WO2012176925A1

WO2012176925A1 - Configuration management device, configuration management method, and program-recording medium storing configuration management program

Info

Publication number: WO2012176925A1
Application number: PCT/JP2012/066307
Authority: WO
Inventors: 育大網代; 一男矢野尾
Original assignee: 日本電気株式会社
Priority date: 2011-06-20
Filing date: 2012-06-20
Publication date: 2012-12-27
Also published as: JP5949764B2; JPWO2012176925A1

Abstract

In a configuration management device which derives combinations of a plurality of types of apparatuses having two different indices, a combination of apparatuses in which the total of one of the indices is greater than or equal to a predetermined value, and the total of the other index is minimized, is derived with a lower amount of computation by performing a search using the branch-and-bound method.

Description

[Name of invention determined by ISA based on Rule 37.2] Configuration management device, configuration management method, and program recording medium for storing configuration management program

The present invention relates to a configuration management device, a configuration management method, and a configuration management program for deriving a combination of devices, and more particularly to a configuration management device, a configuration management method, and a configuration management program for deriving a combination of devices that satisfy a predetermined condition. .

Patent Document 1 describes a system proposal method and a system proposal device that outputs a configuration that satisfies performance requirements and price requirements within a predetermined error range as a configuration to be proposed. According to this document, the system configuration definition unit of the system proposing device accepts the components of the performance evaluation target system, the connection relationship of the components, the behavior of the application, the specifications and parameters of each resource, the performance requirements, and the price requirements. Examples of the component include a database management system, a server side application, and a client side application. The connection relationship of the component indicates how the component is connected to the network. The behavior of the application is, for example, the number of processing steps, the number of times of issuing a DB (Database) search function and a DB update function. The specifications and parameters of each resource are the number of CPUs (Central Processing Units) of the computer, MIPS (Million Instructions Per Second) values, disk seek time, setting parameters of the database management system, and the like. System performance is response time. The performance requirement is a value representing the performance of the system that outputs the configuration by the system proposing device, such as a response time of 5 seconds. The price requirement is a value representing a price of a system for which the system proposing device outputs a configuration, such as 10 million yen.
The optimal configuration calculation unit of the system proposing device of this document formulates a configuration combining the components of the performance evaluation target system, the connection relationship of the components, the behavior of the application, the specifications and parameters of each resource, and the response to each configuration The time is calculated by the performance simulator. In addition, the optimum configuration calculation unit calculates the total price of the hardware and software of each formulated configuration. Based on the calculated response time and total price of each configuration, the optimal configuration calculation unit outputs a configuration that satisfies the performance requirement and the price requirement within a predetermined error range. Thereby, the structure which satisfy | fills performance requirements and price requirements can be output in a short time.
Patent Document 2 describes a system construction mechanism that extracts components (parts) that match preset conditions, determines the system configuration, and outputs the system configuration. The component is, for example, a CPU, a memory, or a disk. The conditions for extracting parts are system specifications such as the number and number of CPU clocks, memory capacity, and disk capacity, and user requirements such as budget and delivery date. The system construction mechanism of this document includes an extraction function and a determination function.
The extraction function extracts components such as a CPU, memory, and disk that satisfy the system specifications from the component specification table. The judgment function excludes parts that cannot be delivered before the delivery deadline in the user requirements from the parts extracted by the extraction function. In addition, the judgment function excludes parts that would not satisfy the price limit when adopted. The system construction mechanism outputs information on the parts that have not been removed.
Patent Document 3 describes a configuration derivation method and a configuration derivation device for calculating a system configuration in which the probability that a transaction processing time exceeds a predetermined response time is equal to or less than a predetermined value and the price is minimum. . The configuration deriving device of Patent Document 3 obtains a system configuration in which a performance function, which is a function representing performance, satisfies a constraint condition and minimizes a price function as an objective function, using Lagrange's undetermined multiplier method. The performance function and the price function are functions having system configuration data as variables. The constraint condition is that the probability that the transaction processing time is equal to or longer than the response time A is equal to or lower than B when executing the performance guarantee target job. The system configuration parameters are CPU performance S, CPU number NCPU, total memory capacity M, I / O (Input / Output) processing speed I, and number of jobs that can be executed simultaneously NJ.
The configuration deriving device of Patent Document 3 repeats the following process while increasing the number of CPUs by 1 from 1 to the maximum number of CPUs that can be added. The configuration deriving device of Patent Document 3 first calculates the optimum system parameters (S, M, I) when the number of CPUs NCPU is fixed by the Lagrange's undetermined multiplier method. Next, the configuration deriving device of Patent Literature 3 searches the price data for an actual system parameter set having the smallest distance to the optimum system parameter obtained by calculation.
The configuration deriving device of Patent Document 3 is a parameter set of a system configuration that minimizes the price from an optimized system configuration for each number of CPUs derived by increasing the number of CPUs from 1 to the maximum number of CPUs that can be added. (NCPU, S, M, I) is determined. Next, the configuration deriving device of Patent Document 3 calculates the number NJ of jobs that can be executed simultaneously for the calculated system variables (S, NCPU, M, I).

JP 2002-183416 A JP 2002-020568 A JP 2004-030292 A

The system proposing device of Patent Document 1 formulates a plurality of configurations each combining system components, connection relationships between the components, application behavior, use of each resource, and parameter specification. And this system proposal apparatus calculates the response time and price of all the structure which were formulated, and outputs the structure which satisfy | fills a performance requirement and a price requirement in the range of a predetermined | prescribed error.
Therefore, the technique of Patent Document 1 has a problem that it is not possible to output the cheapest configuration that satisfies a certain level of performance.
The system construction mechanism of Patent Document 2 extracts parts for each type that satisfy the conditions defined for each type of part, and the total price of the parts selected for each type from the extracted parts is less than a predetermined value Select the combination of parts to become.
In the technique of Patent Document 2, a combination of a plurality of types of components having two types of indexes, and the total of the other indexes among the combinations in which the total value of one of the components included in the combination is equal to or greater than a predetermined value. There was a problem that a combination having the smallest value could not be selected.
The system configuration deriving device of Patent Document 3 calculates optimal system configuration parameters for all the CPU numbers from 1 to the maximum number of CPUs that can be added, and selects a system configuration parameter set with the lowest price from among them. It was.
The technique of Patent Document 3 has a problem that it is assumed that the performance of each CPU is the same.
An object of the present invention is to calculate a combination of a plurality of types of devices having two different indexes with a small number of device combinations in which the total of one index is equal to or greater than a predetermined value and the total of the other index is minimum. The object is to provide a configuration management device derived by quantity.

The configuration management apparatus of the present invention is configured of devices associated with the first index value P1 and the second index value P2 for each of a plurality (J) of types, and a set of devices in which the total of P1 is equal to or greater than the necessary value N Is the satisfaction set, and the minimum value in the total satisfaction set of P2 is the minimum total, the device storage means for storing P1 and P2 for each device type, and the initial value is a value equal to or greater than the minimum total Temporary solution storage means for storing information representing a provisional solution and a provisional set in which the sum of P2 is a set of devices that is the provisional solution, and 1st to jth (j is an integer from 1 to J-1) Search range storage means for storing n (1) to n (j), which is the number of devices of the same type, and N, and n (j + 1) is changed from 0 to n (1) to n (j ) Only the 1st to jth types of devices and the j + 1th type of devices form a satisfaction set. N (j + 1) is stored in the search range storage means while sequentially increasing to a predetermined value (j + 1) that is less than or equal to the minimum number p (j + 1) of the j + 1th type of devices, and N and 1 to j + 1th The search that is a collection of satisfying sets in which the number of the first to j + 1 type devices from P1 and P2 of the types of devices is n (1) to n (j + 1) stored in the search range storage means. The lower limit value of the sum of P2 of the set included in the range (j + 1) is calculated by removing the restriction that the number of devices is an integer, and when the lower limit value is smaller than the provisional solution, the search range (j + 1) ) From which the sum of P2 is smaller than the provisional solution, and the information indicating the sum of P2 of the extracted fulfillment set and the extracted satisfaction set is stored in the provisional solution storage means as a provisional solution and provisional set. Store Search execution means that does not search the search range (j + 1) when the lower limit value is larger than the provisional solution; and output means that outputs the provisional set as a minimum set after the search is executed by the search execution means; including.
The configuration management method according to the present invention includes a set of devices each of which is associated with the first index value P1 and the second index value P2 for each of a plurality of types (J), and the total of P1 is equal to or greater than the necessary value N. Is the satisfaction set and the minimum value in the total satisfaction set of P2 is the minimum total, P1 and P2 are stored in the device storage means for each type of device, and the initial value is equal to or greater than the minimum total. Information representing a provisional solution and a provisional set in which the sum of P2 is a set of devices for which the total of P2 is the provisional solution is stored in the provisional solution storage means, and 1st to jth (j is an integer from 1 to J-1) N (1) to n (j), which are the number of devices of the type, are stored in the search range storage means, N is input, n (j + 1) is changed from 0 to n (1) to n (j ) When a sufficiency set is composed of only the 1st to jth types of devices and the j + 1th type of devices While sequentially increasing to a predetermined value (j + 1) that is less than or equal to the minimum number p (j + 1) of the j + 1th type devices, n (j + 1) is stored in the search range storage means, and N and the 1st to j + 1th type devices From the devices P1 and P2, the search range (1 to j + 1) is a collection of satisfying sets in which n (1) to n (j + 1) are stored in the search range storage means. j + 1) is calculated by removing the lower limit value of the total of P2 of the set included in the set, and the number of devices is an integer, and when the lower limit value is smaller than the provisional solution, from the search range (j + 1) , P2 is extracted with a sum smaller than the provisional solution, and the information indicating the sum of P2 and the extracted satisfaction set is stored as a provisional solution and provisional set in the provisional solution storage means. ,in front If the lower limit value is greater than the interim solution, without searching of the search range (j + 1), after the execution of the search, and outputs the provisional set as the minimum set.
The configuration management program of the present invention is composed of devices associated with the first index value P1 and the second index value P2 for each of a plurality (J) of types, and a set of devices in which the total of P1 is equal to or greater than the necessary value N Is the satisfaction set, the minimum value in the total satisfaction set of P2 is the minimum total, the computer is stored for each device type, device storage means for storing P1 and P2, and the initial value is equal to or greater than the minimum total Temporary solution storage means for storing information representing a provisional solution that is a value, and a provisional set that is a set of devices for which the sum of P2 is the provisional solution, and 1st to jth (j is any one of 1 to J-1) Search range storage means for storing n (1) to n (j), which is the number of devices of type (integer), and N, and n (j + 1) is changed from 0 to n (1), respectively. n (j) 1st to jth type devices and j + 1th type devices Storing n (j + 1) in the search range storage means while sequentially increasing to a predetermined value (j + 1) which is less than or equal to the minimum number p (j + 1) of j + 1 type devices when a satisfaction set is configured with From P1 and P2 of N and 1 to j + 1 type devices, the number of 1 to j + 1 type devices is n (1) to n (j + 1) stored in the search range storage means The lower limit value of the total of P2 of the set included in the search range (j + 1) that is a set of sets is calculated by removing the restriction that the number of devices is an integer, and the lower limit value is smaller than the provisional solution. In this case, a satisfaction set in which the sum of P2 is smaller than the provisional solution is extracted from the search range (j + 1), and information indicating the sum of P2 of the extracted satisfaction sets and the extracted satisfaction set is obtained as the provisional solution and the provisional set. As before If the lower limit is larger than the provisional solution, the search execution means that does not search the search range (j + 1) and the temporary set is minimized after execution of the search by the search execution means. Operate as output means for outputting as a set.

