beneficial effects:

[0049] This application proposes a method for rebalancing and scale optimization of a vehicle sharing system. Existing technologies often ignore the issues of vehicle charging and dispatcher rest. This application takes into account the dispatcher's rest needs, and takes into account vehicle charging, vehicle and dispatcher coupling rebalancing. A scheduling solution is designed for this complex optimization problem. This application can use fewer vehicles and dispatchers to meet passengers' riding needs. This application does not need to collect data from each vehicle and is more suitable for systems with different characteristics. Traditional control methods schedule the system by controlling each vehicle, and the difficulty of solving the problem increases exponentially with the scale of the system, making it difficult to apply to large-scale vehicle sharing systems. The difficulty of solving the problem in this application is independent of the scale of the system, and its scope of application is wider. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 A flow chart of a vehicle sharing system rebalancing and scale optimization method according to an embodiment of the present application;

[0051] Figure 2 A flow chart of a method for rebalancing and scaling optimization of a vehicle sharing system according to an embodiment of the present application;

[0052] Figure 3 A dynamic change model of customers, vehicles and dispatchers in the vehicle sharing system of the embodiment of the present application;

[0053] Figure 4 A schematic diagram of the relationship between the system scale and the number of sites in an embodiment of the present application;

[0054] Figure 5 A schematic diagram of the relationship between the proportion of employees performing different tasks and the number of sites in an embodiment of the present application;

[0055] Figure 6 A schematic diagram showing the relationship between the system scale and the charging rate in an embodiment of the present application;

[0056] Figure 7 A schematic diagram showing the relationship between the system scale and the employee rest time in an embodiment of the present application;

[0057] Figure 8 A schematic diagram showing the change of the number of waiting passengers over time in the non-dispatching system of an embodiment of the present application;

[0058] Fig. 9A schematic diagram showing the change of the number of waiting passengers in the system over time after scheduling is added in an embodiment of the present application;

[0059] Fig.10 A schematic diagram of the maximum number of waiting customers at each station in an embodiment of the present application;

[0060] Fig.11 Schematic diagram of the distribution of vehicles performing different tasks in an embodiment of the present application. DETAILED DESCRIPTION

[0061] The specific implementation of the present application is further described in detail below in conjunction with the drawings and examples.

[0062] At present, the research on rebalancing and scheduling of vehicle sharing systems is mostly focused on autonomous driving vehicles. There are few studies on vehicle scheduling performed manually, and most existing studies ignore the charging needs of vehicles and the rest needs of personnel. Although some studies consider the dispatcher to perform vehicle rebalancing and charging scheduling, they ignore the dispatcher's rest needs. Other studies schedule the dispatcher's rebalancing and rest but ignore the vehicle's charging needs, and are not suitable for large-scale systems.

[0063] This application uses a fluid model to describe the movement of customers, dispatchers, and vehicles, and uses M / M / s and M / D / s queuing theories to model vehicle charging and dispatcher rest processes, and further combines the fluid model and queuing theory to construct a dynamic model of the vehicle sharing system. Subsequently, a method for coupling rebalancing and scale optimization of the vehicle sharing system is given, and the optimal fleet size and number of dispatchers that meet passenger needs are determined. In addition, this application uses a probabilistic model for modeling, which is more suitable for large-scale systems.

[0064] Example:

[0065] This embodiment proposes a method for rebalancing and optimizing the scale of a vehicle sharing system. Figure 1 , Figure 2 As shown, including:

[0066] Step S1: In each passenger station, a migration dynamics equation of passengers, vehicles and dispatchers is created, wherein the migration dynamics equation is created according to the operation mechanism of the vehicle sharing system and takes into account vehicle charging scheduling, dispatcher rest scheduling, and vehicle and dispatcher coupled rebalancing scheduling;

[0067] Step S2: using the M / M / s queuing theory for the vehicle charging process to obtain the charging queue stability condition and the average charging time;

[0068] Step S3: Adopting the M / D / s queuing theory for the dispatcher's rest process model, the stability condition of the dispatcher's rest queue and the average rest time of the dispatcher are obtained;

