JP2001034880A - Delivery schedule supporting device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a delivery schedule supporting device capable of planning an optimum delivery schedule by calculating plural saving matrixes for every start place and return place, retrieving the combination candidate of delivery destinations whose saving values are maximum and allocating a joint candidate to a truck whose start place/return place are nearest to the delivery destination. SOLUTION: A saving value calculation part 5 calculates saving matrixes for respective trucks different in start places/return places based on a distance data base and a truck data base. A retrieval part 7 retrieves the combination of delivery destinations whose saving values in the saving matrixes are the largest. The block of nodes that one truck works are left as they are and allocation is changed to the truck whose start place/return place is the nearest to the block. A joint judgment part 8 judges whether respective restriction conditions are satisfied when the nodes of a merger candidate is merged with the truck.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a delivery plan support system for a plurality of trucks (trucks) having different departure points and return points in a system for solving the planning of a delivery plan by a saving algorithm. In addition,
The term “delivery” in the present invention is a general term for collective delivery and delivery such as collecting packages and collecting empty containers, in addition to delivering packages.

[0002]

2. Description of the Related Art In a delivery plan support system, an algorithm called a simulated annealing method, a genetic algorithm, or a saving method is used. Above all, the saving method is generally used because an accurate solution can be obtained relatively easily. However, in the saving method, all the trucks have one type of departure place and return place (that is, a plurality of trucks depart from the same place of departure and return to the same place of departure. Location may be used).
For this reason, when the departure place and the return place differ depending on the truck, for example, when a rental car other than the private car belonging to the distribution center is used, the A truck is the A departure place and the return place, and the B truck is the B departure. In the case of a place or return point, it is difficult to optimize the delivery plan by the saving method considering the conditions of the place of departure and place of return since each place of departure and return point differs depending on the vehicle. Met. As a method of simply making the saving method correspond to a plurality of departure points and return points, a method of preparing a saving matrix for each departure point and return point and searching for a merger candidate by searching them all is conceivable. However, in this method, each truck is in charge of a destination farther from its own departure point and return point because of the characteristic of the saving method that a destination farther to a certain track is merged with priority. The irrationality that will happen. For this reason, it has been difficult to formulate an optimal delivery plan in consideration of the difference between the departure place and the return place, such as giving priority to the delivery destination near the return place of the truck.

[0003]

As described above, by applying the conventional saving method as it is, a delivery plan for delivering to a plurality of delivery destinations using a plurality of trucks having different departure and return points is prepared. In such a case, there arises an irrational situation that a delivery destination farther from the departure place and the return place may be assigned, and there is a problem that an optimal delivery plan cannot be made.

The present invention has been made in order to solve the above-mentioned problems of the prior art, and relates to a delivery plan for delivering to a plurality of delivery destinations using a plurality of trucks having different departure and return points. Basically, to provide a delivery plan support device that can formulate an optimal delivery plan in which each truck preferentially delivers a delivery destination near its own departure place and return place while applying the saving method. With the goal.

[0005]

Means for Solving the Problems In order to achieve the above object, the present invention is configured as described in the claims. That is, in the present invention,
A plurality of saving matrices are calculated for each destination, and a search is made for a combination of delivery destinations that maximizes the saving value in the matrix, and the truck with the smallest saving value, that is, the departure place and the destination, are searched. Assigns the merger candidate to the truck closest to the destination. As a result, while performing the same dispatch as the conventional saving method, the merger candidate,
Automatically assigned to trucks with closer departure and return locations. Then, when the merging restriction condition (for example, a condition that does not exceed the maximum load capacity of the truck) is satisfied, the routes to which the respective delivery destinations belong are merged. In this way, by simply expanding the conventional saving method to multiple saving matrices for each departure point and return point, optimal allocation in the case of including multiple trucks with different departure points and return points based on saving values Becomes possible. Note that a plurality of trucks having different departure and return locations means that the departure and return locations are different for each truck, and the departure and return locations of a certain track are the same or different (for example, different from the departure location). (When there is a garage returning to a different position).