In the present invention, a combination of a plurality of types of devices having two different indices that satisfies the condition that the sum of one index is equal to or greater than a predetermined value and that minimizes the other index is minimized. There is an effect that it can be derived.

It is a block diagram showing the example of a structure of the configuration management apparatus 1 of 1st Embodiment. It is a flowchart showing the example of operation | movement in which the structure management apparatus 1 of 1st Embodiment acquires the parameter | index of an apparatus. It is a flowchart showing the example of operation | movement of the configuration management apparatus 1 of 1st Embodiment. It is a flowchart showing the example of operation | movement of the search of the structure management apparatus 1 of the 2nd modification of 1st Embodiment and 1st Embodiment. It is a figure showing the 1st example of apparatus information. It is a figure showing the example of index ratio It is a figure showing the example of a provisional solution and a provisional set. It is a figure showing the 1st example of the search tree which illustrated the search process of the solution by the configuration management apparatus 1 of 1st Embodiment. It is a flowchart showing the example of the operation | movement in which the configuration management apparatus 1 of the 1st modification of 1st Embodiment calculates an initial solution. It is a flowchart showing the example of the operation | movement of the search of the configuration management apparatus 1 of the 1st modification of 1st Embodiment. It is a flowchart showing the example of the operation | movement which the structure management apparatus of the 1st and 2nd modification of 1st Embodiment calculates a lower limit. It is a figure showing the 2nd example of the search tree which illustrated the search process of the solution by the configuration management apparatus of 1st Embodiment. It is a figure showing the 2nd example of apparatus information. It is a flowchart showing the example of the operation | movement in which the configuration management apparatus 1 of the 2nd modification of 1st Embodiment calculates an initial solution. It is a block diagram showing the example of a structure of the configuration management apparatus 1 of 2 embodiment.

