Construction machinery delivery plan creation support system, delivery plan creation method, and delivery plan creation processing program

Abstract

This invention provides a support system, a delivery plan creation method, and a processing program that enable a construction equipment rental company to create an efficient delivery plan to deliver the desired construction equipment to the desired location by the desired date and time, in accordance with the user's request. [Solution] The system comprises databases 33-38 where basic information is registered, a first storage means 32 that stores pre-set delivery conditions associated with the basic information, a second storage means 31 that stores request information related to the user, a processing means 21 that generates a delivery plan pattern by allocating construction machinery for a pre-set number of days to vehicles and contractors based on the request information and basic information, an evaluation means 21 that evaluates the generated delivery plan pattern, and a determination means 22 that determines the superiority or inferiority of the delivery plan pattern based on the evaluated evaluation results. The processing means executes the processing using a genetic algorithm.

Description

Technical Field

[0001] The present invention relates to a construction machinery delivery plan creation support system, a delivery plan creation method, and a delivery plan creation processing program. In particular, in a rental company that temporarily lends construction machinery to a construction company, based on the request of the construction company, a system, a creation method, and a creation processing program for supporting the creation of a delivery plan for delivering specific construction machinery within a specified period are provided.

Background Art

[0002] Support systems for processing delivery-related information and generating delivery plans are generally those that create stacking rules for loading goods on delivery vehicles (see Patent Document 1) or generate delivery routes (see Patent Document 2) in companies specializing in delivery services or delivery departments within the same company.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Patent Document 2

Summary of the Invention

Problems to be Solved by the Invention

[0004] The technology disclosed in the above-mentioned Patent Document 1 is a support system for loading and delivering uncertain goods on a delivery vehicle. Whether the goods can be loaded on the delivery vehicle is based on performance data. Therefore, it is configured to generate a delivery vehicle loading rule and create a delivery plan when it matches this loading rule. On the other hand, the technology disclosed in the above-mentioned Patent Document 2 is to generate a new delivery route when an additional delivery order is received after the generation of the delivery route.

[0005] However, when a rental company delivers rental items to a location specified by the user at the user's request, while detailed information about the vehicle used for delivery and the rental items can be obtained in advance, if the rental company prepares a large number of rental items to meet the user's needs, a shortage of delivery vehicles for the number of rental items can become a problem.

[0006] In particular, construction machinery is often packaged in small quantities, making it difficult to stack and transport. Large construction machines may require a separate vehicle for each unit, meaning that efficiency cannot be improved by transporting multiple items simultaneously. Furthermore, some construction machinery requires vehicles equipped with special functions (such as cranes and winches) (such as Unic trucks) for delivery, so efficiency could not be improved simply by the number of items to be delivered relative to the number of vehicles. In addition, when renting construction machinery, different sizes and types of machinery are required to be delivered to each construction site depending on the scale, progress, and content of the work. To satisfy the requests of numerous users (users working at many construction sites in total), it was necessary to effectively utilize the time available for the company's own vehicles to move. Moreover, when the company's own vehicles were unavailable, it was important to meet the construction work schedule by outsourcing delivery only. Considering the arrangements and costs of outsourcing, it was also important to minimize requests to outsourcing, requiring the creation of a delivery plan to that end.

[0007] The present invention has been made in view of the above points, and its object is to provide a support system that enables a construction machinery rental company to create an efficient delivery plan in order to deliver a desired construction machinery to a desired location by a desired date and time, in response to a user's request, and to provide a delivery plan creation method and a processing program for that purpose. [Means for solving the problem]

[0008] Therefore, the first configuration of the present invention relating to a construction machinery delivery plan creation support system is a delivery plan creation support system for delivering multiple construction machines by multiple vehicles based on a request from a user, comprising: a database in which basic information is registered, which includes at least user information relating to the user who will be the source of the delivery, vehicle information relating to the vehicle used by the user to deliver the construction machines, construction machine information relating to the construction machines to be delivered, and contractor information relating to the contractor when delivery is outsourced; a first storage means for storing delivery conditions that are associated with the basic information and are set in advance to realize the delivery; a second storage means for storing request information relating to the user, including the content of the request made by the user; a processing means for generating a delivery plan pattern by allocating a predetermined number of days of construction machines to the vehicles and the contractor based on the request information and the basic information; an evaluation means for evaluating the delivery plan pattern generated by the processing means; and the evaluation means The system comprises a determination means for determining the superiority or inferiority of delivery plan patterns based on evaluation results evaluated in stages, wherein the processing means generates multiple delivery plan patterns by a genetic algorithm, the selection of delivery plan patterns to perform the genetic operation involves sequentially selecting a predetermined number of delivery plan patterns from among the multiple delivery plan patterns generated by the genetic algorithm, starting with those with the largest evaluation values ​​as evaluated by the evaluation means, the genetic algorithm is executed until delivery plan patterns for a predetermined generation are generated, the evaluation means assigns the lowest evaluation value to delivery plan patterns that do not satisfy the delivery conditions, and assigns a high evaluation value to delivery plan patterns that result in a smaller allocation to the consignee, and the determination means determines the delivery plan pattern with the largest evaluation value as evaluated by the evaluation means from among the multiple delivery plan patterns generated in a predetermined generation as the final delivery plan.

