JP5974716B2 - Car sharing system operation planning system, operation planning method - Google Patents

Description

  The present invention relates to an operation planning system for a car sharing system.

  In recent years, car sharing in which a vehicle is shared by a plurality of users has become widespread. For car sharing that can be used for a short period of time, an important service index is “to be able to ride when a user wants to ride”. This is because the profit is lost if the vehicle cannot be supplied when the user wants to ride. In addition, with the recent popularization of electric vehicles, the problem of securing charging time is also added. If the vehicle is being charged, it cannot be rented to the user, and the profit is lost as well. Therefore, adjustment of the vehicle operation plan, such as how to arrange vehicles at each station, is an important issue for operators.

  In many car sharing systems, a user takes a boarding procedure at a station and returns it to the same station after use. On the other hand, there is a desire to return the vehicle to a station different from the boarding station. This form, commonly referred to as drop-off, is not popular in car sharing systems. This is because in car sharing, there are many stations, and reservations, changes, and cancellations are possible until just before, so it is difficult to make a vehicle forwarding plan.

In the car sharing system, technologies for facilitating vehicle management include the following.
The car sharing support system described in Patent Document 1 has a feature that the current position of the vehicle can be easily grasped. According to the present invention, the management server grasps in real time the arrangement status of the vehicle and the vacant status of the parking space, and notifies the user who wants to use it to improve convenience.

  Moreover, the vehicle management server described in Patent Document 2 has a feature that it can automatically manage a charging schedule of an electric vehicle. In the present invention, the management server can manage the battery state of each vehicle and automatically adjust the vehicle charging schedule.

JP 2004-178385 A JP 2010-231258 A

  As described above, there is a desire to use a vehicle between arbitrary stations in car sharing. However, the above-described technique can assist the management of the vehicle, but cannot adjust the vehicle operation plan. For example, if you know that there is a lot of demand from station A to station B in the morning on weekdays and that there is a lot of demand from station B to station C in the afternoon of the same day, The vehicle can be routed from station C to station A at night on weekdays. However, none of the inventions described in Patent Documents 1 and 2 can realize this.

Such demand prediction can be established to some extent by performing data mining on past usage record data. However, in data mining, although it is possible to obtain a movement demand between stations, it is not possible to obtain information on what kind of vehicle operation plan should be made to satisfy the request. For example, on holidays when there are many users, we want to arrange the vehicle so that the waiting time of the user is the shortest, and on weekdays when there are few users, we want to arrange the vehicle so that the operation rate of the entire vehicle is equal to or higher than a predetermined value. Even if there was a request, it could not be realized by the conventional technology.

  The present invention has been made in consideration of the above-described problems, and an object of the present invention is to provide an operation planning system capable of determining a vehicle operation plan suitable for conditions in a car sharing system.

  An operation planning system for a car sharing system according to the present invention includes demand amount acquisition means for acquiring demand amount data, which is a movement demand amount between stations under a specific environmental condition, and network configuration data representing a connection relationship between stations. Network storage means for storing, and model template storage means for storing a model generation template that is a template for generating a mathematical model representing movement of a vehicle between stations.

  The operation planning system for a car sharing system according to the present invention inputs a requirement condition input means for inputting a requirement condition that is a constraint for determining a vehicle operation plan, and an optimum solution condition that is a condition for an optimum solution. Mathematical expression representing vehicle movement between stations by applying the network configuration data and the movement demand in the specific environmental condition acquired by the demand acquisition unit to the optimal solution condition input unit and the model generation template Model generation means for generating a movement model, which is a model, and data calculation means for determining a vehicle operation plan that satisfies the input requirement and optimal solution conditions using the generated movement model.

The vehicle operation plan refers to a plan for at least a part of parameters that can be adjusted by a business operator to provide services, such as the number of vehicles arranged at each station, usage charges, and vehicle forwarding schedules. The operation planning system according to the present invention is a system for determining the vehicle operation plan.
The demand amount data is data representing the movement demand amount between stations. Specifically, it is data representing the amount of users generated by departure and destination in specific environmental conditions (day of the week, weather, time, fee setting, etc.). If the condition is given to the demand amount acquisition means, for example, “from 1 to 2 pm on a rainy Sunday,” the demand amount such as “the movement demand amount from station A to station B is 3 people per 15 minutes”. Data can be obtained. The demand amount data may be obtained by performing data mining on past usage record data, usage intention questionnaire results, and the like.

  The network configuration data is data representing the structure of a station and a route connecting the stations. That is, the data represents the topology of the network through which the vehicle can move. The network configuration data may include a distance between stations, an average traveling speed, an average travel time, and the like in addition to the number of stations and a connection relationship between the stations.

The number of vehicles in a specific state such as “moving between stations A and B” or “parking in station A” can be represented by a variable. The relationship between variables can be expressed using a mathematical model. The model generation template is data in which rules for constructing the mathematical model are described. For example, it can be defined that “the number of vehicles currently in station A is the number of vehicles that have been parked, plus the number of vehicles that have arrived, and the number of vehicles that have left the vehicle subtracted”. . By applying the network configuration data to the model generation template, a mathematical model in the network can be generated.

