JP2021111237A - Route calculation program, route optimization system, and route calculation method - Google Patents

Abstract

To provide a route calculation program, a route optimization system, and a route calculation method, that can solve a large-scale vehicle routing problem in a short time, and can present a practical route.SOLUTION: In work for vehicles 12 departing a depot 11 to collect and carry loads of a plurality of clients C into a factory 13, a route in which the vehicles 12 depart the depot 11, go round all the clients C, and return to the depot 11, and that is a route in which the clients are assigned to each of the vehicles 12 is calculated by using a method for a vehicle routing problem (VRP). The obtained route is divided based on restrain conditions of load capacities and an estimated accumulation value of load amounts, a node for the factory 13 is added to each of the divided routes, and then each of the divided routes is optimized by using a method for a traveling salesman problem (TSP).SELECTED DRAWING: Figure 1

Description

The present invention relates to a route calculation program, a route optimization system, and a route calculation method related to the transportation route problem.

There are a transportation route problem or a vehicle route problem (VRP) and a traveling salesman problem (TSP) as a problem of calculating a route following a plurality of customers distributed in a specific area by a computer under various constraints. The calculation of the route by the computer is a calculation for searching, for example, the route having the shortest total route length from the innumerable combinations of the order of tracing customers, and the calculation performed for each combination of the order of tracing customers is repeated. Is. When these problems are large-scale problems with a large number of customers, it is very difficult to find an exact solution in a realistic time (called NP-hard), and various approximation methods and empirical rules are proposed. Has been done.

A material handling route problem (VRP) is, for example, when multiple vehicles depart from a depot, which is a depot, collect or deliver packages to multiple customers distributed in a specific area, and return to the depot. It is a problem of finding a route that minimizes the cost. The cost (objective function) is, for example, a travel cost such as total time, total distance, and total cost required for travel, and a material cost such as the required number of vehicles. Constraints include, for example, restrictions on vehicles such as load capacity restrictions, restrictions on routes such as route width (roadway width) with respect to vehicle width and one-way traffic, and time constraints such as returning to the depot within a predetermined time. ..

The traveling salesman problem (TSP) is, for example, the problem of finding a route that minimizes the travel cost when one salesman departs from the starting point, visits each customer once, and returns to the starting point. This traveling salesman problem corresponds to the case where there is only one vehicle and there are no various restrictions in the transportation route problem.

As an application example of the transportation route problem, a delivery plan generation method for solving the route problem including the work of putting the delivery package in the delivery container, delivering it to the customer, and returning the empty container to the manufacturer after the delivery of the package is completed has been proposed. (See, for example, Patent Document 1).

In addition, in order to solve a large-scale transportation route problem involving multiple depots under load capacity constraints, each customer is subgrouped by associating each customer with the most recent depot and dividing the customer into depots. A method has been proposed. This method also provides that the dividing lines radiating from each depot ensure that each customer is included in one of the subgroups associated with the depot and that each subgroup is the number of customers that meets the load constraint. Split each customer. Each small group of customers at this stage is solved by the traveling salesman problem method (see, for example, Non-Patent Document 1).

Japanese Unexamined Patent Publication No. 2019-28992

Masashi Ito, Shinya Watanabe, Kazunori Sakakibara, "Proposal of a new search framework based on partial problematization and adaptive problem integration for large-scale Vehicle Routing Problem" Journal of the Evolutionary Computation Society Vol. 6, No. 3 (2015)

However, although the solution of the transport path problem as shown in Patent Document 1 described above is shown as an application example, large-scale correspondence is not mentioned. Further, although the solution of the transportation route problem as shown in Non-Patent Document 1 described above can cope with a large scale, its effectiveness is limited by the actual spatial mutual arrangement between the depot and each customer. Therefore, it is considered to be limited.

By the way, in the field related to the large-scale transportation route problem, for example, there is a field in which household waste in daily life is stored in a packer vehicle and carried into a waste incinerator. In this field, for example, multiple packer vehicles (vehicles with load capacity restrictions) departing from a cleaning office (depot) collect and deliver waste from each household (customer) and bring it to a waste incinerator (factory). Is repeated, and finally, the work of returning to the depot is performed. The amount of waste discharged varies depending on the season, the presence of holidays and events, the day of the week, etc., and the waste collection work must be completed within a specified time. It is necessary to solve and calculate the optimum route.

In this waste collection transportation route problem, the factory plays a role in place of the depot in Non-Patent Document 1 described above, and the cleaning office is set as the depot which is the starting point and the ending point of the route. There are many large-scale transportation route problems that fall into the category of combinations of such depots, customers, and factories, not limited to garbage collection, and practical solutions to such large-scale transportation route problems are required. ..

The present invention solves the above problems, and is a route calculation program, a route optimization system, and a route calculation method that can easily solve a large-scale transportation route problem in a short time and present a practical route. The purpose is to provide.

In order to achieve the above object, in the route calculation program of the present invention, one or a plurality of vehicles departing from a depot which is a vehicle base collects the luggage of a plurality of customers and makes it to a factory designated for the vehicle. A route calculation program that calculates a route for carrying in, and is a distance matrix for the distance between each of the nodes based on the position information of the nodes corresponding to the locations of the depot, the customer, and the factory. The vehicle uses the distance matrix generator, the distance matrix, and the number of working groups composed of the vehicle and the worker to terminate the depot at a predetermined stage without obtaining the best solution. The customer allocation unit that solves the transportation route problem (VRP) for the route that departs from, collects luggage around the customer, and then returns to the depot, and the customer allocation unit allocates the customer for each working group. A route division unit that calculates a division route that divides the route based on the load capacity constraint information of the vehicle and the baggage information of each customer, and the work group from the depot or the factory for each division route. The route to the factory via the customer included in the division route is made to function as a route determination unit for solving the traveling salesman problem (TSP).

In this route calculation program, the customer allocation unit uses the transportation route problem (VRP) under a time constraint that the travel time from the vehicle leaving the depot to returning to the depot is within a predetermined time. You may solve.

In this route calculation program, the predetermined time is calculated for each vehicle in consideration of the travel time from the customer who has the longest route from the designated factory to the factory and the load capacity constraint information of the vehicle. The factory delivery time calculated based on the number of deliveries to the factory may be the time obtained by subtracting the preset working time.

In this route calculation program, the customer allocation unit has a time constraint that the variance of the set having the travel time from the departure of each vehicle in the plurality of vehicles to the return to the depot as an element is equal to or less than a predetermined value. The transport path problem (VRP) may be solved under certain conditions.

In this route calculation program, the route dividing unit may not be divided between the customers whose order is continuous within a predetermined distance.

Further, the route optimization system of the present invention includes a computer for executing any of the above route calculation programs, and one or more vehicles departing from the depot, which is a vehicle base, collects the luggage of a plurality of customers. The vehicle is characterized in that a route for carrying it into a designated factory is obtained.

Further, in the route calculation method of the present invention, one or a plurality of vehicles departing from a depot, which is a vehicle base, collect the luggage of a plurality of customers by a programmed computer and send the vehicle to a factory designated for the vehicle. It is a route calculation method for calculating a route for carrying in, and generates a distance matrix for the distance between each of the nodes based on the position information of the nodes corresponding to the positions of the depot, the customer, and the factory. The vehicle departs from the depot on the premise that the distance matrix generation step, the distance matrix, and the number of working groups composed of the vehicle and the worker are used to terminate the process at a predetermined stage without obtaining the best solution. For the route that returns to the depot after collecting luggage around the customer, the customer allocation process that solves the transportation route problem (VRP) and the route that the customer is assigned by the customer allocation process for each work group. , A route division step of calculating a division route divided based on the load capacity constraint information of the vehicle and the baggage information of each customer, and the division by the working group from the depot or the factory for each division route. The route to the factory via the customer included in the route is characterized by including a route determination step of solving a traveling salesman problem (TSP).

