KR20220104970A - Routing group search optimization scheduling optimization method to search for optimum solution of parallel delivery using vehicles and drones and the system thereof - Google Patents

Abstract

The present invention relates to a method and a system for optimizing RGSO scheduling, which is a scheduling and pathfinding method that minimizes the last time (or makespan) until all products are delivered to all customer locations by collaborating a transportation means and a drone. The method comprises: a first step of inputting location data of a customer; a second step of obtaining delivery location data by which the transportation means and the drone can deliver products from among the location data of the customer, and calculating the distance and time from a depot to the delivery location data; a third step of generating an initial solution for the paths of the transportation means and the drone, and performing a producing process, a scouring process, a ranging process, and a finding process for a producer of the initial solution; a fourth step of calculating the evaluation values of all solutions according to execution results and updating the producer with a solution representing the minimum makespan; and a fifth step of obtaining an optimal solution for parallel delivery using the transportation means and the drone according to update results. Therefore, the method can minimize the total makespan of a delivery system that uses only an existing transportation means, by using parallel optimal scheduling using the transportation means and the drone.

Description

RGSO Scheduling Optimization Method and System for Exploring the Optimal Solution for Parallel Delivery Using Vehicles and Drones

The present invention relates to an RGSO scheduling optimization method and system for searching for an optimal solution for parallel delivery using a transportation means and a drone. It relates to a scheduling and movement route search method that minimizes the last time (or total required time, Makespan).

The Traveling Salesman Problem (TSP) or Vehicle Routing Problem (VRP), which has been used in various ways in the existing logistics industry, has multiple cities (or customers) and the cost (distance or time) of moving from one city to another. etc.), it is to find the order of least-cost movement that visits (or delivers) all cities only once and returns to the original starting point (hereinafter referred to as a 'depot'). In other words, it was to present a travel route and a scheduling solution that would minimize the time until the truck had finished delivering all parcels to all customers. This combinatorial optimization problem is NP-hard, and heuristic algorithms such as Simulated Annealing and Genetic Algorithm are used as the optimal solution search method for traveling salespeople or vehicles to the optimal path.

Recently, a drone that is faster than a truck and has no delay due to congestion has appeared and technology has been developed and applied to various fields. In particular, the need for research on Vehicle & Drone Combined Operations that can deliver parcels to customers in the fastest time by mixing the advantages of vehicles and drones is emerging.

As a system in which trucks and drones cooperate, Murray & Chu (2015) proposed Flying Sidekick TSP (FSTSP) and Parallel Drone Scheduling TSP (PDSTSP) methods. The proposed FSTSP and PDSTSP are composed of trucks, drones and customers, and the goal is to minimize the makespan of delivery to customers. One drone can only deliver to one customer, and if the weight of the delivery parcel is heavier than the standard, it cannot be delivered by drone and can be delivered only by truck. The truck must leave the depot and return to the depot. In the case of FSTSP, a truck can serve as a connection point for drones while carrying drones, and the truck and drone must be synchronized. In contrast, PDSTSP operates independently without the need for synchronization of trucks and drones. Drones must always leave the depot to deliver the parcel to the customer and then return to the depot, only parcels within the flight time due to battery limitations and within the weight the drone can carry. Trucks deliver to customers that drones cannot or are close to on their route.

Murray, C.C., Chu, A., 2015. The flying sidekick traveling salesman problem: Optimization of drone-assisted parcel delivery. Transport. Res. Part C: Emerg. Technol. 54, 86-109.

It is an object of the present invention to find the optimal solution to the parallel delivery PDSTSP problem using transportation means and drones, and the total time required until delivery of all deliveries to all customer locations through cooperation between transportation means and drones (Makespan) We intend to provide a scheduling optimization method for RGSO (Routing Group Search Optimization), which is a scheduling and movement path search method that minimizes

In the Routing Group Search Optimization (RGSO) scheduling optimization method for searching for an optimal solution for parallel delivery using a transportation means and a drone according to an embodiment of the present invention, the first step of inputting customer location data, the customer location A second step of obtaining delivery location data that can be delivered by a means of transport and the drone from among the data, and calculating the distance and required time from a depot to the delivery location data, the route of the means of transport and the drone The initial solution is generated, and the third step of performing the producing process, the scrolling process, the ranging process, and the finding process for the producer of the initial solution. a fourth step of updating a solution representing a total required time (Makespan) to the producer; and a fifth step of obtaining an optimal solution for parallel delivery using the vehicle and the drone according to the update result.

In the RGSO scheduling optimization system for performing a Routing Group Search Optimization (RGSO) scheduling optimization method for searching for an optimal solution for parallel delivery using a transportation means and a drone according to an embodiment of the present invention, an input unit for inputting customer location data , a data processing unit that acquires delivery location data that can be delivered by a transportation means and the drone from among the location data of the customer, and calculates a distance and required time from a depot to the delivery location data, the transportation means and the drone Generates an initial solution for the path of and a producer update unit that updates the producer with a solution representing the minimum total required time (Makespan) to the producer, and an acquisition unit that obtains an optimal solution for parallel delivery using the transportation means and the drone according to the update result.