[0069] Step S4: according to the migration dynamics equation, the stability condition of the charging queue, the average charging time, the stability condition of the dispatcher's rest queue and the average rest time of the dispatcher, the conditions for reaching equilibrium, the minimum vehicle scale for reaching equilibrium and the minimum dispatcher scale for reaching equilibrium are obtained;

[0070] Step S5: According to the conditions for reaching equilibrium, the minimum vehicle scale for reaching equilibrium, and the minimum dispatcher scale for reaching equilibrium, a scheduling strategy for vehicle charging, dispatcher rest, and rebalancing is designed, and the scheduling strategy for vehicle charging, dispatcher rest, and rebalancing is used as the coupled rebalancing and scale optimization result of the vehicle sharing system.

[0071] In step S1, in each passenger station, migration dynamics equations of customers, vehicles and dispatchers are created, including: migration dynamics equations of customers in the passenger station, migration dynamics equations of vehicles in the passenger station and migration dynamics equations of dispatchers in the passenger station;

[0072] (1) The dynamic changes of customers at passenger station i (i.e., the migration dynamics equation of customers in passenger station i) are:

[0073] ;

[0074] Define the Heaviside function:

[0075] ;

[0076] Applying the Heaviside function can convert the customer dynamic change rate into It can be expressed as:

[0077] ;

[0078] in, is the customer arrival rate at passenger station i, The rate at which vehicles are assigned to customers at passenger station i, is the idle vehicle existing in passenger station i at the current moment, is the dynamic change rate of customers at the passenger station, is the number of passengers existing in passenger station i at the current moment, and H() is the Heaviside function.

[0079] (2) Vehicle change rate at passenger station i (i.e., the migration dynamics equation of vehicles in the passenger station) consists of the following parts

[0080] 1) The rate at which passenger vehicles leave from passenger station i:

[0081] ;

[0082] 2) The rate at which passenger vehicles arrive from other passenger stations j:

[0083] ;

[0084] in, and Respectively The number of idle vehicles and passengers at passenger station j before time, is the travel time between site j and site i.

[0085] 3) The rate at which the dispatcher drives a low-battery vehicle from passenger station i to vehicle charging station k:

[0086] ;

[0087] in, is the rate at which low-battery vehicles arrive at station i, represents the proportion of low-battery vehicles from station i to vehicle charging station k, Represents the number of idle dispatchers in site i at the current moment.

[0088] 4) The rate at which the dispatcher drives the vehicle from the vehicle charging station k to the passenger station i:

[0089] ;

[0090] in, represents the proportion of vehicles going from vehicle charging station k to passenger station i, and Respectively

[0091] The number of idle vehicles and dispatchers at passenger station j before time, is the average stay time of vehicles in vehicle charging station k.

[0092] 5) The rate at which the dispatcher drives the rebalanced vehicle away from passenger station i:

[0093] ;

[0094] in, is the rebalancing rate of vehicles from passenger station i to passenger station j.

[0095] 6) The rate at which the dispatcher drives the rebalanced vehicle to the passenger station i:

[0096] ;

[0097] in, Respectively The number of idle dispatchers at passenger station j before time.

[0098] By adding the above 6 parts, we can get the dynamic change of the vehicle at passenger station i (that is, the migration dynamics equation of the vehicle at the passenger station):

[0099] ;

[0100] To summarize, the migration dynamics equation of the passenger station vehicle is composed of the sum of the rate at which passenger vehicles leave passenger station i, the rate at which passenger vehicles arrive from other passenger stations j, the rate at which the dispatcher drives the low-power vehicle from passenger station i to vehicle charging station k, the rate at which the dispatcher drives the vehicle from vehicle charging station k to passenger station i, the rate at which the dispatcher drives the rebalancing vehicle to leave passenger station i, and the rate at which the dispatcher drives the rebalancing vehicle to arrive at passenger station i.

[0101] (3) Dispatcher change rate at passenger station i (i.e., the migration dynamics equation of the passenger station dispatcher) consists of the following parts:

[0102] Dynamic changes of dispatchers The four parts in are the same as 3) to 6) in the vehicle dynamics, because each rebalancing vehicle and vehicle going to or leaving charging contains a dispatcher. The rest are as follows

[0103] 1) The rate at which the dispatcher who needs to rest travels from passenger station i to rest station m:

[0104] ;

[0105] in, is the proportion of dispatchers who need to rest when arriving at passenger station i.