According to the route calculation program, the route optimization system, and the route calculation method of the present invention, the transportation route problem is solved on the premise that the route is terminated at a predetermined stage without obtaining the best solution, and the obtained route is used as the load capacity constraint information and the load. Since it is divided based on information and the divided route is solved as a traveling salesman problem, a practical route can be presented in a shorter time.

(A) is a conceptual diagram of the application target of the route calculation method adopted in the route calculation program according to the embodiment of the present invention, and (b) is the route to which the customer of (a) is assigned by the customer allocation process of the same method. (C) is an explanatory diagram of the route to the factory calculated by dividing the route of (b). Flow chart of the same route calculation method. (A) to (d) are conceptual diagrams for sequentially explaining the processes in the same route calculation method. (A) is a conceptual diagram showing the positional relationship between the road node and the customer and the customer node and the road distance matrix, and (b) is a conceptual diagram showing the customer node and the customer distance matrix incorporated between the road nodes. A block configuration diagram of a program and data for executing the same route calculation method. A block configuration diagram of a computer that executes the same route calculation method. The conceptual diagram which shows the example in the case where a plurality of depots can correspond to the customer to which the same route calculation method is applied. The figure which shows the example of the condition input screen composition of the computer which executes the same route calculation method. The figure which shows the example of the result output screen composition of the computer which executes the same route calculation method. The figure which shows the operation example of the same result output screen. The figure of the computer screen which shows the composition example of the input data about a customer in the same route calculation method. The figure which developed the computer screen which shows the configuration example of the input data. (A) is a diagram for explaining the concept of S work in the same route calculation method, and (b) is a diagram for explaining the concept of the W work. The figure which shows the route and the customer node on the map and explains about the road network in the same route calculation method. (A) is an enlarged view of the M1 portion of FIG. 14, and (b) is an enlarged view of the M2 portion of the same. It is an enlarged view of the M3 part of FIG. 14, and is the figure which shows and explains the route of a large vehicle and a customer node on a map. (A) is a diagram for explaining the movement between customer nodes when the route has no direction constraint in the same route calculation method, and (b) is a diagram for explaining the movement between customer nodes when the route has a direction constraint. (A) is a diagram for explaining the setting of the customer node when there are no restrictions on the approach to the left and right of the road in the same route calculation method and both sides can be collected, and (b) is a diagram for explaining the movement via the customer node. .. (A) is a diagram for explaining the setting of the customer node when the approach to the left and right of the road is restricted in the same route calculation method and both sides cannot be collected. The figure explaining the movement which goes through by the origin. FIG. 18 (b) is a diagram illustrating an image of a road that is a target of movement and collection on both sides. FIG. 19 (b) is a diagram illustrating an image of a road that is a target of movement and one-sided collection. The figure explaining the addition of the road node to the map data in the same route calculation method. The figure which added the road node and the route to FIG. 22. The figure which shows the example of the position coordinate of each of the road nodes at both ends of a road and the road nodes at both ends of a road added to the road in digital map data. The figure explaining the addition of the walking route to the map data in the same route calculation method. The figure which added the stop position and the walking route to FIG. The figure which shows the example of the position coordinate of each of the customer position to be walked, the road node at both ends of the road, and the stop position added to the road in the digital map data. The figure which shows the example of the delivery destination factory registration screen composition in the same route calculation method. The figure which shows the example of the screen composition of the factory distribution designation in the same route calculation method. (A) and (b) are diagrams showing an example of a screen configuration for explaining the process of calculating the number of units in the same route calculation method.

Hereinafter, a route calculation program, a route optimization system, and a route calculation method according to an embodiment of the present invention will be described with reference to the drawings.

(Outline of route calculation method)
First, an outline of the route calculation method adopted in the route calculation program of the present invention will be described with reference to FIG. In this method, as shown in FIG. 1A, one or more (two in this example) vehicles 12 depart from the depot 11 which is a vehicle base for the luggage of a plurality of customers c distributed in a specific area 10. This is a method of automatically calculating the movement route of the vehicle 12 in the work of collecting and carrying it into the factory 13. In the present embodiment, the customer c is a garbage collection point, the luggage is household waste, the vehicle 12 is a packer vehicle that compresses and transports the waste, the factory 13 is a waste incinerator, and the depot 11 manages and operates the waste collection and transportation business. I will explain with the setting that it is an office to do. Area 10 is, for example, in a town where the depot 11 is in charge of business. These settings are exemplary and the present invention is not limited to these settings.

As shown in FIG. 1 (b), a route is calculated in which the two vehicles 12 each depart from the depot 11 and return to the depot 11 around all the customers c. In the example of this figure, the route in which one vehicle 12 goes around the customers c11 to c17 in order and the other vehicle 12 goes around the customers c21 to c27 in order is calculated. A customer is assigned to each vehicle 12 by calculating a series of routes from the depot 11 to the depot 11 via the customer c.

Although not shown in this figure, there is a road leading to each customer c, and road nodes that define the road are set for each important point of the road, and each road node can be selected and traced. You can drop in at customer c (customer node). In this specification, a line connecting adjacent nodes is called a route, and a set of continuous routes is also called a route. In addition, a commercially available digital road map database can be used to calculate the route, in which the road is a route connecting the road node and each road node (called a link, a road link, a road, etc.). It is defined. Therefore, in the present application, a route and a road are used without particular distinction.

A method of the transport route problem (VRP) can be used for the problem of finding such a route. If the best solution is not desired, the route assigned to each vehicle 12 by the customer can be easily obtained in a short time.

The weight of the luggage of each customer c can be set to an expected value by a preliminary survey. Therefore, along the route from the depot to the depot, the integrated value of the amount of luggage for each customer on the route can be obtained. Vehicles generally have load capacity restrictions. Therefore, when the integrated value of the luggage amount approaches the constraint value, the integrated value is returned to zero and the integrated value is restarted from the next customer so that the load capacity does not exceed the constraint value. For example, for one vehicle 12, the integrated value is set to zero at the stage P1 via the customer c13, and the integrated value starting from the customer c14 is within the constraint after passing through the customer c17 before returning to the depot. Similarly, for the other vehicle 12, the integrated value is set to zero only once at the stage P2 via the customer c22.

In such steps P1 and P2, each one-round route is divided, and each of the customers c13, c14 and customers c22, c23 located on both sides of the division point is associated with the factory 13. Further, in order to give each vehicle 12 a return route, the factory 13 is associated with the depot 11. The route obtained by the division is (depot, customer, factory) or (factory, customer, factory, depot) route, and generally (factory, customer, factory) or (depot, customer, factory, depot). ) Or (depot, factory, depot to drop depots, customers, workers).

Each route divided in this way is not necessarily an optimized route from the viewpoint of shortening the travel time or shortening the distance, and there is room for optimization. The search for such an optimized route is performed, for example, by optimizing the order of the customer c. The order optimization is a search for newly determining which road node (that is, a route) is to be taken for a road that does not appear in the figure.