According to an embodiment of the present invention, by using the RGSO scheduling optimization method applicable to the existing PDSTSP method, the total required time (Makespan) of the delivery system using only the existing transportation means with parallel optimal scheduling of the transportation means and the drone can be minimized.

In addition, according to an embodiment of the present invention, in order to search for an optimal scheduling solution of the existing PDSTSP problem, routing is performed in the producing process, the scrolling process, and the ranging process of the existing Group Search Optimization (GSO) method. An efficient RSGO scheduling optimization method can be provided by adding a finding process that applies a mechanism that can effectively search the solution of the problem (searching for a connection point considering the relative distance of each node when searching for a routing solution).

1 is a diagram to explain a conventional PDSTSP scheduling method.
2 is a flowchart illustrating an RGSO scheduling optimization method according to an embodiment of the present invention.
3 is a diagram for explaining an example of calculating a solution expression and a total required time (Makespan) according to an embodiment of the present invention.
4 and 5 are diagrams for explaining a producing process according to an embodiment of the present invention.
6 and 7 are diagrams for explaining a scrubbing process according to an embodiment of the present invention.
8 and 9 are diagrams for explaining a ranging process according to an embodiment of the present invention.
10 to 12 are diagrams for explaining a finding process according to an embodiment of the present invention.
13 is a block diagram illustrating a detailed configuration of an RGSO scheduling optimization system according to an embodiment of the present invention.

Advantages and features of the present invention and methods of achieving them will become apparent with reference to the embodiments described below in detail in conjunction with the accompanying drawings. However, the present invention is not limited to the embodiments disclosed below, but will be embodied in various different forms, and only these embodiments allow the disclosure of the present invention to be complete, and common knowledge in the art to which the present invention pertains It is provided to fully inform those who have the scope of the invention, and the present invention is only defined by the scope of the claims.

The terminology used herein is for the purpose of describing the embodiments, and is not intended to limit the present invention. In this specification, the singular also includes the plural, unless specifically stated otherwise in the phrase. As used herein, “comprises” and/or “comprising” refers to the presence of one or more other components, steps, operations and/or elements mentioned. or addition is not excluded.

Unless otherwise defined, all terms (including technical and scientific terms) used herein may be used with the meaning commonly understood by those of ordinary skill in the art to which the present invention belongs. In addition, terms defined in a commonly used dictionary are not to be interpreted ideally or excessively unless clearly specifically defined.

Hereinafter, preferred embodiments of the present invention will be described in more detail with reference to the accompanying drawings. The same reference numerals are used for the same components in the drawings, and repeated descriptions of the same components are omitted.

1 is a diagram to explain a conventional PDSTSP scheduling method.

In the following, the Parallel Drone Scheduling TSP (PDSTSP) scheduling repair model used in the cooperative system of trucks and drones proposed by Murray & Chu (2015) is applied and explained using examples.

C = {1, 2, … , c} is the set of all customers, and C’⊆C is the set of customers that can be delivered to customers by drone because the weight of the delivery is not heavy. In addition, among the customers, the set of customers in the radius area where the drone can deliver from the depot is denoted as C’’⊆C’.

는 운송 수단이 노드 i∈N₀에서 노드 j∈N₊까지 이동하는데 소요되는 시간을 나타내며,

는 드론이 이동하는데 소요되는 시간을 나타낸다. There is one depot that is the departure point or warehouse of the transportation means, and the transportation means starts from the depot node 0 and returns to the depot node c+1. N = {0, 1, 2, … , c+1} represents all nodes existing in the network. Also, N ₀ = {0, 1, ... , c} represents the set of nodes from which the means of transport depart, N ₊ = {1, 2, ... , c+1} represents a set of modes in which the transportation means can visit.

represents the time it takes for the transport to move from node i∈N ₀ to node j∈N ₊ ,

represents the time it takes for the drone to move.

In this case, the transportation means may be a truck such as a truck, but the type is not limited. Hereinafter, the transport means will be referred to as a truck.

이고, 그렇지 않으면

이다. 실시예에 따라서, 고객 i가 디포에서 배달 가능한 지역이고, 배달물의 중량이 드론으로 가능할 경우에는

이고, 드론 v가 고객 i를 방문할 경우, 이진 의사 결정 변수는

이고, 그렇지 않으면

이다. As an example, if a truck moves from customer i to customer j, the binary decision variable is

and otherwise

to be. According to an embodiment, if customer i is an area that can be delivered from the depot and the weight of the delivery is possible with a drone,

, and if drone v visits customer i, the binary decision variable is

and otherwise

to be.

The mathematical model of PDSTSP, a cooperative system between trucks and drones, can be expressed by [Equation 1], which is an objective expression, and [Equation 2] to [Equation 6], which are restrictive expressions. [Equation 1] is to minimize the last time that trucks and drones return to the depot after serving customers, that is, the total required time (Makespan). [Equation 2] and [Equation 3] are the lower bounds of z according to customer allocation of trucks and drones. [Equation 4] indicates that a truck or drone visits each customer only once, [Equation 5] indicates that the truck departs from the depot once and returns to the depot, and [Equation 6] indicates that the truck visits the customer node. and indicates that it starts from that node.

