KR101462337B1 - Method and apparatus of scheduling electric vehicle relocation using genetic algorithms - Google Patents

Abstract

A method and apparatus for scheduling an electric vehicle relocation using a genetic algorithm in a method for relocating a vehicle provided for car sharing. The method includes obtaining information related to a vehicle relocation plan, generating a plurality of vectors including integer elements based on the obtained information, and generating a plurality of vectors from the vectors using a genetic algorithm, And generating a plan. The method also includes a method of encoding a relocation schedule into integer vectors to apply genetic algorithm operations.

Description

[0001] METHOD AND APPARATUS OF SCHEDULING ELECTRIC VEHICLE RELOCATION USING GENETIC ALGORITHMS [0002]

The following embodiments relate to a method and apparatus for scheduling relocation plans for electric vehicles in car sharing for electric vehicles. More particularly, to a method and apparatus for efficiently scheduling relocation plans for electric vehicles using genetic algorithms.

Car sharing not only provides convenience to service users, but also reduces emissions of greenhouse gases by reducing the number of cars. If you share an electric car, you can overcome the high purchase cost and long charge time of the electric car. Through electric vehicle sharing, electric vehicles can be introduced into our daily lives. The introduction of electric vehicle sharing requires a reduction in service provision costs.

There is also research on genetic algorithms apart from car sharing.

Genetic algorithms use a chromosome-like data structure to encode potential solutions as a specific problem solution. Applying a recombination operator to these data structures preserves conclusive information. Individuals with high fitness for the environment will be able to reproduce the population, and the population will be able to adapt to the environment and evaluate the adaptability with the evaluation function.

Embodiments of the present invention use a genetic algorithm to schedule a vehicle relocation plan. Effective vehicle relocation is essential for reducing service delivery costs. By scheduling the most effective vehicle relocation plan through the genetic algorithm, electric vehicle sharing can be efficiently operated.

A method for relocating a vehicle provided in car sharing according to an exemplary embodiment of the present invention includes: acquiring information related to a vehicle relocation plan using an electric vehicle relocation scheduling method using a genetic algorithm; Generating a plurality of vectors including integer elements based on the obtained information, and generating a vehicle relocation plan with a minimum relocation distance from the vectors using a genetic algorithm.

The information related to the vehicle relocation plan may include at least one of a population size, a number of repetitions, a number of service staff or personnel performing vehicle relocation, a number of vehicles to be relocated, Number of stations or the number of stations.

The vehicle relocation plan using the genetic algorithm can be expressed as an independent chromosome.

An electric vehicle relocation scheduling method using a genetic algorithm comprises grouping stations based on local adjacencies and storing and managing the vehicles provided for car sharing based on local proximity and generating the vehicle relocation plan for the grouped stations can do.

When the vehicle relocation plan is generated for stations belonging to different groups among the stations, an intermediary station - a vehicle relocation through a station defined to belong to a plurality of groups for vehicle relocation between groups, You can create a plan.

The vectors may be encoded to include information about the overflow station - the station to which the vehicle to be relocated belongs and the underfloor station - the station to which the vehicle is to be moved.

The vectors may comprise as many integer elements as the number of vehicles to be relocated and the positions in the vectors of the integer elements may represent information about an overflow station, Can be displayed.

The vectors may be encoded such that the vehicle of the overflow station corresponds to a location or index in the vectors of the integer elements, and the value of the integer elements may be encoded such that the vehicle of the underflow station corresponds.

Copying the overflow station to the mapping vector by the number of excess vehicles of the overflow station to map the position or index in the vectors of the integer elements to the vehicle of the overflow station, And the underflow station may be copied to the mapping vector by the number of the deficient vehicles so as to correspond to the vehicle of the underflowing station.

The genetic algorithm may use a fitness function based on the relocation distance of the vehicles to be relocated.

The information associated with the vehicle relocation plan may include information on the number of service staff performing vehicle relocation, and the fitness function may adjust the relocation distance based on the number of service staff performing the vehicle relocation.

The vectors may be randomly generated within a valid range.

The genetic algorithm can perform a selection operation of a parent chromosome using a roulette wheel selection technique.

The genetic algorithm can perform a replacement operation to exclude the same chromosome in a population by using a replacement for duplicated genes technique.