Since the load capacity of each divided route is set to be equal to or less than the constraint value, the patrol route can be optimized without considering the constraint conditions. Such optimization of the patrol route can be easily performed in a short time for a small number of customer nodes by using the method of the traveling salesman problem (TSP). FIG. 1 (c) shows the optimum route calculated by using the above method.

(Detailed explanation of route calculation method)
Next, the route calculation method of the present invention will be described in detail with reference to FIGS. 2, 3, and 4. As the route calculation method, a plurality of types of vehicles 12 having different vehicle widths and load capacities are designated, taking into consideration the road distance, road width, traffic restrictions, restrictions on the load capacity of the vehicle 12, the time required for collection, and the like. It is configured to transport (collect) the customer's demand (baggage) and find the route with the minimum cost (for example, the shortest distance) to bring it to the designated factory 13. As shown in FIG. 2, this method includes a distance matrix generation step S1, a customer allocation step S2, a route division step S3, a factory insertion step S4, and a route determination step S5.

(Distance matrix generation step S1)
In this step S1, a distance matrix for the distance between each of the nodes is generated based on the position information of the nodes corresponding to the positions of the depot 11, the customer c, and the factory 13. In this step S1, not only the generation of the distance matrix, but also the setting of various constraint conditions necessary for the route calculation, the initial setting (for example, the setting of the expected value of the amount of luggage for each customer c) and the like are performed. Further, the term of distance matrix is used in a broad sense including a matrix or an array of data necessary for route calculation, and can also include, for example, a time matrix, a cost matrix, and the like.

The distance matrix representing the distance itself includes a road distance matrix representing the mutual distance between the road nodes set at each key point of the road on which the vehicle 12 travels, and a road distance matrix representing the mutual distance between the road nodes and the road closest to the customer c corresponding to the customer c. There is a customer distance matrix that represents the mutual distance between the customer nodes set in.

When this method is used as a route optimization system at a baggage collection management site (depot, etc.) and a route is calculated, various customer data are used as inputs for the optimization calculation. These data are used as constraints for route calculation or for improving service to customers. Such customer data is collected in advance and registered in the route optimization system in this step S1.

For each customer, the customer data includes, for example, the customer's location, address, site photo, customer type (characteristics such as whether the site collects luggage on foot or a large site), and the collection direction (two-way road). In some cases, whether or not the customer can drop in in either direction), the type of vehicle to be assigned to that customer, the type of luggage, the collection day for each type of luggage, and each day of the week The amount of luggage (weight kg), etc. The customer's position and address can be automatically acquired by taking a picture of the site by using the GPS function of the camera or the like.

Here, the generation of the road distance matrix and the customer distance matrix will be described. As shown in FIG. 4A, a case where three road nodes nr1, nr2, nr3 generate two roads in this order and the nearest customers c1 and c2 are located adjacent to each road. Consider (customer nodes nc1 and nc2 corresponding to the customer's position are set on the road between the nearest road nodes). Rows and columns are defined by each road node, and a road distance matrix is set with the distance between each node as the value of the matrix element. The road distance matrix is a symmetric matrix in which the value of the diagonal element is 0.

The road node is read directly from the road information of commercially available digital map data, for example, and the distance between the nodes is calculated. The road distance matrix also includes the distances for the nodes of factory 13 and depot 11. Considering the case where a plurality of types of vehicles 12 are used for collecting luggage, for example, a road network A representing all roads including narrow roads that large vehicles cannot pass through and large vehicles can pass through. Two patterns of road distance matrix with the road network B having a width may be generated. These road distance matrices may be generated as a large distance matrix combined into one.

For the customer, as shown in FIG. 4B, the customer's position is moved on the road using the road information of the map data and reconstructed as the customer nodes nc1 and nc2 on the road, and the distance between the customer nodes. Is set as the value of the matrix element, and the customer distance matrix is set. The distance between customer nodes is, for example, the shortest distance obtained by accumulating the distances between adjacent nodes, starting from one customer node, following a road node, and reaching another customer node. This matrix is also a symmetric matrix with zero diagonal elements. The process of reconstructing the customer's position on the road may be performed manually by an operator, fully automated by a computer, or partially fully automated.

(Customer allocation process S2)
In this step S2, the vehicle 12 departs from the depot 11 and the customer uses the distance matrix and the number of working groups composed of the vehicle 12 and the worker on the premise that the vehicle 12 is terminated at a predetermined stage without obtaining the best solution. The transport route problem (VRP) is solved to calculate the route back to the depot 11 after collecting the luggage around c. The calculated route is a series of routes that are determined for each work group and leave the depot and return to the depot. If the number of work groups is 2, two series of routes are calculated. The number of working groups is usually the same as the number of vehicles.

Here, the term "working group" (also simply referred to as "group") is, for example, a group of two workers (excluding the driver) who work with one vehicle 12 and depot. Represents a concept linked to a series of routes out and back to the depot. There is a one-to-one correspondence between working groups for this series of routes. However, it may be permissible for the vehicle to change along the way for this series of routes. For example, two vehicles may be involved in this series of routes (see S work and W work in FIG. 13). Under the unit of this working group (in other words, for a set of customers belonging to a series of routes leaving the depot and returning to the depot), all the packages of a large number of customers are brought to the factory 13 in order to carry them to the factory 13. It is expected to be carried in multiple times.

In addition, the term "efficiency" will be introduced as a concept that divides the work performed by the power group temporally and spatially. Efficiency refers to a group of operations from when one vehicle 12 is empty to when it collects luggage and brings it to the factory. Efficiency refers to each split route that is the subject of the Patrol Salaryman Problem (TSP), which is obtained by splitting a series of routes for the vehicle to leave the depot and return to the depot. The term "efficiency" has a one-to-one correspondence with the carry-in work to the factory 13. Customers included in the area covered by Depot 11 are included in any efficiency. The work that Depot 11 is in charge of is a set of efficiencies.

Individual efficiencies are all identified by which vehicle, which worker (ie, which power group) performed the factory delivery work, and how many times. The sequence of routes (more precisely, the route after splitting and inserting the factory) that one vehicle obtains in the customer allocation process S2 leaves the depot and returns to the depot (that is, in the case of normal S work) Defines a set of efficiencies characterized by its common vehicle. In the set, the first efficiency, the second efficiency, and the like are ordered in the order of the route.

When calculating the route, it is possible to impose constraints such as the day of the week specified by the user of this method, the type of vehicle 12, the number of vehicles, the number of times of efficiency, the factory 13 of the delivery destination for each efficiency, and the work constraint time. Under the constraint condition, the transportation route problem (VRP) based on the total number of vehicles 12 is solved. When a plurality of vehicles are used, constraints may be imposed on the types (vehicle types) of the vehicles 12 that pass through each customer. The route of each vehicle 12 is searched under various constraints, for example, based on the distance matrix generated in step S1 and considering the vehicle width of each vehicle 12. Various methods have been proposed for solving the transport route problem (VRP), for example, to achieve either the shortest total mileage, the shorter route calculation time, or both. Has been done.

When calculating a route under a time constraint, for example, a transportation route problem (a transportation route problem) under a time constraint that the travel time from the vehicle leaving the depot to returning to the depot is within a predetermined time. VRP) is solved. This can prevent large variations in travel distance and working time between vehicles when a plurality of vehicles are involved.