  Further, the request condition received by the request condition input means is a condition that should be satisfied in the operation of the car sharing system. The requirement may be any event that can be expressed by a mathematical model. Examples include “at least one vehicle at each station”, “the time from when a user arrives until the vehicle becomes available within 5 minutes”, and the like. The above-described variables are determined within a range that satisfies the requirements.

  The optimum solution condition accepted by the optimum solution condition input means is an optimum solution condition obtained within a range that satisfies the above-mentioned requirement conditions. For example, “maximize profit for a specified period”, “maximize vehicle operation rate”, “minimize the worst value of user waiting time”, and the like. The optimum solution condition is to maximize or minimize a specific value, and may be any value as long as the specific value can be expressed by a mathematical model. By specifying the optimal solution condition, the value of the variable representing the vehicle operation plan can be specified.

  The model generation means can generate a mathematical model representing movement of the vehicle under the environmental condition by applying the network configuration data to the model generation template and further applying the demand amount data under the specific environmental condition. . The data calculation means adds the input requirement condition to these mathematical models, and determines a solution that satisfies the specified requirement condition and the optimal solution condition by mathematical programming. And the value regarding a vehicle operation plan is extracted from the determined variable.

  As described above, the operation planning system according to the present invention can acquire the optimum value related to the vehicle operation plan based on the demand amount data, the network configuration data, the request condition, and the optimum solution condition. For example, “How many vehicles should be placed at which station during the previous day in order to make the user's waiting time within 5 minutes and maximize profits”, “ It is possible to calculate a solution to the question “how many minimum number of vehicle forwarding personnel should be arranged to avoid the maximum vehicle availability”.

  The network configuration data may include a time restriction when the vehicle moves between stations. For example, the travel time of the vehicle can be accurately grasped by including the required time, distance, average speed, and the like when the vehicle moves between stations.

  The movement model may include a mathematical expression that represents the number of vehicles that are stopped at each station or that are moving between the stations for each predetermined time step. A time step is a time expressed at regular intervals. Thus, it is desirable that the movement model is an extension of the model representing the relationship between spaces in the time axis direction.

  The movement model may further include a mathematical expression representing the movement of the user. That is, by representing the user who drives the vehicle by a mathematical model, the flow of the vehicle in the network can be represented more accurately.

The movement model includes a mathematical expression that represents the number of users at each station for each predetermined time step, and the model generation unit calculates the number of users generated at the station at each predetermined time step. The determination may be made using quantity data.
Users at the station can also be represented by mathematical models. Further, the number of users generated at the station, that is, the number of users arriving at the station can be obtained from the demand amount data obtained by the demand amount obtaining means.

Further, the demand amount acquisition unit acquires a plurality of demand amount data corresponding to a plurality of environmental conditions, the model generation unit generates a plurality of mathematical models using the plurality of demand amount data, and the data The calculation means may be characterized by performing an operation on the plurality of mathematical models.
In this way, by performing the calculation a plurality of times using the demand amount data under different environmental conditions, the parameters constituting the environmental conditions can be calculated as values related to the vehicle operation plan. For example, when there is “charge unit price per hour” as a parameter constituting the environmental condition, an optimal charge unit price can be determined by performing a plurality of calculations while changing the charge unit price.

  In addition, the request condition input unit receives an input about a target time zone for calculating a vehicle operation plan, the model generation unit generates a movement model for a time step corresponding to the input time zone, and the data The calculating means may be characterized by performing an operation on the input time zone. By doing in this way, since it can calculate only about a required range, a vehicle operation plan can be obtained in a shorter time.

  In addition, this invention can be specified as an operation planning system of a car sharing system including at least a part of the above means. The present invention can also be specified as an operation planning method including at least a part of the above processing. The above processes and means can be freely combined and implemented as long as no technical contradiction occurs.

  ADVANTAGE OF THE INVENTION According to this invention, the operation planning system which can determine the vehicle operation plan suitable for conditions can be provided in a car sharing system.

It is a figure showing the relationship between a station and a link in a car sharing system. It is a figure which simplifies and demonstrates the relationship between a station and a link. It is a figure which simplifies and demonstrates the movement model which concerns on 1st embodiment. It is a figure explaining the movement model in case the link between three stations is looping. It is a system configuration figure of the vehicle operation planning device concerning a first embodiment. It is a figure explaining demand prediction data concerning a first embodiment. It is a process flowchart of the vehicle operation planning apparatus which concerns on 1st embodiment. It is an example of the network configuration data which concerns on 1st embodiment. It is a process flowchart of the vehicle operation planning apparatus which concerns on 2nd embodiment. It is an example of restriction data concerning a modification of an embodiment.

(First embodiment)
<Outline of car sharing system>
Prior to the description of the embodiment, a car sharing system which is a premise of the present embodiment will be described. FIG. 1 is a diagram illustrating the arrangement of stations and the movement path between stations. A station is a parking lot where a vehicle is arranged, and a user visits the station, gets on the vehicle, and moves. After using the vehicle, the user returns the vehicle to the station at the destination. At this time, the user is charged a usage fee according to the travel distance or time. As described above, the car sharing system that is the premise of the present embodiment is a system that allows a vehicle to be discarded.