Next, embodiments of the present invention will be described in detail with reference to the drawings.
(First embodiment)
FIG. 1 is a block diagram showing the configuration of the configuration management apparatus 1 according to the first embodiment of the present invention.
Referring to FIG. 1, the configuration management apparatus 1 of the present embodiment includes a device storage unit 10, a provisional solution storage unit 11, a search range storage unit 12, a search execution unit 13, and an output unit 14. The configuration management device 1 may include an initial solution calculation unit 15 and an index acquisition unit 16. A terminal 2 is connected to the configuration management apparatus 1. A server 3 including a file 31 may be connected to the configuration management device 1. Note that the server 3 does not exist, and the configuration management device 1 may include the file 31.
The configuration management apparatus 1 according to the present embodiment can be realized by a computer and a program for controlling the computer, dedicated hardware, or a combination of the computer and the program for controlling the computer and dedicated hardware. The device storage unit 10, the provisional solution storage unit 11, and the search range storage unit 12 can be realized by a memory or a hard disk device included in the computer. Further, the search execution unit 13, the output unit 14, the initial solution calculation unit 15, and the index acquisition unit 16 are dedicated for realizing the function of each unit read into a memory from a recording medium that stores a program, for example. And a processor that executes the program. Alternatively, some or all of the device storage unit 10, the provisional solution storage unit 11, the search range storage unit 12, the search execution unit 13, the output unit 14, the initial solution calculation unit 15, and the index acquisition unit 16 Also, it can be realized by a dedicated circuit for realizing the function of each unit.
The device storage unit 10 stores, for each type, two indexes of a plurality of types (J types) of devices having at least two different indexes. The device is, for example, a calculation device such as a server, a CPU, or the like. Of the two indicators, the first indicator (P1) is, for example, an indicator representing the performance of the device. The index indicating the performance is, for example, an index such as a throughput, a MIPS value, or a FLOPS (Floating point number Operations Per Second) value. The first index may be an index representing reliability, availability, and the like, for example. Of the two indicators, the second indicator (P2) is, for example, an indicator representing the cost of the device. The index representing the cost is an index that affects the cost of system construction and operation, such as price, power consumption, amount of exhaust heat, weight, installation space, and the like. Further, the second index may be a value such as the number of parts constituting the device, for example. These two indices for a plurality of devices are the sum of the two indices for each device.
For example, when the device is a computing device, the first index represents performance, and the second index is price, the performance of the computer system configured by a set of one or more devices selected from a plurality of types of devices Is represented by the sum of the first indices of the devices included in the set. The price of this computer system is the total of the second index of the devices included in the set.
In the following description, the set in which the sum of the second indexes is the smallest among the set of devices (satisfaction set) in which the sum of the first indexes exceeds a predetermined value (necessary value) is the minimum set. The sum of the second index of the devices belonging to the minimum set is the minimum total. The necessary value may be, for example, a value input to the device configuration apparatus 1 by the user via an input unit (not shown) provided in the terminal 2. In addition, a set in which the sum of the second indices is the smallest among the satisfaction sets in which the search execution unit 13 described later has already performed a search described later is a provisional set. And the sum total of the 2nd parameter | index of the apparatus which belongs to a provisional set is a provisional solution.
In the present embodiment, the number of devices of each type that can be included in a set of devices is not limited. In general, the problem of deriving the optimal combination of devices under a predetermined condition without limiting the number of devices of each type that can be included in the set of devices is called an unlimited problem. For example, when an information processing system composed of one or more computers selected from a plurality of types of computers is newly introduced, if there is no limit on the number of devices of each type to be introduced, the predetermined performance is satisfied, and the most The problem of deriving a low-cost computer is an unlimited problem.
The 1st parameter | index and 2nd parameter | index of each apparatus which the apparatus memory | storage part 10 memorize | stores should just be given previously. Alternatively, the index acquisition unit 16 may acquire these indexes from the file 31 as described later. Further, for example, the user may input these indices into the configuration management apparatus 1 via the terminal 2 and store them in the device storage unit 10.
The initial solution calculation unit 15 calculates the initial value (initial solution) of the provisional solution using, for example, a method (heuristics) for obtaining a reasonably good solution with high probability. Heuristics does not necessarily have to be a way to calculate the minimum total. The heuristics may be any method that can extract a set of devices in which the total of the first index exceeds the required value in a shorter time than the method of always calculating the minimum total. Since heuristics is not necessarily a method for calculating the minimum total, the initial value of the provisional solution calculated by the initial solution calculation unit 15 is a value equal to or greater than the minimum total. An example of heuristic used by the initial solution calculation unit 15 will be described later. Examples of general heuristics include genetic algorithms, annealing methods, and algorithms using neural networks.
The initial solution calculation unit 15 stores the calculated initial value of the temporary solution and information indicating a corresponding set of devices (initial value of the temporary set) in the temporary solution storage unit 11.
The provisional solution storage unit 11 stores provisional solutions and information representing provisional sets. As described above, the provisional solution stored in the provisional solution storage unit 11 in the initial state and the information representing the provisional set represent, for example, the initial value of the provisional solution calculated by the initial solution calculation unit 15 and the initial value of the provisional set. Any information may be used.
In the initial state, the provisional solution storage unit 11 may not store the initial value of the provisional solution calculated by the initial solution calculation unit 15, but may store in advance a sufficiently large value that can be expressed on a computer as a provisional solution. In this case, the provisional solution storage unit 11 may not store a provisional set corresponding to the provisional solution in the initial state. Furthermore, in this case, the initial solution calculation unit 15 may not exist. However, in this case, there is a high possibility that the search efficiency is reduced.
The search range storage unit 12 stores n (1) to n (j), which is the number of devices of the first to jth (j is an integer of 1 to J-1) type. The search execution unit 13 described later can read the number of devices stored in the search range storage unit 12 and set a set of devices determined based on the read number of devices as a search target. For example, when the search range storage unit 12 stores n (1) to n (j), which is the number of devices of the 1st to jth types, the search execution unit 13 selects the 1st to jth type of devices. A set of devices whose number of devices is n (1) to n (j), respectively, may be searched. In this case, n (1) to n (j) stored in the search range storage unit 12 represent a search range that is a set of devices to be searched.
The search execution unit 13 selects, from a search range that is a set of sets (the above-described satisfaction set) in which the sum of the first indices is equal to or greater than a necessary value among a set of devices composed of a plurality of types of devices, A search is performed to extract a combination of devices whose total index is equal to or less than the provisional solution. The search execution unit 13 divides the search range into a plurality of search ranges during the search. And the search execution part 13 determines the presence or absence of possibility of including the minimum set for every divided | segmented search range. The search execution unit 13 excludes a search range that may not include the minimum set as a result of the determination from the search target. The search will be described in detail in the description of the operation of the present embodiment. The provisional set after the search execution unit 13 completes the search for all the search ranges that are not excluded from the search targets is the minimum set.
After the search by the search execution unit 13 is completed, the output unit 14 outputs information representing the temporary set stored in the temporary solution storage unit 11 to the terminal 2 as information representing the minimum set.
The terminal 2 displays information representing the minimum set output by the output unit 14. The display by the terminal 2 may be in any format as long as it can be understood how many types of devices are included in the minimum set.
The index acquisition unit 16 extracts, for example, the first index and the second index of each device included in the file 31 stored in the server 3 and stores them in the device storage unit 10.
The server 3 may be the same device as the configuration management device 1, for example, may be another device that is communicably connected via a network.
The file 31 includes, for example, one or a plurality of files that make up one or a plurality of web pages in which specifications of each device are disclosed on the Internet by a vendor of each device, and specifications of each device. It is a file that makes up the database that contains it.
For example, the index acquisition unit 16 acquires the file 31 in accordance with a user instruction that specifies a URI (Uniform Resource Identifier) that identifies a web page or the like via an input unit (not illustrated) of the terminal 2. The index acquisition unit 16 uses the first index and the second index of each device included in the file 31 as a key, for example, a character string such as “performance”, “price”, “power consumption”, “number of parts”, etc. Extract. The index acquisition unit 16 stores the extracted first index and second index of each device in the device storage unit 10. In addition, the index acquisition unit 16 may acquire the first index and the second index from the database according to an instruction that specifies a query to the database or the like via an input unit (not illustrated) of the terminal 2. The index acquisition unit 16 may extract information for obtaining an index (for example, the CPU and the number of clocks of the device) from the file 31. Then, the index acquisition unit 16 may acquire the first index and the second index from the database in which the index is stored based on the extracted information.
The index acquisition unit 16, the server 3, and the file 31 may not exist. In that case, the device storage unit 10 may store in advance two indexes of a plurality of types (J types) of devices having at least two different indexes for each type.
Next, the operation of the configuration management apparatus 1 of this embodiment will be described in detail with reference to the drawings.
FIG. 2 is a flowchart illustrating an example of an operation in which the configuration management apparatus 1 according to the present embodiment acquires a device index.
Referring to FIG. 2, the index acquisition unit 16 first receives an instruction via, for example, an input unit (not shown) of the terminal 2 and acquires a file 31 including a first index and a second index for each type of device. (Step S61).
Next, the index acquisition unit 16 extracts the first index and the second index for each type of device from the acquired file 31 (step S62).
The index extraction unit 16 stores the extracted first index and second index for each device type in the device storage unit 10 in association with the device type (step S63).
FIG. 3 is a flowchart showing an overall example of an operation in which the configuration management apparatus 1 of the present embodiment extracts a minimum set (solution).
Referring to FIG. 3, the search execution unit 13 of the configuration management apparatus 1 first receives the necessary value N via an input unit (not shown) of the terminal 2 (step S1).
Next, the initial solution calculation unit 15, for example, based on the necessary value N received from the search execution unit 13 and the first index and the second index of each device type received from the device storage unit 10. The initial value of the provisional solution is calculated using tics (step S2). The initial solution calculation unit 15 stores the temporary solution storage unit 11 together with information representing the temporary set corresponding to the temporary solution.
Examples of heuristics in which the initial solution calculation unit 15 of the present embodiment calculates the initial value of the provisional solution are as follows. For example, the initial solution calculation unit 15 may calculate a ratio (index ratio) of the first index to the second index for each type of device. Then, the initial solution calculation unit 15 may calculate the minimum number of devices when the satisfaction set is configured by only the types of devices having the largest index ratio. The initial solution calculation unit 15 may set an initial value of the provisional set to a set composed of the types of devices having the smallest calculated index number and the largest index ratio. Then, the initial solution calculation unit 15 may set the sum of the second indices of the initial values of the temporary set as the initial value of the temporary solution. When the first index represents performance and the second index represents cost, the index ratio represents cost performance (cost performance ratio).
FIG. 5 is a diagram illustrating an example of device information (first index and second index for each type of device) stored in the device storage unit 10. In the example of FIG. 5, the device storage unit 10 stores a first index and a second index of three types of devices, model A to model C.
The initial solution calculation unit 15 reads these indexes from the device storage unit 10 and calculates an index ratio for each type of device.
FIG. 6 is a diagram illustrating an index ratio calculated from the example of the device information illustrated in FIG. Referring to FIG. 6, in this case, model C has the largest index ratio, and the index ratio is 0.88.
For example, when the required value N is 15, the first index of the model C is 7, and 15/7 is about 2.14. Therefore, in order to satisfy the required value with only the model C, three models C are required. is necessary. The total of the second index of the set composed of the three models C is 24. In this case, the initial solution calculation unit 15 may store 24 as the temporary solution in the temporary solution storage unit 11. Then, the initial solution calculation unit 15 may store information representing a set including three models C as a temporary set in the temporary solution storage unit 11.
FIG. 7 is a diagram illustrating information representing the provisional solution and the provisional set stored in the provisional solution storage unit 11 when the provisional solution is 24 and the provisional set is a set composed of three models C.
Next, the search execution unit 13 includes a plurality of types of devices in combination, and the minimum set in which the total of the second index is the minimum from the set in which the total value of the first index is equal to or greater than the necessary value N A search, which is a process for extracting, is performed (step S3).
The search execution unit 13 performs a search based on the branch and bound method. Although details will be described in the description of FIG. 4, the search based on the branch and bound method is roughly as follows.
The search execution unit 13 divides the search range into a plurality of search ranges before exhaustively searching the search range that is a set of devices. Then, the search execution unit 13 determines whether or not there is a possibility that the minimum set is included in the divided search range. The search execution unit 13 divides the search range by setting a set of devices that are included in the search range and the same number of devices of a certain type of device as a new search range. Good.