[0006]

According to the present invention, an accurate solution is basically obtained relatively easily with respect to a delivery plan for delivering to a plurality of delivery destinations using a plurality of trucks having different departure and return locations. It is possible to formulate an optimal delivery plan that gives priority to delivery destinations near the departure point and return point of the truck itself, which could not be formulated with the conventional saving method while using the saving method that is possible Is obtained.

Further, in the present invention, since the saving method is basically used, it is possible to perform a vehicle allocation in consideration of a departure point and a return point only by moving costs between nodes such as distances, and The conventional search device can be used as it is for searching for a saving value, and the present invention can be realized by simply providing a plurality of saving matrices for each track having different starting and returning points and expanding the search engine for a plurality of saving matrices. Since it can be implemented, it can be easily implemented.

[0008]

FIG. 1 is a block diagram showing a configuration of an embodiment of the present invention, and FIG. 2 is a diagram showing an example of a delivery route. In FIG. 1, reference numeral 1 denotes a customer database memory for storing a customer database. The customer database is the customer information of the delivery destination, and the location of the delivery destination (town name, location,
Address, latitude, longitude, etc.), time constraints (delivery time, etc.), and the amount of cargo are stored. The distance database calculation unit 2 calculates a moving cost between a plurality of customers based on the above-mentioned customer database, and stores it in the distance database memory 3 as a distance database. The moving cost is a value including the moving distance between customers, the moving time, the fuel cost, the labor cost, and the like. However, for simplification, for example, the moving cost can be represented only by the moving distance between the customers. Reference numeral 4 denotes a track database memory for storing a track database, and stores information such as a maximum load capacity, positions of departure and return points, operation time, and departure time for each track. The above-mentioned customer database and track database are stored in advance in each memory.

The above-mentioned distance database calculation unit 2 and saving value calculation unit 5, search unit 7, and merge determination unit 8 described later.
Corresponds to the function of the delivery plan support device 11 installed in the delivery center 10 or the like, for example, as shown in FIG. The delivery plan support device 11 is composed of, for example, a computer. The customer database memory 1, distance database memory 3, track database memory 4, and saving matrix memory 6 described below correspond to, for example, a hard disk device in the computer. In FIG. 2, white circles indicate destinations (customers), and black lines connecting the destinations indicate delivery routes.

Next, the saving value calculation unit 5 calculates a saving matrix for each track having different departure and return points based on the distance database and the track database.

Next, the search unit 7 searches (primary search) for a combination of delivery destinations that gives the maximum saving value in the saving matrix. For example, when Smn is the maximum as the saving value in the saving matrix, the delivery destination node m and the node n are merging candidates. However, when a plurality of trucks having different departure points and return points are included, if the above-described node m and node n are used as merge candidates as in the past, each of the delivery destinations farther from the departure point and the return point is determined. Inconsistent results of being in charge.
For this reason, it is necessary to change the assignment to a track whose departure point and return point are closer to the block while leaving the block of nodes in charge of one truck as it is. The fact that the departure place and the return place are close means that the saving value is small. Therefore, as in the past, the nodes m and n of Smn that give the maximum saving value are taken as merging candidates, but the trucks actually in charge of the delivery of the node are determined by the secondary search in the combination of the nodes m and n. The track having the smallest saving value is set. As a result, a solution is obtained in which each truck is in charge of a delivery node group near each departure point and return point, and each lump is a proper combination as in the conventional saving method.

Next, the merge judging section 8 judges whether or not each of the constraint conditions is satisfied when merging candidate nodes are merged into the track. For example, it is checked whether the maximum loading capacity of the truck is exceeded due to the merger. If the constraints are respected, the merger is actually performed. Otherwise, do not merge. Regardless of whether or not merging is performed, all saving values of the merging candidates are reset to 0 so that the merging candidates will not be merged candidates from the next time.