  In this example, it is assumed that there are seven stations from station A to station G. The actual route between stations includes an infinite number of routes depending on the arrangement of roads. In the description of the embodiment, the route between stations is represented by one link representing the shortest route.

When abandonment is adopted in a car sharing system, there is a possibility that the arrangement of vehicles is biased depending on how it is used. For example, an event may occur in which the station A and the station B are short of vehicles, or the station E and the station F are full and cannot enter. In order to reduce vehicle bias, it is necessary to appropriately determine a vehicle operation plan. The vehicle operation plan refers to a plan for at least a part of parameters that can be operated by the sharing operator for service provision. Here, the following three are exemplified. (1) Number of vehicles arranged at each station at a certain point (2) Number of dispatched personnel at each station at a certain point (3) Usage charge setting for each section

The number of vehicles arranged is the number of vehicles arranged at the station. For example, when there is a lot of movement from station A to station B in the time zone from 6 o'clock to 8 o'clock, it is conceivable to move the vehicle from station B to station A by 6 o'clock.
Further, the forwarding personnel are personnel for forwarding the vehicle between stations. The forwarding personnel forwards the vehicle between the corresponding stations when the vehicle is biased. It is preferable that the forwarding personnel should have enough people to eliminate the vehicle bias.

The usage fee is a fee charged to the user when using the sharing vehicle. The charge method may be a time system or a distance system. In the present embodiment, it is assumed that a section fee system is adopted in which a fee is added each time one link shown in FIG. 1 is traveled. For example, it may be 150 yen per section.
Also, the usage fee may be negative. For example, when the station A is frequently used from the station A, the charge from the station B to the station A is made negative, and the charge is returned to the user who has traveled in this section. By doing so, the user can be assisted in forwarding the vehicle.

  In this way, by appropriately adjusting the vehicle operation plan, it is possible to supply vehicles to a larger number of users as compared with the case where adjustment is not performed. In addition, although there exist many parameters about a vehicle operation plan other than what was illustrated, in this embodiment, in order to simplify description, a vehicle operation plan will be demonstrated focusing on said (1).

<Calculation based on mathematical model>
Here, the method of determining the vehicle operation plan by the vehicle operation planning apparatus according to the present embodiment is the example of FIG. 2, that is, the case where there are two paths connecting the stations A, B, and C to each station. A brief explanation. Hereinafter, a path connecting the stations is referred to as a link.

  In this embodiment, a model representing the number of vehicles and the number of users is constructed, and a vehicle operation plan is calculated using the model. FIG. 3 is a diagram showing a model constructed for the network of FIG. The horizontal axis is a time axis, and in this example, the state of the station and the link at regular intervals is expressed. For example, when time 0 is 0: 0, time 1 can be 0: 5 and time 2 can be 0:10. Each engraved time is referred to as a time step. The step size of the time step is 5 minutes in this example, but may be set in any manner.

Stations and links for each time step, that is, circles in the figure are referred to as nodes in the following description. The number of vehicles existing in the node can be expressed by a variable. For example, the number of vehicles present at station A (node 101) at time 0 can be expressed as v _{(A, 0),} and is also moving from station A to station B, ie, present at link AB (node 102). The number of vehicles to be played can be expressed as v _{(AB, 0)} .

  Moreover, the relationship between variables can be expressed by a mathematical expression. For example, the total number of vehicles present at station A at time 0 is the number of vehicles that remain at station A until the next time step, that is, the number of vehicles that transition to time 1 via reference numeral 103; This is the total of the number of vehicles that have flowed out to the link AB, that is, the number of vehicles that transition to the time 1 via the reference numeral 104. Similarly, the number of vehicles present at station A (node 105) at time 1 is the number of vehicles present at station A at time 0 and staying at station A until the next time step. This is the total number of vehicles that exist on the link AB and have entered the station A until time 1.

In this way, by representing the number of vehicles existing at each node by the number of vehicles that have flowed out or entered from each node, all of the models shown in FIG. 3 can be represented by mathematical expressions. The mathematical formula group is a movement model in the present invention. The necessary variables are the following three types.
(1) v _{(X, t)} : Number of vehicles existing at station or link X at time t (2) out _{(X, Y, t)} : Between time t and t + 1, Number of vehicles flowing out of station or link X and heading to destination station Y (3) in _{(X, Y, t)} : From time t to t + 1, starting from station Y The number of vehicles flowing into the station or link X, that is, the number v _{(B, t + 1)} of vehicles existing at the station B at the time t + 1 can be expressed by Equation 1.

Similarly, the number v _{(AB, t + 1)} of vehicles existing on the link AB at time t + 1 can be expressed by Equation 2.

Here, a condition regarding movement between stations is further added. For example, if 2 time steps are required to move from station A to station B and 3 time steps are required to move from station C to station B, the number of vehicles flowing into station B can be expressed by Equation 3. Can do.

  By applying the equation (equalities) described above to the three time steps shown in FIG. 3, the stations A, B, C and the links AB, BC between time 0 and time 2 are used. The relationship of vehicles existing in can be expressed by mathematical formulas.