Further, for the predetermined time under the above time constraint condition, for example, for the customer who has the longest distance to each designated destination factory 13 in the designated area, the travel time from that customer to the factory 13 is calculated. , Considering the number of times of efficiency, from the time frame set by the constraint condition, the total time of delivery to the factory 13 (total time of round trip excluding the time of delivery to the last factory and the time of delivery of the last one way) The time is calculated by subtracting the sum).

It also solves the shortest path problem (VRP) under time constraints where the variance of the set, which is based on the travel time of individual vehicles from the depot to the depot, is less than or equal to a predetermined value. You may do so. This can prevent large variations in travel distance, working time, and efficiency frequency between vehicles.

The main purpose of this step S2 is to allocate a customer c in charge of collection for each vehicle 12. In the calculation of the route for each vehicle 12 in this method, various methods based on the method of solving the transport route problem (VRP) or an application or modification of the method may be used according to this main purpose. .. The number of vehicles 12 is not limited to a plurality of vehicles and may be one, which is different from the general shortest path problem (VRP) for a plurality of vehicles.

FIG. 3A shows how each route R of two vehicles 12 departing from the depot 11 and returning to the depot 11 after passing through the customer is calculated by using the method of the transportation route problem (VRP). ..

(Route division step S3)
In this step S3, for each work group, the route to which the customer c is assigned by the customer allocation step S2 is divided based on the load capacity constraint information of the vehicle 12 and the baggage information of each customer c, and "efficiency" is defined. Will be done. The load capacity constraint information of the vehicle 12 and the luggage information of each customer c are registered as data used in this method in the distance matrix generation step S1. As a result of this division, a division route is calculated in which the number of customers is less than the number of customers on the route before the division. Each of the split routes corresponds to an individual efficiency. The series of routes from depot to depot is divided according to efficiency.

FIG. 3B shows how the route R is divided at each points P1, ..., P4 for one of the two vehicles 12, and the efficiency (divided route) of the first to fifth times is calculated. ..

(Factory insertion process S4)
In this step S4, the node of the factory 13 designated for each vehicle 12 is inserted into each division route. That is, preparations are made for solving the traveling salesman problem (TSP) in the next route determination step S5. This step may be included in the route dividing step S3 or may be included in the route determining step S5.

FIG. 3C shows how the path R inserts the node of the factory 13 at each point P1, ..., P4. In the present embodiment, the final efficiency (fifth efficiency) of delivery to the factory is set so that the worker stops at the depot and only the vehicle and the driver go to the factory. The method of carrying in the factory at the final efficiency is not limited to the example of the present embodiment, and may be arbitrarily set according to, for example, the location conditions of the factory and the depot.

(Route determination step S5)
In this step S5, the traveling salesman problem (TSP) is solved for the route from the depot 11 or the factory 13 to the factory 13 via the customer c included in the division route for each division route. In this step S5, for each divided route, the road node, the customer node, the factory 13 node, and the depot node if the depot 11 node is included in the route are included in these nodes. The set will be re-solved as a traveling salesman problem. By this resolving, the total route length is shortened and the route is optimized.

FIG. 3D shows how the route optimization is performed by resolving the traveling salesman problem for each efficiency (divided route).

As described above, this method is a method of calculating the optimum route in which the mileage of the vehicle 12 is minimized as much as possible while suppressing the calculation time to a short time. A better solution should be obtained if the calculation of the VRP solution method is continued for a long time, but in consideration of usability, the calculation time is cut off in a short time, and then TSP is executed in a very short time. , The route optimization is streamlined. As described above, the calculation process of the TSP can be completed in a short time because the calculation process by the TSP is performed for a small number of customer nodes in each division path.

The number of sets or the number of vehicles used when calculating the optimization route may be arbitrarily specified by the user. For that purpose, the number of sets input by the user should be registered in the setting file for VRP execution together with the information of each vehicle 12, and the destination factory for each efficiency should be set for the number of sets. good.

The setting file for VRP execution is, for example, a CSV file (text data file) in which conditions including options are recorded for setting conditions when solving the transportation route problem (VRP) in the customer allocation process S2. be. In addition to being referred to when solving the transport route problem (VRP), the setting file for executing VRP is used to set conditions when executing the route optimization calculation that solves the traveling salesman problem (TSP) in the route determination step S5. It is also used for such purposes, and its contents are referred to. Therefore, the setting file for VRP execution is also the setting file for TSP execution. You may think of VRP in a broad sense and include TSP in the concept of VRP. In the following, the term "configuration file for VRP execution" is used in this sense.

Further, the number of times of efficiency when calculating the optimization route in the route determination step S5 may be arbitrarily specified by the user. For that purpose, the designated delivery destination may be set in the setting file for VRP execution for the number of times of efficiency input by the user. In the factory insertion step S4, the delivery destination for each group and each efficiency is set.

The vehicle that departs from the point called Depot 11 returns to Depot 11 via the customer. Here, the office designated as the depot 11 may be arbitrarily designated by the user. Therefore, a CSV file in which information (latitude and longitude, etc.) of the office that is expected to be specified is registered is saved in the input data storage directory referred to by the program that executes this method. By registering the office specified by the user in the setting file for VRP execution, the registered office can be set as the depot 11 and the VRP can be solved.

If it is expected that a plurality of depots 11 will be switched and used, prepare a plurality of similar input data storage directories and specify the directory to be used as an option when executing the route calculation program. You can change it. It should be noted that returning to the office designated as Depot 11 is not for each efficiency, but after the work of the vehicle or group is completed.

(Route calculation program and route optimization system)
Next, with reference to FIGS. 5 and 6, a route calculation program, which is a program for executing the route calculation method, data and a program used for route calculation and result display, and further, this route calculation method is executed. A computer system, that is, a route optimization system will be described.

As shown in FIG. 5, the route calculation program 1 according to the present invention includes a distance matrix generation program 3 and a route calculation program 4. The route calculation program 4 includes a customer allocation program, a route division program, a factory insertion program, and a route determination program.

The distance matrix generation program 3, the customer allocation program, the route division program, the factory insertion program, and the route determination program constituting the route calculation program 1 are the distance matrix generation process S1 and the customer allocation process S2 in the route calculation method, respectively. It is used to execute the route dividing step S3, the factory insertion step S4, and the route determining step S5. The route calculation program 1 may further include a result creation program 5 for outputting the result, and by displaying the display data created by the result creation program 5 as the screen display 6, the route can be interactively routed. The calculation can proceed and the optimization can be improved.

The route calculation program 1 executes route generation by referring to the position information, the amount of luggage, and other information of each customer recorded in the customer database DBc and the road information recorded in the map database DBm. Various constraints used for route calculation are also recorded in these databases DBc and DBm.

The distance matrix generation program 3 uses the data of the databases DBc and DBm to generate the distance matrix 3a, the time matrix 3b, the customer CSV file 3c, the road data file 3d, and the graph 3e. The route calculation program 4 calculates an optimized route by referring to the distance matrix 3a, the time matrix 3b, and the customer CSV file 3c. The calculation result by the route calculation program 4 is saved as the route information 4a, the execution time 4b, and the setting 4c.

The result creation program 5 creates, for example, four types of files by referring to the data of the route information 4a and the execution time 4b, which are the calculation results by the route calculation program 4, and the data generated by the distance matrix generation program 3. The files are a cost list CSV file 5a, a calculation time graph image file 5b, a JSON format text file for each route JSON file 5c, and a route-specific shape file 5d. The result creation program 5 creates the data of the screen display 6 by using the data of these files and the data directly obtained from the route calculation program 4.