[Formula 1]

[Equation 2]

[Equation 3]

[Equation 4]

[Equation 5]

[Equation 6]

,

를 나타낸다. here,

,

indicates

Referring to Figs. 1 and 2, referring to Figs. 1 and 2, Fig. 1 (a) shows one depot (0) location (20, 20) and customer locations 1 to 9. As shown in Figs. Customer 1 (18, 14), Customer 2 (11, 23), Customer 3 (11, 16), Customer 4 (36, 36), Customer 5 (32, 13), Customer 6 (25) , 4), customer 7 (38, 22), customer 8 (20, 25), customer 9 (0, 40). In the depot, customers 1-3 locations are within the flight time of the drone, and the weight of the delivery can also be carried by the drone. It is assumed that customer locations of customers 4 to 9 can only be delivered by truck.

와 드론이 디포 0에서 고객 i에 배달하고 다시 디포 c+1로 복귀하는 소요 시간

이 계산된다. 트럭에 할당된 고객 서비스 시간 z의 하한(lower bound)인 [수식 2]와 드론에 할당된 고객 서비스 시간 z의 하한인 [수식 3]을 나타낸다. [수식 4]와 같이 고객 9, 8, 4, 7, 6, 5는 트럭으로, 고객 1, 2, 3은 드론으로 1회만 방문하여 배달물을 배달한다. [수식 5]는 트럭이 디포(0)에서 한번 출발하여 고객 9, 8, 4, 7, 6, 5를 거쳐 다시 디포(0)으로 돌아오는 것을 나타내며, [수식 6]은 트럭이 고객 9, 8, 4, 7, 6, 5에 각각 도착하고 출발한다는 것을 나타낸다. FIG. 1( b ) may be displayed as a path in which the truck returns to the depot ( 0 ) → 9 → 8 → 4 → 7 → 5 → 6 → depot ( 0 ). Since trucks travel on the road, the distance is calculated as the Manhattan distance. The drone can be displayed in the following paths: depot (0) → 2 → depot (0), depot (0) → 3 → depot (0), depot (0) → 3 → depot (0). At this time, it is assumed that the drone moves on the ground without obstacles and is calculated as a straight-line distance (Euclidean distance). The delivery time for each truck and drone is calculated by dividing the perpendicular distance of the truck and the straight-line distance of the drone by their respective speeds, and the time required for the truck to move from customer i to customer j

and the time it takes the drone to deliver from depot 0 to customer i and back to depot c+1

This is calculated [Equation 2], which is the lower bound of the customer service time z allocated to the truck, and [Equation 3], which is the lower bound of the customer service time z allocated to the drone. As shown in [Equation 4], customers 9, 8, 4, 7, 6, and 5 are delivered by truck, and customers 1, 2, and 3 are delivered only once by drone. [Equation 5] indicates that the truck departs from the depot (0) once, passes through customers 9, 8, 4, 7, 6, and 5, and returns to the depot (0). Indicates arrival and departure at 8, 4, 7, 6, and 5 respectively.

2 is a flowchart illustrating an operation of an RGSO scheduling optimization method according to an embodiment of the present invention. did it

The method of FIG. 2 is performed by the RGSO scheduling optimization system according to the embodiment of the present invention shown in FIG. 13 .

Referring to FIG. 2 , in step S210, location data of a customer is input.

In step S210, the customer's location data (x, y), the speed of the transportation means, the speed of the drone, and the weight of the deliverable by the drone are input, and the deliverable weight node of the drone and the delivery range of the drone can be input. .

In step S220, from the location data of the customer, delivery location data that can be delivered by a transportation means and a drone is acquired, and a distance from the depot to the delivery location data and a required time are calculated.

Step S220 may include obtaining delivery location data that can be delivered by the drone by checking a delivery location where the drone cannot deliver due to excess weight of the delivery item and a delivery location that can be delivered by the drone from the depot among the customer's location data.

Accordingly, step S220 calculates the right angle distance and required time of the transportation means between all customer locations including the depot from among the delivery location data, calculates the linear distance and the required time of the drone between all customer locations in the depot, and calculates the transportation means and the drone It is possible to define the evaluation function of the total required time (Makespan) as the required time. At this time, in step S220, since the transportation means moves on the road, it is calculated as a right angle distance (Manhattan distance), and the drone is calculated as a straight distance (Euclidean distance) assuming that it moves on the ground without obstacles. The delivery time required for each of the transportation means and the drone is calculated by dividing the perpendicular distance of the transportation means and the straight-line distance of the drone by the respective speeds, which is the same as the existing PDSTSP scheduling method described above.

In step S230, an initial solution for the path of the vehicle and the drone is generated, and a producing process, a scrolling process, a ranging process, and a finding process are performed for a producer of the initial solution.