The vehicle operation planning apparatus according to the first embodiment obtains the number of vehicles arranged for each station, that is, the number of vehicles to be arranged in stations A, B, and C as part of the vehicle operation plan. The number of vehicles to be obtained, that is, the number of vehicles arranged at stations A, B, and C at time 0 can be represented by variables v _{(A, 0)} , v _{(B, 0)} , and v _{(C, 0)} , respectively. it can.

Next, in order to express when the vehicle starts to move, information about the usage demand for car sharing is added.
Before describing the usage demand, a method for expressing a user by a mathematical expression will be described. The number of users who are at the station X at the time t and whose destination is the station Y can be expressed as a variable C _{(X, Y, t)} . The user who arrives at the station gets on the vehicle if there is a vehicle, starts moving, and waits in the station if there is no vehicle. That is, it is necessary to express the following if-then condition by a mathematical expression. The following conditions represent the relationship between the number of vehicles, the number of users, and the number of vehicles departing from a certain station at a certain time step. v is the number of vehicles, C is the number of users, and out is the number of vehicles leaving. if v−C ≦ 0 then out = v else out = C
That is, if the number of vehicles is equal to or less than the number of users, the number of departures is the number of vehicles, and if there are more vehicles than users, the number of departures is the number of users.

In order to make the above conditional expression into a form that can be handled by a mathematical model, Expression 4 is introduced. The constant M is a number sufficiently larger than the number of parked stations and the number of users who can wait in the station, and the variable z takes a value of 0 or 1. In Formula 4, even if the number of vehicles v and the number of users C change, the variable z is determined so that the formula always holds. The following description is an example in which only the user who goes to the station B in the station A is considered.

When the variable z is used to represent the if-then condition, the four equations shown in Equation 5 can be configured. In the case of v−C ≦ 0, from Equation 4, z = 0 needs to be satisfied. As a result, since Equations 5 (3) and (4) always hold, Equations 5 (1) and (2) are satisfied. In addition, it is necessary that out−v = 0. Similarly, in the case of v−C> 0, z = 1 needs to be satisfied, and Equations 5 (1) and (2) always hold. Therefore, in order to satisfy Equations (3) and (4), out−C = Must be zero.

From Equation 4 and Equation 5, the number of vehicles leaving the station can be specified from the number of users at the station and the number of vehicles existing at the station. The number of users who are at station A and whose destination is station B can be expressed by Equation 6. Here, D _{(A, B, t)} is the number of users who arrive (arrive) at station A at time t and whose destination is station B.

  In other words, if there is data about the time when the user arrives at the station and the destination (demand data in the present invention), the number of users at the station can be determined for each time step. Equations 4 to 6 are equations that focus on the movement from station A to B, but the movement between other stations can also be expressed by equations in the same manner.

The constraint conditions exemplified above are constraints that must be satisfied in accordance with the laws of physics, but any other constraint that is necessary for operation can be imposed. The arbitrary constraints may be any as long as they can be expressed by mathematical expressions. For example, “There are one or more vehicles in station A at all time steps”. This exemplified condition can be expressed by an inequality of Equation 7. In the present invention, an arbitrary restriction necessary for operation as shown in Expression 7 is referred to as a requirement.

In mathematical programming, it is necessary to add a condition to specify the optimal solution. This condition is called the optimum solution condition. The optimum solution condition is for maximizing or minimizing a specific value, and may be anything as long as the specific value can be expressed by a mathematical expression. For example, “maximize sales” “minimize cost”. The optimal solution condition is composed of two pieces of information: “a specific value represented by a mathematical expression” and “what to do with the specific value”. In the present embodiment, the former is called an optimization value, and the latter is called an optimization condition.
As an example, the sales at a certain time step when the stations are arranged as shown in FIG. 2 can be expressed by Expression 8 (in the case of 150 yen per section). That is, Expression 8 represents the optimization value, and the optimization condition is “maximization”.

The vehicle operation planning apparatus according to the first embodiment solves the optimization problem based on a plurality of mathematical formulas described above by mathematical programming, so that the number of vehicles at stations A, B, and C at time 0 in the vehicle operation plan ( That is, the variables v _{(A, 0)} , v _{(B, 0)} , and v _{(C, 0)} can be determined.
In order to obtain the vehicle operation plan by the mathematical programming method, it is necessary to generate a large number of mathematical expressions. However, the vehicle operation planning apparatus according to the present invention can automatically generate the mathematical expressions using the stored template. A detailed method for generating mathematical expressions will be described later.

  In the example of FIG. 3, the relationship between the stations A, B, and C is shown only for three time steps. However, in the vehicle operation planning apparatus according to the first embodiment, the relationship between a plurality of stations is calculated. Only the time step is defined. For example, if the link connecting stations A, B, and C forms a loop and the target time step is 10 steps, the movement model to be constructed is as shown in FIG. In FIG. 4, the links connecting the two stations are described as if they were one, but actually two links are defined for each direction.