A method for relocating a vehicle provided in car sharing according to an embodiment of the present invention includes an information obtaining unit for obtaining information related to a vehicle relocation plan using an electric vehicle relocation scheduling apparatus using a genetic algorithm; Generating a plurality of vectors including integer elements based on the obtained information, and generating a relocation plan that generates a vehicle relocation plan having a minimum relocation distance from the vectors using a genetic algorithm operation .

The information associated with the vehicle relocation plan may include at least one of a population size, a number of repetitions, a number of service personnel performing vehicle relocation, a number of vehicles to be relocated, a number of vehicles to be relocated, And the like.

The vehicle relocation plan using the genetic algorithm can be expressed as an independent chromosome.

The vectors may be encoded to include information about an overflow station and information about an underflow station.

The vectors may comprise as many integer elements as the number of vehicles to be relocated and the positions in the vectors of the integer elements may represent information about an overflow station, Can be displayed.

The vectors may be encoded such that the vehicle of the overflow station corresponds to a location or index in the vectors of the integer elements, and the value of the integer elements may be encoded such that the vehicle of the underflow station corresponds.

Copying the overflow station to the mapping vector by the number of excess vehicles of the overflow station to map the position or index in the vectors of the integer elements to the vehicle of the overflow station, And the underflow station may be copied to the mapping vector by the number of the deficient vehicles so as to correspond to the vehicle of the underflowing station.

Embodiments of the present invention use a genetic algorithm to schedule a vehicle relocation plan. Effective vehicle relocation is essential for reducing service delivery costs. By scheduling the most effective vehicle relocation plan through the genetic algorithm, electric vehicle sharing can be efficiently operated.

1 is a view showing a place where a station of a car sharing service can be installed and a predicted route of a vehicle provided for car sharing.
Fig. 2 is a view showing a path in which cars provided for car sharing are moved from an overflow station to an underflow station.
FIG. 3 is a diagram showing stations classified into an overflow station and an underflow station in order to represent a relocation plan as a vector.
FIG. 4 is a diagram showing the relocation plan represented by a vector, the relocation pairs included in the relocation plan, and the change of relocation pairs according to the number of service staff.
FIG. 5 is a graph showing a vehicle relocation distance according to the size of a solution set and the number of service steps.
6 is a graph showing the distance of relocation of the vehicle according to the number of electric vehicles and the number of service steps.
7 is a graph showing a vehicle relocation distance according to the number of vehicles moved and the number of service steps.
FIG. 8 is a graph showing a vehicle relocation distance according to the number of underflow stations and the number of service steps.
9 is a flowchart illustrating a method of relocating an electric vehicle using a genetic algorithm in a method of relocating a vehicle provided in car sharing according to an embodiment of the present invention.
10 is a block diagram illustrating an apparatus for relocating an electric vehicle using a genetic algorithm in an apparatus for relocating a vehicle provided in car sharing according to an embodiment of the present invention.

Hereinafter, embodiments will be described in detail with reference to the accompanying drawings.

Car sharing reduces the number of cars and emissions of greenhouse gases. Sharing an electric car can overcome the high cost and long charge time of the electric car. As a result, electric vehicles can be introduced into our daily lives in the near future. In an electric vehicle sharing service, electric vehicles are parked and managed at a station.

The present invention utilizes a genetic algorithm to efficiently schedule vehicle relocation. Genetic algorithms are one of the techniques for solving optimization problems and are global optimization techniques. Genetic algorithms are examples of evolutionary operations that mimic the evolution of living organisms. They include variations (mutation), crossing (mating) operations, and many other terms applied to living organisms. They also use terms such as generation and population .

Most genetic algorithms have a population of solutions that consists of a fixed number of solutions. The genetic algorithm begins with an initialization step that randomly generates n solutions. Each of these solutions is made through selection, crossover, and mutation steps. These k solutions are replaced with k solutions in the solution group. This process is performed until an arbitrary stop condition is satisfied, and then the best solution among the solutions remaining in the solution group is taken as an answer. The constant k above determines k / n as the number of generations to determine how much the population of the sun is displaced at one time.

Among the above operators, the selection operator is an operator that selects an arbitrary solution from the solution group for the intersection, and operates in such a way as to increase the probability of selecting a good solution. This selected solution is called the parent. An intersection is an operator that produces an offspring from two parent solutions. Intersection is a representative operator of genetic algorithms and has a great influence on the performance of genetic algorithms.