[0106] 2) The rate at which the dispatcher returns to passenger station i after a break:

[0107] ;

[0108] in, and Respectively The number of idle vehicles and dispatchers at station j before time, is the average stay time of the dispatchers in rest stop m.

[0109] 3) Rebalance the rate at which the dispatcher leaves passenger station i:

[0110] ;

[0111] in, is the rebalancing rate of the dispatcher from site i to site j.

[0112] 4) Rebalance the rate at which the dispatcher arrives at passenger station i:

[0113] ;

[0114] By adding 3) to 6) in (2) and 1) to 4) 8 in (3), we can get the dynamic change of the dispatcher at passenger station i as follows:

[0115] ;

[0116] To summarize, the migration dynamics equation of the dispatcher in the passenger station i is composed of the sum of the rate at which the dispatcher drives the low-power vehicle from the passenger station i to the vehicle charging station k, the rate at which the dispatcher drives the vehicle from the vehicle charging station k to the passenger station i, the rate at which the dispatcher drives the rebalancing vehicle to leave the passenger station i, the rate at which the dispatcher drives the rebalancing vehicle to arrive at the passenger station i, the rate at which the dispatcher who needs to rest goes from the passenger station i to the rest station m, the rate at which the dispatcher returns to the passenger station i after resting, the rate at which the rebalancing dispatcher leaves the passenger station i, and the rate at which the rebalancing dispatcher arrives at the passenger station i.

[0117] In step S2, the M / M / s queuing theory is used for the vehicle charging process to obtain the charging queue stability condition and the average charging time; wherein, the charging queue stability condition and the average charging time are given by:

[0118] (1) Consider the rate at which low-battery vehicles arrive at vehicle charging station k It obeys Poisson distribution and charging time obeys exponential distribution, so the M / M / s queuing theory can be used to analyze the charging queue. In the M / M / s queuing theory (M / M / sQueueing Model), ‌M‌ indicates that the arrival interval of charging vehicles obeys exponential distribution (Poisson distribution), ‌M‌ indicates that the charging time obeys exponential distribution, and ‌s‌ indicates the number of charging piles in the charging station.

[0119] (2) Assume that the vehicle charging station k contains Charging piles, each charging pile has a service rate of , the total service rate of vehicle charging station k To stabilize the charging queue, the vehicle arrival rate at the charging station must be strictly less than the total service rate, that is:

[0120] ;

[0121] in, is the dispatch rate from passenger station i to vehicle charging station k, is the rate at which low-battery vehicles arrive at passenger station i, The probability that a low-battery vehicle at passenger station i chooses to go to vehicle charging station k for charging.

[0122] (3) According to the M / M / s queuing theory, the average time a vehicle spends at charging station k can be obtained as:

[0123] ;

[0124] in, is the sum of the waiting time in queue and the charging time of the vehicle. is the vehicle arrival rate at vehicle charging station k, l ranges from 0 to An integer.

[0125] In step S3, the M / D / s queuing theory is used for the dispatcher rest process model to obtain the dispatcher rest queue stability condition and the dispatcher's average rest time. In the M / D / s queuing theory (M / D / s Queueing Model), M indicates that the arrival interval of the fatigued dispatcher follows an exponential distribution, D indicates that the fatigued dispatcher's rest time follows a fixed-length distribution, and s indicates the number of spaces in the rest station that can be used by the dispatcher to rest at the same time. Among them, the dispatcher rest queue stability condition and the dispatcher's average rest time are given by:

[0126] (1) Consider the rate at which fatigued dispatchers arrive at rest stop m Follows Poisson distribution, rest time Fixed, so the service rate In order to obey the deterministic distribution, the M / D / s queuing theory can be used to analyze the rest queue.

[0127] (2) Assume that rest stop m contains rest space, and each dispatcher has a rest time of , the service rate is , if the rest queue is to be stable, the dispatcher arrival rate at the rest station must be strictly less than the total service rate, that is:

[0128] ;

[0129] in, is the rate at which passenger station i sends dispatchers to rest stop m, is the probability that a fatigued dispatcher at passenger station i chooses to go to rest station m for rest.