<System configuration>
The vehicle operation planning apparatus according to the first embodiment will be described with reference to FIG. 5 which is a system configuration diagram. The vehicle operation planning device 10 according to the first embodiment is for determining the number of vehicles to be placed in each station in advance in order to satisfy the specified requirement condition and the optimal solution condition for the specified time zone. It is a computer. The vehicle operation planning device 10 may be a combination of a plurality of computers.

The input / output unit 11 is a means for acquiring various conditions for obtaining the vehicle operation plan from the user, and a means for presenting the obtained vehicle operation plan to the user. Consists of a liquid crystal display, keyboard, touch panel, and the like.
The network information storage unit 12 is a means for storing network configuration data of a target car sharing system, that is, data representing a connection relationship between stations. The network configuration data of the target system is created and stored in advance.

  The demand prediction data 13 is data in which past usage records in the target car sharing system are recorded. An example of the demand forecast data is shown in FIG. The demand forecast data is raw data that has not been aggregated. The demand prediction data 13 may not be past usage record data as long as the demand amount for the user's car sharing can be defined. For example, as shown in FIG. 6B, data based on a questionnaire result may be used.

The demand amount acquisition unit 14 is a means for acquiring a predicted demand amount under specific conditions (demand amount data in the present invention) from the demand prediction data 13 that is raw data. The specific condition is referred to as an environmental condition. The environmental conditions consist of a plurality of items such as days of the week, weather, time zones, and usage charges. When an arbitrary environmental condition is given, the demand amount acquisition unit 14 can acquire a predicted demand amount based on the environmental condition.
For example, when four “Month”, “Day of the week”, “Weather”, and “Time” are used as environmental conditions, it is possible to specify “Sunday in May (weather: fine) from 13:00 to 14:00”. The environmental condition may be any item as long as the demand amount acquisition unit 14 can process it. For example, the condition may be “when it is Saturday and the section fee is set to 150 yen”.

  The demand amount acquisition unit 14 estimates the demand amount in the designated environmental condition using a known data mining method. Specifically, regression analysis or the like is performed on the demand forecast data, and the estimated arrival interval of the user when a specified environmental condition is assumed is acquired. For example, information such as “when a specified environmental condition is applied, a user moving from station A to station B arrives at station A every 15 minutes” can be obtained. Then, the estimated arrival time of the user is determined from the acquired estimated arrival interval. The estimated arrival time may be obtained by simply assigning the obtained estimated arrival interval from the start time, or may be obtained using a Poisson distribution or the like. Further, any data mining technique used by the demand amount acquisition unit 14 may be used as long as it is a known technique. For example, a Bayesian filter, a neural network, a decision tree, linear regression, or the like can be used.

  The model generation unit 15 is model generation means for generating a movement model in the present invention. The model generation unit 15 stores a model template. By applying the network structure acquired from the network information storage unit 12 to the model template, it is possible to generate a mathematical expression group that represents the movement of the vehicle. Moreover, it has the function which replaces the demand amount data which the demand amount acquisition part 14 acquired, the requirement conditions and optimization value which were acquired from the user through the input-output part 11 with a numerical formula, and adds it to a numerical formula group.

Here, a specific example of the model template stored in the model generation unit 15 will be described. A model template is information representing a generation method of a movement model regardless of a specific network configuration. For example, it is information representing the following three constraint generation methods.
(Condition 1) The number of vehicles existing in a node is calculated by adding the number of vehicles that have flowed in from other nodes to the current time step to the number of vehicles in the previous time step, and The number is subtracted.
(Condition 2) A vehicle that has flowed out of a certain node flows into the destination node after a predetermined time.
(Condition 3) When the vehicle is available, the user who arrives at the station gets on, drives the vehicle, and heads for the destination. The user who arrives at the destination leaves the station.

  Among these, (Condition 1) and (Condition 2) are conditions that should be used. (Condition 1) and (Condition 2) are conditions (flow rate conservation law) that guarantees that the total number of vehicles does not change when a node enters and leaves, that is, the vehicle does not move out of the network. Condition 1 corresponds to Equations 1 and 2 described above, and Condition 2 corresponds to Equation 3 described above. First, a method of generalizing (Condition 1) and (Condition 2) using mathematical expressions will be described. In the description of the mathematical expression, X is a target station or link, and Yi is a station other than the target station or link X.

Condition 1 is a condition that represents the number of vehicles existing in a station or link. The number of vehicles existing in a station or link, as exemplified by Equations 1 and 2, can be generalized in the form of Equation 9. That is, Equation 9 is stored as a model template, and after acquiring specific network configuration data, all stations, links, and time steps may be expanded.

Condition 2 is a condition representing travel time. The relationship between the number of outflows from the node and the number of inflows into the node, as exemplified by Equation 3, can be generalized in the form of Equation 10. That is, Equation 10 is stored as a model template and developed for all stations and time steps. Note that TravelTime (X, Yi) is a value representing the required travel time (number of time steps) from station X to station Yi. The required travel time can be obtained from network configuration data and substituted when the mathematical expression is expanded.