[0009] According to the above configuration, based on basic information, the processing means can generate a first-generation delivery plan pattern, and then generate a next-generation delivery plan pattern through genetic manipulation (such as crossover and mutation). For each generation of delivery plan patterns, the evaluation means derives an evaluation value, and a predetermined number of delivery plan patterns selected using this evaluation value are used for the next-generation genetic manipulation. By generating delivery plan patterns in appropriate generations, a good delivery plan pattern can be generated. By pre-setting the generations in which genetic manipulation should be performed, and using the determination means to select the delivery plan pattern with the highest evaluation value from among the delivery plan patterns generated in the last generation, the creation of a suitable delivery plan can be supported.

[0010] In this invention, a genetic algorithm refers to an algorithm for deriving an optimal conclusion by repeatedly performing transformations, selections, crossovers, mutations, etc., on given data, and a genetic operation refers to executing the above algorithm. The genetic algorithm in this invention is configured to repeatedly perform selections, crossovers, mutations, and repairs, and the evaluation results from an evaluation means are referred to before selection. If road information is included in the basic information, in the repair process, when the arrival time is predicted by referring to the road information, construction machinery that cannot reach the scheduled time can be assigned to the contractor.

[0011] Furthermore, the road information mentioned above is a road map used for route detection during delivery, and it includes information that identifies congested sections by time and area. Various types of road information are available, but it is possible to use pre-created road information for a specific area (delivery area).

[0012] A second configuration of the present invention relating to a system for creating delivery plans for construction machinery is that, in the first configuration, the delivery conditions include a condition in which a construction machine containing a specific element among the construction machines registered as construction machine information is delivered by a vehicle that satisfies a specific condition among the vehicles registered as vehicle information.

[0013] According to the above configuration, it becomes possible to generate delivery plan patterns in which vehicles that meet specific conditions (e.g., a crane-equipped truck) can be used to deliver construction machinery that includes specific elements (e.g., requiring a crane, etc.). The specific elements related to the construction machinery may include weight and size, while the specific elements related to the vehicle may include load capacity and other equipment.

[0014] A third configuration of the present invention relating to a construction machinery delivery plan creation support system is that, in the second configuration, the request content includes a delivery deadline which is the date and time on which the user will use the desired construction machinery or the date and time on which delivery should be completed, and the delivery conditions further include a condition that the delivery must be made by the delivery deadline.

[0015] According to the above configuration, the concept of a delivery deadline can be introduced as a delivery condition. By delivering to the delivery location before the specified date and time (for example, one day or three days before), the range of delivery plan patterns generated by the processing means can be expanded. By evaluating the delivery plan patterns generated within this expanded range, it becomes possible to create low-budget delivery plans that reduce the amount allocated to the contractors.

[0016] A fourth configuration of the present invention relating to a system for creating delivery plans for construction machinery is that, in the third configuration, the evaluation means can differentiate evaluation values ​​based on a predetermined priority for each vehicle, within the scope in which the distribution of vehicles in the delivery plan pattern is to be carried out by vehicles used by the user.

[0017] According to the above configuration, the frequency of assigning drivers who require special licenses or qualifications to operate the vehicles can be reduced, while vehicles that do not require special licenses can be used more frequently, thus reducing the number of considerations regarding the arrangement (availability) of such drivers.

[0018] A fifth configuration of the present invention relating to a construction machinery delivery plan creation support system is as follows: In any of the first to fourth configurations, the delivery plan patterns generated by the genetic algorithm consist of 100 or more individuals, and the delivery plan patterns for genetic manipulation in the next generation consist of 10% of the individuals selected from the generated delivery plan patterns, and the selected delivery plan patterns are used to generate eight times the number of delivery plan patterns through crossover and one time the number of delivery plan patterns through mutation.

[0019] According to the above configuration, the next generation of delivery plan patterns generated by genetic manipulation will have a ratio of 1:8:1 for the number of individuals in the selected delivery plan pattern (which is set to "1"), and the ratio of the number of individuals in the delivery plan patterns generated by crossover and mutation will be 1:8:1. As the number of delivery plan patterns gradually improved by crossover increases, 10% of those with the highest overall evaluation values ​​are selected and provided for the next genetic manipulation, where they are further refined. The number of individuals to be selected is 100 or more, so it could be 500 or 1,000, but it should be adjusted to a range where the evaluation value converges when compared with the number of times the genetic algorithm is repeated (number of generations).

[0020] Therefore, the sixth configuration of the present invention relating to a construction machinery delivery plan creation support system is adjusted in the fifth configuration so that when the number of delivery plan patterns generated by the genetic algorithm is 1000, the genetic operation is performed within a range of 40 to 50 generations.

[0021] According to the above configuration, initially, 100 delivery plan patterns are selected, representing the top 10% based on evaluation results. These 100 individuals are then used as a baseline, and genetic manipulation increases the number by 900 to a total of 1000. The top 100 individuals are then selected again based on the re-evaluation results, which is equivalent to evaluating one million delivery plan patterns in one generation. By performing this genetic manipulation for 40 to 50 generations, the evaluation values ​​can be converged at a high level. In fact, experimentally, the increase in evaluation values ​​slows down after about 35 generations, and it has been experimentally confirmed that they converge around 40 generations. Although there is a slight difference in evaluation values ​​between 40 and 50 generations, the difference is minimal, and the processing time is about 5 minutes for 40 generations and about 6 minutes for 50 generations. Therefore, genetic manipulation beyond 50 generations only prolongs the processing time, and 50 generations can be set as the upper limit.