As described above, when the search range storage unit 12 stores n (1) to n (j), the search execution unit 13 searches for n (j) from n (1). What is necessary is just to read from the range memory | storage part 12. And the search execution part 13 should just define a search range by n (j) read from n (j). The search execution unit 13 may set a set of devices having the number of first to jth types of devices from n (1) to n (j) as a search range (search range (j)).
For the search range (j) determined in this way when the search range storage unit 12 stores n (1) to n (j), the search performed by the search execution unit 13 is “search (j)”. It is. In the search (j), the search execution unit 13 divides the search range (j) read from the search range storage unit 12 by n (1) to n (j) into a plurality of search ranges. Then, the search execution unit 13 determines whether or not there is a possibility that the minimum set is included in the divided search range. For example, the search execution unit 13 may divide the search range (j) into a collection of sets in which the number of devices of the (j + 1) th type is the same. For example, when the search range (j) is divided and a set of devices having a certain value n (j + 1) of the number of j + 1th type devices is set as a new search range, the search execution unit 13 uses the value. A certain n (j + 1) may be stored in the search range storage unit 12. In this case, the search range determined by n (1) to n (j + 1) stored in the search range storage unit 12 is the search range (j + 1). The search performed by the search execution unit 13 for the search range (j + 1) is the search (j + 1).
Search (0) is a search performed by the search execution unit 13 for the search range when the search range storage unit 12 does not store the number of devices of any type. The search range when the search range storage unit 12 does not store the number of devices of any type is the initial value of the search range (search range (0)). The search range (0) will be described later.
The search execution unit 13 may determine whether or not there is a possibility that the search range includes the minimum set by comparing the theoretical lower limit value of the second index of the device included in the search range with the provisional solution. The search execution part 13 should just calculate this lower limit, for example using a relaxation method. When the lower limit value is larger than the provisional solution, there is no possibility that the search range includes a set in which the sum of the second indices is smaller than the provisional solution. In this case, the search execution unit 13 excludes the search range from the search target (pruning) and does not perform a search for the search range. If the lower limit value is smaller than the provisional solution, these sets may include a set in which the sum of the second indices is a smaller value than the provisional solution. In this case, the search execution unit 13 executes a search for these sets. This determination process for determining whether to perform a search is a limited operation.
The search execution unit 13 may further divide the search range determined to require search into a plurality of search ranges and determine whether a search is necessary or unnecessary in each of the divided search ranges. The search execution part 13 should just perform a search only in the search range determined to require a search among the further divided search ranges.
The search execution unit 13 does not have to include in the search range a sufficiency set in which the sum of the second indices is clearly not the minimum sum. For example, when the satisfaction set is composed of only the first type of device, the total of the second index of the satisfaction set including the first type of devices more than the minimum number of devices p (1) is Obviously not the minimum sum. The search execution unit 13 may exclude such a set from the search range. For example, the search execution unit 13 calculates the minimum number p (1) of devices when a satisfaction set is configured by only the first type of device, and the first device is 0 or more and not more than p (1). A collection of a certain satisfaction set may be used as an initial value of the search range (search range (0)). The search execution unit 13 reads the first index and the second index of the first type of device from the device storage unit 10, and based on the necessary value N and the read first index and second index. , P (1) may be calculated. Note that the order of device types may be any order, but search efficiency may be improved if the index ratio is increased or decreased.
The process of step S3 will be described in detail in the description of FIG.
After the search by the search execution unit 13 is completed, the output unit 14 reads information representing the temporary set stored in the temporary solution storage unit 11 and outputs the information to the terminal 2 (step S4).
For example, the provisional solution stored in the provisional solution storage unit 11 is 20, and the information indicating the provisional set represents that “model A is one, model B is one, and model C is one”. If it is information, the output unit 14 may output information representing the provisional set to the terminal 2 in a predetermined format, for example. For example, even if the output format is data such as “((A, 1), (B, 1), (C, 1))”, “model A: 1 unit, model B: 1 unit, model A character string such as “C: 1” may be used. The output unit 14 may output a provisional solution.
FIG. 4 is a flowchart showing an example of the operation of the configuration management apparatus 1 during the search in step S2.
Referring to FIG. 4, the search execution unit 13 is given a j indicating the order of the device types (j is any integer from 1 to J−1) and starts the search process.
The search range storage unit 12 stores the number of first to j-th types of devices (n (1) to n (j), respectively) at the start of the search operation. The set of devices to be searched in search (j) is a set of devices in which the number of first to jth devices is n (1) to n (j), respectively. Therefore, the search range (j), which is a collection of devices to be searched in the search (j), is stored in each type of device stored in the search range storage unit 12 at the start of the search (j) operation. The number n (1) to n (j) is determined. However, when j is 0, the number of devices is not stored in the search range storage unit 12.
When j is 0 (step S11, N), the search range is an initial value (search range (0)). The search (0) in step S3 in FIG. 3 means that the search execution unit 13 performs the process shown in FIG. 4 on the search range (0).
When j is larger than 0 (step S11, Y), the search execution unit 13 reads the number n (1) to n (j) of the first to j−1th types of devices from the search range storage unit 12. (Step S12). In this case, the search range is a collection of a set of devices in which the number of first to jth devices is n (1) to n (j), respectively.
Next, the search execution unit 13 sets the minimum of the (j + 1) th device when the satisfying set is composed of the 1st to jth types of devices having the number n (1) to n (j) and the (j + 1) th device. The number p (j + 1) is calculated (step S13). The search execution unit 13 calculates p (j + 1) from the necessary value N, the values from n (1) to n (j), and the first index of the first to j + 1th types of devices read from the device storage unit 10. ) May be calculated. In this way, the minimum number p (j + 1) of j + 1 type devices when a satisfaction set is configured by only the 1st to j + 1 type devices according to the values from n (1) to n (j). The method of calculating is “fine method”. Specifically, the search execution unit 13 may first calculate the sum of the first indices of n (1) to n (j) 1st to jth types of devices. Next, the search execution part 13 should just calculate the difference of the sum of the required value N and the calculated 1st parameter | index. The search execution unit 13 may divide the calculated difference by the value of the first index of the (j + 1) th type device and set the value obtained by rounding up the fractional part to p (j + 1). The search execution unit 13 is a set in which the number of first to jth type devices is n (1) to n (j), respectively, and the number of j + 1th type devices is p (j + 1) +1 or more. Need not be the target of the search.
Next, the search execution unit 13 calculates the maximum number r (j + 1) of the (j + 1) th type devices included in the search target set (step S14).
In the present embodiment, since the number of each type of device is not limited, the search execution unit 13 may set p (j + 1) to r (j + 1). When the number of j + 1 type devices that can be included in the set is limited, the search execution unit 13 determines the smaller one of p (j + 1) and the maximum number of j + 1 type devices as r (j + 1). ). In this case, the value of r (j + 1) is less than or equal to p (j + 1). The case where the number of devices that can be included in the set is limited will be described in detail in Modification 1 and Modification 2 described later.
Alternatively, the search execution unit 13 may calculate the minimum number of devices when the satisfaction set is configured by only the j + 1th device, and may be set to r (j + 1). As described above, the minimum number when the satisfaction set is configured only by the (j + 1) th type of device is calculated and set as a search target regardless of the number of the first to jth types of devices constituting the device set. The method of setting the maximum number of j + 1 type devices in the set of devices is the “rough method”. When calculating r (j + 1) by a rough method, the search execution unit 13 divides the necessary value by the first index of each model, and if it is not divisible, the value obtained by rounding up after the decimal point is set to r (j + 1). Good. When the number of j + 1 types of devices that can be included in the set is limited, the search execution unit 13 determines the minimum number of devices and the j + 1 type of devices when the satisfaction set is configured by only the j + 1th device. It is sufficient that the smaller one of the maximum number is set to r (j + 1). When calculating r (j + 1) by a rough method, the search execution part 13 does not need to perform the process of step S13.
When r (j + 1) is calculated by the fine method described above, the value of r (j + 1) varies depending on the number of first to jth types of devices. However, when r (j + 1) is calculated by a rough method, the value of r (j + 1) is constant regardless of the number of devices of the first to jth types. For example, if the required value is 15, and the j + 1 type of device is the above-mentioned model A, r (j + 1) is 15/3 = 5, so the number of devices of the first to jth types is It is 5 regardless. Similarly, if the j + 1 type device is the above-described model B, r (j + 1) is 3 because it is 15/5 = 3, regardless of the number of the first to jth types of devices. If the (j + 1) th type device is the above-mentioned model C, r (j + 1) is 15/7 = 2.14. . . Therefore, it is 3 regardless of the number of devices of the first to jth types. R (j + 1) calculated by the rough method is a value equal to or larger than r (j + 1) calculated by the fine method.
By the way, it is possible to illustrate a set of devices to be searched by the search execution unit 13 in the form of a search tree. The search execution unit 13 can perform processing by representing a set of devices to be searched as data of an existing search tree in an arbitrary format. The search execution unit 13 may appropriately expand the search range into a search tree and store the search range in the search range storage unit 12, for example.
FIG. 8 is a diagram illustrating a part of an example of a search tree in the example of the device illustrated in FIG. The model A, model B, and model C in FIG. 8 correspond to the first to third types of devices, respectively.
In the search tree of FIG. 8, circles represent nodes, and lines connecting the nodes represent branches. This search tree has four hierarchies. The node at the top (first hierarchy) is a root node, the node at the lowest (fourth hierarchy) is a leaf node, and the nodes between them are intermediate nodes.
The second to fourth layers correspond to the first to third types of devices (in order, model A, model B, and model C).
The numbers written on the nodes in each layer other than the top represent the number of corresponding models. Each path from the root node to the leaf node corresponds to a set of models. For example, the route from the root node to the number 0 of model A, the number 0 of model B, and the number 3 of model C in FIG. 5 represents a set of three models C. In the following description, for convenience, a node corresponding to “0 model A” is represented as (A, 0), and a set of devices is represented by ((A, 0), (B, 0), (C, 3). ).
For example, the node (A, 0) corresponds to a set of devices whose number n (1) of the first type (model A) is 0. That is, the node (A, 0) represents a search range represented by n (1) = 0. Similarly, nodes ((A, 0), (B, 0)) corresponding to “0 model A, 0 model B” are represented by n (1) = 0 and n (2) = 0. A search range in which the number n (1) of the first type (model A) is 0 and the number n (2) of the second type (model B) is 0 is represented.
The search execution unit 13 operates from the root node corresponding to the set of devices to the root node corresponding to the last (J-th) type device from the root node. This corresponds to extracting the minimum set by sequentially searching the route to the leaf node. For example, in the case of the example of FIG. 5, if the search execution unit 13 expands all the nodes, first, the nodes (A, 0) to (A, 5), which are the lower-level nodes of the root node, are expanded. To do. Next, the search execution unit 13 is the node (A, 0), which is the next lower node ((A, 0), (B, 0)) to ((A, 0), (B, 3)). Expand. The search execution unit 13 performs a search by expanding the node in the lower layer because the node in the lower layer of the node ((A, 0), (B, 0)) is a leaf node. Similarly, the search execution unit 13 performs a search by expanding the nodes in the lower layer of ((A, 0), (B, 3)) from ((A, 0), (B, 1)). When the expansion and search of all the nodes on the route including the node (A, 0) are completed, the search execution unit 13 similarly expands and searches all the nodes on the route including the node (A, 1). . However, the search execution part 13 of this embodiment does not necessarily expand and search all nodes, as will be described below.
The search execution unit 13 calculates the lower limit value of the second index of the set of devices corresponding to all the routes passing through the node in each expanded node, and when the lower limit value exceeds the provisional solution, “Pruning” is performed to exclude all routes passing through the node from the search target. The search execution unit 13 calculates the total of the second index of the set of devices corresponding to the route to the leaf node when the leaf node is expanded without performing pruning at the intermediate node, and is compared with the provisional solution. To do. When the total of the calculated second index falls below the provisional solution, the search execution unit 13 sets the total as a new provisional solution and the corresponding set of devices as a new provisional set.
In the above operation, step S14 corresponds to calculating the number of branches to the node in the lower layer when each node expands the node in the lower layer.
The number of branches from each node to the node one level lower is the branching coefficient. Since the number of devices for each type may be zero, for example, the branch coefficient in the node corresponding to the j−1th device is searched from the number of each of the 1st to j−1th devices corresponding to the node. This is a value obtained by adding 1 to r (j) calculated by the unit 13.
When the branch coefficient is calculated by a rough method, the nodes of the number of branch coefficients of the device type corresponding to the node one level below always branch from each upper node. Therefore, when the branch coefficient is calculated by a rough method, in the search tree corresponding to each device shown in FIG. 5, the number of leaf nodes corresponding to the total number of combinations of devices is 6 × 4 × obtained by adding the branch coefficients of each model. 4 = 96.
On the other hand, in the detailed method, the branching coefficient of the node below the third hierarchy corresponding to the second or more types of devices is calculated by the number of corresponding devices corresponding to the nodes on the route from the root node to the node. This is a method of calculating based on the difference between the total of the first index and the required value.