Similarly, for (Condition 3), it is only necessary to store a template that can obtain Formulas 4 to 6 by expanding Formulas for each station. The content of (Condition 3) may be changed in accordance with a desired condition. For example, “A user who arrives at a station gets on the vehicle when a vehicle that can carry a predetermined number of people is available and heads for the destination.
Arbitrary conditions may be added. For example, a condition that “a user who could not get on the vehicle moves to a node representing lost demand” may be added.
Further, the model generation unit 15 stores a template for replacing the requirement condition and the optimization value with a mathematical expression. Details will be described in the description of the processing flowchart.
Here, an example in which the template is expressed by a mathematical formula has been described. However, the template may be expressed in an arbitrary format as long as the model generation unit 15 can process the template. For example, it may be expressed by a programming language.

  The data calculation unit 16 is a means for calculating a value that satisfies the optimization condition by mathematical programming for the movement model generated by the model generation unit 15, that is, the mathematical expression group. As long as the data calculation part 16 is a solver (optimization solver) which can solve a mathematical programming problem, what kind of method may be used.

<Process flowchart>
Next, a processing method performed by the vehicle operation planning apparatus according to the present embodiment will be specifically described. FIG. 7 is a flowchart of processing performed by the vehicle operation planning apparatus according to the present embodiment.
First, in step S11, the input / output unit 11 acquires a calculation condition from the user. The calculation conditions are the following five data.
(1) Time zone conditions (2) Environmental conditions (3) Requirements (4) Optimization values (5) Optimization conditions

  Each will be described in turn. The time zone condition is information representing a start time and an end time that are targets for calculating the vehicle operation plan. It is desirable to match the time zone conditions with the environmental conditions input to the demand amount acquisition unit 14. For example, when the demand amount acquisition unit 14 can acquire demand amount data by specifying a day of the week and a time, the start time and the end time are specified by the day of the week and the time. If environmental conditions can be specified by date and time, specify the date and time. By specifying the time zone condition, the total time step used for the calculation is determined. As the length of one time step, a value held in advance by the apparatus is used. For example, one time step may be 5 minutes or 1 minute.

  The environmental condition is a condition for the demand amount acquisition unit 14 to obtain demand amount data. As described above, the environmental condition may be input in any way as long as it can be processed by the demand amount acquisition unit 14. For example, when five “Month”, “Day of the week”, “Weather”, “Time zone”, and “Section fee” can be processed, only “Month”, “Day of the week”, and “Time zone” may be input.

The requirements are conditions that must be satisfied in the operation of the car sharing system, and are conditions specified by the system user, that is, the car sharing company. The request condition may be any event as long as it can be expressed by a mathematical expression. Here, the following three are exemplified.
(A) “All time steps, at least one vehicle at all stations” (B) “At all stations, the time for zero vehicles to be less than 3 steps”
(C) “The number of users waiting at all time steps and all stations (total number of people by destination) must be less than three”

Among the illustrated requirements, (A) can be expressed by expanding the above-described Expression 7 for all stations and all time steps. For (B), Formula 11 may be expanded for all stations and all time steps. For (C), Formula 12 may be expanded for all stations, all destinations, and all time steps.

The optimization value and the optimization condition are elements constituting the optimal solution condition. The optimum solution condition maximizes or minimizes a specific value as described above. The specific value is an optimization value, and information indicating “maximization” and “minimization” is an optimization condition.
Take an example. For example, taking the sales S obtained per time step for a certain station X as an example, Equation 13 is obtained. In Equation 13, the section fee is represented by F, and the number of sections from the station X to Yi is represented by n (X, Yi). If Formula 13 is developed for all stations and all time steps and totaled, the total sales will be obtained. The total sales is an optimization value, and information that “maximizes” the total sales is an optimization condition.

Among the input calculation conditions, (1), (3), and (4) are transmitted to the model generation unit 15, (2) is transmitted to the demand amount acquisition unit 14, and (5) is transmitted to the data calculation unit 16.
In the present embodiment, the system user creates and inputs the mathematical expression representing the requirement condition and the optimization value as described above, and the model generation unit 15 expands the input mathematical expression to generate the mathematical expression. . The above mathematical formula may be stored in advance. When the mathematical formula is stored in advance, the mathematical formula used by the user may be selected, or a value to be substituted into the mathematical formula may be designated as necessary.

  In step S <b> 12, first, the model generation unit 15 acquires network configuration data from the network information storage unit 12. FIG. 8 shows network configuration data representing the network shown in FIG. The network configuration data defines the connection relationship between the stations and links constituting the network and the number of time steps required for movement between the stations. For example, if the time required for moving from station A to station B is 4 time steps and 1 time step is 5 minutes, the time required for movement is 20 minutes.

  Then, a mathematical formula is generated using the stored model template and the acquired network configuration data. The part of TravelTime in Condition 2 (Formula 10) described above can be obtained from the network configuration data. In the example of FIG. 8, since there are seven stations, the equation shown in Equation 1 is generated for 7 stations × target time step. In this example, since there are six destinations that can go from station A, there are six variables out and in, respectively. Similarly, the mathematical expressions 2 to 6 are generated in a number corresponding to the number of stations and links and the target time step. The generated mathematical formula is temporarily stored by the model generation unit 15.