As shown in FIG. 6, the route optimization system 2 according to the present invention includes a CPU (central processing unit) 20, a hard disk 21, a communication unit 22, which is various peripheral devices, and a random access memory (RAM) 23. It is composed of a computer including an input operation unit 24 and a display 25. Although FIG. 6 shows an example in which the route optimization system 2 is configured by one computer, the route optimization system 2 may be configured by a plurality of computers.

The CPU 20 includes a distance matrix generation unit 2a, a customer allocation unit 2b, a route division unit 2c, and a route determination unit 2d. Each of these parts 2a, 2b, 2c, and 2d has a distance matrix generation process (S1), a customer allocation process (S2), a route division process (S3) and a factory insertion process (S4), and a route determination process (S5), respectively. To execute.

On the hard disk 21, a computer 2 is operated to allow one or more vehicles departing from the depot, which is a vehicle base, to collect the luggage of a plurality of customers and bring it to a factory designated for the vehicle. The route calculation program 1 for calculating the above is stored. More specifically, the route calculation program 1 uses the CPU 20 of the computer 2 as a distance matrix generation unit 2a for executing the distance matrix generation step (S1), a customer allocation unit 2b for executing the customer allocation process (S2), and a route division. This is a program that functions as a route dividing unit 2c that executes the process (S3) and the factory insertion process (S4), and a route determining unit 2d that executes the route determining process (S5).

As described above, according to the route calculation program, the route optimization system 2, and the route calculation method according to the embodiment of the present invention, the transportation route problem is solved on the premise that the route calculation method is terminated at a predetermined stage without obtaining the best solution. Since the divided route is divided based on the load capacity constraint information and the luggage information and the divided route is solved as the traveling salesman problem, a practical route can be presented in a shorter time.

In the following, a modification of the above-described embodiment and a specific example using a computer screen display example will be described mainly with the route optimization system 2 in mind.

(If there are multiple depots)
As shown in FIG. 7, the application target of the route calculation method may include a plurality of (two in this example) depots 11a and 11b. In this example, the customer c is further separated into the jurisdiction areas 10a and 10b of the depots 11a and 11b. In such a case, since the depots 11a and 11b can deal with each other independently for each of the areas 10a and 10b, they can be solved as separate route calculation problems and the route optimization can be performed.

Unlike the example of FIG. 7, for example, when the two depots 11a and 11b jointly respond to the customer c in the area where the areas 10a and 10b are merged (referred to as the area 10), or when the area 10 is divided into areas 10. Depending on the day of the week and the type of luggage to be collected, the two depots 11a and 11b (collectively referred to as depot 11) may take turns.

For such a case, the position information and node information of each depot 11a and 11b are registered in the data referred to in the distance matrix generation step S1 and automatically by the user or according to a preset schedule. , Depot 11 may be selectable.

(Condition input screen)
FIG. 8 shows a configuration example of the condition input screen 61 in the computer that executes the route calculation method. The screen 61 can be operated interactively using a pointer such as a normal mouse, and the user can input, select, browse, change the screen, and the like by operating the screen 61. The left pane (window frame) of the screen 61 is provided with menus for optimization execution 6a, result reference 6b, data registration 6c, and download 6d.

The screen 61 in this figure is a screen when the menu of the optimization execution 6a is selected. A plurality of condition fields 61a are provided in the center of the right pane of the screen 61, a delivery factory field 61b is provided below the condition field 61a, a time setting field 61c is provided on the right side, and a calculation start button 61d is provided on the lower right side. The distribution selection button 61e and the number calculation button 61f will be described later with reference to FIGS. 29 and 30, respectively.

(Setting of route calculation conditions)
The condition column 61a is a column for inputting or selecting a condition for performing route optimization, and includes columns for an office, that is, a depot, a type of luggage, a day of the week, a vehicle type, a number of groups, a work classification, and a number of times of efficiency. From the pull-down menu in each column, you can select a pre-registered option or enter a new one on the spot. The S work in the work category column will be described later (see FIG. 13).

(Specify the destination factory for each efficiency)
In the carry-in factory column 61b, the carry-in factory 13 can be specified for each efficiency. In the setting file for VRP execution, the information of the factory 13 of the expected delivery destination is registered in advance. The user calls this setting file from the carry-in factory column 61b, selects and sets the carry-in destination factory 13 to be specified for each efficiency time. As a result, it is possible to easily specify the delivery destination factory for each efficiency time.

(Specify the start time and end time of work and break time)
In the time setting field 61c, a time (time) constraint condition for the work of collecting from the customer and carrying in the factory to be executed is set together with the break time. The time constraint condition (that is, the constraint time h) is set by subtracting the break time and the time t related to the delivery to the factory 13 from the work start time and the work end time input by the user.

This restricted time h is an upper limit of the total working time required for all vehicles 12 to depart from the depot 11, complete the collection, and carry in the factory, and return to the depot 11. When the constraint time h is set, the route must be calculated so that all the work is completed within the upper limit of the total work time. This constraint time h is applied in both the customer allocation process S2 and the route determination process S5. Further, the constraint time h2 in the step S2 and the constraint time h5 in the step S5 may be made different from each other, and for example, the latter may be stricter (h2> h5).

The constraint time h is described in the configuration file for VRP execution. For example, when the work time input by the user is from 8:00 to 15:00 and the break is one hour, when it is not necessary to consider the delivery time to the factory 13 (t = 0), (15-8)-. According to 1 = 6, it is described in the setting file as the constraint time h = 6 (hours).

When there is delivery to the factory 13 (t ≠ 0), the time t is the time t from the position of the customer c farthest to the designated factory 13 in the designated area, where n is the efficiency input by the user. The maximum travel time T to the factory 13 is calculated by calculating the mileage up to and dividing by 80% of the maximum speed of the road traveling that distance, and then calculating by the following formula.
t = (n-1) × 2 × (maximum travel time T to factory 13).

In addition, the one-way travel time for the final delivery to the factory 13 is also calculated and further subtracted from the constraint time h, so that the constraint is set more severely. The break time may be input, for example, set in advance to 1 hour when the working time is 5 hours or more and 30 minutes when the working time is less than 5 hours, and the system may automatically select the break time.

(Collection time per customer)
By registering the collection time (for example, initial value 20 seconds) in the setting file for VRP execution, the collection time per customer can be reflected in the constraint time. The number of seconds involved in this collection is the work time for a worker to pick up a bag of garbage and put it in a packer vehicle when the baggage is household waste. This value can be registered from the user interface (interactive screen), and this value will be updated if it is changed. This collection time is reflected in the optimization calculation by the method of the transportation route problem (VRP) in the customer allocation process S2.

(Working time at the factory)
By registering the time required for work at the factory 13 (for example, the initial value of 5 minutes) in the setting file for VRP execution, the work time for each delivery can be reflected in the constraint time. .. The working time at this factory is, for example, the amount of entry clock and exit clock of the packer vehicle, and the time required for dump processing from the packer vehicle. This value can be registered from the user interface and will be updated if changed. This collection time is reflected in the optimization calculation by the method of the transportation route problem (VRP) in the customer allocation process S2.

(Result output screen)
9 and 10 show a configuration example of the result output screen 62 in a computer that executes the route calculation method, and an operation example of the screen. The screen 62 is a screen when the menu of the result reference 6b is selected. There is a condition display field 62a at the top of the center pane of the screen 62, a map display field 62b at the bottom, a specification summary field 62c of the calculation result in the upper right pane, and a customer address field 62d in the lower right pane. The condition display column 62a shows the conditions for route calculation, and the map display column 62b shows the route calculated on the map of the route calculation target area and the node of each customer.