Hereinafter, the contents of the above-described saving matrix creation, search, and merger will be described in detail. 3A and 3B are diagrams for explaining a method of calculating a saving value by a normal saving method. FIG. 3A shows a case where the departure place and the return place are the same place, and FIG. 3B shows a case where the departure place and the return place are different. Is shown. First, as shown in FIG. 3A, when the departure place and the return place are the same, the case where the delivery is performed from the departure place node 0 (node 0) to the delivery destinations nodei (node i) and nodej (node j). Think. Each nod
Assuming that the distances between e are d0i, d0j, and dij, the saving distance in the case of traveling in the order of node0 → nodei → nodej in the case of piston transportation (simply go from the place of departure to each destination and return) is as follows: It is represented by the following (Equation 1). Sij = (d0i + d0j) × 2− (d0i + d0j + dj) = d0i + d0j−dj (Equation 1) The value of (Equation 1) is called a saving value.

If a saving matrix in which saving values are calculated for all combinations of nodei and nodej is created and nodes are merged in descending order of the saving value, a vehicle that is nearly optimal can be allocated. For example, when Max (S) = Sij, nodei, n
odej will be merged into a route that can be carried on one truck. At this time, if the conditions of the truck (maximum loading capacity, etc.) cannot be observed due to the merger, the merger will not be performed.
As a merger candidate.

As shown in FIG. 3B, the saving value when the departure point and the return point are different can be expanded as shown in the following equation (2). In FIG. 3B, n
odem (node m) is a return point different from the departure point node0. s1 = d0i + dmi + d0j + dmj- (d0i + dij + dmj) = dmi + d0j-djs s2 = d0i + dmi + d0j + dmj- (d0j + dj + dmi) = dj + d0i-dij is a place where the starting place is different from that of the sigma where the starting place is different from the sig where the starting place is different from that of the sj = 2, and the starting place is different from the starting place where the starting place is different from the starting place where the starting place is different from the starting place where the starting place is different from the starting place where the starting place is different from the starting place where the starting place is different from the starting place where the starting place is different from that of the starting place. It is a saving matrix when there is only one type of return place. However, when the departure place and the return place are different for each track, the calculation formula of the saving value is different for each track, so that the track cannot be used as it is.

As described above, in the conventional saving method, when Max (S) = Sijk, that is, when the saving value becomes maximum when the track k circulates between nodei and nodej, nodei and nodej are used.
Is incorporated in the route to be delivered by truck k. In this way, as can be seen from the saving value equation, a cluster of nodes that are farther and closer together are merged into one route. Therefore,
When a plurality of tracks having different return destinations are used, for example, when there is an extreme node bias as shown in FIG. 4, an inconsistent result is produced in which each track is in charge of a far area. In FIG. 4, a plurality of black dots indicate a destination node, and circles A and B indicate a departure place and a return place. That is, truck A departs from departure point A and delivers in area BB and returns to return point A, and truck B departs from departure point B and delivers in area AA and returns to return point B. .

In order to solve the above-mentioned problem, the nod
Since the method of combining ei and nodej into a lump is appropriate, the departure point near the route of nodeei and nodej
This node is assigned to the track having a return point, that is, the track having the minimum saving value. For example, as shown in FIG. 5, the area AA is handled by the truck A and the area BB is handled by the truck B, so that each node is left as it is, and each of the nodes is delivered by a truck near the departure place. .

FIG. 6 is a flowchart showing an embodiment of the calculation method. In FIG. 6, first, step S
In step 1, a customer database and a track database are created and stored in a memory. Next, step S2
Then, a distance database is created based on the customer database. The distance database is, to be precise, the travel cost including the travel distance between customers, travel time, fuel cost, labor cost, etc., but for simplification, it can be represented only by the travel distance between customers, for example. .

Next, in step S3, based on the distance database and the track database, a saving matrix is calculated for each track having different departure and return points. FIG. 7 is a diagram conceptually showing one saving matrix for each track. As shown in FIG. 3B, the saving value is calculated, for example, as follows: departure point is node0, return point is nodem, and each destination is no.
dei and nodej, and the moving costs (distance data) between the respective nodes are represented by d0i, dij, dm as shown in the figure.
If j, d0j, and dmi, nodeei and node
The saving value Sij of ej is as shown in the following (Equation 3). Sij = Max (dmi + d0j-dij, dmj + d0i-dij) (Equation 3) Next, as shown in FIG. 8, as shown in FIG. Next search). For example, when Smn is the maximum as the saving value in the saving matrix, the delivery destination node m and the node n are merging candidates.