  Next, in step S13, the demand amount acquisition unit 14 acquires the demand amount data corresponding to the environmental condition acquired in step S11 by performing a data mining process on the demand prediction data 13. As described above, the demand amount data can be obtained in a format such as “when the user arrives at what time” for each departure place and destination. The acquired data is transmitted to the model generation unit 15 and is substituted into a variable D corresponding to the time when the user arrives among a plurality of mathematical expressions represented by the mathematical expression 6.

  Step S14 is a process of adding a mathematical expression representing the required condition and a mathematical expression representing the optimization value. First, the model generation unit 15 generates a mathematical expression representing the request condition input in step S11. The required conditions such as Formula 11 and Formula 12 are expanded according to the number of stations (or the number of destinations and the number of time steps) to generate a formula. Further, an optimization value such as Expression 13 is similarly developed and stored in the model generation unit 15. The optimization condition is transmitted to the data calculation unit 16.

Next, in step S15, the formula group stored in the model generation unit 15 is transmitted to the data calculation unit 16, and the data calculation unit 16 mathematically calculates the optimization problem formulated by the formula group and the optimization condition. Solve by programming. As described above, the data calculation unit 16 is a solver that can solve a mathematical programming problem, and can obtain an optimal solution for all the defined variables.
Then, a variable to be presented to the user, that is, a variable representing a desired vehicle operation plan is extracted and presented to the user through the input / output unit 11. In the present embodiment, since the number of vehicles at stations A to G at time 0 is a calculation target, seven variables (variable v _{(A, 0)} to variable v _{(G, 0)} ) are presented.

  The vehicle operation planning device according to the first embodiment represents the number of vehicles in a specific state and the number of users by variables, and describes the relationship between the variables by mathematical expressions, thereby representing the movement of vehicles and people. be able to. Since all mathematical expressions can be described by equations and inequalities, a vehicle operation plan that is optimal within given conditions can be determined by executing mathematical programming.

In the first embodiment, as the vehicle operation plan to be presented to the user, the number of vehicles to be placed in the station has been described. However, the vehicle operation plan can be expressed by a mathematical formula and is Any device that can be operated may be used.
For example, it may be the number of personnel who forward the vehicle instead of the user. In this case, the number of forwarding personnel per station is defined as a variable, and a formula that expresses the number of forwarding vehicles that depart (arrive at each station) and the number of vehicles that exist at a station is the default number. By adding a mathematical expression that represents the condition “start forwarding to the corresponding station when the number is less than”, the forwarding of the vehicle can be expressed. Also here, if-then needs to be expressed by an equation, the methods of equations 4 and 5 can be used.

  Further, a forwarding fee (negative usage fee) and a demand for vehicle forwarding for each fee may be defined in the demand prediction data, and may be calculated simultaneously with the demand for vehicle usage. Thereby, the assistance of the vehicle forwarding by a user can be expressed. In the first embodiment, the vehicle that has flowed out from the departure node is assumed to flow into the destination node after a predetermined time (condition 2 exemplified in the model template). Alternatively, a movement model considering the waypoint may be generated.

(Second embodiment)
In the first embodiment, after specifying a specific environmental condition in step S11, the value of the variable corresponding to the vehicle operation plan is obtained. However, the value that the user wants to obtain may be the environmental condition itself. For example, there is a case where an optimum value of “charge per section” is desired. Since the section fee is an environmental condition for obtaining demand data, in the first embodiment, it is not possible to obtain a value such as “section fee with maximum profit”. In order to cope with this, the second embodiment is a form in which an environmental condition is satisfied by repeatedly performing an operation while changing the environmental condition. In the present embodiment, “section fee” which is one of the environmental conditions is a calculation target.

  Since the system configuration of the vehicle operation planning device according to the second embodiment is the same as that of the first embodiment, description thereof is omitted. FIG. 9 is a process flowchart of the vehicle operation planning apparatus according to the second embodiment. Processes other than the steps described below are the same as those in the first embodiment.

In step S11, when the environmental condition is designated, the section fee is not designated.
In step S13, one of a plurality of section charges included in the demand forecast data is selected. That is, the optimum solution obtained in step S15 is based on the section fee selected in step S13.
In step S21, the data calculation unit 16 confirms whether there is a section fee not selected in step S13. If there is an unselected section charge, the process proceeds to step S13, a different section charge is selected to obtain demand data, and the optimum solution is obtained again.
In step S16, a plurality of patterns of solutions are obtained for each selected section fee. Here, the apparatus selects a solution that best matches the input optimum solution condition and presents it to the user. For example, a result such as “the section fee for maximizing the profit is 200 yen and there are five vehicles to be placed in the station A” can be obtained.

  As described above, in the second embodiment, even when the value to be obtained is a part of the environmental condition, a value satisfying the condition can be calculated. In the second embodiment, an example in which the section fee is a calculation target has been described. However, the calculation target may be anything as long as it represents an environmental condition.

(Modification)
The description of each embodiment is an exemplification for explaining the present invention, and the present invention can be implemented with appropriate modifications or combinations without departing from the spirit of the invention.