[0130] (3) According to the M / D / s queuing theory, the average time a dispatcher spends at rest stop m is:

[0131] ;

[0132] in, The busyness of the site, is the dispatcher arrival rate at rest stop m, is the probability that there is no customer queue, is the average rest time of the dispatcher, l represents the time from 0 to An integer.

[0133] In step S4, based on the migration dynamics equation, the charging queue stability condition, the average charging time, the dispatcher rest queue stability condition and the dispatcher average rest time, the conditions for reaching equilibrium, the minimum vehicle size for reaching equilibrium and the minimum dispatcher size for reaching equilibrium are obtained.

[0134] The system equilibrium state and the conditions when the system reaches equilibrium are given by:

[0135] (1) When the system reaches equilibrium, the change rate of customers, vehicles, and dispatchers at each passenger station is 0, that is, , and Established at the same time.

[0136] (2) Substituting into the customer dynamics equation, we get

[0137] ;

[0138] because , the condition for the above equation to have a unique solution is:

[0139] ;

[0140] That is, in equilibrium, there are available vehicles at every station and no waiting customers.

[0141] (3) Substituting the solution conditions of the above customer dynamics equation into the vehicle dynamics equation, we can get

[0142] ;

[0143] in, , is the dispatch rate of vehicle charging station k to passenger station j, is the proportion of vehicles returning to passenger station j from vehicle charging station k, is the vehicle arrival rate at vehicle charging station k. The condition for the above formula to be true is Substituting this condition into the above formula, we can get:

[0144] ;

[0145] (4) Substituting the solution conditions of the above customer dynamics equation and vehicle dynamics equation into the dispatcher dynamics equation, we can get:

[0146] ;

[0147] in, is the dispatcher rate sent from rest stop m to passenger stop j, The proportion of dispatchers in rest stop m who choose to go to passenger stop j. is the probability that a passenger in passenger station j goes to passenger station i, is the dispatcher's rebalancing rate from passenger station i to passenger station j, is the rebalancing rate of the dispatcher from passenger station j to passenger station i, The rate at which dispatchers are sent to passenger stop j for rest stop m.

[0148] (5) The equations in (3) and (4) of step S4 above are the scheduling strategies. The conditions that must be met for the system to reach equilibrium.

[0149] The minimum size of the system at which it reaches equilibrium is given by:

[0150] In the equilibrium state, when the system size is the smallest, all vehicles and dispatchers are busy.

[0151] (1) The vehicle size is calculated as follows:

[0152] The number of passenger vehicles and vehicles performing rebalancing tasks is:

[0153] ;

[0154] The number of vehicles heading for charging and currently charging is:

[0155] ;

[0156] The number of vehicles heading to the station after completing charging is:

[0157] ;

[0158] Adding the above three parts together, the minimum vehicle size in the system is:

[0159] ;

[0160] (2) The dispatcher size is calculated as follows:

[0161] The number of vehicle rebalancing and self-rebalancing performed by the dispatcher is:

[0162] ;

[0163] The number of dispatchers performing vehicle charging tasks is:

[0164] ;

[0165] The number of dispatchers who went to the rest stop to rest and then went to the station after resting was:

[0166] ;

[0167] Adding the above three parts together, the minimum dispatcher size in the system is:

[0168] ;

[0169] To summarize, the minimum vehicle size required to achieve equilibrium is composed of the sum of the number of passenger-carrying vehicles and vehicles performing rebalancing tasks, the number of vehicles going to charging and charging, and the number of vehicles going to passenger stations after completing charging; the minimum dispatcher size required to achieve equilibrium is composed of the sum of the number of dispatchers performing vehicle rebalancing and self-rebalancing, the number of dispatchers performing vehicle charging tasks, the number of dispatchers going to rest stops to rest, and the number of dispatchers going to passenger stations after resting.

[0170] In step S5, according to the conditions for reaching equilibrium, the minimum vehicle size for reaching equilibrium, and the minimum dispatcher size for reaching equilibrium, the dispatching strategy for vehicle charging, dispatcher rest, and rebalancing is designed, such as Figure 3 As shown, the scheduling strategy of vehicle charging, dispatcher rest and rebalancing is used as the result of coupling rebalancing and scale optimization of the vehicle sharing system.