(Display of route map)
The result creation program 5 (see FIG. 5) calculates the shortest route for each efficiency for each day of the week and the type of vehicle, and outputs various results as follows. That is, in addition to the JSON file 5c of vehicle group data and the shape file 5d that holds route and customer information, a CSV that holds the route length, elapsed time, total customer demand, number of customers passing through, etc. at each efficiency. Files etc. By designating the vehicle type and efficiency on the result output screen 62, based on these data, as shown in FIGS. 9 and 10, the screen shows the route, the position of the customer, etc., and the required efficiency. The time and total distance traveled are displayed.

(Time to stop by each customer, time required for each efficiency)
In the setting file for VRP execution, the average speed during the visit to the customer as the initial value, the time required for one collection as the initial work value, the time required for one delivery at the factory 13 as the factory work, As the initial value maximum speed, the maximum speed at the time of round trip to the factory 13 is registered. By adding the average degree of congestion of each road to the contents of these items, the time to stop by each customer c, the time required for each efficiency time, and the like are automatically calculated, and the user can refer to them. For example, as shown in FIG. 10, by clicking the customer node n in the map display field 62b, the stop-by time and the like for the customer are displayed and can be referred to.

The speed when moving to or from the factory 13 is calculated as, for example, 80% of the speed set as the initial maximum maximum speed. This ratio may be changeable. For example, an item called an average speed rate may be provided in the same file, and by changing this value, the speed of movement to and from the factory 13 may be arbitrarily set.

(Structure of customer input data)
FIG. 11 shows a screen 63 for displaying information about the customer, and FIG. 12 shows a structure of input data of the information. In the screen 63 of FIG. 11, there is a customer list in the left pane, vehicle information 63a such as a vehicle type in the upper pane, and an input data field 63b for each customer in the center pane.

(Assign a car model to each customer.)
The solution of the transportation route problem (VRP) in the customer allocation step S2 is executed by inputting the customer CSV file 3c (see FIG. 5) that holds the information of each customer generated in the distance matrix generation step S1. The file holds attributes such as latitude and longitude of each customer, and as one of the attributes, the type of vehicle 12 that passes through the competent office (depot) and each customer is also defined. When executing the route calculation calculation, the user selects the type of the office or the vehicle 12. The route optimization system 2 (FIG. 6) refers to the office and vehicle type specified by the user, and executes the customer allocation process S2 with the set of customers having the specified office and vehicle type as attributes as the calculation target. do. Therefore, when different customers facing the same road have vehicles of different vehicle types as attributes, they are assigned to different routes.

In the input data field 63b for each customer on the screen 63, in addition to the customer's address information, the types of luggage A, B, C (for example, types of combustible waste, bins, cans, recyclable waste, etc.) for each day, and the types thereof. There are input items such as expected weight. The estimated weight is integrated for each customer, and the integrated value is subject to comparison with the load capacity constraint condition.

Furthermore, as customer characteristics, there are selection check items such as a walking site, b large site, c condominium, d building key required, and e group collection. For example, a site where a packer car walks more than a predetermined distance to collect garbage, b large garbage other than general household garbage, c garbage in a large garbage box, d gate and a site where the garbage storage place is locked, e It corresponds to the collective collection of recyclable waste. In addition, there are two types of collection directions, two-sided collection and one-sided collection (see later, FIGS. 18 and 19).

(S work and W work)
As shown in FIG. 13 (a), for one vehicle 12, the worker m who rides on this vehicle and moves together forms a work group up to the end of the entire route of the vehicle, and performs the collection work. Is assumed as normal work, and this work is called S work.

As shown in FIG. 13 (b), with respect to the two vehicles 12, the worker m who transfers between these vehicles gets off from one vehicle before starting the movement to the factory 13, and the other. The work of changing to a vehicle and continuing the collection is called W work. This W work is efficient in that when the distance to the delivery destination factory 13 is long, the collection work can be continued without spending the time of the work vehicle m on the travel time to the factory. In the case of this W work, it is the worker m who continuously moves from the beginning to the end along a series of routes leaving the depot and returning to the depot, not the vehicle 12. In this respect, it differs from the normal transport path problem (VRP).

In the route calculation program, the route optimization system, and the route calculation method, the user can arbitrarily select any work category of these S work and W work. The work category selection setting can be performed, for example, through the screen 61 of FIG. Further, since the distribution of the work division by the S work and the W work can be arbitrarily set, the movement to the factory 13 is long, for example, for each division route in the series of routes leaving the depot and returning to the depot, that is, for each efficiency. In the case of distance, W work may be selectively set.

When solving the transportation route problem (VRP) in the customer allocation process S2, since this process S2 is the customer allocation process, the difference in work classification between the S work and the W work differs in time constraints. There is only. For example, in the case of W work, another vehicle 12 will be picked up immediately after collecting the customer's luggage, so when setting the time limit, it is not possible to subtract the time required for delivery to the factory 13 from the total time limit. do not.

(Loading rate of each model)
The loading rate can be registered in the setting file for VRP execution. The load factor is a ratio to be multiplied by the legal maximum allowable load capacity determined for the vehicle, and the amount obtained by the multiplication is the upper limit of the load capacity set for the vehicle 12 in the route optimization system. This value can be registered from the user interface and will be updated if changed.

(Loading capacity constraint)
Based on the load capacity constraint condition, it is necessary to carry the luggage into the factory 13 so that the load capacity of each vehicle 12 does not exceed the specified load capacity upper limit. When the route calculated by the method of the transport route problem (VRP) in the customer allocation process S2 is divided in the route division step S3, the route is divided so that the load capacity at each efficiency is equal. This is to prevent the load capacity from becoming extremely small, for example, in the efficiency of the final round.

In addition, the load capacity at each efficiency may slightly exceed the upper limit of the setting depending on the situation, but in order to eliminate this as much as possible, set a margin value of the load capacity (maximum load capacity x load rate ×). The amount of luggage (weight) obtained by subtracting this value from the number of times of efficiency is taken as the total load capacity at full efficiency. This process is performed for each vehicle.

(Equalize the load capacity of each efficiency)
At each efficiency time, the load capacity of each efficiency in one set is leveled so that there are no cases where the load capacity is extremely small. This leveling is the maximum efficiency by allocating a part of the amount of waste to other efficiencies for the efficiency with the highest demand in one set, for example, by allocating to the efficiencies before and after that efficiency. Try to reduce the amount of garbage that was the value. However, by setting a distance that is not divided into efficiency in the setting file for VRP execution, customers who are in a continuous order of stopping and whose positions are close to each other (within a predetermined distance) Regarding, adjust so that it is not divided into different efficiencies.

(Road width and vehicle width, traffic restrictions)
Restrictions on road width and vehicle width will be described with reference to FIGS. 14, 15, and 16. The route diagram M0 of FIG. 14 includes two patterns of routes, a road network A including a road that cannot be passed by a large vehicle as a movement route, and a road network B having a road having a width that a large vehicle can pass. I'm out. The route map M1 (FIG. 15 (a)) and the route map M2 (FIG. 15 (b)) correspond to the road network A, and FIG. 16 corresponds to the road network B.