[0022] On the other hand, the present invention relating to a method for creating a delivery plan for construction machinery is a method for creating a delivery plan for distributing multiple construction machines by multiple vehicles based on a request from a user, and includes the steps of: registering basic information in a database which includes at least user information relating to the user who will be the source of the delivery, vehicle information relating to the vehicle used by the user to deliver the construction machinery, construction machinery information relating to the construction machinery to be delivered, and contractor information relating to the contractor when delivery is outsourced; storing delivery conditions associated with the basic information and set in advance to realize delivery in a first storage means; storing request information relating to the user, including the content of the request made by the user, in a second storage means; and processing steps which process the information stored in the database, the first storage means and the second storage means, wherein the processing steps include a processing step which generates a delivery plan pattern by allocating a predetermined number of days of construction machinery to the vehicles and the contractor based on the request information and the basic information, and an evaluation of the delivery plan pattern generated by the processing steps The processing step comprises an evaluation step and a determination step that determines the superiority or inferiority of the delivery plan patterns based on the evaluation results evaluated by the evaluation means, wherein the processing step generates a plurality of delivery plan patterns by a genetic algorithm, the selection of the delivery plan patterns to perform the genetic operation involves selecting a predetermined number of delivery plan patterns from the plurality of delivery plan patterns generated by the genetic algorithm, in order from those with the largest evaluation values ​​based on the evaluation results evaluated by the evaluation means, the genetic algorithm is executed until a delivery plan pattern for a predetermined generation is generated, the evaluation step assigns the lowest evaluation value to delivery plan patterns that do not satisfy the delivery conditions saved by the first saving step, and assigns a high evaluation value to delivery plan patterns that result in a smaller allocation amount to the consignee, and the determination step determines the delivery plan pattern with the largest evaluation value evaluated by the evaluation step from among the plurality of delivery plan patterns generated in a predetermined generation as the final delivery plan.

[0023] According to the above configuration, by previously storing basic information, delivery conditions, and request information, and processing each step through a processing process based on this information, it is possible to appropriately generate a delivery plan pattern and create a delivery plan with a high evaluation value.

[0024] Further, the present invention related to a program for generating a delivery plan for construction machinery is characterized in that each step in the processing steps described in the above delivery plan generation method is executed by a computer. According to such a configuration, the delivery plan generation method for construction machinery can be automated by a computer.

Effect of the Invention

[0025] According to the delivery plan generation support system of the present invention, while using a pattern with a high evaluation value among the delivery plan patterns generated by genetic operations for the next generation of genetic operations, a pattern with a high evaluation value is selected as the delivery plan from the delivery plan patterns generated in a predetermined generation. Therefore, those that do not meet the delivery conditions are excluded, and a delivery plan with high delivery efficiency can be obtained. As a result, a dealer renting construction machinery can create an efficient delivery plan for delivering a desired construction machinery to a desired location by a desired date and time in response to a user's request.

[0026] Further, according to the delivery plan generation method of the present invention, by registering predetermined information in advance, the generation of the delivery plan pattern and the creation of the delivery plan can be automated. And according to the program for generating a delivery plan processing of the present invention, the above automated processing can be executed by a computer.

Brief Description of the Drawings

[0027] [Figure 1] It is an explanatory diagram showing an outline of an embodiment related to a delivery plan generation support system. [Figure 2] It is a conceptual diagram showing a rough flow until a delivery plan is created based on various information. [Figure 3]This is a model diagram showing a single individual (delivery plan pattern). [Figure 4] This is a processing flow that shows the entire processing steps. [Figure 5] This is a processing flow for initializing an individual that has not yet been assigned a vehicle. [Figure 6] This is the processing flow for the evaluation step. [Figure 7] (a) is the processing flow for the selection step, and (b) is the processing flow for the crossover step. [Figure 8] This is the processing flow for the mutation step. [Figure 9] This is the processing flow for the repair step. [Modes for carrying out the invention]

[0028] The embodiments of the present invention will be described below with reference to the drawings. First, the overall structure of the delivery plan creation support system will be described, and then the individual elements will be described.

[0029] <Overall structure of the delivery plan creation support system> Figure 1 is a schematic diagram showing an embodiment of a delivery plan creation support system. As shown in this figure, the delivery plan creation support system 1 of this embodiment is composed of a general personal computer including a processing unit 2 and a storage device 3, to which an input / output device 10 is connected. The input / output device 10 is equipped with a keyboard for inputting necessary information and is equipped with a monitor for outputting processing results to the monitor.

[0030] The processing unit 2 is a so-called processor that can process information input from the input / output device 10 and information stored in the storage device 3 according to a predetermined program. This processing unit 2 generally comprises a delivery plan pattern generation processing unit (processing means) 21, an evaluation value calculation unit (evaluation means) 22, and a determination unit (determination means) 23.

[0031] The storage device 3 consists of storage such as an HDD or SSD, as well as memory such as ROM or DRAM. Information (data) can be stored in the storage and memory as needed, and this information can be provided during processing by the processing device 2. The information stored in this storage device is classified into information about the user (system user) and information about the requester (customer), and fixed information is stored in a database. Therefore, the storage device 2 generally consists of a request information storage unit (second storage means) 31 and a delivery condition storage unit (first storage means) 32, as well as a user information database 33, a vehicle information database 34, a construction machinery information database 35, a contractor information database 36, a road information database 37, a customer information database 38, and so on. In addition, storage areas necessary for processing in the processing device 2 are also provided as appropriate.