For example, in the case of a set corresponding to a route passing through a node represented by (A, 3), the total of the first indices of three models A corresponding to the node is 9. The branch coefficient of this node for which the branch count was calculated by the rough method is 15/5 = 3 (the model B of the model B calculated by the rough method described above) because the first index of the model B which is the second device type is 5. 4 obtained by adding 1 to r (j)). However, the sum of the second indexes of the nodes after the node (A, 3) is 6 or more, which is the difference between the necessary value 15 and the total 9 of the first indexes of the node (A, 3). It is. The branching coefficient of the node represented by (A, 3) calculated by a fine method is a value obtained by adding 1 to the value obtained by rounding up the value obtained by dividing 6 by 5 which is the first index of model B. It is 3.
The detailed method has a feature that it is necessary to calculate a branching coefficient for each node, and a large amount of calculation is required to calculate the total number of combinations when there are many layers (when there are many types of devices). The fine method is further characterized in that the number of nodes can be reduced and search efficiency can be improved. There is no possibility that the path corresponding to the minimum set passes through nodes other than the branch node based on the branch coefficient calculated by the fine method. Therefore, it is necessary to search only a route that passes through a branch node generated based on a branch coefficient calculated by a fine method.
On the other hand, the branch coefficient calculated by the rough method is a value equal to or greater than the branch coefficient calculated for each node by the fine method. Since the rough method calculates the branch coefficient for each type of device instead of calculating the branch coefficient for each node, the calculation amount for calculating the branch coefficient is smaller than that of the fine method. However, the branch coefficient calculated by the rough method becomes a value equal to or greater than the branch coefficient calculated for each node by the fine method, and usually the number of nodes is increased and the search efficiency is poor as compared with the fine method.
Next, the search execution unit 13 performs the following processing while sequentially increasing n (j + 1) representing the number of the j + 1th type of devices from 0 to r (j + 1) (step S15).
First, the search execution unit 13 stores n (j + 1) representing the number of j + 1-th type devices in the search range storage unit 12 (step S16). The search execution unit 13 searches for n (j + 1) in a form in which it is possible to determine that n (j + 1) represents the number of devices of the j + 1th type, for example, storing n (j + 1) in association with j + 1. What is necessary is just to store in the range memory | storage part 12. The number of devices of each type already stored in the search range storage unit 12 and n (j + 1) represent the search range (search range (j + 1)).
Next, the search execution unit 13 sets the lower limit of the total value of the second index of the set included in the search range (j + 1) represented by n (1) to n (j + 1) stored in the search range storage unit 12. A value is calculated (step S17).
The search execution part 13 should just calculate a lower limit by the relaxation method as mentioned above. The relaxation method is a method of calculating an optimal solution for a relaxation problem that eases the restriction of the original problem and makes it easy to explain. The optimal solution of the relaxation problem is generally a better solution than the optimal solution of the original problem. For example, when the original problem is a problem of calculating a minimum value, the minimum value that is the optimal solution of the relaxation problem of the problem is generally smaller than the minimum value that is the solution of the original problem. Therefore, the solution of the relaxation problem can be the lower limit value of the solution of the original problem. A typical mitigation problem is, for example, when the restriction that the device is used in units of one and the number of devices is an integer is imposed on the original problem, the restriction is relaxed and one device is Is a problem that allows the number of devices to be real numbers.
The relaxation of the restriction for the search execution unit 13 of the present embodiment to calculate the lower limit value may be to remove the restriction that the number of devices is an integer and to allow a real number. However, even when the number of devices is not an integer, the sum of the first index and the second index is the same as the case where the number of devices is an integer. It is the value multiplied by the number.
In the case where the number of devices is a real number, the minimum value of the total of the second index of the set of devices in which the total of the first index is equal to or greater than the predetermined value is the index ratio in which the total of the first index is the predetermined value. Is obtained by calculating the sum of the second index of the set of devices composed of only the types of devices having the largest. The sum of the second index of any set of devices in which the sum of the first index is equal to or greater than a predetermined value does not fall below this minimum value.
The number of 1st to j + 1th types of devices in the set included in the search range (j + 1) is n (1) to n (j + 1). In the following description, the remaining required value is a value obtained by subtracting the total of the first index of the 1st to j + 1th types of devices from n (1) to n (j + 1) from the required value. Also, the type of the device having the largest index value among the j + 2 to Jth types of devices is the model X.
The search execution unit 13 is included in the search range (j + 1), and a model of a satisfaction set composed of the first to j + 1th types of devices and the remaining types of devices (model X) having the largest index ratio. The minimum number of Xs may be calculated by removing the restriction that the number of models X is an integer. The search execution unit 13 calculates the total of the second index of the first to j + 1 type devices from n (1) to n (j + 1) units and the calculated minimum number of models X as the search range (j + 1). The lower limit value of the total of the second indices may be used.
For this purpose, first, the search execution unit 13 may read the first index of the first to j + 1th types of devices from the device storage unit 10. Moreover, the search execution part 13 should just read n (1) to n (j + 1) which is the number of the 1st thru | or j + 1 type apparatus from the search range memory | storage part 12. FIG. The search execution unit 13 may add a value obtained by multiplying the first index of each type of device read from the device storage unit 10 by the number of corresponding types of devices read from the search range storage unit 12. In this way, the search execution unit 13 may calculate the sum of the first indices of the first to j + 1 type devices from n (1) to n (j + 1).
Next, the search execution part 13 should just read each parameter | index of the kind of j + 2 thru | or the J-th kind apparatus from the apparatus memory | storage part 10, and should just calculate an index ratio. Alternatively, the search execution unit 13 may calculate the index ratio of each type of device in advance, and store it in the device storage unit 10 in association with the type of device, for example. Then, the search execution unit 13 may read the index ratios of the j + 2 to Jth type devices from the device storage unit 10 when selecting the type of device having the largest index ratio. The search execution unit 13 selects a device having the largest index ratio among the devices of the j + 2 to Jth types.
The search execution unit 13 calculates a real value (relaxed number) obtained by dividing the difference between the necessary value and the calculated total of the first index by the first index of the device type having the largest index ratio.
The search execution unit 13 sets the second index of the first to j + 1 type devices from n (1) to n (j + 1) units, and the j + 2 type device having the calculated real value as the number. Calculate the total. The search execution unit 13 sets the calculated sum as the lower limit value of the sum of the second index of the search range (j + 1). The search execution unit 13 multiplies the value obtained by multiplying the calculated real value by the second index of the j + 2nd type device to the sum of the second index of the j + 2th type device using the calculated real value as the number of units. do it.
For example, in the intermediate node (A, 0) in FIG. 8, when the required value is 15, the number of model A is 0, so the remaining required value is 15. Of the remaining device types, the device type with the largest index ratio is model C. In the case where the number is allowed to be a real number, the number of models C in which the total of the first index is 15 is 15/7. The total of the second index of 15/7 model C is 15/7 × 8≈17.14. Since the number of the model A is 0, this is the lower limit value regarding the sum of the second index obtained by the search after the intermediate node (A, 0). When model A is zero, it is theoretically impossible to obtain a solution smaller than 17.14 with respect to the total of the second index. Therefore, if the provisional solution is smaller than the calculated lower limit value of 17.14, there is no possibility that the sum of the second indices of the nodes in the route passing through the intermediate node (A, 0) is less than the provisional solution. . In this case, the search execution unit 13 does not search for a route that passes through the intermediate node (A, 0). On the other hand, when the provisional solution is larger than the calculated lower limit value of 17.14, there is a possibility that a route whose total of the second index exceeds the provisional solution exists in any of the routes that pass through the intermediate node (A, 0). There is sex. In this case, the search execution unit 13 searches for a route that passes through the intermediate node (A, 0).
Next, the search execution unit 13 compares the provisional solution read from the provisional solution storage unit 11 with the lower limit value of the total of the second indexes of the calculated search range (j + 1).
When the search range is represented by the intermediate node (A, 0) in FIG. 8, if the provisional solution stored in the provisional solution storage unit 11 is 24, the lower limit value of the second index of the search range is about 17.14. So the lower limit is smaller than the provisional solution.
As a result of the comparison, when the provisional solution falls below the lower limit (step S18, N), there is no possibility that the search range (j + 1) includes a sufficient set in which the sum of the second indices is less than the provisional solution. In this case, the search execution unit 13 does not search in the search range (j + 1) and returns to step S15.
For example, when the search range is represented by the intermediate node (A, 2) in FIG. 8, the sum of the first indices of the two models A is 3 × 2 = 6. When the required value is 15, the remaining required value calculated by the search execution unit 13 is 9. The second index of the model A is 5, the first index of the model C having the largest index ratio among the remaining device types (model B and model C) is 7, and the second index is 8. Therefore, the lower limit value calculated by the search execution unit 13 is approximately 20.29 (= 10 + 8 × 9/7). When the provisional solution stored in the provisional solution storage unit 11 is 20, and the corresponding provisional set is a set of devices represented by ((A, 1), (B, 1), (C, 1)), search execution is performed. The unit 13 determines in step S18 that the provisional solution is below the lower limit (step S18, N). In this case, since the search range represented by the node (A, 2) may not include the minimum set, the search execution unit 13 performs a search for the search range represented by the node (A, 2). Absent. Similarly, the search execution unit 13 does not perform a search for the search range represented by each of the nodes (A, 3), (A, 4), and (A, 5).
As a result of the comparison, when the provisional solution falls below the lower limit value (step S18, Y), there is a possibility that the search range (j + 1) includes a sufficient set in which the sum of the second indices is less than the provisional solution. The search execution unit 13 searches for the minimum set in the search range (j + 1). The search method for the search set is arbitrary.
For example, the search execution unit 13 calculates the sum of the first indices of all the sets included in the search range (j + 1), and if the sum of the first indices is equal to or greater than the necessary value, the sum of the second indices A search may be performed by calculating and comparing with a provisional solution. In generating all sets included in the search range (j + 1), for example, the search execution unit 13 may first calculate p (j + 2) by the same process as in step S13. The search execution unit 13 may set an integer from 0 to p (j + 2) as the number of devices of the j + 2nd type in the set included in the search range (j + 1). The search execution unit 13 may calculate p (j + 3) in the case where the number of devices of the j + 2nd type is 0 to p (j + 2) by the same process as in step S13. The search execution unit 13 may perform the same processing as in step S13 in this way until all combinations of the numbers of j + 2 to Jth types of devices are derived. In this way, the search execution unit 13 can generate all sets included in the search range (j + 1).
Moreover, the search execution part 13 may perform a search as follows.
When j is smaller than J-2 (step S19, Y), the search execution unit 13 increments j by 1 and performs the search operation of FIG. 4 recursively (step S20). That is, if there is a possibility that the search range (j + 1) includes a provisional set in which the total of the second indices is less than the provisional solution, if there are a plurality of types of devices whose number is not stored in the search range storage unit 12 The search execution unit 13 further divides the search range (j + 1) in the same manner as described above. Then, the search execution unit 13 includes a set in which the total of the second index is less than the provisional solution in each search range (j + 2) obtained by dividing the search range (j + 1) according to the number of devices of the j + 2nd type. To determine whether or not Then, the search execution unit 13 searches in the search range (j + 2) according to the determination result. In this way, the search execution unit 13 may perform the search operation of FIG. 4 recursively while increasing j by 1 until j becomes J-2.
For example, when the search range is represented by the intermediate node (A, 0) of the search tree described above, the lower limit value is smaller than the provisional solution, and j is 0 and J is 3, so j is J-2. Smaller than. Therefore, the search execution part 13 should just perform the process of FIG. 4 from the beginning with respect to the search range (1) represented by a node (A, 0). In this case, the search execution unit 13 divides the search range (1) represented by the node (A, 0) according to the number of second device types (model B) ((A, 0). , (B, 0), ((A, 0), (B, 1)),...,...
When j is J−2 (step S19, N), the search execution unit 13 determines, in the search range (j + 1), if there is a set of devices whose total of the second index is less than the provisional solution. Let the sum of the second index be the new provisional solution and the set be the new provisional set. When a new provisional solution and provisional set are found, the search execution unit 13 stores the new provisional solution and information representing the new provisional solution in the provisional solution storage unit 11, and updates the provisional solution and provisional set (step) S21).
The search range (j + 1) is a set of devices (where j (J-1) is n (1) to n (J-1)) when j is J-2. Search range (J-1)). In step S21, the search execution unit 13 may first calculate p (J) in the same manner as in step S13, and calculate r (J) in the same manner as in step S14.
When p (J) is calculated by the above-described fine method, p (J) is a set included in the search range (J-1), and the Jth of the sufficient set that minimizes the sum of the second indices. Number of types of equipment. In this embodiment, since there is no limit to the number of devices, r (J) is p (J). Therefore, the search execution unit 13 calculates the total of the second index when the number of J-th type devices is r (J), and the calculated total of the second index is the provisional solution storage unit 11. Compare with the provisional solution read from.