There are two typical conditions for stopping the genetic algorithm. One of the two most typical cases is to perform a repeat-until loop a certain number of times and then stop. When the diversity of the solutions in the population falls below a certain level And stopping it. It is common to see that most of the chromosomes in the population of the sea (for example, 70%) are identical in order to judge the diversity.

Even with a genetic algorithm designed to perform a repeat-until loop a certain number of times and then stop, there must be an empirical guess that the solutions will converge to some degree.

1 is a view showing a place where a station of a car sharing service can be installed and a predicted route of a vehicle provided for car sharing.

The service designer chooses the location of the sharing station considering demand, space, and accessibility. Customers can rent an electric car in the space where the electric car is needed. As shown in FIG. 1, stations may be installed in airports, shopping malls, residential areas, tourist areas, and commercial areas. The tenant can borrow an electric car from the warehouse at the starting point and return the electric car to the destination store.

The one-direction lathing service is most convenient for the customer, but the unbalance of the stock is likely to occur due to the asymmetric movement pattern. This imbalance can degrade the quality of service and inconvenience consumers. Therefore, in order to solve this problem, electric vehicles must be relocated from an overflow station to an underflow station through a service step or a large transportation means. It is appropriate to relocate by service staff when electric cars are spread across multiple stations. The scheduling method and apparatus of the present invention can reduce the travel distance and time of a given service employee.

If there are multiple relocation teams, it is necessary to set up a relocation path for each team, and the routing is the same as the multiple traveling sales-man problem (mTSP) problem. For such a relocation plan, a cluster-first and route-second approach can be used. Relocating the vehicle throughout the service area can be very long driving distance. A set of distribution criteria can be defined including local adjacency and temporal dependencies. Above all, stations can be distributed according to specific system goals such as access priority, commute support, and so on. If so, the relocation plan is reduced to a relocation team operational plan. For electric vehicle relocation, each station set assigned to one team does not necessarily have to be separated. Duplicated stations may be used for inter-cluster route adaptation.

In essence, time complexity is very important because the number of electric vehicles in the sharing system can increase significantly. In this regard, the present invention is based on a genetic algorithm that can provide a reasonable solution to an electric vehicle relocation plan within a controllable time. One embodiment of the present invention assumes that the sharing stations are classified into a plurality of groups, and each group is assigned a relocation team consisting of two to five service staff and one relocation service vehicle. Each relocation scheme is represented as a vector with integer values to perform a genetic operator such as crossover, selection, reproduction, and mutation. The fitness function also estimates the quality of the relocation plan, taking into account the number of electric vehicles that can move between the two stations.

For electric vehicle relocation in car sharing systems, a single vehicle relocation strategy and a team-based vehicle relocation strategy are possible. In a single vehicle relocation operation, two staff members can drive one service vehicle to the overflow station. One of them drives an electric car to the assigned Underfloor station, and the other one follows a service vehicle. Thereafter, the two persons drive the service vehicle again and move to the next overflow station. The relocation design determines the overflow and underflow stations by calculating how many electric vehicles need to be moved to bring the current vehicle placement to the target placement. After that, electric vehicles located at the overflow station are moved to the Underfloor station based on the stable marriage problem model. Matching results can effectively reduce the relocation distance, but this design can not be applied to more than one service step.

The system operator usually relocates at a time that is expected to be small. In addition, target placement, which is the deployment status of electric vehicles after relocation, can be determined by appropriate strategies combined with future demand forecasting. With a relocation team consisting of m people, the stations can be classified into several groups, mainly by geographical conditions, and each group can be assigned a relocation team. Electric cars can be reallocated into groups to reduce the overall relocation distance. It may be necessary for electric cars to be moved to another group. Such relocation between groups can be performed through an intermediary station belonging to a plurality of groups. The intermediate station may be an overflow station of the underflow group.