[0032] The client information storage unit 31 and the delivery condition storage unit 32 in the storage device 3 are overwritten and saved when initial information is input by the user via the input / output device 10, or when modified information is input as appropriate. In addition, the information stored in the various databases 33 to 38 can also be modified and overwritten via the input / output device 10 as needed.

[0033] <database> In this embodiment, multiple databases 33 to 38 are constructed, and information is stored separately according to its nature. For example, the user information database 33 registers information such as the location of the user's base (head office and branch offices, etc.) and the areas where delivery is possible (delivery area). In conjunction with the vehicle information described later, the type and number of vehicles owned may also be registered. The vehicle information database 34 registers information for all vehicles, including their model, year of manufacture, vehicle inspection expiration date, load capacity, and the presence and type of special equipment. The construction machinery information database 35 registers information for the construction machinery to be delivered, including its type, model, and size, as well as the presence or absence of special equipment necessary for delivery, and information such as the functions and capabilities of the construction machinery. In addition, the current location of the construction machinery is registered, and spatial information such as whether it is at the client's location (if installed at a construction site) or waiting at the user's branch office, etc., is registered. This information is appropriately changed and overwritten according to the rental period. The contractor information database 36 contains information about contractors to whom delivery is outsourced. For each user location, information such as the types and number of construction machines that can be delivered is registered for contractors that can be outsourced. The road information database 37 primarily registers information that has been digitized as a road map, and information such as congested road sections and times can also be registered in advance. The customer information database 38 contains information about customers who make repeat orders, and only information that does not change frequently is registered in advance.

[0034] <Request Information and Delivery Conditions> The request information stored in the request information storage unit 31 includes, for example, a case number to identify the case related to the request, and if there are multiple cases, a sub-number (or consecutive number) of the case number is registered as an index. For each case number, basic information such as the date and time the request was received, the name of the requester, the requester's name and address, and other information specific to the requester (customer number if already registered in customer information) is registered, as well as delivery destination information such as the delivery address, type of work, and name of the construction site manager. Information regarding the requested product is registered such as the product number to identify the construction machinery (if there is no product number, information that identifies it by type, size, etc.) and the delivery period (delivery deadline). If necessary, information such as whether the delivery time must be strictly adhered to, the delivery status (e.g., on-board delivery or mobile installation delivery), and whether a delivery vehicle is specified may also be registered.

[0035] The delivery conditions recorded in the delivery conditions storage unit 32 are used exclusively by the evaluation means (evaluation value calculation unit 22), and the criteria for determining whether delivery is possible or not for each combination of vehicle and construction machinery are stored. For example, delivery conditions may include conditions such as the total weight of the construction machinery to be delivered being within the vehicle's load capacity, and conditions such as construction machinery requiring special equipment being loaded onto a vehicle equipped with special equipment. In addition, depending on the delivery deadline, conditions such as the delivery date and time not exceeding the deadline may also be included.

[0036] <Processing Method> The delivery plan creation support system 1 of this embodiment generates delivery plan patterns based on the various basic information stored as described above, and the processing device 2 processes the data to ultimately create a good delivery plan. A genetic algorithm is used as the processing method here. The conceptual details of the processing method using the genetic algorithm in this embodiment will now be explained.

[0037] Figure 2 is a conceptual diagram showing the general flow of how a delivery plan is created based on various information, and Figure 3 is a model diagram showing a single individual (delivery plan pattern). As shown in Figure 2, the processing method of this embodiment is divided into pre-operation and genetic operation. Pre-operation is the operation of generating a set of individuals (delivery plan patterns) (for example, 1,000 individuals) to be used for genetic operation.

[0038] The preliminary operation generates a vehicle list based on information about vehicles available for delivery, selected from various sources. Since the vehicle list contains information about vehicles available to the user, it is a fixed list of information. However, if there are insufficient vehicles for delivery, the work will be outsourced, and the outsourced vehicle will be included in the vehicle list. Because this outsourced vehicle only requires ordering from the outsourced company, it is included in the vehicle list under the assumption that it can deliver construction machinery indefinitely as a single vehicle.

[0039] Furthermore, genes are generated based on information about the construction machinery to be delivered, selected from various sources. These genes encompass a single, identified construction machine, along with the delivery conditions from its current location to the destination, and the delivery deadline. Since the delivery deadline includes a certain degree of temporal flexibility, it can be modified according to the condition of the vehicle being used.

[0040] The combinations generated to ensure all genes (requests) are delivered by vehicles from the vehicle list constitute an individual (delivery plan pattern). In the preliminary operation, individuals (delivery plan patterns) are generated by randomly combining multiple genes and multiple vehicles, and the preliminary operation is completed when a predetermined number (e.g., 1000) individuals (delivery plan patterns) are formed into a population.

[0041] As shown in Figure 3, each individual (delivery plan pattern) has genes assigned to each vehicle (vehicle 1 to vehicle n) for each day of the delivery period (e.g., one week or one month). The assigned genes are a specific single construction machine and the delivery conditions and other details (elements) related to that construction machine, so it is possible to assign multiple genes to each individual vehicle (vehicle 1 to vehicle n). If it is before the deadline, it is possible to move the delivery to the previous day, so the planning period is set to cover multiple days. Genes that cannot be processed within a specific day (construction machines that cannot be delivered) are included in the individual (delivery plan pattern) and planned to be delivered using the contractor's vehicles. The contractor's vehicles are grouped together, and it can be assumed that multiple genes will be delivered using an appropriate number of vehicles at the contractor's discretion. Note that Figure 3 is a graphical representation to aid in understanding the individual (delivery plan pattern), and the details may be represented in a different form in the data.