On the other hand, when r (J) is calculated by the above-described rough method, the minimum value p (J) of the J-th type device in the satisfaction set included in the search range (J-1) is equal to or less than r (J). One of the values. In this case, the search execution unit 13 sequentially increases the number n (J) of the Jth type device from 0 while being included in the search range (J-1), and the number of the Jth type devices is n. The sum of the first index of the set (J) is calculated. The search execution unit 13 compares the calculated total of the first index with the necessary value, and first calculates the total of the second index of the set in which the total of the first index exceeds the necessary value. The search execution unit 13 may compare the calculated total of the second indices with the provisional solution read from the provisional solution storage unit 11. If n (J) reaches r (J) without the total of the first index exceeding the required value, the search execution unit 13 does not update the provisional solution and provisional set, and the process of step S21 Can be terminated.
As a result of the comparison, when the total of the calculated second index is less than the provisional solution, the search execution unit 13 may set the calculated total of the second index as a new provisional solution. In addition, the search execution unit 13 stores the search range storage unit 12 from n (1) to n (J), which is the number of the first to J-1th types of devices representing the search range (J-1). −1) and the set represented by the number of J-th type devices r (J) may be a new provisional set. The search execution unit 13 stores information representing the new provisional solution and provisional set in the provisional solution storage unit 11.
As a result of the comparison, if the calculated total of the second indices exceeds the provisional solution, the search execution unit 13 does not update the information representing the provisional solution and the provisional set, and ends the process of step S21.
For example, when the search range is the search range (2) represented by the nodes ((A, 0), (B, 1)) of the above-described search tree in step S19, j (= 1) is J-2. Since it is not smaller than (1) (step S19, N), the search execution part 13 performs the process of step S21.
The search execution unit 13 calculates ((A, 0), (B, 1), (C, 0)), ((A, 0), (B, 1) when calculating r (j) by a rough method. ), (C, 1)), the total of the first index of the set of devices represented by (C, 1)) and the comparison with the required value are sequentially performed. The first index and the second index of these devices stored in the device storage unit 10 are as in the example of FIG. Therefore, the sum of the first index of the set of these devices is smaller than the required value. Next, the search execution unit 13 calculates the sum of the first index of the set of devices represented by ((A, 0), (B, 1), (C, 2)). Since the obtained total 19 is larger than the necessary value 15, the search execution unit 13 calculates the sum of the second index of the set of devices. When the provisional solution stored in the provisional solution storage unit 11 is 24, 23, which is the total of the calculated second indexes, is smaller than the provisional solution. In this case, the search execution unit 13 uses the calculated second index total 23 corresponding to the new provisional solution ((A, 0), (B, 1), (C, 2)). Is a new provisional set. The search execution unit 13 stores information representing the new provisional solution and provisional set in the provisional solution storage unit 11.
After the search by the search execution unit 13 is completed, the output unit 14 reads out information representing the temporary set stored in the temporary solution storage unit 11 as described in step S3 of FIG. Just output.
In the present embodiment described above, a combination of a plurality of types of devices having two different indices that satisfies the condition that the sum of one index is equal to or greater than a predetermined value and the sum of the other index is minimized. There is an effect that it can be derived with a small amount of calculation.
The reason is that the search execution unit 13 determines whether or not there is a possibility that the search range includes the minimum set for each search range represented by the number of some types of devices. Then, the search execution unit 13 does not perform a search in the search range when there is no possibility that the minimum set is included.
(First modification of the first embodiment)
Next, a first modification of the first embodiment of the present invention will be described.
In the first embodiment described above, the number of devices for each type of device is not limited, whereas the number of devices for each type of device of this modification is one or zero (problem 01). The configuration management apparatus 1 according to the present modification outputs a set that is a satisfiable set that does not include two or more devices of the same type and that has the minimum sum of the second indices.
First, the configuration of the configuration management apparatus 1 according to this modification will be described in detail with reference to the drawings.
FIG. 1 is a diagram illustrating a configuration of a configuration management apparatus 1 according to the present modification. The configuration of the configuration management device 1 according to this modification is the same as the configuration of the configuration management device 1 according to the first embodiment. Therefore, the description regarding the configuration is omitted.
FIG. 3 is a flowchart showing an example of the operation of the overall configuration for extracting the minimum set by the configuration management apparatus 1 of the present modification.
An example of the operation in which the configuration management apparatus 1 of the present modification extracts the minimum set is an example of the operation in which the configuration management apparatus 1 of the first embodiment extracts the minimum set and the calculation of the initial solution in step S2. The method differs from the search method in step S3.
Moreover, the search execution part 13 should just make the collection of satisfying sets whose 1st apparatuses are 0 or more and 1 or less into the initial value (search range (0)) of a search range.
Since other operations of the configuration management apparatus 1 of the present modification are the same as the operations of the configuration management apparatus 1 of the first embodiment, description thereof will be omitted.
FIG. 9 is a flowchart illustrating an example of an operation in which the configuration management apparatus 1 according to the present modification calculates an initial solution.
Referring to FIG. 9, the initial solution calculation unit 15 reads out the first index and the second index of all device types from the device storage unit 10 (step S31).
Next, the initial solution calculation unit 15 calculates an index ratio of all device types (step S32). The initial solution calculation unit 15 may sort the device types in descending order of the calculated index ratio.
Next, the initial solution calculation unit 15 selects one type of device having the largest index ratio among the types of unselected devices (step S33).
For example, the initial solution calculation unit 15 may read information on the temporary set stored in the temporary solution storage unit 11. The initial solution calculation unit 15 may update the provisional set by adding information on one selected type of device to the read provisional set information. The initial solution calculation unit 15 may store information on the updated temporary set in the temporary solution storage unit 11. If the provisional solution storage unit 11 does not store information on the provisional set, the initial solution calculation unit 15 may generate information on the provisional set composed of one selected type of device. The initial solution calculation unit 15 may store the information on the provisional set in the provisional solution storage unit 11.
The initial solution calculation unit 15 calculates the total of the first index of the type of the device that has already been selected (step S34).
Note that the initial solution calculation unit 15 may calculate the cumulative total of the second index when executing the operation of step S34. Then, the initial solution calculation unit 15 may store the calculated cumulative total of the second indices in the temporary solution storage unit 11 as a temporary solution. For example, the initial solution calculation unit 15 may read the provisional solution that is the cumulative total of the second indices from the provisional solution storage unit 11. Then, the initial solution calculation unit 15 may add the second index of the type of device newly selected in step S33 to the read provisional solution. Then, the initial solution calculation unit 15 may store the combined provisional solution in the provisional solution storage unit 11. In this case, if the provisional solution is not stored in the provisional solution storage unit 11, the second index of the device of the type newly selected in step S33 may be stored in the provisional solution storage unit 11 as a provisional solution.
The initial solution calculation unit 15 compares the accumulated value of the first index of the selected device type with the required value. The initial solution calculation unit 15 calculates the cumulative total of the second indexes of the selected device type, which is the initial value of the provisional solution, when the sum of the first indexes is equal to or greater than the necessary value (step S35, Y). (Step S36).
The initial solution calculation unit 15 stores the calculated cumulative total of the second indices in the temporary solution storage unit 11 as the initial value of the temporary solution (step S37).
In step S34, when the cumulative total of the second index is stored in the temporary solution storage unit 11 as a temporary solution, in step S36, the cumulative total of the second index of the selected device type is the initial value of the temporary solution. There is no need to store it as a value.
When the cumulative total of the first index is not greater than or equal to the required value (step S35, Y), when all the device types have not been selected and there are device types that have not yet been selected (step S38, N), The process returns to step S33.
When all the device types have been selected and there are no device types that have not yet been selected (step S38, Y), the initial solution calculation process is terminated. In this case, even if the set of devices includes all devices in which each index is stored in the device storage unit 10, the sum of the first index of the set is less than the required value, so there is no sufficient set.
FIG. 10 is a flowchart illustrating an example of a detailed operation of the search of the configuration management apparatus 1 according to this modification.
Comparing the flowcharts representing the search operation of the first embodiment in FIG. 10 and FIG. 4, the present modification is different in that step S13 and step S14 are not included, and step S22 is included instead of step S15. In the present modification, the maximum number of devices of each type is 1, so r (j + 1) is always 1. Therefore, step S13 and step S14 for calculating r (j + 1) are unnecessary. Step S22 corresponds to step S15 in the case where r (j + 1) is always 1.
Further, the operation of the configuration management device 1 of this modification is different from the operation of the configuration management device 1 of the first embodiment in that the lower limit value is calculated by the relaxation method in step S17 and the provisional solution in step S21. And the operation of updating the provisional set is different. Since other points are the same as the operation of the first embodiment, the description is omitted.
FIG. 11 is a flowchart illustrating an example of an operation in which the search execution unit 13 of the configuration management device 1 of the present modification calculates the lower limit value in step S17.
Referring to FIG. 11, the search execution unit 13 calculates the sum of the first indexes of the first to j + 1th devices based on the first index and the second index read from the device storage unit 10 ( Step S41).
Next, the search execution unit 13 calculates a required remaining value obtained by subtracting the calculated first index of the 1st to (j + 1) th devices from the required value N (step S42).
Next, the search execution unit 13 selects the device type having the largest index ratio calculated from the first index and the second index read from the device storage unit 10 from the types of the j + 2 to Jth unselected devices. Select (step S43).
The search execution unit 13 calculates the sum of the first indices for the maximum number of devices of the selected type (step S44). In the present modification, since the maximum number of all types of devices is 1, the sum of the calculated first index is the first index itself of the type of device selected in step S43.
If the calculated sum of the first indices is smaller than the necessary remaining value (step S45, N), the search execution unit 13 has already selected all the types of the j + 2 to Jth devices and has not yet been selected. If there is no such type (step S49), the process may be terminated. In this case, even if all the 1st to Jth types of devices are combined, the total of the first index does not reach the required value.
If the calculated sum of the first indices is smaller than the required remaining value (step S45, N), the search execution unit 13 determines that there is a device type that has not yet been selected among the j + 2 to Jth device types. The remaining necessary value is updated (step S50). The search execution unit 13 may update the remaining required value by subtracting the sum of the first indices calculated in step S44 from the remaining required value to obtain a new required remaining value.
After performing the operation of step S50, the search execution unit 13 returns to step S43.
When the calculated sum of the first indices is larger than the necessary remaining value (step S45, Y), the search execution unit 13 calculates a real value obtained by dividing the necessary remaining value by the first index of the selected type of device. (Step S46). The calculated real value is the number of devices of the selected type (relaxed number) calculated by removing the restriction that the number of units is an integer so that the value of the first index becomes the remaining necessary value. In the present modification, the maximum number of devices of each type is 1, so the number of mitigation calculated by the search execution unit 13 is a real value included in the range from 0 to 1.
Next, the search execution part 13 calculates the 2nd parameter | index for the calculated mitigation number of the apparatus of the selected kind (step S47). In the present modification, since the maximum number of devices of each type is 1, the second index for the number of mitigation calculated by the search execution unit 13 is 0 or more, and the second index per device of the selected type is 2nd. The value is below the index.
The search execution unit 13 sets the second number of all selected devices except the last selected device and the first to j + 1 type devices, and the last selected device for the calculated number of mitigation, respectively. Calculate the sum of the indicators. The search execution unit 13 sets the calculated total to the lower limit value (step S48). As described above, in this modification, the maximum number of each device is one.
In the above processing, for example, when the first index represents performance and the second index represents cost, the devices in order from the type with the best cost performance until the sum of the first indexes is equal to or greater than the required value. This is equivalent to selecting the type. The above processing further divides the last selected device so that the total of the first index is exactly the required value when the total of the first index exceeds the required value, and the already selected device is selected. This is equivalent to calculating the sum of the second indices of all the devices that have been used and setting the lower limit of the second index.
Next, the operation of step S21 in FIG. 10 of the configuration management apparatus 1 according to this modification will be described.
In step S <b> 21, the search execution unit 13 firstly, among the satisfaction sets included in the search range (J−1), the number p (J) of the J-th type devices in the satisfaction set that minimizes the second index. J) may be calculated. Then, when p (J) is equal to or less than the maximum number of J-th devices (0 or 1 in the present modification), the search execution unit 13 may update the provisional solution / provisional set. When p (J) is larger than 1, which is the maximum number of J-th devices, the search range (J-1) does not include a sufficient set, so the search execution unit 13 updates the provisional solution / provisional set. Not performed.