When the n number of the called station, the station set _{S = {S 1, S 2} , ..., S n}, the current placement _{C = {C 1, C 2} , ..., C n}, the target The arrangement is defined as T = {T ₁ , T ₂ , ..., T _n }. The relocation vector {V _i } is C _i ? _i . < / RTI > If V _i is positive, S _i is an overflow station with an excess electric vehicle that must be moved to another station. Conversely, if V _i is negative, S _i is the underflow station that wants to receive the electric vehicle. If E electric vehicle that _k is to be moved from its current located station, called S _i to another station of S _j, which may be expressed as (S _i, S _j), this is (E _k, S _j) and same. If there are u electric vehicles to be relocated, the relocation plan consists of u pairs, which are | {( _Si , _Sj )} | = u. Here, although the electric vehicles are all different, the underflow station can emerge as many as the number of electric vehicles to be distributed. The relocation distance is the sum of the distances of all station pairs present in the relocation plan. To calculate the relocation distance, you need to know the distance between all pairs of stations.

To find an efficient relocation plan using genetic algorithms it is necessary to encode the relocation plan into a vector with the corresponding integer value (chromosome). The encoding criterion can be explained with reference to FIGS. 2 to 4. FIG.

Fig. 2 is a view showing a path in which cars provided for car sharing are moved from an overflow station to an underflow station.

Referring to FIG. 2, there are five stations from S ₁ to S ₅ , of which S ₁ , S ₂ and S ₃ are overflow stations. Since seven electric vehicles must be relocated, each relocation plan has seven relocation pairs. FIG arrow 2 indicates the relocation root, relocation pair corresponding to each of arrows _{(S 1, S 4),} (S 1, S 4), (S 1, S 4), (S 1, S 5) , (S ₂ , S ₅ ), (S ₂ , S ₅ ), (S ₃ , S ₅ ). The rearrangement pairs can be represented by 3 (S ₁ , S ₄ ), 2 (S ₁ , S ₅ ), (S ₂ , S ₅ ), (S ₃ , S ₅ ) If the vehicle is relocated from the same overflow station to the same underfloor station, it is possible to simultaneously relocate the same vehicles with the same relocation path if the number of relocation service steps is sufficient. For example, in order to perform rearrangement according to 3 (S ₁ , S ₄ ) with four or more rearrangement service staffs, the rearrangement is performed not only three times between S ₁ and S ₄ .

FIG. 3 is a diagram showing stations classified into an overflow station and an underflow station in order to represent a relocation plan as a vector.

In order to represent the relocation plan as a vector having an integer value, the overflow station 320 is sequentially displayed in duplicate. Each index 310 is now coupled to a specific overflow station 320. Next, each underflow station 330 is displayed in the index 310 according to the number of electric vehicles that are insufficient. As shown in Figure 3, since the S ₄ requires a three electric vehicles are 0, 1 and 2 were given to S _4.

A relocation pair can be represented by a number located at the index and index, and the relocation plan can be represented by a sequence of numbers within the range of the index. For example, in [0 1 5 2 3 4 6], the number 0 appears at the position of index 0. This means a relocation pair that moves the electric vehicle to the underflow station S ₄ corresponding to the position of the number 0 in the overflow station S ₁ associated with the index 0. Also, the number 4 appears at the position of index 5. Index 5 joins with overflow station S _2, and number 4 joins to underflow station S ₅ corresponding to the position of number 4.

Now, the relocation problem results in finding the optimal sequence with the least cost, such as a TSP (Traveling Salesman Problem) solution. Genetic operations such as crossover, selection, reproduction, and mutation can be performed through the encoding criteria. Even if the relocation pair increases, the genetic algorithm can adjust the iterative loop length to schedule the reproduction plan in a reasonable time span.

The genetic algorithm loop computes and improves the fitness of each relocation plan. In relocation of an electric vehicle, the relocation distance defined by the distance traveled by the relocation vehicle is the most important criterion. The relocation distance has the greatest impact on energy savings and service staff efficiency. The shorter the relocation distance, the better the relocation plan. The relocation distance can be calculated by the sum of the distances of stations existing in each relocation pair. However, if the relocation team consists of m service staffs, m-1 electric vehicles can be simultaneously moved from the same overflow station to the same underflow station.

FIG. 4 is a diagram showing the relocation plan represented by a vector, the relocation pairs included in the relocation plan, and the change of relocation pairs according to the number of service staff.

First, one staff member drives a relocation service vehicle and moves m steps to the overflow station. m-1 staff moves the m-1 electric vehicles to the underfloor station respectively, and the driver of the relocation service vehicle moves along with them in the relocation service vehicle. Thereafter, the m steps may move together to the next overflow station. The m-1 electric vehicles are simultaneously moved by the above method. However, m may be limited depending on the capacity of the relocation service vehicle.