[0042] Then, as shown in Figure 2, after the preliminary operations are completed, genetic operations are performed. Genetic operations are performed using a population of a predetermined number (e.g., 1000) individuals (delivery plan patterns) formed by the preliminary operations. Genetic operations always involve selecting a certain percentage of superior individuals (delivery plan patterns) from the predetermined number (1000) individuals (delivery plan patterns) that make up the population, and then performing crossover and mutation based on these individuals to form a new population of a predetermined number (e.g., 1000) that includes individuals (delivery plan patterns) different from those selected before the genetic operations. After repeating this operation a predetermined number of times (a predetermined number of generations), the single most superior individual from the final population is determined to be the delivery plan.

[0043] Crossover can be performed using methods such as one-point crossover or two-point crossover, but in this embodiment, the one-point crossover method is employed. Specifically, within the same day, genes assigned to a contractor are moved to a vehicle that meets the conditions (is deliverable). Mutation, on the other hand, involves movement across days, where randomly selected genes are moved to other vehicles that meet the conditions. If there are interchangeable genes in the destination vehicle, they are exchanged; if there are no interchangeable genes, only one of the genes is moved.

[0044] Here, prior to selection, evaluation is performed using an evaluation tool, and an evaluation value is calculated for each (1000) individual (delivery plan pattern). A certain percentage of individuals (delivery plan patterns) with high evaluation values ​​are selected. Furthermore, after crossover and mutation, repair is performed. Repair is an operation in which multiple construction machines loaded on the same vehicle are compared with their delivery dates, and construction machines that are late on their delivery dates are eliminated (contracted).

[0045] By performing the genetic manipulation described above, a set of superior (highly rated) individuals (delivery plan patterns) is formed in each generation, and these can be evolved generation by generation. After performing this operation for a predetermined number of generations, the individual (delivery plan pattern) that is ultimately estimated to be the most superior will be output as the delivery plan.

[0046] <Summary> As described above, the embodiment of the delivery plan creation support system is such that it can generate delivery plan patterns that meet the predetermined delivery conditions while eliminating those that do not, and then select a delivery plan pattern with a high evaluation value from among them. The evaluation value indicator will be set considering the user's priorities, but if the focus is on cost reduction, delivery plan patterns that allocate fewer genes to subcontractors (resulting in reduced outsourcing costs) will receive a higher evaluation value, and delivery plan patterns with shorter travel distances (reduced fuel costs) and shorter travel times (reduced labor costs) will receive a higher evaluation, thereby enabling the creation of cost-efficient delivery plans.

[0047] The above embodiment is merely an example of the present invention relating to a delivery plan creation support system. Therefore, the present invention is not limited to the above embodiment, and each element can be modified as appropriate, and other elements may be added. For example, in calculating evaluation values ​​by the evaluation means, when the construction machinery to be delivered is delivered by a vehicle used by the user, a difference in evaluation values ​​may be set based on a predetermined priority. In this example, priority was given to cost, but it may also be set to give a higher evaluation to delivery plan patterns that reduce the frequency of use of special vehicles (vehicles requiring special licenses).

[0048] Furthermore, while 1000 individuals (delivery plan patterns) were shown as an example for the individuals (delivery plan patterns) generated by the genetic algorithm, if there are more than 100 individuals, it is possible to generate good individuals (delivery plan patterns) in a reasonable amount of time through several dozen generations of genetic manipulation. In the next generation, the selection of individuals (delivery plan patterns) for genetic manipulation should be limited to about 10% of the total number of individuals in the generated set. By generating eight times the number of individuals (delivery plan patterns) through crossover and one time the number through mutation from the selected individuals (delivery plan patterns), and forming a set of the same number of individuals as the previous generation along with the selected individuals (delivery plan patterns), the evolution by the genetic algorithm can be accelerated. In this case, when 1000 individuals (delivery plan patterns) are created by the genetic algorithm, performing genetic manipulation within the range of 40 to 50 generations can be experimentally completed in about 5 minutes using a computer with a typical processing speed.

[0049] <Example of a delivery plan creation method> Next, an embodiment of the invention relating to the delivery plan creation method will be described. The delivery plan creation method of this embodiment is a processing method when using the delivery plan creation support system described above. As a preliminary step, basic information is registered in the database, delivery conditions are saved, and request information is saved. Furthermore, as necessary elements for processing, the number of individuals (delivery plan patterns) included in the set is set, the evaluation method is set, the number of selections, crossovers, and mutations is set, and the number of generations for which genetic operations are performed is set.

[0050] The database will register various basic information, including user information about the user who will be delivering the goods, vehicle information about the vehicle used by the user to deliver the construction machinery, construction machinery information about the construction machinery to be delivered, contractor information about the contractor if delivery is outsourced, road information about road conditions in the delivery area, and customer information. Delivery conditions are predetermined to enable delivery and include conditions that make delivery impossible, which are linked to the basic information (mainly vehicle information and construction machinery information). Request information consists of order conditions requested by the user and mainly includes information about the construction machinery, delivery date and delivery location, as well as individual requests.