[0171] Design a coupled scheduling strategy for vehicle charging, dispatcher rest and rebalancing to minimize the scheduling cost and system scale. The specific plan is as follows:

[0172] (1) The goal is to find an optimal rebalancing strategy Minimize the number of empty vehicles and dispatchers on the way, that is, minimize the rebalancing cost:

[0173] and

[0174] ;

[0175] Constraints:

[0176] ;

[0177] ;

[0178] ;

[0179] in, is the customer arrival rate of station passengers at station i, is the rate at which low-battery vehicles arrive at passenger station i, is the proportion of dispatchers who need to rest when arriving at passenger station i, is the customer arrival rate at passenger station j, is the probability that a passenger in passenger station i goes to passenger station j, is the probability that a passenger in passenger station j goes to passenger station i, is the rebalancing rate of vehicles from passenger station i to passenger station j, is the rebalancing rate from passenger station j to passenger station i, is the dispatch rate from passenger station i to vehicle charging station k, is the dispatch rate of vehicle charging station k to passenger station j, is the dispatcher's rebalancing rate from passenger station i to passenger station j, is the dispatcher's rebalancing rate from passenger station j to passenger station i, is the rate at which the dispatcher at passenger station i sends messages to the rest stop m, is the probability that a fatigued dispatcher at passenger station i chooses to go to rest station m for rest, is the probability that the dispatcher of rest stop m goes to passenger stop j, is the sum of the waiting time in queue and the charging time of the vehicle. is the sum of the dispatcher’s waiting time in queue and rest time. is the number of rest stops m that can provide rest for dispatchers at the same time, is the service rate of rest stop m, is the travel time between passenger station i and passenger station j, is the travel time between passenger station j and passenger station i, is the travel time from passenger station i to vehicle charging station k, is the travel time from vehicle charging station k to passenger station j, is the travel time from passenger station i to rest stop m, is the travel time from rest stop m to passenger stop i, N is the set of passenger stops, M is the set of vehicle charging stations, P is the set of rest stops, is the number of rest spaces, is the service rate for each dispatcher, For each charging pile service rate, is the number of charging piles.

[0180] (2) Since charging vehicles and fatigue dispatchers can only participate in system dispatch after they have completed charging or rested, a certain number of vehicles and dispatchers can be deployed in advance at each vehicle charging station and rest station, and dispatched according to the dispatch strategy immediately after the system starts running. Send out vehicles and dispatchers to reduce the delay of charging and rest. Therefore, the number of additional vehicles that should be deployed at each vehicle charging station is:

[0181] ;

[0182] The number of additional dispatchers that should be deployed at each rest stop is:

[0183] ;

[0184] At this point, the rebalancing and scale optimization method of the vehicle sharing system is completed, and the optimal scheduling strategy can be used. Vehicles and dispatchers are sent outward, and the dispatching strategies of vehicle charging, dispatcher rest and rebalancing are used as the coupling rebalancing and scale optimization results of the vehicle sharing system.

[0185] This application proposes a rebalancing and scale optimization method for a vehicle sharing system, which fully considers the joint scheduling process of vehicle charging, dispatcher rest and coupled rebalancing, models the system based on the fluid model, M / M / s and M / D / s queuing theory, and gives a system scale calculation method based on this. The system is modeled based on a probability model and an optimization problem is established. The difficulty of solving it is independent of the system scale and is more suitable for large-scale systems.

[0186] The following is a rebalancing and scale optimization method for a vehicle sharing system, using the simulation software MATLAB and the solution tool LINGO to verify the effectiveness of the proposed method.