In step S230, based on the result values of the distance and required time calculated in step S220, an initial solution for each route of the transportation means and drone is generated from the delivery location data, and a population of the initial solution is generated, and the total You can calculate the required time (Makespan).

Referring to FIG. 3, the solution expression and the total required time (Makespan) calculation will be described. FIG. 3(a) shows the depot (0) and 9 customer location courier scheduling solutions of FIG. 1 . In the case of drones, customers 1 to 3 locations in the depot can be delivered with drones within flight distance and less than the weight of the deliverables. Drone delivery is marked with ‘1’, and the remaining locations that are not delivered by drone are marked with ‘0’. Truck, which is a means of transportation, is displayed in the order of 9→8→4→7→5→6 routes except for customer 1-3 locations.

Referring to FIG. 3(b), the right angle distance (0-9-8-4-7-5-6-0) of each truck is 40, 25, 17, 16, 15, 16, 15, 16, 21. Calculated. The straight-line distances (0-1-0, 0-2-0, 0-3-0) of the drone are calculated as 6.32 round trip distances of 12.64, 9.85 round trip distances 19.7, and 9.49 round trip distances 18.98, respectively. At this time, it is assumed that the truck speed is 40 km/h and the drone speed is 60 km/h, and the required time is 3.75 hours by dividing the total perpendicular distance (150 km) of the truck by the speed of the truck, 51.32 km) is divided by the speed of the drone to calculate 0.855 hours. Therefore, the total required time (Makespan) is 3.75 hours.

Referring back to FIG. 2 , in step S230, a producing process, a scrolling process, a ranging process, and a finding process are performed for a producer of the initial year, and the producing process, the scrolling process, the ranging process, and the finding process are performed. You can determine the proportions of the course.

More specifically, step S230 may perform a producing process of searching for neighboring solutions in a producer and updating them to a new producer based on a reference probability of a transportation means and a reference probability of a drone. The producer represents the best solution, searches the neighboring solutions of the current best solution, and when a better solution is found, it converges while updating the solution with a new producer.

In addition, step S230 selects about 60% excluding the new producer as the previous scrounger, searches for a solution according to the new producer, and performs a scrubbing process of updating to the new scrounger In addition, about 10% excluding new producers are selected as previous rangers, and various solutions of random neighboring solutions can be searched for and a ranging process of updating to a new ranger can be performed. The scrounger follows the producer to generate a neighboring solution and converges while updating to the new solution regardless of the evaluation value of the new solution. In addition, a ranger performs various searches by randomly generating solutions.

In addition, in step S230, about 30% excluding new producers are selected as previous finders, and a finding process of searching for a solution through a mechanism suitable for routing and updating to a new finder is performed. can

The RGSO scheduling optimization method according to an embodiment of the present invention is routed to the producing process, the scrolling process, and the ranging process of the existing Group Search Optimization (GSO) method in order to search for the optimal scheduling solution of the existing PDSTSP problem. Routing), an efficient RSGO scheduling optimization method can be provided by adding a finding process that applies a mechanism that can effectively search the solution of the problem (searching for connection points by considering the relative distance of each node when searching for a routing solution). The producing process, the scrolling process, the ranging process, and the finding process will be described in detail with reference to FIGS. 4 to 12 below.

In step S240, evaluation values of all solutions according to the execution results are calculated and a solution indicating the total required time (Makespan) of the minimum value is updated to the producer.

Step S240 calculates the total required time (Makespan) for all N solutions in the producing process, the scrolling process, the ranging process, and the finding process, and selects a new producer showing the minimum evaluation value among them, and You can update producers by comparing new producers.

In step S241, it is determined whether the termination condition according to the update result is satisfied, and in step S250, an optimal solution for parallel delivery using a transportation means and a drone is obtained according to the update result.

Step S250 ends the RGSO scheduling optimization method when the preset termination condition is met with no change in the producer for a certain generation, sets the best solution among the solutions found so far as the producer, and then performs parallel delivery. can be derived as an optimal solution for For example, in step S250, when the evaluation value of the new producer is greater than the evaluation value of the existing producer, the need to update the existing producer is eliminated, and thus it may be determined that a preset termination condition in which there is no change of the producer for a certain generation is satisfied.

4 and 5 are diagrams for explaining a producing process according to an embodiment of the present invention.

A production process of the RGSO scheduling optimization method for parallel delivery of trucks and drones as transportation means will be described with reference to FIGS. 4 and 5 . The previous producer in the previous stage shown in Fig. 4 delivered only customer 2 by drone, and the truck delivered to the depot (0), 1, 3, 9, 5, 4, 7, 8, 6, and depot (0). Returning to , the evaluation value total required time (Makespan) becomes 6.05. Here, it is assumed that random numbers R1 and R2 (0~1) are determined with random probabilities, the reference probability of the truck is 0.1, and the reference probability of the drone is 0.7.

Only customer locations 1, 2, and 3 can be delivered by drone, but random probabilities 0.52 and 0.64 are smaller than the reference probability 0.7, so the solution expression '0' for drone customers 1 and 2 is changed to '1', as shown in FIG. Likewise, the year of the new producer will be changed to deliver customers 1, 2, and 3 all by drone.