As can be seen from the above example, assuming that the same relocation plan is performed, the relocation distance can be reduced when there are two or more staff members. If the overflow station or the underflow station are all different, each electric vehicle must be moved one at a time, so that the effect of reducing the relocation distance according to the number of steps can not be obtained. As the sharing station group is subdivided, the number of stations included in one group decreases. This means that there is a high possibility that the same relocation pair will be created in the relocation plan. Relocation between groups is performed through an intermediary node belonging to a part or group of the group. This relocation creates many relocation pairs. Even though their final destination is a station in the underflow group, the station in the overflow group can become the intermediate station.

5 to 8 show the results of measuring the performance of relocation plans.

The initial population of chromosomes is arbitrarily selected and the operation of the genetic algorithm is performed by the Roulette wheel selection method. For more diverse populations, redundant genes are eliminated. The redundant gene is replaced with any new one during the mutation operation. For better efficiency, the fitness value of the relocation plan is calculated when the relocation plan is first produced. The relocation distance may be included in the fitness. The dielectric parameters, such as the number of repetitions and the size of the population, can be chosen freely according to the experiment. In a relocation scheme, a fitness function merges relocation pairs with the same overflow station and the same underfloor station. Accordingly, the m-1 pairs that can be moved by one step at a time are grouped into a single pair. In this experiment, the number of stations is 10 and the distance between stations is set to 3 km on average.

The performance matrix is a relocation distance, and the performance parameters consist of the size of the population, the number of electric vehicles, the number of mobile stations and the number of underflow stations. Underflow and overflow stations are arbitrarily determined according to the number of underflow stations given among 10 stations. In the experiment, the number of excess electric vehicles in the overflow station and the number of electric vehicles required in the underflow station are arbitrarily determined, respectively. Any number is determined by the system clock in each experiment. For each parameter setting, 30 sets are generated and the results are averaged. The relocation distance is calculated for two, three, four, and five staff members in the relocation team, respectively. As mentioned above, m-1 electric vehicles can be moved simultaneously when there are m service staffs. Here, the inter-group route is not considered.

FIG. 5 is a graph showing a vehicle relocation distance according to the size of a solution set and the number of service steps.

The experiment of FIG. 5 measures the effect of the size of the population on the relocation distance. The larger the size of the population, the better the diversity, but not necessarily the better solution. Even if the size of the sea population is large, a low-quality relocation plan can be caused by a variety of initial age groups. Here, the number of electric vehicles is 100, and 5 out of 10 stations are underflow stations, and each relocation plan is composed of 30 pairs of relocation pairs. As shown in Fig. 5, when the size of the sea group was 60, the relocation distance was the shortest. Fig. 5 shows that the size of the sea group has little influence on the relocation distance, especially when there are five staff members. However, if there are two staff members, the difference in relocation distance reaches 11.3 km. If the size of the sea group is 60 and the number of service staff increases from 2 to 3, the relocation distance is reduced from 64.6 km to 37.7 km. This corresponds to a decrease of 31.3%. However, if the number of service staff increases from four to five, the performance improvement will remain at 9.8%.

6 is a graph showing the distance of relocation of the vehicle according to the number of electric vehicles and the number of service steps.

In Fig. 6, the number of underflow stations is five. Here, 30% of electric cars are required to be relocated. The size of the population is 50, and the number of genetic loop repetitions is 1000. The results are shown in Fig. Basically, according to the proposed design, the relocation distance increases as the number of vehicles increases for all cases. If the number of staff increases from two to three, the performance is improved by 22% when 20 electric cars are used, but the performance is improved by 39.1% when 200 electric cars are used. If the number of staff increases from four to five, the relocation distance is reduced by 0.1% for 20 electric cars and 7.9% for 200 electric cars. Increasing the number of service staff from 2 to 5 reduces the relocation distance by an average of 38%, 17.2% and 10.5%.

7 is a graph showing a vehicle relocation distance according to the number of vehicles moved and the number of service steps.