[0051] After the above-mentioned preliminary information is entered, the processing steps are executed. The processing steps are divided into a processing step, an evaluation step, and a decision step. The delivery plan patterns generated by the processing step are evaluated, other delivery plan patterns are generated and evaluated again, and finally a single delivery plan is selected by the decision step.

[0052] Therefore, the specific details of the processing steps will be explained based on the processing steps. Figures 4 to 9 are flowcharts showing the processing steps at each stage.

[0053] Figure 4 shows the overall process. As shown in this figure, the processing steps first involve creating a number of genes for each request based on the request information and construction machinery information (S101), then creating "initial individuals" from all the genes (S102), and finally creating a "set" which is a collection of these individuals (S103). At this point, the "initial individuals" do not yet have a delivery plan pattern and consist only of genes that do not include vehicles (information on construction machinery associated with the desired delivery date). After collecting these initial individuals and constructing a provisional set, the initial individuals are initialized in the initialization step (S200), thereby assigning vehicles and completing primitive individuals (delivery plan patterns). The initialization process will be described later.

[0054] As described above, primitive individuals (delivery plan patterns) are used as a basis and are evolved through genetic manipulation. The genetic manipulation process involves looping predetermined processing steps for a predetermined number of generations (S104, S105). The genetic manipulation process through loops involves deriving evaluation values ​​through a step of evaluating the set (individuals) (S300), selecting individuals with high evaluation values ​​through a step of selecting individuals (S400), changing each individual through a step of performing crossover (S500) and a step of performing mutation (S600) from the selected individuals, and repairing the individuals through a repair step (S700). Details of these steps will be described later, but finally, after completing the loop for a predetermined number of generations, the individuals (delivery plan patterns) that form the set are evaluated by an evaluation step (S300), and based on the resulting evaluation values, a final single delivery plan pattern is identified as the delivery plan in a determination step (S800).

[0055] Here, we will explain the content of each processing step. Figure 5 shows the processing flow for initializing an initial individual (an individual that has not yet been assigned a vehicle). This shows the process of forming a primitive individual (a primitive delivery plan pattern) from the initial individual.

[0056] First, the desired delivery date is obtained from the information contained in all genes (S201), and from that desired delivery date, the nearest desired date (the smallest number of days counting from the processing date) and the latest desired date (the largest number of days counting from the processing date) are obtained (S202). This process is used to create a delivery plan for the number of days obtained. Note that when creating a delivery plan for a limited period (for example, one week or one month), it is also possible to set the upper limit of the latest desired date to the end of the period (e.g., 7 days or 30 days).

[0057] Once the period for creating the delivery plan is determined, available vehicles are provisionally allocated each day during that period (from the most recent day to the final day) in preparation for allocation (S203). Subsequently, contracted vehicles for all days are also provisionally allocated (S204). At this point, the vehicles to which genes can be allocated are ready.

[0058] Next, the genetic information is read, and a random delivery date (the date on which delivery is scheduled) is assigned within the range of desired delivery dates (S205). The genes are then randomly assigned to vehicles that are scheduled for those dates (S206). In this state, at least the construction machinery, which is information belonging to the genes, can be assumed to be delivered by one of the vehicles on one of the days up to the desired delivery date. However, at this stage, no comparison with delivery conditions has been made, nor has it been evaluated. Therefore, this state has formed a primitive individual (delivery plan pattern) that can be subjected to genetic manipulation.

[0059] Figure 6 shows the processing flow of the evaluation process. The evaluation process evaluates primitive individuals (delivery plan patterns), but it is not limited to that and is also a processing step used in genetic manipulation. The purpose is to calculate evaluation values ​​for primitive and generationally generated individuals (delivery plan patterns) using a predetermined evaluation method.

[0060] As shown in Figure 6, the evaluation process acquires information on the genes assigned to each vehicle (including contracted vehicles) (S301), and detects genes that do not match the dispatch conditions by comparing them with the dispatch conditions (S302). It also detects the difference (in days) between the gene delivery date and the desired delivery date (S303), and further detects the genes assigned to contracted vehicles, and calculates an estimate of the contract costs when using those contracted vehicles. The information detected or calculated in each of these steps is temporarily stored and used when calculating the evaluation value.

[0061] The evaluation value is calculated based on genes that do not meet the above conditions (genes that do not fit the conditions), the difference (in days) from the desired delivery date, and the commission fee (S305). In the figure, each is shown as an "index," and the evaluation value is the reciprocal of the sum of these indices. Therefore, if the value of each index is large (total commission fee, difference in days from the delivery date, etc. is large), the evaluation value is set to be small. For example, the total commission fee index can be set to 1 / 00 ​​of the total cost (in yen), and the total number of days difference from the desired delivery date index can be set to 1 / 100 of the calculated number of days (in days), thereby creating an evaluation value that places more emphasis on the amount of commission fee than on the difference in days. Also, in the case of a condition not being met, for example, simply setting the value to 10 million would result in a much larger number even if the commission fee was 1 million yen, so when the reciprocal is used, it would result in an extremely low evaluation value, which can be eliminated in the later selection step (selection of excellent delivery plan patterns).