In addition, as shown in FIG. 4 , since the random selection probability 0.03 of the 8th node is smaller than the reference probability 0.1, it is changed to the 5th node randomly selected among nodes 1 to 9 excluding itself. In addition, the random probability of nodes 1 and 3 is lower than the reference probability of 0.7, so the expression of the drone is also changed from 0 to 1. In addition, the random probability 0.72 of node 2 is higher than the reference probability 0.7 of the drone, and thus the current drone expression 1 is maintained. The drone random probability of node 6 is 0.72, which is greater than the reference probability of 0.7, but the expression 0 is maintained because the drone cannot reach it.

The new producer created through the above-mentioned production process is delivered by drones to customers 1, 2, and 3, and the trucks are depots (0), 9, 8, 4, 7, 5, and 6 Returning to the depot (0), the evaluation value total required time (Makespan) became 3.75, and the solution improved.

6 and 7 are diagrams for explaining a scrubbing process according to an embodiment of the present invention.

6 and 7, a scrolling process in which the scrollers follow the best solution (or producer, producer) for the solution search of the RGSO scheduling optimization method for parallel delivery of trucks and drones, which are transportation means, will be described. .

The previous scrounger as shown in Fig. 6 delivered only customer 2 by drone, and the truck as a means of transportation was depot (0), 4, 5, 6, 9, 8, 1, 3, 7, Returning to the depot (0), the total time required (Makespan) becomes 6.15. At this time, the change probabilities R1 and R2 (0 to 1) are randomly determined, and it is assumed that the reference probability of the truck is 0.3 and the reference probability of the drone is 0.7.

In Fig. 6, the previous scrounger's solutions have random probabilities of 0.12, 0.16, and 0.08 of nodes 6, 9, and 8 are lower than the reference probability of 0.3, so they resemble the solutions of the producer's solutions 9, 8, and 4 . Accordingly, in the previous scrounger, the 2nd node of the drone does not change because the probability 0.75 is greater than the reference probability 0.7, and the 1 and 3 nodes are 0.16, 0.08, which is smaller than the reference probability 0.7, so the producer's 1, According to the expression 1 of the 3 nodes, it is changed to 1 of the new scrounger solution as shown in FIG. 7 .

The new scrounger created through the above-mentioned scrubbing process is delivered by drones to customers 1, 2, and 3, and the trucks to depots (0), 6, 5, 9, 8 , 4, 7, returned to the depot (0), and the evaluation value total required time (Makespan) became 4.35, improving the solution.

8 and 9 are diagrams for explaining a ranging process according to an embodiment of the present invention.

A ranging process for variously searching for solutions of the RGSO scheduling optimization method for parallel delivery of trucks and drones as transportation means will be described with reference to FIGS. 8 and 9 .

For the solution of the previous ranger as shown in FIG. 8, in the case of a truck, nodes 6, 2, 7, and 9 are changed to random probability R1, and in the case of a drone, nodes 2 and 3 are changed to a random probability R2. Accordingly, the node that is changed to the solution of the new ranger shown in FIG. 9 is randomly changed to a node between 1 and 9. Accordingly, the expression of the drone is also identified as the node is changed.

10 to 12 are diagrams for explaining a finding process according to an embodiment of the present invention.

A finding process for searching for a finder solution suitable for routing in an RGSO scheduling optimization method for parallel delivery of trucks and drones as transportation means will be described with reference to FIGS. 10 to 12 .

In the previous finder as shown in Fig. 10, only customer 3 is delivered by drone, and the truck returns to the depot (0), 6, 1, 4, 2, 5, 9, 8, 7 depot (0). The evaluation value total required time (Makespan) becomes 7.05. At this time, it is assumed that the change probabilities R1 and R2 (0~1) are randomly determined, and the truck-based probability and the drone-based probability are 0.7.

In the solution of the previous finder of FIG. 10 , 0.66 of No. 2 and 0.49 No. 3, belonging to the drone change probability R2, are smaller than the reference probability 0.7, so 0 is changed to 1 and 1 is changed to 0, and the solution is converted to the same solution as in FIG. 11A .

After the drone route is changed, in FIG. 11B , the No. 1 node can be changed because the truck No. 0.24 is smaller than the reference probability of 0.7. Accordingly, when changing node 1, the perpendicular distances (17, 23, 26, 16, 31, 36, 61), and the perpendicular distances (40, 38, 45, 27, 16, 17, 40) from node 4 to 1, 2, 3, 5, 7, 8, 9 to calculate the two distances. Calculate the sum (57, 61, 71, 45, 47, 53, 101) of Accordingly, as shown in FIG. 11C , node 5 with the smallest value of 45 among the sums of distances is selected and exchanged with node 1.