The number of movements of the vehicle corresponds to the number of rearrangement pairs. If an electric car does not come in and out of a relocation group, the number of relocation pairs is equal to the number of electric cars in excess, which is equal to the number of electric cars in shortage. The relocation plan includes an intermediary station that makes all numbers the same. Here, the number of electric vehicles is set to 100, and the underfloor station is set to 5 out of 10 stations. In contrast to the case of Fig. 6, the increase in the performance due to the change in the number of movements is not large for all the number of steps. When the number of electric vehicles is constant, the graph of Fig. 7 is more linear than the graph of Fig. Except for 10 moves, the increase amounts to 38.1%, 17.2% and 12.7%, respectively. The performance improvement is more influenced by the number of electric vehicles.

FIG. 8 is a graph showing a vehicle relocation distance according to the number of underflow stations and the number of service steps.

If there are nine underfloor stations, then there is one overfloor station that needs to relocate excess electric cars to the other nine stations. The number of electric cars is set at 100, and the number of moving cars is set at 30. The number of underflow stations is adjusted between three and nine. When the number of underflow stations approaches the number of overflow stations, the relocation distance becomes shorter. If the number of underfloor stations and overflow stations is unequal, a larger number of identical relocation pairs can be merged into the same relocation pair, resulting in a significant performance difference between two adjacent points. When the number of service staff is increased from two to three, the relocation distance is reduced to 42.9%, and when the number of service staff is increased from four to five, the relocation distance is reduced to 17.1%.

9 is a flowchart illustrating a method of relocating an electric vehicle using a genetic algorithm in a method of relocating a vehicle provided in car sharing according to an embodiment of the present invention.

Referring to FIG. 9, an electric vehicle relocation scheduling method 900 using a genetic algorithm according to an embodiment of the present invention acquires information related to a vehicle relocation plan (910).

The information associated with the vehicle relocation plan may include, for example, population size, number of repetitions, number of service personnel performing vehicle relocation, number of vehicles to be relocated, number of vehicles to be relocated, Information. The information is information necessary for scheduling vehicle relocation.

The size of the population can include the number of genes initially generated in the genetic algorithm or the number of final genes. The larger the size of the population, the better the diversity, but not necessarily the better solution. Even if the size of the sea population is large, a low quality relocation plan can be created due to the influence of the initial sea population.

The number of iterations can be the number of times the genetic algorithm repeats the process of obtaining the child solution from the parent solution through the selection, crossover, and mutation steps. The genetic algorithm derives various solutions through the iteration of the genetic operation and can obtain a relocation plan with the minimum relocation distance through evolution of various solutions.

The vehicle relocation distance can be adjusted according to the number of service personnel performing the vehicle relocation. In particular, the number of service staff can have a significant impact when relocating multiple vehicles from the same overflow station to the same underflow station. When stations are grouped, it is possible to increase the probability that the overflow station and the underflow station become identical to each other.

The number of vehicles to be relocated affects the relocation distance. The greater the number of vehicles to be relocated, the more likely the vehicle relocation distance will increase. The number of vehicles to be relocated is equal to the number of relocation pairs. Each of the integer elements of the vector can be represented by a relocation pair according to the above-mentioned encoding criterion and the first element of the relocation pair can represent the vehicle to be relocated.

The vehicle relocation distance can be adjusted according to the number of movements of the vehicle to be relocated. It is also possible to calculate the vehicle relocation distance according to a specific number of movements and to group the stations to specify the number of movements.

The information on the number of stations includes the number of overflow stations, the number of underflow stations, and the number of all stations. If there is a large difference between the number of underflow stations and the number of overflow stations, there is an increased likelihood that the relocation paths of the vehicles to be relocated are the same. Accordingly, the possibility of reducing the relocation distance according to the number of service steps can also be increased.

The electric vehicle relocation scheduling method 900 using a genetic algorithm according to an embodiment of the present invention generates a plurality of vectors including integer elements based on the acquired information (920).

A plurality of vectors including integer elements each correspond to a relocation plan. A plurality of vectors are randomly generated within a valid range. Also, a plurality of randomly generated vectors correspond to chromosomes first generated in the genetic algorithm. The vectors may be randomly generated as much as the size of the acquired initial solution population. The set of vectors corresponds to an initial population.

The method 900 for scheduling an electric vehicle relocation using a genetic algorithm according to an embodiment of the present invention generates a vehicle relocation plan having a minimum relocation distance from the vectors using a genetic algorithm (930).