The road network A is a mixed road network of all vehicle types, and small vehicles can naturally pass through it, but when large vehicles pass through, it is necessary to select a passable road. The road network B is a road network through which large vehicles can pass, and naturally small vehicles can pass through. In the distance matrix refining step S1, two patterns of road distance matrices corresponding to these two road networks A and B are generated, respectively.

(Judgment of passability)
The width W of the largest vehicle is assumed to be W = 249 cm (large dump truck). When the width D of the road is D = W + 40 = 289 cm, which is 40 cm larger than W, the road is considered to be passable even by a large vehicle. If the width D of the road is less than that value and D <289 cm, a large vehicle is set as an impassable road.

By registering a large vehicle type (that is, a vehicle type that can only pass on a road of D = 289 cm or more) in the setting file for VRP execution, it is possible to give the customer a large vehicle type as an attribute. Customers with such attributes can be dealt with by imposing a constraint condition that the route is calculated using only the road network B that even a large vehicle can pass through.

(Roads whose width information is not registered in the map data)
The map data referred to in the distance matrix generation step S1 is obtained based on commercially available map data. There are two types of road information files of commercially available map data, roads with and without road width records. Based on such commercially available map data, there are two patterns of road distances: a road network A that includes narrow roads that large vehicles cannot pass through, and a road network B that has a width that large vehicles can pass through. When generating a matrix, process as follows.

(Road with width record)
(1) In the case of a road with two or more lanes, the length of the road is set to both road networks A and B because there is a width that allows large vehicles to pass.
(2) If the road is a one-lane road and a specific width value is set, if the road width is 289 cm or more, the length of the road is set to both road networks A and B, and it is less than 289 cm. If there is, the length of the road is set to the road network A, and the road network B is set to infinity (impossible to pass).
(3) If the width of a one-lane road is "uninvestigated" or "unpaved", the length of the road is set to road network A, and road network B is infinite (impossible to pass). And set.

(Road without width record)
The length of the road is set in the road network A so that the road width is less than 289 cm, and infinity (passless) is set in the road network B.

(Neighboring customers stop by with the same vehicle and efficiency)
For customers located in neighboring areas, the same group will be in charge by grouping (clustering) customers in order to improve the efficiency of collection work and make it easier for workers to understand. For the same reason, customers located in neighboring areas will not be divided into efficiency. This setting is realized by registering the distance that is not divided into efficiency in the setting file for VRP execution. The route optimization system checks the distance between adjacent customers based on the route order, and if they are located within the distance set here, adjusts so that the adjacent customers do not separate from each other with different efficiencies.

(Handling of expressways and toll roads)
Generally, the expressway is not used while visiting customers, but some designated expressways are targeted as the delivery route to the factory 13. Commercially available map data is divided into road files according to the type of road, such as national roads, general roads, narrow roads, and highways. When reading the map data at the beginning of the process, all these road files are to be read, but at that time, in the case of the expressway file, only the expressway specified by referring to the name is read.

Toll roads registered in the map data as general roads are excluded if they are not used. There are toll roads among general roads, and toll roads specified not to be used should not be read as excluded when reading the road file of general roads.

(About one-way street)
As shown in FIG. 17A, when there is no direction constraint on the route, it is possible to move in both directions between the adjacent customer nodes nc1 and nc2. In the distance matrix generation step S1, when the road distance is first calculated, the information on whether or not the road is one-way is referred to from the road information in the map data. A general road can be treated as a directed graph with the same distance cost in both directions, but in the case of one-way traffic, it has a finite distance cost only in one direction, and infinite distance cost in multiple directions (impossible to pass). ).

As shown in FIG. 17B, in the case of a one-way route having a direction constraint on the route, in order to reach the node nc2 from the node nc1 between the adjacent customer nodes nc1 and nc2, the road nodes nr1 and nr4 , Nr3, nr2 is a detour movement. When calculating the road distance, such a route is calculated with reference to the distance cost.

(Double-sided collection and one-sided collection)
Two-sided collection and one-sided collection according to the characteristics of the road will be described with reference to FIGS. 18 to 21. As shown in FIGS. 18 (a) and 18 (b), and in the case of not a large street, when collecting (collecting) the luggage of a customer along the road, only the customer on the left side of the road with respect to the direction of travel of the vehicle. Instead, it is possible to collect (collect) luggage from customers on the right side of the road on both sides.

The customer nodes nc1, nc2, and nc3 of the customers c1, c2, and c3 located on both sides (left and right) of the road distinguish the route directions on the route between the road nodes nr1 and nr2. Can be set without. In other words, roads that can be recovered in both directions are not subject to one-way restrictions on the road.

As shown in FIGS. 19 (a) and 19 (b) and 21, in the case of a customer along the main street, the vehicle 12 is not stopped in the opposite lane for collection and the worker does not cross the road, but one side. It is a one-sided collection that collects one by one. In this case, the routes between the road nodes nr1 and nr2 are two directed routes that do not overlap each other. Here, for example, the left and right sides of the road are determined with reference to the direction from the road node nr1 to the road node nr2.

The customer nodes nc1 and nc3 of the customers c1 and c3 located on the left side of the road are set on the route from the road node nr1 to the road node nr2. Further, the customer node nc2 of the customer c2 located on the right side of the road is set on the route from the road node nr2 to the road node nr1. In this way, one-way collection is subject to the same restrictions as one-way roads.

In the distance matrix generation step S1, when inputting customer data, the information on whether or not both sides can be collected is referred to from the road information of the map data, and the collection method of the customer is set in the customer data registered in advance. back. In the route optimization system, a directed graph is adopted, the information of two-sided collection (boulevard) or one-sided collection (one-way or narrow road) is reflected in the route between road nodes, and that information is used for the optimum route calculation. And.

(Addition of road information and road nodes)
The addition and registration of roads that are not registered in the map data will be described with reference to FIGS. 22, 23, and 24. If there is a customer c who is the target of route creation along a road that is not registered in the map data, such as a narrow road or a new road, the information of that road is additionally registered in the system. Since narrow roads such as private roads are often not registered in commercially available road data, the user can visually determine the road to be added as road data from the map screen and add a new one.

As shown in FIG. 22, customers c1, c2, and c3 are located at positions away from the set route R, and although they are not registered as map data, there are roads accessible to each customer c. A case is assumed. In this case, on the screen, the road nodes nr1 and nr2 on the route R are pointed (clicked) to be selected, and the existing route R to which the newly added road is connected is specified. Next, two points, which are the start point p1 and the end point p2 of the new route to be added, are pointed and set.

Then, as shown in FIG. 23, a new route R1 adjacent to the customer c1 is added. Similarly, by adding new routes R2 and R3 to the route R1, routes that can approach the customers c2 and c3, respectively, can be obtained. By such an additional operation by the user, the route optimization system 2 registers the latitude and longitude of the designated existing route R, the start point node and the end point node of the newly added routes R1, R2, and R3 in the CSV file. FIG. 24 shows an example of data registered in the CSV file. This CSV file is a file referenced by the distance matrix generation program in the distance matrix generation step S1.

(Addition of walking route)
If it is not appropriate for the vehicle to move to the location of the customer node, it may go to pick up on foot to the location of the customer node, and the addition and registration of such walking routes are described in FIGS. 25, 26, 27. Will be described with reference to. For a customer facing a narrow road or the like, the customer's attributes can be specified so that the vehicle 12 does not move to the site by the vehicle 12, the vehicle 12 is stopped on a nearby road, and the customer stops on foot. In such a case, the route is optimized after registering the specified stop position in the system.