After exchanging node 5 with node 1, in FIG. 11D , node no. 4 of truck has a change probability of 0.64, which is smaller than the reference probability of 0.7, so it is necessary to decide which node to exchange with. Accordingly, the previous node of node 4 is node 5, and the next node is node 2, but since node 2 moves by drone, node 1, which moves by truck, becomes the next node of node 4. Therefore, the distance (31, 24, 27, 16, 15, 34, 59) from node 5 to the previous node 5 of node 4 and 2, 3, 4, 6, 7, 8, 9 is calculated, and the node next to node 4 is calculated. Calculate the distances (16, 9, 40, 17, 28, 24, 44) from node 1 and 2, 3, 4, 6, 7, 8, 9 to the sum of the two distances (47, 33, 67, 33 , 43, 58, 103). As a result, among the nodes 3 and 6 of 33, which is the smallest among the sum of distances, 3 is selected as shown in FIG. 11E and exchanged with the node 4.

A new finder as shown in FIG. 12 is created by changing nodes 9, 8, and 7 of FIG. 11F in the same manner as described above. In the new Finder created through the above-mentioned finding process, Customer 2 delivers the delivery by drone, and the truck uses Depot (0), 6, 5, 9, 7, 3, 1, 8, 4, Depot. Returning to (0), the evaluation value total required time (Makespan) became 6.65, which improved the solution.

13 is a block diagram illustrating a detailed configuration of an RGSO scheduling optimization system according to an embodiment of the present invention.

Referring to FIG. 13 , the RGSO scheduling optimization system according to an embodiment of the present invention cooperates with transportation means and drones to deliver the last time (or total required time, Makespan) until all deliveries are delivered to all customer locations. We propose a method of minimizing scheduling and movement path search.

To this end, the RGSO scheduling optimization system 1300 according to an embodiment of the present invention includes an input unit 1310, a data processing unit 1320, an execution unit 1330, a producer update unit 1340, an acquisition unit 1350, and a control unit ( 1360). In this case, the controller 1360 according to the embodiment of the present invention may control the performing unit 1330 and the producer updating unit 1340 to be repeatedly performed according to whether the termination condition is met.

The input unit 1310 inputs customer location data.

The input unit 1310 inputs the customer's location data (x, y), the speed of the transportation means, the speed of the drone, and the weight of the deliverable of the drone, and the delivery range of the drone deliverable weight node and the drone. can

The data processing unit 1320 obtains delivery location data that can be delivered by transportation means and drones from among the location data of the customer, and calculates the distance and required time from the depot to the delivery location data.

The data processing unit 1320 may obtain delivery location data that can be delivered by a drone by checking a delivery location where a drone cannot deliver due to excessive weight of a delivery item and a delivery location that can be delivered by a drone from a depot among the location data of the customer.

Accordingly, the data processing unit 1320 calculates the orthogonal distance and required time of transportation means between all customer locations including the depot from among the delivery location data, calculates the straight-line distance and the required time of the drone between all customer locations in the depot, and transports The evaluation function of the total required time (Makespan) can be defined as the time required for the means and the drone. At this time, the data processing unit 1320 calculates as a right angle distance (Manhattan distance) since the transportation means moves on the road, and it is calculated as a straight distance (Euclidean distance) assuming that the drone moves on the ground without obstacles. The delivery time required for each of the transportation means and the drone is calculated by dividing the perpendicular distance of the transportation means and the straight-line distance of the drone by the respective speeds, which is the same as the existing PDSTSP scheduling method described in FIG. 1 .

The performer 1330 generates an initial solution for the route of the vehicle and the drone, and performs a producing process, a scrolling process, a ranging process, and a finding process for a producer of the initial solution.

The performing unit 1330 generates an initial solution for each route of the transportation means and the drone from among the delivery location data based on the result values of the distance and the required time calculated by the data processing unit 132, and the population of the initial solution (Population) ) and calculate the total required time (Makespan).

The performing unit 1330 performs a producing process, a scrolling process, a ranging process, and a finding process for the producer of the initial year, and the composition ratio of the producing process, the scrolling process, the ranging process, and the finding process can be decided

More specifically, the performing unit 1330 may perform a producing process of searching for neighboring solutions in a producer and updating them to a new producer based on a reference probability of a transportation means and a reference probability of a drone. The producer represents the best solution, searches the neighboring solutions of the current best solution, and when a better solution is found, it converges while updating the solution with a new producer.

In addition, the performing unit 1330 selects about 60% of excluding the new producer as the previous scrounger, searches for a solution according to the new producer and updates it with the new scrounger. process, selecting about 10% excluding new producers as previous rangers, searching for various solutions of random neighboring solutions, and updating to new rangers can be performed. The scrounger follows the producer to generate a neighboring solution and converges while updating to the new solution regardless of the evaluation value of the new solution. In addition, a ranger performs various searches by randomly generating solutions.

In addition, the executor 1330 selects about 30% excluding the new producer as the previous finder, searches for a solution through a mechanism suitable for routing, and updates the finding process with the new finder. can be performed.