[0187] Verification 1:

[0188] First, we explore the relationship between the minimum number of vehicles and dispatchers required to achieve system stability. We randomly generate 10-100 passenger stops in a 100X100 unit environment and calculate the travel time between stops based on the Euclidean distance. , the customer arrival rate of each station is randomly selected between [0.6,1]. The proportion of charging vehicles at each station is 10% of the vehicles arriving at the station, and the proportion of fatigued dispatchers is 5% of the dispatchers arriving at the station. Assume that each vehicle charging station is equipped with 3 charging piles, and the charging pile service rate is [0.1,0.3]. Each rest area is designed to accommodate three dispatchers at the same time, and the dispatcher's designated rest time is set to 10 time units. Figure 4The relationship between the minimum required system size and the number of stations is shown. The number of vehicles and dispatchers increases approximately linearly with the number of stations. The relationship between the proportion of dispatchers performing different tasks and the number of passenger stations is given by Figure 5 It can be seen that as the number of passenger stops increases, the proportion of dispatchers performing vehicle rebalancing and charging trips increases, while the proportion of dispatchers performing rest and self-rebalancing trips decreases. Figure 6 It is shown that with a rest period of 10 units, the required system size decreases as the charging speed increases. Figure 7 The relationship between the minimum required system size and the dispatcher's rest time is shown when the charging rate is 0.2.

[0189] Verification 2:

[0190] The actual traffic data from Didi Chuxing in Chengdu is used to verify the effectiveness of the proposed method. 15468 travel records recorded during the morning peak period of 8:00-10:00 on a certain day are selected. The data set includes the order number, the start and end timestamps of the order, and the longitude and latitude of the pickup location. The K-means clustering algorithm is used to determine a set of passenger stations N, and the pick-up and drop-off locations of the order are clustered into 42 passenger stations. Based on the 42 clustering stations, the required parameters such as customer arrival rate, inter-station routing probability, and actual travel time between each OD station pair can be obtained. In order to ensure the stability of the system, it is set that each station is allocated an average of 43 vehicles and 15 dispatchers at the initial moment. Two systems with and without scheduling are simulated. Figure 6 and attached Figure 7 It shows how the number of customers waiting at each station changes over time. Figure 8 shows the system without intervention, where it is clear that some stops have no waiting customers while others have a large accumulation of customers, this difference occurs because some locations are more popular, resulting in a buildup of vehicles at some locations and a lack of available vehicles at others. This highlights the need for scheduling within the system. In contrast, Fig. 9 The evolution of the system with scheduling is described, showing that scheduling has a significant suppressive effect on the number of waiting passengers, with the number of waiting passengers at each station significantly reduced after scheduling compared to the system without scheduling. Fig.10 is the peak number of people waiting at each station. It can be clearly seen that the peak number of people waiting with the dispatching system is significantly lower than that without the dispatching system. This shows that this method effectively suppresses the number of customers waiting at each station, thereby ensuring the stable operation of the system. Fig.11 Figure 2 shows the distribution of vehicles performing different tasks in a 2-hour simulation experiment. The figure shows that the strategy effectively reduces the number of waiting customers by using fewer rebalancing trips.

[0191] Based on such understanding, the technical solution of the present application, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a computer program product.

[0192] The various embodiments in the present application are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

[0193] The protection scope of the present application is not limited to the above-mentioned embodiments. Obviously, those skilled in the art can make various changes and modifications to the present disclosure without departing from the scope and spirit of the present disclosure. If these changes and modifications fall within the scope of the claims of the present disclosure and their equivalents, the intention of the present disclosure also includes these changes and modifications.

Claims

1. A method for rebalancing and optimizing the scale of a vehicle sharing system, characterized in that: include: In each passenger station, a migration dynamics equation of passengers, vehicles and dispatchers is created. The migration dynamics equation is created based on the operation mechanism of the vehicle sharing system and takes into account vehicle charging scheduling, dispatcher rest scheduling, and vehicle and dispatcher coupled rebalancing scheduling; The M / M / s queuing theory is used for the vehicle charging process to obtain the stability conditions of the charging queue and the average charging time. The M / D / s queuing theory is used for the dispatcher rest process model to obtain the stability conditions of the dispatcher rest queue and the average rest time of the dispatcher. According to the migration dynamics equation, the stability condition of the charging queue, the average charging time, the stability condition of the dispatcher's rest queue and the average rest time of the dispatcher, the conditions for reaching equilibrium, the minimum vehicle size to reach equilibrium and the minimum dispatcher size to reach equilibrium are obtained; According to the conditions for reaching equilibrium, the minimum vehicle scale to reach equilibrium, and the minimum dispatcher scale to reach equilibrium, the dispatching strategy for vehicle charging, dispatcher rest, and rebalancing is designed, and the dispatching strategy for vehicle charging, dispatcher rest, and rebalancing is used as the result of coupled rebalancing and scale optimization of the vehicle sharing system; The migration dynamics equations of customers, vehicles and dispatchers are created in each passenger station, including: the migration dynamics equations of customers at the passenger station, the migration dynamics equations of vehicles at the passenger station and the migration dynamics equations of dispatchers at the passenger station; The migration dynamics equation of the passenger station customers is calculated as follows: ;in, is the customer arrival rate at passenger station i, The rate at which vehicles are assigned to customers at passenger station i, is the idle vehicle existing in passenger station i at the current moment, is the dynamic change rate of customers at the passenger station, is the number of passengers existing in passenger station i at the current moment, and H() is the Heaviside function.