A vehicle relocation plan is generated through a genetic operation based on a plurality of randomly generated vectors. The relocation plan corresponds to a plurality of vectors including integer elements. Genetic operations may include crossover, selection, reproduction, and mutation. Genetic operations generate various solution sets from the initial solution set. The solution set evolves according to the genetic algorithm, so that a more efficient vehicle relocation plan can be created. In particular, it is possible to derive a group of solutions that minimizes the relocation distance, which is the most important factor in an efficient vehicle relocation plan. This process is performed until any stop condition is satisfied. When the stop condition is satisfied, the most desirable year is selected as the answer. For example, the year group with the shortest relocation distance can be the answer.

A fitness function is an objective function used to calculate a figure of merit that evaluates how close the individual solutions derived by the algorithm are to the ultimate goal of the algorithm. The fitness function in the genetic algorithm can suggest the direction of the evolution, that is, the direction of the optimal solution in performing the genetic algorithm.

The vehicle relocation distance refers to the distance between the stations, so in creating the vehicle relocation plan, the location of the station must be considered. Also, relocating the vehicle over the service area can group a plurality of stations because the relocation distance can be very long. Grouping stations also has the effect of increasing the likelihood of relocating the vehicle from the same overflow station to the same underfloor station. Therefore, the relocation distance can be reduced according to the number of service staff.

However, it may be necessary for electric cars to be moved to another group. Such relocation between groups can be performed through an intermediary station belonging to a plurality of groups. The intermediate station may be an overflow station of the underflow group.

10 is a block diagram illustrating an apparatus for relocating an electric vehicle using a genetic algorithm in an apparatus for relocating a vehicle provided in car sharing according to an embodiment of the present invention.

Referring to FIG. 10, an electric vehicle relocation scheduling apparatus 1000 using a genetic algorithm includes an information obtaining unit 1010, a vector generating unit 1020, and a relocation plan generating unit 1030.

The information obtaining unit 1010 obtains information related to the vehicle relocation plan. Information related to the vehicle relocation plan includes information on the population size, the number of repetitions, the number of service personnel performing vehicle relocation, the number of vehicles to be relocated, the number of vehicles to be relocated, . ≪ / RTI >

The vector generation unit 1020 generates a plurality of vectors including integer elements based on the acquired information. A plurality of vectors are randomly generated within a valid range. The set of multiple vectors corresponds to an initial population.

The relocation plan generation unit 1030 generates a vehicle relocation plan with a minimum relocation distance from the vectors using a genetic algorithm. The contents of the genetic algorithm used by the rearrangement plan generation unit include the above-mentioned contents, and a detailed description will be omitted.

The method according to an embodiment may be implemented in the form of a program command that can be executed through various computer means and recorded in a computer-readable medium. The computer-readable medium may include program instructions, data files, data structures, and the like, alone or in combination. The program instructions to be recorded on the medium may be those specially designed and configured for the embodiments or may be available to those skilled in the art of computer software. Examples of computer-readable media include magnetic media such as hard disks, floppy disks and magnetic tape; optical media such as CD-ROMs and DVDs; magnetic media such as floppy disks; Magneto-optical media, and hardware devices specifically configured to store and execute program instructions such as ROM, RAM, flash memory, and the like. Examples of program instructions include machine language code such as those produced by a compiler, as well as high-level language code that can be executed by a computer using an interpreter or the like. The hardware devices described above may be configured to operate as one or more software modules to perform the operations of the embodiments, and vice versa.

While the present invention has been particularly shown and described with reference to exemplary embodiments thereof, it is to be understood that the invention is not limited to the disclosed exemplary embodiments. For example, it is to be understood that the techniques described may be performed in a different order than the described methods, and / or that components of the described systems, structures, devices, circuits, Lt; / RTI > or equivalents, even if it is replaced or replaced.

Therefore, other implementations, other embodiments, and equivalents to the claims are also within the scope of the following claims.