The RGSO scheduling optimization system 1300 according to an embodiment of the present invention is a producing process, a scrolling process, and a ranging process of the existing group search optimization (GSO) method in order to search for an optimal scheduling solution of the existing PDSTSP problem. An efficient RSGO scheduling optimization method can be provided by adding a finding process that applies a mechanism that can effectively search the solution of the routing problem to have.

The producer update unit 1340 calculates evaluation values of all solutions according to the execution result and updates the solution indicating the minimum total required time (Makespan) to the producer.

The producer update unit 1340 calculates the total required time (Makespan) for all N solutions in the producing process, the scrolling process, the ranging process, and the finding process, and selects a new producer representing the minimum evaluation value among them. You can update the producer by comparing the producer with the new producer.

The acquisition unit 1350 acquires an optimal solution for parallel delivery using a transportation means and a drone according to the update result.

The acquisition unit 1350 terminates the RGSO scheduling optimization method when the preset termination condition is met with no change in the producer for a certain generation, and sets the best solution among the solutions found so far as the producer, It can be derived as an optimal solution for parallel delivery. As an example, when the evaluation value of the new producer is greater than the evaluation value of the existing producer, the acquiring unit 1350 determines that it meets the preset termination condition in which there is no change of the producer for a certain generation because the need to update the existing producer is eliminated. can

Although the description of the system of FIG. 13 is omitted, it is apparent to those skilled in the art that the system according to the present invention may include all the contents described with reference to FIGS. 1 to 12 .

The system or apparatus described above may be implemented as a hardware component, a software component, and/or a combination of the hardware component and the software component. For example, the devices and components described in the embodiments may include, for example, a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable gate array (FPGA). , a programmable logic unit (PLU), microprocessor, or any other device capable of executing and responding to instructions, may be implemented using one or more general purpose or special purpose computers. The processing device may execute an operating system (OS) and one or more software applications running on the operating system. The processing device may also access, store, manipulate, process, and generate data in response to execution of the software. For convenience of understanding, although one processing device is sometimes described as being used, one of ordinary skill in the art will recognize that the processing device includes a plurality of processing elements and/or a plurality of types of processing elements. It can be seen that may include For example, the processing device may include a plurality of processors or one processor and one controller. Other processing configurations are also possible, such as parallel processors.

The software may comprise a computer program, code, instructions, or a combination of one or more thereof, which configures a processing device to operate as desired or is independently or collectively processed You can command the device. The software and/or data may be any kind of machine, component, physical device, virtual equipment, computer storage medium or device, to be interpreted by or to provide instructions or data to the processing device. , or may be permanently or temporarily embody in a transmitted signal wave. The software may be distributed over networked computer systems and stored or executed in a distributed manner. Software and data may be stored in one or more computer-readable recording media.

The method according to the embodiment may be implemented in the form of program instructions 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 recorded on the medium may be specially designed and configured for the embodiment, or may be known and available to those skilled in the art of computer software. Examples of the computer-readable recording medium include magnetic media such as hard disks, floppy disks and magnetic tapes, optical media such as CD-ROMs and DVDs, and magnetic such as floppy disks. - includes magneto-optical media, and hardware devices specially configured to store and execute program instructions, such as ROM, RAM, flash memory, and the like. Examples of program instructions include not only machine language codes such as those generated by a compiler, but also high-level language codes 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.

As described above, although the embodiments have been described with reference to the limited embodiments and drawings, various modifications and variations are possible by those skilled in the art from the above description. For example, the described techniques are performed in a different order than the described method, and/or the described components of the system, structure, apparatus, circuit, etc. are combined or combined in a different form than the described method, or other components Or substituted or substituted by equivalents may achieve an appropriate result.

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

Claims (20)