When updating the provisional solution / provisional set, the search execution unit 13 calculates the total of the second index when the number of J-th type devices is r (J), as in the first embodiment. What is necessary is just to calculate. Then, the search execution unit 13 may compare the calculated total of the second indices with the provisional solution read from the provisional solution storage unit 11. The provisional solution / provisional set is updated when p (J) is equal to or less than the maximum number of J-th devices (1 in this modification), and r (J) is the maximum number of J-th types. In this case, r (J) is p (J).
As a result of the comparison, when the total of the calculated second index is less than the provisional solution, the search execution unit 13 may set the calculated total of the second index as a new provisional solution. In addition, the search execution unit 13 stores the search range storage unit 12 from n (1) to n (J), which is the number of the first to J-1th types of devices representing the search range (J-1). −1) and the set represented by the number of J-th type devices r (J) may be a new provisional set. The search execution unit 13 stores information representing the new provisional solution and provisional set in the provisional solution storage unit 11.
FIG. 12 is a diagram illustrating a binary search tree that represents a set of devices to be searched by the search execution unit 13 of this modification.
12 differs from FIG. 8 in that the branching coefficient of each node is all 2 and the maximum number of devices of each type is limited to 1 in the search tree of FIG.
In this modification described above, when the maximum number of devices of each type is 1, the condition that the sum of one index is equal to or greater than a predetermined value is satisfied, and the sum of the other index is minimized. There is an effect that a combination of a plurality of types of devices having two indexes can be derived with a small amount of calculation.
The reason is that the search execution unit 13 minimizes the search range for each search range represented by the number of some types of devices under the condition that the maximum number of devices of each type is 1. This is because it is determined whether or not there is a possibility that a set is included. Then, the search execution unit 13 does not perform a search in the search range when there is no possibility that the minimum set is included.
(Second modification of the first embodiment)
Next, a second modification of the first embodiment will be described in detail with reference to the drawings.
In the second modification, there is a maximum number for each type of device. The configuration management apparatus 1 according to the present modification outputs a set in which the total of the second index is the smallest in a sufficient set that does not include the same type of devices more than the maximum number of devices of that type.
FIG. 1 is a diagram illustrating a configuration of a configuration management apparatus 1 according to the present modification. The configuration of the configuration management device 1 according to this modification is the same as the configuration of the configuration management device 1 according to the first embodiment.
However, in addition to the first index and the second index, the device storage unit 10 according to the present modification stores the maximum number for each device type in association with the device type. The first index, the second index, and the maximum number of each type of device may be stored in the device storage unit 10 in advance. Alternatively, the index acquisition unit 16 may extract the first index, the second index, and the maximum number of each type of device included in the file 31 and store them in the device storage unit 10.
FIG. 13 is a diagram illustrating an example of the first index, the second index, and the maximum number stored in the device storage unit 10 according to the present modification.
In the example of FIG. 13, the maximum number of model A is four, the maximum number of model B is three, and the maximum number of model C is two.
The other components of the configuration management apparatus 1 of the present modification are the same as the corresponding components of the first embodiment, and thus description thereof is omitted.
Next, the operation of this modification will be described in detail with reference to the drawings.
FIG. 3 is a flowchart showing an example of the entire operation of extracting the minimum set by the configuration management apparatus 1 of the present modification.
An example of the operation in which the configuration management apparatus 1 of the present modification extracts the minimum set is compared with the example of the operation in which the configuration management apparatus 1 of the first embodiment extracts the minimum set. The calculation method is different from the search method in step S3.
In addition, the search execution unit 13 calculates, for example, the minimum number of devices p (1) when a satisfaction set is configured by only the first type of device, and p (1) and the maximum of the first type of device. The smaller of the number is set to r (1). And the search execution part 13 should just make the collection of sufficient sets whose 1st apparatuses are 0 or more r (1) or less into the initial value (search range (0)) of a search range.
Since other operations of the configuration management apparatus 1 of the present modification are the same as the operations of the configuration management apparatus 1 of the first embodiment, description thereof will be omitted.
FIG. 14 is a flowchart illustrating an example of an operation in which the configuration management apparatus 1 according to the present modification calculates an initial solution.
The operation of calculating the initial solution of the configuration management apparatus 1 of the present modification shown in FIG. 14 is compared with the operation of calculating the initial solution of the configuration management apparatus 1 of the first modification shown in FIG. The difference is that the operation of step S39 is performed during S34. In addition, the first modification in which the operation of calculating the total of the first index and the total of the second index in step S34 and the operation of calculating the initial value of the provisional solution in step S36 of the present modification correspond to each other. Different from the example behavior.
The initial solution calculation unit 15 of the present modification example selects the device type having the largest index ratio among the unselected device types (step S33), and then sets the maximum number of selected device types to the device storage unit. 10 (step S38).
In step S 34, the initial solution calculation unit 15 reads out the maximum number of the already selected device types from the first index and the maximum number of the already selected device types read from the device storage unit 10. The cumulative total of 1 index is calculated.
In step S36, the initial solution calculation unit 15 selects all types of devices that have already been selected other than the last selected type of device and the maximum number of devices of each type and the number of devices of the last selected type. The minimum number of devices of the last selected type in the satisfaction set composed of is calculated. The initial solution calculation unit 15 has the maximum number of devices of each type, all the types of devices already selected other than the type of the last selected device, and the type of the last selected type of the calculated minimum number of devices. A set composed of devices is set as an initial value of the provisional set. The initial solution calculation unit 15 calculates the sum of the second index of the initial value of the provisional set and sets it as the initial value of the provisional solution.
Other operations of the configuration management apparatus 1 of this modification in FIG. 14 are the same as the operations of the steps denoted by the same reference numerals in the configuration management apparatus 1 of the first modification of FIG. .
Next, the operation at the time of searching for the minimum set of the configuration management apparatus 1 according to this modification will be described in detail with reference to the drawings.
FIG. 4 is a flowchart showing the operation at the time of searching for the minimum set of the configuration management apparatus 1 of the present modification. The flowchart representing the operation at the time of searching for the minimum set of the configuration management device 1 of the present modification is the same as that of the first embodiment, but the method of calculating r (j + 1) in step S14 and the lower limit of step S17. The value calculation method is different from the provisional solution / temporary set update method in step S21.
In step S14, the search execution unit 13 compares p (j + 1) calculated in step S14 with the maximum number of j + 1 type devices read from the device storage unit 10, and sets the larger one to r (j + 1). .
FIG. 11 is a flowchart showing an operation in which the search execution unit 13 calculates the lower limit value in step S17 of the present modification. The flowchart showing the operation of the search execution unit 13 of this modification calculating the lower limit value is the same as that of the first modification. However, while the maximum number of each type of device is one in the first modification, the maximum number of each type of device in this modification is not limited to one, and is associated with the type of device. The difference is that it is stored in the device storage unit 10.
In step S44, the search execution unit 13 reads the maximum number of devices of the selected type from the device storage unit 10, and calculates the sum of the first indices for the maximum number of devices of the selected type.
In step S46, the relaxed number calculated by the search execution unit 13 is a real value included in a range of 0 or more and the maximum number of selected devices read from the device storage unit 10.
In step S48, the search execution unit 13 determines the maximum number of devices read from the device storage unit 10, all the selected devices except the last selected device, the first to j + 1 type devices, and the calculated mitigation number. Calculate the sum of the second index of the last selected device in minutes. The search execution unit 13 sets the calculated total as the lower limit value.
Next, the operation of step S21 in FIG. 4 of the configuration management apparatus 1 according to this modification will be described.
In step S <b> 21, the search execution unit 13 firstly, among the satisfaction sets included in the search range (J−1), the number p (J) of the J-th type devices in the satisfaction set that minimizes the second index. J) may be calculated. Then, when p (J) is equal to or less than the maximum number of devices of the Jth type, the search execution unit 13 may update the provisional solution / provisional set. If p (J) is larger than the maximum number of devices of the J-th type, the search range (J-1) does not include a sufficient set, so the search execution unit 13 updates the provisional solution / provisional set. Absent.
When updating the provisional solution / provisional set, the search execution unit 13 calculates the total of the second index when the number of J-th type devices is r (J), as in the first embodiment. What is necessary is just to calculate. Then, the search execution unit 13 may compare the calculated total of the second indices with the provisional solution read from the provisional solution storage unit 11. The provisional solution / provisional set is updated when p (J) is less than or equal to the maximum number of J-th devices, and r (J) is either the maximum number of the J-th type or p (J). In this case, r (J) is p (J) because it is the smaller value.
As a result of the comparison, when the total of the calculated second index is less than the provisional solution, the search execution unit 13 may set the calculated total of the second index as a new provisional solution. In addition, the search execution unit 13 stores the search range storage unit 12 from n (1) to n (J), which is the number of the first to J-1th types of devices representing the search range (J-1). −1) and the set represented by the number of J-th type devices r (J) may be a new provisional set. The search execution unit 13 stores information representing the new provisional solution and provisional set in the provisional solution storage unit 11.
In this modification described above, when the maximum number of each type of device is determined, the condition that the sum of one index is equal to or greater than a predetermined value is satisfied, and the sum of the other index is minimized. There is an effect that a combination of a plurality of types of devices having two different indexes can be derived with a small amount of calculation.
The reason is that the search execution unit 13 performs a search represented by the number of some types of devices under the condition that the maximum number of devices of each type is the number stored in the device storage unit 10. This is because, for each range, it is determined whether or not there is a possibility that the minimum set is included in the search range. Then, the search execution unit 13 does not perform a search in the search range when there is no possibility that the minimum set is included.
Next, a second embodiment of the present invention will be described in detail with reference to the drawings.
FIG. 15 is a diagram illustrating the configuration of the configuration management apparatus 1A according to the present embodiment.
Referring to FIG. 15, the configuration management device 1A of the present embodiment is configured with devices associated with the first index value P1 and the second index value P2 for each of a plurality (J) types, and the total of P1 is required. A device storage unit 10 that stores P1 and P2 for each type of device when the set of devices equal to or greater than the value N is a satisfaction set, and the minimum value in the total satisfaction set of P2 is the minimum total, and an initial value Is a provisional solution storage unit 11 for storing information representing a provisional solution whose value is greater than or equal to the minimum sum, and a provisional set in which the sum of P2 is a set of devices that is the provisional solution, and 1st to jth (j is 1 To (n-1), the search range storage unit 12 stores n (1) to n (j), which is the number of types of devices, and N is input, and n (j + 1) is set to 0. From n (1) to n (j) 1st to jth types of devices and j + 1th N (1 + 1) is stored in the search range storage unit 12 while sequentially increasing to a predetermined value (j + 1) that is equal to or less than the minimum number p (j + 1) of the j + 1th type of devices when a satisfaction set is configured with only types of devices. Then, from P1 and P2 of N and the 1st to j + 1th type devices, the number of 1st to j + 1th type devices is changed from n (1) to n (j + 1) stored in the search range storage unit 12. The lower limit value of the sum of P2 of the set included in the search range (j + 1) that is a collection of satisfying sets is calculated by removing the restriction that the number of devices is an integer, and the lower limit value is calculated from the provisional solution. If it is smaller, a satisfying set in which the sum of P2 is smaller than the provisional solution is extracted from the search range (j + 1), and information indicating the sum of P2 of the extracted fulfillment set and the extracted satisfaction set is used as the provisional solution and provisional set. In the storage unit 11 If the lower limit value is larger than the provisional solution, the search execution unit 13 that does not search the search range (j + 1), and the output unit 14 that outputs the provisional set as a minimum set after the search execution unit 13 performs the search. including.
In the present embodiment described above, a combination of a plurality of types of devices having two different indices that satisfies the condition that the sum of one index is equal to or greater than a predetermined value and the sum of the other index is minimized. There is an effect that it can be derived with a small amount of calculation.
The reason is that the search execution unit 13 determines whether or not there is a possibility that the search range includes the minimum set for each search range represented by the number of some types of devices. Then, the search execution unit 13 does not perform a search in the search range when there is no possibility that the minimum set is included.
The present invention has been described above with reference to the embodiments, but the present invention is not limited to the above embodiments. Various changes that can be understood by those skilled in the art can be made to the configuration and details of the present invention within the scope of the present invention.
This application claims the priority on the basis of Japanese application Japanese Patent Application No. 2011-136487 for which it applied on June 20, 2011, and takes in those the indications of all here.