2. The vehicle sharing system rebalancing and scale optimization method according to claim 1, characterized in that: The migration dynamics equation of vehicles at the passenger station i is composed of the sum of the rate at which passenger vehicles leave the passenger station i, the rate at which passenger vehicles arrive from other passenger stations j, the rate at which the dispatcher drives the low-power vehicle from the passenger station i to the vehicle charging station k, the rate at which the dispatcher drives the vehicle from the vehicle charging station k to the passenger station i, the rate at which the dispatcher drives the rebalancing vehicle to leave the passenger station i, and the rate at which the dispatcher drives the rebalancing vehicle to arrive at the passenger station i.

3. The vehicle sharing system rebalancing and scale optimization method according to claim 1, characterized in that: The migration dynamics equation of the dispatcher at the passenger station i is composed of the sum of the rate at which the dispatcher drives the low-power vehicle from the passenger point i to the vehicle charging station k, the rate at which the dispatcher drives the vehicle from the vehicle charging station k to the passenger station i, the rate at which the dispatcher drives the rebalancing vehicle to leave the passenger station i, the rate at which the dispatcher drives the rebalancing vehicle to arrive at the passenger station i, the rate at which the dispatcher who needs to rest goes from the passenger station i to the rest station m, the rate at which the dispatcher returns to the passenger station i after resting, the rate at which the rebalancing dispatcher leaves the passenger station i, and the rate at which the rebalancing dispatcher arrives at the passenger station i.

4. The vehicle sharing system rebalancing and scale optimization method according to claim 1, characterized in that: The charging queue stability condition is that the vehicle arrival rate of the vehicle charging station k is less than the total service rate, and the calculation formula is as follows: ;in, is the dispatch rate from passenger station i to vehicle charging station k, is the rate at which low-battery vehicles arrive at passenger station i, The probability that a low-battery vehicle at passenger station i chooses to go to vehicle charging station k for charging, is the number of charging piles, For each charging pile service rate, is the total service rate of all charging piles in vehicle charging station k; The average charging time is the sum of the vehicle waiting time and the charging time, and is calculated as follows: ;in, is the sum of the waiting time in queue and the charging time of the vehicle. is the vehicle arrival rate at vehicle charging station k, l ranges from 0 to An integer.

5. The vehicle sharing system rebalancing and scale optimization method according to claim 1, characterized in that: The dispatcher rest queue stability condition is that the dispatcher arrival rate is less than the total service rate, and the calculation formula is as follows: ;in, is the rate at which the dispatcher at passenger station i sends messages to the rest stop m, is the probability that a fatigued dispatcher at passenger station i chooses to go to rest station m for rest, is the proportion of the dispatcher who needs to rest when arriving at passenger station i, and rest station m contains rest space, and each dispatcher has a rest time of , the service rate of rest stop m is ; The average rest time of the dispatcher is calculated as follows: ;in, The busyness of the site, is the dispatcher arrival rate at rest stop m, is the probability that there is no customer queue, is the average rest time of the dispatcher, l represents the time from 0 to An integer.