[0062] Therefore, as shown in Figure 7(a), the selection step is performed by sequentially acquiring a predetermined number of individuals (S401) based on the evaluation values ​​calculated in the evaluation step above, starting with those with the highest evaluation values. In the selection step, these individuals are added to a new set (S402) to be used for genetic manipulation. Specifically, the crossover and mutation steps are performed based on the individuals included in the new set in the selection step above.

[0063] Here, the crossover step is performed using a one-point crossover method, as shown in Figure 7(b). First, the individuals selected in the selection step are copied (S501), and then, for the new individuals, genes that satisfy the delivery conditions are moved from among the genes assigned to the contracted vehicles to other vehicles. This process allows for combinations of moving multiple genes, resulting in the generation of more individuals than the number of selected individuals (for example, eight times the number of individuals).

[0064] Furthermore, the mutation step, as shown in Figure 8, involves an operation to exchange genes assigned to vehicles other than those used by the contracted service (vehicles used by the user). Since the reference individual is the selected individual, first, that individual is copied (S601), and the copied individual is used. Using the copied individual, information on a randomly selected gene (here referred to as gene (a)) assigned to a randomly identified vehicle (here referred to as vehicle (A)) on a randomly selected date is obtained (S602). From the information on gene (a), the desired delivery date is obtained, and information on the vehicles (for all vehicles) that are positioned (prepared) on the same day as that desired date is obtained (S603). The system determines whether there is a vehicle (referred to as vehicle (B) in this case) among the vehicles on that day that matches the conditions of gene (a) (S604). If such a vehicle exists, the system obtains the gene information assigned to that vehicle (B) (S605). The system then determines whether the obtained gene matches the conditions of the original vehicle (A) (referred to as gene (b) in this case) (S607). If the conditions match, the system performs a process to exchange both genes (a) and (b) with each other (S608).

[0065] On the other hand, after determining whether or not a vehicle (B) that matches the conditions of gene (a) exists (S604), if vehicle (B) does not exist, another vehicle is identified and the same process is performed. If vehicle (B) does not exist after repeating this 10 times, the gene (a) is assigned to a contracted vehicle (S606).

[0066] Furthermore, after determining whether or not a vehicle (B) that matches the conditions of gene (a) exists (S604), if vehicle (B) exists, and gene (b) obtained from vehicle (B) does not match the conditions, the same process is performed on the other genes, and if gene (b) does not exist even after repeating this 10 times, gene (a) is assigned to a new vehicle (B) (S609).

[0067] The mutation step is completed when the allocation of gene (a) is changed or exchanged with gene (b). The number of individuals generated by this mutation can be fewer than the number generated by crossover (for example, the same number as the number of selected individuals). Furthermore, by processing the mutation in parallel with crossover, all generated individuals can be repaired.

[0068] The repair step is a process that adjusts the delivery times of vehicles. As shown in Figure 9, this repair step is performed for all vehicles, and the delivery times of each vehicle are obtained. Then, the process begins by rearranging the genes in order of these times (S701). Following the rearranged order, information is obtained for the first gene with the earliest delivery time (referred to as gene (a) here) and the next highest-ranking gene (referred to as gene (b) here) (S702). A time setting is then applied to the first gene (a) (S703), and a time setting is also applied to the next highest-ranking gene (b) (S704). Time setting involves calculating and setting the time required for delivery from road information. As a result, it is determined whether it is possible to deliver these two genes (a) and (b) (S705). If it is determined that delivery is not possible, the priority of the two is compared, and the one with the higher priority is kept (S706). If delivery is possible, the same process is performed for the next highest-ranking gene (S707).

[0069] The priority mentioned above can be determined by factors such as how close the desired delivery date is or the customer's priority, but other factors may also be used. If one gene remains after comparing priorities, the other gene is allocated to the contracted vehicle (S709, S710). In this way, by allocating the other gene to the contracted vehicle, both genes (a) and (b) can be delivered for the time being.

[0070] Once the above process is completed, the same process is performed on the next-ranked gene (S711), and this process continues for all genes in the same vehicle until it is determined that processing has been completed for all genes in the vehicle (S712). Furthermore, the same process is performed on each individual vehicle until the above process has been completed for all vehicles, and once it is determined that processing has been completed for all vehicles (S713), the repair process is terminated.

[0071] <Summary> As described above, by executing each processing step, it is possible to generate delivery plan patterns using genetic manipulation, and by selecting the delivery plan pattern with the highest evaluation value from among the delivery plan patterns generated in a given generation as the final delivery plan, it becomes possible to create an efficient delivery plan. Since each step in the above process can be executed by a computer, it is also possible to construct a program for this purpose.

[0072] The embodiment of the delivery plan creation method described above is merely an example of the present invention; therefore, the present invention is not limited to the above embodiment, and each element can be modified as appropriate, and other elements may be added. For example, in the above embodiment, the evaluation process is deliberately separated from the selection process, but as in the case of a general genetic algorithm, the evaluation process may be included in the selection process as a preprocessing step. Furthermore, the method for calculating the evaluation values ​​(especially each index) may be modified as appropriate according to their importance. [Explanation of Symbols]

[0073] 1. Delivery plan creation support system 2 Processing Unit 3 Storage device 10 Input / Output Devices 21 Delivery Plan Pattern Generation Processing Unit 22 Evaluation Value Calculation Unit (Evaluation Means) 23 Judgment unit (judgment means) 31. Client Information Storage Unit (Second Storage Means) 32 Delivery condition storage unit (first storage means) 33 User Information Database 34 Vehicle Information Database 35 Construction Machinery Information Database 36. Database of Contractors 37 Road Information Database 38 Customer Information Database