Claims (20)

A method of relocating a vehicle provided for car sharing,
Obtaining information related to a vehicle relocation plan;
Generating a plurality of vectors including integer elements based on the obtained information; and
Generating a vehicle relocation plan with a minimum relocation distance from the vectors using a genetic algorithm;
Lt; / RTI >
Wherein the plurality of vectors are encoded to include information about an overflow station-the station to which the vehicle to be relocated belongs and a station to which the underfloor station-
A Method of Scheduling Electric Vehicle Relocation Using Genetic Algorithm. The method according to claim 1,
The information related to the vehicle relocation plan
The population size of the population, the number of repetitions, the number of service staff or employees performing the vehicle relocation, the number of vehicles to be relocated, the number of vehicles to be relocated, or the number of stations A method of scheduling an electric vehicle relocation using a genetic algorithm. The method according to claim 1,
A method of scheduling an electric vehicle relocation using a genetic algorithm represented by an independent chromosome, the vehicle relocation plan using the genetic algorithm. The method according to claim 1,
Station-grouping based on local adjacency, where to store and manage vehicles provided for car sharing, and scheduling an electric vehicle relocation using a genetic algorithm to generate the vehicle relocation plan for the grouped stations . 5. The method of claim 4,
When the vehicle relocation plan is generated for stations belonging to different groups among the stations, an intermediary station - a vehicle relocation through a station defined to belong to a plurality of groups for vehicle relocation between groups, A method for scheduling relocation of electric vehicles using genetic algorithms to generate plans. delete The method according to claim 1,
Said vectors comprising integer elements as many as the number of vehicles to be relocated,
Wherein the location of the integer elements in the vectors represents information about an overflow station and wherein the value of the integer elements represents information about an underflow station. The method according to claim 1,
The vectors
The position or index in the vectors of the integer elements is encoded so as to correspond to the vehicle of the overflow station,
And a genetic algorithm encoded such that the value of the integer elements corresponds to a vehicle of an underfloor station. 9. The method of claim 8,
Copying the overflow station to the mapping vector by the number of excess vehicles of the overflow station to map the location or index in the vectors of the integer elements to the vehicle of the overflow station,
And copying the underflow station to a mapping vector as many as the number of deficient vehicles in order to correspond the value of the integer elements to the vehicle of the underflow station. The method according to claim 1,
The genetic algorithm
And a genetic algorithm using a fitness function based on a relocation distance of the vehicles to be relocated. 11. The method of claim 10,
The information related to the vehicle relocation plan
Information on the number of service staffs performing vehicle relocation,
The fitness function
And adjusting the relocation distance based on the number of service steps performing the relocation of the vehicle. The method according to claim 1,
Wherein the vectors are randomly generated within a valid range. The method according to claim 1,
The genetic algorithm
An electric vehicle relocation scheduling method using a genetic algorithm that replaces redundant genes with arbitrary new ones. An apparatus for rearranging a vehicle provided for car sharing,
An information acquisition unit for acquiring information related to a vehicle relocation plan;
A vector generating unit for generating a plurality of vectors including integer elements based on the obtained information,
A rearrangement plan generation unit for generating a rearrangement plan having a minimum rearrangement distance from the vectors using a genetic algorithm;
Lt; / RTI >
The plurality of vectors being encoded to include information about an overflow station and information about an underflow station,
An Electric Vehicle Relocation Scheduling Device Using Genetic Algorithm. 15. The method of claim 14,
The information related to the vehicle relocation plan
The population size of the population, the number of repetitions, the number of service staff or employees performing the vehicle relocation, the number of vehicles to be relocated, the number of vehicles to be relocated, or the number of stations An Electric Vehicle Relocation Scheduling Device Using Genetic Algorithm. 15. The method of claim 14,
A vehicle relocation plan using the genetic algorithm is an electric vehicle relocation scheduling apparatus using a genetic algorithm represented by an independent chromosome. delete 15. The method of claim 14,
Said vectors comprising integer elements as many as the number of vehicles to be relocated,
Wherein the location of the integer elements in the vectors represents information about an overflow station and wherein the value of the integer elements represents information about an underflow station. 15. The method of claim 14,
The vectors
The position or index in the vectors of the integer elements is encoded so as to correspond to the vehicle of the overflow station,
And a genetic algorithm encoded such that the value of the integer elements corresponds to the vehicle of the underflow station. 20. The method of claim 19,
Copying the overflow station to the mapping vector by the number of excess vehicles of the overflow station to map the location or index in the vectors of the integer elements to the vehicle of the overflow station,
And a genetic algorithm for copying the underflow station to a mapping vector as many as the number of deficient vehicles to associate the value of the integer elements with the vehicle of the underflow station.

Images