Here, the example which further added restrictions on vehicle operation to the first embodiment is given.
For example, if the vehicle is an electric vehicle, a mathematical expression representing the number of vehicles being charged at the station can be provided. The state of charge of the vehicle can also be expressed by a mathematical formula. Expression 14 is an expression representing the number of vehicles being charged at the station A. The variable chg _{(A, t)} is
The number of vehicles being charged at station A at time t, chgout _{(A, t)} is t from time t
Similarly _, the number of vehicles that have been charged during +1, and chgin _{(A, t)} are the number of vehicles that have entered a charged state.

Here, if it is assumed that 10 time steps are uniformly required for charging the vehicle, the time required for charging can be defined as Equation 15. When it is desired that the vehicle being charged cannot be used, the number of usable vehicles, that is, the number of vehicles that the user can get on immediately can be defined according to Equation 16. Information for generating these mathematical expressions can be stored in the model generation unit 15 as one of the templates.

  Thus, if the operation planning system of the car sharing system according to the present invention can be expressed by a mathematical expression, it is possible to obtain an optimum solution by incorporating various events that occur in operation. In addition to this, if a variable representing the number of users who could not get on due to a shortage of vehicles is provided, information on how much lost profit has occurred, which could not be known in the past, can be obtained. .

FIG. 10 is an example in which restrictions on vehicle operation other than those exemplified in the description of the embodiment are represented by data.
The station constraint information is data representing the upper limit of the number of vehicles that can be parked at each station and the number of chargers. If the vehicle arrives at the destination and the number of vehicles in the station is equal to the number of parked vehicles, the vehicle cannot be stored. In such a case, the destination may be changed to the nearest station. In addition, a variable indicating the number of vehicles that are being charged at each station may be provided, and when the charger is insufficient, it may be handled as a vehicle waiting for charging.

  The vehicle constraint information is data defining details about the vehicle. It is possible to define the vehicle capacity, parameters required for charging, and the like. The accounting constraint information is data that defines operational costs and the like. By generating a mathematical expression representing the operation cost, the total profit and the like can be calculated. For example, a variable representing the cost of forwarding the vehicle may be added, and the cost may be added every time the vehicle is forwarded. These pieces of information used at the time of formula generation can also be stored in the model generation unit 15 as one of the templates.

DESCRIPTION OF SYMBOLS 10 Vehicle operation planning apparatus 11 Input / output part 12 Network information storage part 13 Demand prediction data 14 Demand amount acquisition part 15 Model generation part 16 Data calculation part

Claims (8)

An operation planning system for a car sharing system,
Demand amount acquisition means for acquiring demand amount data, which is a movement demand amount between stations under specific environmental conditions,
Network storage means for storing network configuration data representing the connection relationship between stations;
Model template storage means for storing a model generation template which is a template for generating a mathematical model representing movement of a vehicle between stations;
A requirement input means for inputting a requirement that is a constraint in determining a vehicle operation plan;
An optimum solution condition input means for inputting an optimum solution condition which is an optimum solution condition;
By applying the network configuration data and the movement demand amount in the specific environmental condition acquired by the demand amount acquisition unit to the model generation template, a movement model that is a mathematical model representing the movement of the vehicle between the stations is generated. Model generation means for
Using the generated movement model, data calculation means for determining a vehicle operation plan that satisfies the input requirements and the optimal solution;
An operation planning system for car sharing systems. The network configuration data includes time constraints when the vehicle moves between stations,
The operation planning system of the car sharing system according to claim 1. The movement model includes a mathematical expression representing the number of vehicles that are stopped at each station or moving between the stations for each predetermined time step.
The operation planning system of the car sharing system according to claim 1 or 2. The movement model further includes a mathematical expression representing the movement of the user.
The operation planning system for a car sharing system according to any one of claims 1 to 3. The movement model includes a mathematical expression representing the number of users at each station for each predetermined time step,
The model generation means determines the number of users generated in the station at each predetermined time step using the demand data.
The operation planning system of the car sharing system according to claim 4. The demand amount acquisition means acquires a plurality of demand amount data corresponding to a plurality of environmental conditions,
The model generation means generates a plurality of mathematical models using the plurality of demand amount data,
The data calculation means performs an operation on the plurality of mathematical models.
An operation planning system for a car sharing system according to any one of claims 1 to 5. The request condition input means receives an input about a target time zone for calculating a vehicle operation plan,
The model generation means generates a movement model for a time step corresponding to the input time zone,
The data calculation means performs an operation on the input time zone.
The operation planning system for a car sharing system according to any one of claims 1 to 6. An operation planning method performed by an operation planning system of a car sharing system,
Obtaining network configuration data representing a network configuration in which stations are linked together;
Obtaining a model generation template that is a template for generating a mathematical model representing movement of a vehicle between stations;
Obtaining demand data, which is the demand for movement between stations under specific environmental conditions;
Obtaining a requirement that is a constraint in determining a vehicle operation plan;
Obtaining an optimal solution condition that is the optimal solution condition;
Generating a movement model that is a mathematical model representing movement of a vehicle between stations by applying the network configuration data and movement demand in a specific environmental condition to the model generation template;
Using the generated movement model to determine a vehicle operation plan that satisfies the input requirements and the optimal solution;
Operation planning method including

Images