However, if a plurality of trucks having different departure and return locations are included as in the prior art and the above-mentioned nodes m and n are considered as merging candidates, as shown in FIG. The contradictory result is that you are responsible for a destination that is farther from the destination. In this case, it is considered efficient that one truck delivers a group of nodes, but if the truck in charge of the same save method is used, it belongs to a farther departure place / return place Is a problem. Therefore, it is necessary to change the assignment to a truck whose departure point and return point are closer to the block while leaving the block of nodes in charge of one truck as it is. The fact that the departure place and the return place are close means that the saving value is small. Therefore, as in the prior art, the nodes m and n of Smn that provide the maximum saving value
Is a merger candidate. In step S6, a track t having the smallest saving value is searched for in the combination of the node m and the node n. As a result, as shown in FIG. 5, each truck takes charge of a delivery node group near each departure point and return point, and each lump obtains a solution that is a proper combination as in the conventional saving method. Can be

Next, in step S7, it is determined whether or not the respective constraint conditions are satisfied when the merging candidate node is merged with the track. For example, it is checked whether the maximum loading capacity of the truck is exceeded due to the merger.
If the constraints are satisfied, merging is actually performed in step S8. Otherwise, do not merge. Regardless of whether or not the merging is performed, all the saving values of the merging candidates are reset to 0 in step S9, so that the merging candidates will not be merged from the next time. Then, returning to step S4,
The processing after the maximum saving value search is repeated.

In step S5, it is determined whether Smn = 0. That is, all the merging candidates are inspected, and when all the saving values become 0, the vehicle allocation plan is ended, and the result is output in step S10.

If both the nodei and the nodej are already assigned to the tracks, the tracks are merged into the track with the better condition in the case where each track is visited. If the quickness of the return time is added to the condition, the truck closer to the return destination will take charge of the delivery route. If the conditions are exactly the same, a truck with a higher priority (for example, a 4-ton car priority) is used.

The present invention can be applied to all cases where only the departure point is different, only when the return point is different, and when both are different. It should be noted that the present invention can be naturally applied to the conventional case where the departure place and the return place are the same.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of an embodiment of the present invention.

FIG. 2 is a diagram showing an example of a delivery route.

FIG. 3 is a diagram for explaining a method of calculating a saving value.

FIG. 4 is a conceptual diagram for explaining inconsistencies in a delivery plan.

FIG. 5 is a conceptual diagram showing an example of a delivery plan according to the present invention.

FIG. 6 is a flowchart illustrating an embodiment of a calculation method.

FIG. 7 is an example of a saving matrix.

FIG. 8 is a view for explaining a method of searching for a merger candidate from a saving matrix.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Customer database memory 2 ... Distance database calculation part 3 ... Distance database memory 4 ... Track database memory 5 ... Saving value calculation part 6 ... Saving matrix memory 7 ... Search part 8 ... Merging judgment part 10 ... Delivery center 11 ... Delivery planning device

Claims (1)

[Claims] 1. A delivery plan support device for planning a delivery plan for delivering to a plurality of delivery destinations using a plurality of trucks having different departure and return locations, comprising: a customer database storing delivery destination data; A track database storing attributes for each truck used for delivery; a distance database calculating means for calculating a distance database corresponding to a moving cost between customers based on the customer database; and a track database and the distance database. A saving matrix calculating means for calculating a matrix of saving values between delivery destinations for each departure place and return place; and a primary search for a combination of delivery destinations having the largest saving value from the saving matrix, Responsible for starting point and returning point to minimize saving value in combination Search means for performing a secondary search for a truck to be searched, and determining whether or not a constraint condition is satisfied when the routes to which the secondary search destinations belong are merged. And a merging determination unit for merging the routes to which the destinations belong and for setting the element of the saving matrix corresponding to the combination of the primary searched delivery destinations to 0.