The present invention is suitable for an apparatus or system for keeping the cost of a system low in system construction.

1, 1A configuration management device 2 terminal 3 server 10 device storage unit 11 provisional solution storage unit 12 search range storage unit 13 search execution unit 14 output unit 15 initial solution calculation unit 16 index acquisition unit 31 file

Claims

A set of devices each of which is associated with the first index value P1 and the second index value P2 for each of a plurality of types (J) and the total of P1 is equal to or greater than the necessary value N is a sufficient set, and the total of P2 If the smallest value in the sufficient set is the smallest sum,
Device storage means for storing P1 and P2 for each device type;
A provisional solution storing means for storing information representing a provisional solution whose initial value is a value equal to or greater than the minimum total, and a provisional set in which the sum of P2 is a set of devices serving as the provisional solution;
Search range storage means for storing n (1) to n (j), which is the number of devices of the 1st to jth types (j is an integer from 1 to J-1);
Enter N
n (j + 1) is a j + 1-th type when a satisfying set is composed only of n (1) -n (j) 1st to j-th type devices and j + 1-th type devices from 0, respectively. While sequentially increasing to a predetermined value (j + 1) below the minimum number of devices p (j + 1),
n (j + 1) is stored in the search range storage means,
From P1 and P2 of N and 1 to j + 1 type devices, the number of 1 to j + 1 type devices is n (1) to n (j + 1) stored in the search range storage means Calculating the lower limit value of the total P2 of the set included in the search range (j + 1) which is a set of sets, excluding the restriction that the number of devices is an integer;
When the lower limit value is smaller than the provisional solution,
From the search range (j + 1), a satisfaction set in which the sum of P2 is smaller than the provisional solution is extracted. Stored in provisional solution storage means,
When the lower limit value is larger than the provisional solution, search execution means that does not search the search range (j + 1);
A configuration management device comprising: output means for outputting the provisional set as a minimum set after execution of search by the search execution means.
The device information storage means stores the maximum number q (j) (1 ≦ j ≦ J) of each type of device in association with the type of device,
The search execution means determines the smaller one of p (j) and q (j) (1 ≦ j ≦ J) input from the device information storage means as the predetermined value (j) equal to or less than p (j). The configuration management apparatus according to claim 1.
The search execution means includes
An index ratio obtained by dividing P1 stored by the device storage means by P2 is calculated for each type of device,
The total of P1 of the 1st to j (1 ≦ j <J) type devices of the number n (1) to n (j) respectively stored in the search range storage unit is calculated, and the calculated total is calculated. Calculate the remaining required value subtracted from N,
Calculate the mitigation number that is a real value obtained by dividing the remaining required value by P1 of the device (device X) having the largest index value among the devices of the j + 1th to Jth types,
The total of P2 of the 1st to jth types of devices of n (1) to n (j) and the value obtained by adding the value obtained by multiplying P2 of the device X by the relaxed number is the search range ( The configuration management apparatus according to claim 1, wherein j is the lower limit value.
The search execution means includes
An index ratio obtained by dividing P1 stored by the device storage means by P2 is calculated for each type of device,
The total of P1 of the 1st to j (1 ≦ j <J) type devices of the number n (1) to n (j) stored in the search range storage means is calculated as the default required value, The total of P2 is calculated as a provisional lower limit, and the value obtained by subtracting the existing required value from N is set as the remaining required value.
While selecting the device type sequentially from the j + 1st to Jth device types in descending order of the index value,
Calculate the sum of P1 for the maximum number q (S) stored in the device information storage means of the selected type of device (Sth type),
If the total is less than the remaining required value, add the total to the predetermined required value, subtract it from the remaining required value, and add the sum of P2 of q (S) S-type devices Add to the provisional lower limit,
If the total is equal to or greater than the remaining required value, a relaxed number that is a real value obtained by dividing the remaining required value by P1 of the Sth type device is calculated, and P2 of the Sth type device corresponding to the relaxed number Is added to the provisional lower limit value,
The configuration management apparatus according to claim 2, wherein the lower limit value of the search range (j) is set.
A set of devices each of which is associated with the first index value P1 and the second index value P2 for each of a plurality of types (J) and the total of P1 is equal to or greater than the necessary value N is a sufficient set, and the total of P2 If the smallest value in the sufficient set is the smallest sum,
For each device type, P1 and P2 are stored in the device storage means,
Information indicating a provisional solution whose initial value is a value equal to or greater than the minimum total and a provisional set that is a set of devices in which the sum of P2 is the provisional solution is stored in the provisional solution storage unit;
N (1) to n (j), which are the number of devices of the 1st to jth types (j is an integer from 1 to J-1), are stored in the search range storage means,
Enter N
n (j + 1) is a j + 1-th type when a satisfying set is composed only of n (1) -n (j) 1st to j-th type devices and j + 1-th type devices from 0, respectively. While sequentially increasing to a predetermined value (j + 1) below the minimum number of devices p (j + 1),
n (j + 1) is stored in the search range storage means,
From P1 and P2 of N and 1 to j + 1 type devices, the number of 1 to j + 1 type devices is n (1) to n (j + 1) stored in the search range storage means Calculating the lower limit value of the total P2 of the set included in the search range (j + 1) which is a set of sets, excluding the restriction that the number of devices is an integer;
When the lower limit value is smaller than the provisional solution,
From the search range (j + 1), a satisfaction set in which the sum of P2 is smaller than the provisional solution is extracted. Stored in provisional solution storage means,
When the lower limit value is larger than the provisional solution, the search of the search range (j + 1) is not performed,
A configuration management method for outputting the temporary set as a minimum set after execution of the search.
The maximum number q (j) (1 ≦ j ≦ J) of each type of device is stored in the device information storage unit in association with the type of device,
The smaller one of p (j) and q (j) (1 ≦ j ≦ J) input from the device information storage means is set to the predetermined value (j) equal to or less than p (j). Configuration management method.
A set of devices each of which is associated with the first index value P1 and the second index value P2 for each of a plurality of types (J) and the total of P1 is equal to or greater than the necessary value N is a sufficient set, and the total of P2 If the smallest value in the sufficient set is the smallest sum,
Computer
Device storage means for storing P1 and P2 for each device type;
A provisional solution storing means for storing information representing a provisional solution whose initial value is a value equal to or greater than the minimum total, and a provisional set in which the sum of P2 is a set of devices serving as the provisional solution;
Search range storage means for storing n (1) to n (j), which is the number of devices of the 1st to jth types (j is an integer from 1 to J-1);
Enter N
n (j + 1) is a j + 1-th type when a satisfying set is composed only of n (1) -n (j) 1st to j-th type devices and j + 1-th type devices from 0, respectively. While sequentially increasing to a predetermined value (j + 1) below the minimum number of devices p (j + 1),
n (j + 1) is stored in the search range storage means,
From P1 and P2 of N and 1 to j + 1 type devices, the number of 1 to j + 1 type devices is n (1) to n (j + 1) stored in the search range storage means Calculating the lower limit value of the total P2 of the set included in the search range (j + 1) which is a set of sets, excluding the restriction that the number of devices is an integer;
When the lower limit value is smaller than the provisional solution,
From the search range (j + 1), a satisfaction set in which the sum of P2 is smaller than the provisional solution is extracted. Stored in provisional solution storage means,
When the lower limit value is larger than the provisional solution, search execution means that does not search the search range (j + 1);
A program recording medium for storing a configuration management program to be operated as output means for outputting the provisional set as a minimum set after execution of search by the search execution means.
The device information storage means for storing the maximum number q (j) (1 ≦ j ≦ J) of each type of device in association with the type of device;
the search execution means for setting the smaller one of p (j) and q (j) (1 ≦ j ≦ J) input from the device information storage means to the predetermined value (j) equal to or less than p (j); The program recording medium according to claim 7, wherein a configuration management program to be operated is stored.
Computer
An index ratio obtained by dividing P1 stored by the device storage means by P2 is calculated for each type of device,
The total of P1 of the 1st to j (1 ≦ j <J) type devices of the number n (1) to n (j) respectively stored in the search range storage unit is calculated, and the calculated total is calculated. Calculate the remaining required value subtracted from N,
Calculate the mitigation number that is a real value obtained by dividing the remaining required value by P1 of the device (device X) having the largest index value among the devices of the j + 1th to Jth types,
The total of P2 of the 1st to jth types of devices of n (1) to n (j) and the value obtained by adding the value obtained by multiplying P2 of the device X by the relaxed number is the search range ( The program recording medium according to claim 7, wherein a configuration management program that is operated as the search execution unit that sets the lower limit of j) is stored.
Computer
An index ratio obtained by dividing P1 stored by the device storage means by P2 is calculated for each type of device,
The total of P1 of the 1st to j (1 ≦ j <J) type devices of the number n (1) to n (j) stored in the search range storage means is calculated as the default required value, The total of P2 is calculated as a provisional lower limit, and the value obtained by subtracting the existing required value from N is set as the remaining required value.
While selecting the device type sequentially from the j + 1st to Jth device types in descending order of the index value,
Calculate the sum of P1 for the maximum number q (S) stored in the device information storage means of the selected type of device (Sth type),
If the total is less than the remaining required value, add the total to the predetermined required value, subtract it from the remaining required value, and add the sum of P2 of q (S) S-type devices Add to the provisional lower limit,
If the total is equal to or greater than the remaining required value, a relaxed number that is a real value obtained by dividing the remaining required value by P1 of the Sth type device is calculated, and P2 of the Sth type device corresponding to the relaxed number Is added to the provisional lower limit value,
The program recording medium according to claim 8, wherein a configuration management program to be operated as the search execution means for setting the lower limit value of the search range (j) is stored.