Claims

1. A delivery plan creation support system for distributing multiple construction machines among multiple vehicles based on a request from a user, A database in which basic information is registered, including at least user information about the user who will be the source of the delivery, vehicle information about the vehicle used by the user to deliver the construction machinery, construction machinery information about the construction machinery to be delivered, and contractor information about the contractor when delivery is outsourced, A first storage means that stores delivery conditions that are associated with the aforementioned basic information and are set in advance to realize delivery, A second storage means for storing user-related request information, including the content of requests made by the user, A processing means for generating a delivery plan pattern by allocating construction machinery for a predetermined number of days to the vehicle and the contractor based on the request information and the basic information, An evaluation means for evaluating the delivery plan pattern generated by the processing means, A determination means for determining the superiority or inferiority of delivery plan patterns based on the evaluation results evaluated by the evaluation means, Equipped with, The processing means generates multiple delivery plan patterns using a genetic algorithm. The selection of the delivery plan patterns to perform the aforementioned genetic manipulation involves sequentially selecting a predetermined number of delivery plan patterns from among a plurality of delivery plan patterns generated by the genetic algorithm, starting with those that have the highest evaluation values ​​as evaluated by the evaluation means. The aforementioned genetic algorithm is executed until a distribution plan pattern for a given generation is generated. The evaluation means assigns the lowest evaluation value to delivery plan patterns that do not meet the delivery conditions, while assigning a high evaluation value to delivery plan patterns that result in a smaller allocation to the contractor. The determination means selects the delivery plan pattern with the largest evaluation value, as evaluated by the evaluation means, from among a plurality of delivery plan patterns generated in a predetermined generation, and determines it as the final delivery plan. A system for supporting the creation of delivery plans for construction machinery, characterized by the following features.

2. The construction machinery delivery plan creation support system according to claim 1, wherein the delivery conditions include a condition that a construction machine containing a specific element among the construction machines registered as construction machinery information is delivered by a vehicle that satisfies specific conditions among the vehicles registered as vehicle information.

3. The construction machinery delivery plan creation support system according to claim 2, wherein the requested information includes a delivery deadline which is the date and time on which the user will use the desired construction machinery or the date and time on which delivery should be completed, and the delivery conditions further include a condition that the machinery be delivered by the delivery deadline.

4. The construction machinery delivery plan creation support system according to claim 3, wherein the evaluation means can differentiate evaluation values ​​based on a predetermined priority for each vehicle, within the scope in which the distribution of vehicles in the delivery plan pattern is to be carried out by vehicles used by the user.

5. A construction machinery delivery plan creation support system according to any one of claims 1 to 4, wherein the delivery plan patterns generated by the genetic algorithm consist of 100 or more individuals, the delivery plan patterns for genetic manipulation in the next generation consist of 10% of the individuals selected from the generated delivery plan patterns, and the selected delivery plan patterns are used to generate eight times the number of delivery plan patterns by crossover and one time the number of delivery plan patterns by mutation.

6. The construction machinery delivery plan creation support system according to claim 5, wherein, when the delivery plan patterns generated by the genetic algorithm number 1,000 individuals, the genetic manipulation is performed within the range of 40 to 50 generations.

7. A method for creating a delivery plan for distributing multiple construction machines among multiple vehicles based on a request from a user, A step of registering basic information in a database, which includes at least user information about the user who will be the source of the delivery, vehicle information about the vehicle used by the user to deliver the construction machinery, construction machinery information about the construction machinery to be delivered, and contractor information about the contractor when delivery is outsourced. A step of storing delivery conditions, which are associated with the aforementioned basic information and are set in advance to realize delivery, in a first storage means; A step of storing user-related request information, including the content of the request made by the user, in a second storage means, The process includes processing steps for processing the information stored in the database, the first storage means, and the second storage means, The aforementioned processing step is: A processing step to generate a delivery plan pattern by allocating construction machinery for a predetermined number of days to the vehicle and the contractor based on the request information and the basic information, An evaluation step for evaluating the delivery plan pattern generated by the processing step, A determination step in which the superiority or inferiority of the delivery plan pattern is determined based on the evaluation results evaluated by the evaluation means, Equipped with, The processing step involves the generation of multiple delivery plan patterns by a genetic algorithm. The selection of the delivery plan patterns to perform the aforementioned genetic manipulation involves sequentially selecting a predetermined number of delivery plan patterns from among a plurality of delivery plan patterns generated by the genetic algorithm, starting with those that have the highest evaluation values ​​as evaluated by the evaluation means. The aforementioned genetic algorithm is executed until a distribution plan pattern for a given generation is generated. The evaluation step assigns the lowest evaluation value to delivery plan patterns that do not meet the delivery conditions saved in the first saving step, while assigning a high evaluation value to delivery plan patterns that result in a smaller allocation to the contractor. The determination step selects the delivery plan pattern with the highest evaluation value, as evaluated by the evaluation step, from among a plurality of delivery plan patterns generated in a predetermined generation, and determines it as the final delivery plan. A method for creating a delivery plan for construction machinery, characterized by the following features.

8. A program for creating a delivery plan for construction machinery, characterized in that it causes a computer to execute each step in the processing process described in claim 7.

Images