In the RGSO (Routing Group Search Optimization) scheduling optimization method for searching for an optimal solution of parallel delivery using transportation means and drones,
A first step of inputting customer location data;
a second step of acquiring delivery location data that can be delivered by a transportation means and the drone from among the customer location data, and calculating a distance and a required time from a depot to the delivery location data;
a third step of generating an initial solution for the route of the vehicle and the drone, and performing a producing process, a scrolling process, a ranging process, and a finding process for a producer of the initial solution;
a fourth step of calculating evaluation values of all solutions according to the performance result and updating the solution indicating the minimum total required time (Makespan) to the producer; and
A fifth step of obtaining an optimal solution for parallel delivery using the vehicle and the drone according to the update result
RGSO scheduling optimization method comprising a. According to claim 1,
The first step is
RGSO scheduling, inputting the customer's location data (x, y), the transport means and the speed of the drone, and the deliverable weight of the drone, and inputting the deliverable weight node of the drone and the deliverable range of the drone Optimization method. 3. The method of claim 2,
The second step is
Acquiring the delivery location data by identifying a delivery location that the drone cannot deliver due to excess weight of a delivery item and a delivery location that can be delivered by the drone from the depot among the location data of the customer, RGSO scheduling optimization method comprising the steps of: . 4. The method of claim 3,
The second step is
From the delivery location data, calculating the right angle distance and required time of the transportation means between all customer locations including the depot, calculating the straight-line distance and required time of the drone between all customer locations in the depot, and the transportation means and RGSO scheduling optimization method for defining an evaluation function of the total required time (Makespan) as the required time of the drone. 5. The method of claim 4,
The third step is
Based on the result values of the distance and the required time calculated in the second step, an initial solution for each route of the transportation means and the drone is generated from the delivery location data, and a population of the initial solution is generated and calculating the total required time (Makespan), an RGSO scheduling optimization method. According to claim 1,
The third step is
A method for optimizing RGSO scheduling, performing the producing process of searching for neighboring solutions in the producer and updating them to a new producer based on a reference probability of a vehicle and a reference probability of a drone. 7. The method of claim 6,
The third step is
The scrubbing process of selecting about 60% excluding the new producer as previous scrounger, searching for a solution according to the new producer and updating to a new scrounger, and the new producer A method for optimizing RGSO scheduling, in which about 10% excluding , is selected as a previous ranger, and the ranging process of searching for various solutions of random neighboring solutions and updating them with a new ranger is performed. 8. The method of claim 7,
The third step is
About 30%, excluding the new producer, are selected as previous finders, and the finding process of searching for a solution through a mechanism suitable for routing and updating to a new finder is performed. , RGSO scheduling optimization method. According to claim 1,
The fourth step is
Calculate the total required time (Makespan) for all N solutions in the producing process, the scrolling process, the ranging process, and the finding process, and select a new producer showing the minimum evaluation value among them, A method for optimizing RGSO scheduling, wherein the producer is updated by comparing the producer with the new producer. 10. The method of claim 9,
The fifth step is
The RGSO scheduling optimization method of deriving the optimal solution for parallel delivery by terminating the RGSO scheduling optimization method when a preset termination condition in which the producer does not change for a certain generation is met. In the RGSO scheduling optimization system for performing the RGSO (Routing Group Search Optimization) scheduling optimization method to search for the optimal solution of parallel delivery using transportation means and drones,
an input unit for inputting customer location data;
a data processing unit that acquires delivery location data that can be delivered by a transportation means and the drone from among the location data of the customer, and calculates a distance and a required time from a depot to the delivery location data;
an execution unit generating an initial solution for the routes of the transportation means and the drone, and performing a producing process, a scrolling process, a ranging process, and a finding process for a producer of the initial solution;
a producer update unit that calculates evaluation values of all solutions according to the execution result and updates the solution representing the minimum total required time (Makespan) to the producer; and
Acquisition unit that obtains an optimal solution for parallel delivery using the vehicle and the drone according to the update result
RGSO scheduling optimization system comprising a. 12. The method of claim 11,
the input unit
RGSO scheduling, inputting the customer's location data (x, y), the transport means and the speed of the drone, and the deliverable weight of the drone, and inputting the deliverable weight node of the drone and the deliverable range of the drone optimization system. 12. The method of claim 11,
The data processing unit
RGSO scheduling optimization system, characterized in that the delivery location data is obtained by identifying a delivery location that the drone cannot deliver due to excess weight of the delivery item and a delivery location that can be delivered to the drone in the depot among the location data of the customer . 14. The method of claim 13,
The data processing unit
From the delivery location data, calculating the right angle distance and required time of the transportation means between all customer locations including the depot, calculating the straight-line distance and required time of the drone between all customer locations in the depot, and the transportation means and An RGSO scheduling optimization system that defines an evaluation function of the total required time (Makespan) as the required time of the drone. 15. The method of claim 14,
the performing unit
Based on the result values of the distance and the required time calculated by the data processing unit, an initial solution is generated for each route of the transportation means and the drone from among the delivery location data, and a population of the initial solution is generated, , RGSO Scheduling Optimization System, which calculates the total required time (Makespan). 12. The method of claim 11,
the performing unit
An RGSO scheduling optimization system for performing the producing process of searching for neighboring solutions in the producer and updating them to a new producer based on a reference probability of a vehicle and a reference probability of a drone. 17. The method of claim 16,
the performing unit
The scrubbing process of selecting about 60% excluding the new producer as previous scrounger, searching for a solution according to the new producer and updating to a new scrounger, and the new producer An RGSO scheduling optimization system that selects about 10% excluding , as a previous ranger, and performs the ranging process of searching for various solutions of random neighboring solutions and updating them with a new ranger. 18. The method of claim 17,
the performing unit
About 30%, excluding the new producer, are selected as previous finders, and the finding process of searching for a solution through a mechanism suitable for routing and updating to a new finder is performed. which, RGSO Scheduling Optimization System. 12. The method of claim 11,
The producer update department
Calculate the total required time (Makespan) for all N solutions in the producing process, the scrolling process, the ranging process, and the finding process, and select a new producer showing the minimum evaluation value among them, An RGSO scheduling optimization system for updating the producer by comparing the producer with the new producer. 20. The method of claim 19,
the acquisition unit
The RGSO scheduling optimization system for deriving the optimal solution for parallel delivery by terminating the RGSO scheduling optimization method when a preset termination condition in which the producer does not change for a certain generation is met.

Images