As shown in FIG. 25, a walking route is set for the customer c1 away from the route R. First, on the screen, an existing route R to stop by pointing (clicking) on the road nodes nr1 and nr2 on the route R is specified. Next, the stop position p1 is selected, and the position of the customer c1 is further selected.

Then, as shown in FIG. 26, the stop position P is added near the route R, and the walking route RW from the stop position P to the customer c1 is set. By such an operation by the user, the route optimization system 2 registers the latitude and longitude of the set stop position P and the position of the customer c1 to be picked up on foot in the CSV file. FIG. 27 shows an example of data registered in the CSV file. This CSV file is referred to by the distance matrix generation program in the distance matrix generation step S1. In this case, the customer node when generating the customer distance matrix is set to the stop position.

(Manual registration of road width)
For roads for which width information is not registered in the map data, the user can manually register the width on the screen. Since there are many roads for which the road width is not set in the commercially available road data, the width information is registered for such roads by describing the unique number of the road in the CSV file. In the distance matrix generation step S1, the distance matrix generation program generates two patterns of road networks A and B (see FIGS. 14 to 16 and their explanations) with reference to the information in this CSV file.

(Destination factory registration screen)
As shown in FIG. 28, a plurality of factories 13 and their respective acceptable quantities can be registered on the delivery destination factory registration screen 64 when the menu of data registration 6c is selected. By registering such an acceptable amount, when a plurality of factories 13 are designated as carry-in destinations, the factories 13 of the carry-in destinations of each vehicle 12 are appropriately distributed in consideration of each acceptable amount. Can be done.

(Designated for factory distribution)
As shown in FIG. 29, by clicking the distribution selection button 61e on the condition input screen 61, a list 61g of the factories to be carried in is displayed. If the total acceptable quantity of each factory 13 is greater than the total acceptable quantity of the customer's baggage, then when performing the optimization, first refer to the acceptable quantity of each factory 13 shown in FIG. The total amount of luggage is apportioned based on the carry-in amount ratio, and the system designates the factory 13 to which each luggage is carried. Then, a normal route optimization process (route determination step S5) is performed.

(Calculation of number)
As shown in FIG. 30A, by clicking the number calculation button 61f on the condition input screen 61, the screen 65 for calculating the number of vehicles used in the area in charge of the depot (office) is displayed. The required number of vehicles 12 can be obtained by adding a margin rate to information such as the day of the week, vehicle type, and destination factory. By inputting the margin ratio on the screen 65 and clicking the "calculate" button, the required number of vehicles 12 is calculated and displayed by the system as shown in FIG. 30 (b). When all the work in the area is S work, the number of sets and the number of sets are the same, and the number of sets is determined by clicking the "Set to the number of sets" button.

The required number of vehicles 12 is calculated as follows. The detailed information of each baggage is saved in, for example, a garbage CSV file including the estimated amount of baggage (garbage) of each customer, and the information of the depot (office) in charge of the customer area is also registered in this CSV file. NS. Conversely, by designating a depot, all customers targeted by that depot are designated. Therefore, when the screen 65 in which the office (depot) is designated is displayed, the estimated value of the total amount of cargo (total amount of garbage) carried into the factory 13 by the depot can be used for the calculation of the number of units.

Notation of the total amount of garbage G, the margin rate Y that determines the amount to be deducted in order to tighten the maximum load regulation amount of the vehicle, the average value M of the amount of collected garbage per efficiency, the number of times N of efficiency, and the required number of units X When is used, the number X required for the specified vehicle type is obtained by X = G / ((1-Y) · MN). This calculation is to find the number of sets, and in the case of S work, the number of sets matches the number of sets, so that the number of sets X = the number of sets (k if there are k sets to perform W work). Add a stand).

The present invention is not limited to the above configuration and can be modified in various ways. For example, the configurations of the above-described embodiments can be combined with each other.

1 Route calculation program 2 Route optimization system (computer)
2a Distance matrix generation unit 2b Customer allocation unit 2c Route division unit 2d Route determination unit 3 Distance matrix generation program 4 Route calculation program 11 Depot 12 Vehicle 13 Factory c, c1, c11, ..., c27 Customer nc1, nc2, nc3 Customer node nr1, nr2, nr3 Road node R, RW, R1, R2, R3 route (road link)

Claims (7)

A route calculation program in which one or more vehicles departing from a depot, which is a vehicle base, collects the luggage of a plurality of customers and calculates a route for carrying the vehicle into a designated factory.
Computer,
A distance matrix generator that generates a distance matrix for the distance between each of the nodes based on the position information of the nodes corresponding to the positions of the depot, the customer, and the factory.
Using the distance matrix and the number of working groups composed of the vehicle and the worker, the vehicle departs from the depot and visits the customer on the premise that the vehicle is terminated at a predetermined stage without obtaining the best solution. Regarding the route to return to the depot after collecting the luggage, the customer allocation department that solves the transportation route problem (VRP) and
For each work group, a route division unit that calculates a division route that divides the route assigned by the customer by the customer allocation unit based on the load capacity constraint information of the vehicle and the cargo information of each customer, and a route division unit.
For each division route, as a route determination unit for solving the traveling salesman problem (TSP) for a route from the depot or the factory to the factory via the customer included in the division route. Make it work
A route calculation program characterized by this. The customer allocation unit is characterized by solving the transportation route problem (VRP) under a time constraint that the travel time from the vehicle departing from the depot to returning to the depot is within a predetermined time. The route calculation program according to claim 1. The predetermined time is the number of times of delivery to the factory calculated in consideration of the travel time from the customer having the longest route from the designated factory to the factory and the load capacity constraint information of the vehicle for each vehicle. The route calculation program according to claim 2, wherein the factory delivery time calculated based on the above is the time obtained by subtracting the preset working time. The customer allocation unit is subject to a time constraint condition in which the variance of the set having the travel time of each vehicle in the plurality of vehicles from the depot to the depot as an element is equal to or less than a predetermined value. The route calculation program according to any one of claims 1 to 3, wherein the route calculation program is characterized by solving a transportation route problem (VRP). The route calculation program according to any one of claims 1 to 4, wherein the route dividing unit is not divided between the customers whose order is continuous within a predetermined distance. A computer for executing the route calculation program according to any one of claims 1 to 5, and one or more vehicles departing from the depot, which is a vehicle base, collects the luggage of a plurality of customers, and the above-mentioned A route optimization system characterized by finding a route for a vehicle to be delivered to a designated factory. Route calculation using a programmed computer to calculate the route for one or more vehicles departing from the depot, which is the vehicle base, to collect the luggage of multiple customers and bring it to the factory designated for the vehicle. It ’s a method,
A distance matrix generation step of generating a distance matrix for the distance between each of these nodes based on the position information of the nodes corresponding to the positions of the depot, the customer, and the factory.
Using the distance matrix and the number of working groups composed of the vehicle and the worker, the vehicle departs from the depot and visits the customer on the premise that the vehicle is terminated at a predetermined stage without obtaining the best solution. Regarding the route to return to the depot after collecting the luggage, the customer allocation process to solve the transportation route problem (VRP) and the customer allocation process
For each work group, a route division step of calculating a division route in which the route assigned by the customer by the customer allocation process is divided based on the load capacity constraint information of the vehicle and the cargo information of each customer, and a route division step.
For each division route, a route determination step of solving the traveling salesman problem (TSP) for a route from the depot or the factory to the factory via the customer included in the division route. A route calculation method characterized by including.

Images