6. The vehicle sharing system rebalancing and scale optimization method according to claim 1, characterized in that: The condition for reaching equilibrium is calculated as follows: ;in, , is the dispatch rate of vehicle charging station k to passenger station j, is the proportion of vehicles returning to passenger station j from vehicle charging station k, is the vehicle arrival rate at vehicle charging station k, is the probability that a passenger in passenger station j goes to passenger station i, is the dispatcher rate sent from rest stop m to passenger stop j, is the proportion of dispatchers in rest stop m who choose to go to passenger stop j, is the dispatcher arrival rate at rest stop m, is the rebalancing rate of vehicles from passenger station i to passenger station j, is the rebalancing rate from passenger station j to passenger station i, is the dispatcher's rebalancing rate from passenger station i to passenger station j, is the rebalancing rate of the dispatcher from passenger station j to passenger station i, is the customer arrival rate of station passengers at station i, is the customer arrival rate of station passengers at station j, is the rate at which dispatchers are sent from passenger station i to rest stop m, is the dispatch rate from passenger station i to vehicle charging station k.

7. The vehicle sharing system rebalancing and scale optimization method according to claim 1, characterized in that: The minimum vehicle scale for achieving equilibrium is composed of the sum of the number of passenger vehicles and vehicles performing rebalancing tasks, the number of vehicles heading for charging and the number of vehicles being charged, and the number of vehicles heading for the station after completing charging; The minimum size of dispatchers required to achieve equilibrium is composed of the sum of the number of dispatchers who perform vehicle rebalancing and self-balancing, the number of dispatchers who perform vehicle charging tasks, the number of dispatchers who go to rest stops and the number of dispatchers who go to stations after resting.

8. The vehicle sharing system rebalancing and scale optimization method according to claim 1, characterized in that: The scheduling strategy for vehicle charging, dispatcher rest and rebalancing is obtained by solving the following nonlinear programming problem, and the calculation formula is as follows: and ； Constraints: ; ; ;in, is the customer arrival rate of station passengers at station i, is the rate at which low-battery vehicles arrive at passenger station i, is the proportion of dispatchers who need to rest when arriving at passenger station i, is the customer arrival rate at passenger station j, is the probability that a passenger in passenger station i goes to passenger station j, is the probability that a passenger in passenger station j goes to passenger station i, is the rebalancing rate of vehicles from passenger station i to passenger station j, is the rebalancing rate from passenger station j to passenger station i, is the dispatch rate from passenger station i to vehicle charging station k, is the dispatch rate of vehicle charging station k to passenger station j, is the dispatcher's rebalancing rate from passenger station i to passenger station j, is the dispatcher's rebalancing rate from passenger station j to passenger station i, is the rate at which the dispatcher at passenger station i sends messages to the rest stop m, is the probability that a fatigued dispatcher at passenger station i chooses to go to rest station m for rest, is the probability that the dispatcher of rest stop m goes to passenger stop j, is the sum of the waiting time in queue and the charging time of the vehicle. is the sum of the dispatcher’s waiting time in queue and rest time. is the number of rest stops m that can provide rest for dispatchers at the same time, is the service rate of rest stop m, is the travel time between passenger station i and passenger station j, is the travel time between passenger station j and passenger station i, is the travel time from passenger station i to vehicle charging station k, is the travel time from vehicle charging station k to passenger station j, is the travel time from passenger station i to rest stop m, is the travel time from rest stop m to passenger stop i, N is the set of passenger stops, M is the set of vehicle charging stations, P is the set of rest stops, is the number of rest spaces, is the service rate for each dispatcher, For each charging pile service rate, is the number of charging piles.

9. The vehicle sharing system rebalancing and scale optimization method according to claim 8, characterized in that: The scheduling strategy for vehicle charging, dispatcher rest and rebalancing also includes: the number of additionally deployed vehicles and the number of additionally deployed dispatchers, wherein the number of additionally deployed vehicles is calculated as follows: ;in, is the number of additional vehicles deployed, is the dispatch rate from passenger station i to vehicle charging station k, is the sum of the waiting time in queue and the charging time of the vehicle. is the travel time from passenger station i to vehicle charging station k; The number of additional dispatchers deployed is calculated as follows: ;in, is the number of additional dispatchers deployed, is the rate at which the dispatcher at passenger station i sends messages to the rest stop m, is the travel time from passenger station i to rest stop m, It is the sum of the dispatcher's waiting time in queue and rest time.

Images