Information processing device, information processing method, and calculation support program
The hierarchical use of low-, medium-, and high-precision basis functions in the localized wave basis density functional method addresses non-convergence and time issues, enabling efficient and accurate calculation results.
Patent Information
- Authority / Receiving Office
- JP · JP
- Patent Type
- Patents
- Current Assignee / Owner
- RESONAC CORP
- Filing Date
- 2025-07-09
- Publication Date
- 2026-07-29
AI Technical Summary
Existing calculation programs based on the localized wave basis density functional method face challenges such as non-convergence and excessive computation time when attempting high-precision calculations, particularly when using reference basis functions directly for structural optimization.
The system employs a hierarchical approach, starting with low-precision structural optimization, followed by medium-precision, and finally high-precision optimization, using a combination of low-, medium-, and reference-basis functions to gradually achieve target precision, thereby controlling the calculation process to avoid non-convergence and excessive time consumption.
This approach allows for efficient and accurate calculation results by optimizing the structural information in a stepwise manner, reducing computation time and ensuring convergence, thus enhancing user convenience and accuracy.
Smart Images
Figure 0007896799000001 
Figure 0007896799000002 
Figure 0007896799000003
Abstract
Description
Technical Field
[0005] ,
[0001] The present disclosure relates to an information processing apparatus, an information processing method, and a calculation support program.
Background Art
[0002] As one of the first-principles calculation methods, the density functional theory (DFT) is known, and as an electronic property calculation program using DFT, a calculation program based on the localized wave basis density functional method is known.
[0003] According to the calculation program, for example, when the structural information of a target substance is input, while searching for the energy value of the ground state, the structural information of the target substance can be optimized, and the physical property value of the optimized structural information can be calculated.
Prior Art Documents
Patent Documents
[0004]
Patent Document 1
Summary of the Invention
Problems to be Solved by the Invention
[0005] <000An information processing device that executes a calculation program based on the localized wave basis density functional method, The calculation program is provided with a basis function acquisition unit that acquires a reference basis function set for execution at a target accuracy level and a low-precision basis function set for execution at an accuracy level lower than the target accuracy level. When an instruction to execute the calculation program is received regarding the structural information of the target substance, By using the aforementioned structural information as initial structural information and executing the calculation program using the aforementioned low-precision basis function set, low-precision optimized structural information is obtained. By using the aforementioned low-precision optimized structure information as initial structure information and executing the calculation program using the aforementioned reference basis function set, the optimized structure information with the target precision level is obtained. It includes a unit for obtaining calculation results.
[0008] A second aspect of this disclosure is an information processing apparatus described in the first aspect, The number of basis functions included in the low-precision basis function set is less than the number of basis functions included in the reference basis function set.
[0009] A third aspect of this disclosure is an information processing apparatus described in the second aspect, The basis function acquisition unit is, Further, obtain a set of medium-precision basis functions that allows the calculation program to be executed at a precision level lower than the target precision level but higher than the low-precision basis function set. The calculation result acquisition unit, When an instruction to execute the calculation program is received regarding the structural information of the target substance, By using the aforementioned structural information as initial structural information and executing the calculation program using the aforementioned low-precision basis function set, low-precision optimized structural information is obtained. By using the aforementioned low-precision optimized structure information as initial structure information and executing the calculation program using the aforementioned medium-precision basis function set, medium-precision optimized structure information is obtained. By using the aforementioned medium-precision optimized structure information as initial structure information and executing the calculation program using the aforementioned reference basis function set, the optimized structure information with the target precision level is obtained.
[0010] A fourth aspect of this disclosure is an information processing apparatus as described in the third aspect, The basis function acquisition unit is, Obtain the Nth basis function set so that the calculation program is executed at an Nth precision level lower than the target precision level and higher than the (N-1)th precision level (where N is an integer greater than or equal to 2), The calculation result acquisition unit, By using the N-1 precision level of optimized structure information as initial structure information and executing the calculation program with the N basis function set, optimized structure information of the N precision level is obtained.
[0011] A fifth aspect of this disclosure is an information processing device described in any of the first to fourth aspects, It has a storage unit that stores combinations of basis function sets with different levels of precision.
[0012] A sixth aspect of this disclosure is an information processing apparatus described in the fifth aspect, Each basis function set obtained by the basis function acquisition unit is a basis function set selected by the user from among a plurality of basis function sets included in the combination.
[0013] A seventh aspect of this disclosure is an information processing apparatus as described in the fifth aspect, When the basis function acquisition unit acquires a reference basis function set selected by the user from among a plurality of basis function sets included in the combination, it acquires other basis function sets included in the combination that correspond to the reference basis function set.
[0014] The eighth aspect of this disclosure is an information processing device described in any of the first to fourth aspects, When the basis function acquisition unit acquires a set of reference basis functions input by a user, it generates a set of basis functions with a lower accuracy level than the set of reference basis functions based on the set of reference basis functions.
[0015] A ninth aspect of the present disclosure is the information processing apparatus according to the first aspect, When using the structure information as the initial structure information and executing the calculation program using the set of low-accuracy basis functions, the convergence determination value is larger than the convergence determination value when executing the calculation program using the set of reference basis functions with the low-accuracy optimized structure information as the initial structure information.
[0016] A tenth aspect of the present disclosure is the information processing apparatus according to the first aspect, The structure information of the target substance is generated using any one of the molecular mechanics method, the semi-empirical method, and the machine learning method.
[0017] An eleventh aspect of the present disclosure is the information processing apparatus according to the first aspect, When using the structure information as the initial structure information and executing the calculation program using the set of low-accuracy basis functions, the number of structure optimization steps is smaller than the number of structure optimization steps when executing the calculation program using the set of reference basis functions with the low-accuracy optimized structure information as the initial structure information.
[0018] A twelfth aspect of the present disclosure is an information processing method, A computer of an information processing apparatus that executes a calculation program based on the localized wave basis density functional method, A basis function acquisition step of acquiring a set of reference basis functions for executing the calculation program at a target accuracy level and a set of low-accuracy basis functions for executing the calculation program at an accuracy level lower than the target accuracy level; When receiving an execution instruction of the calculation program for the structure information of the target substance, By using the structure information as the initial structure information and executing the calculation program using the set of low-accuracy basis functions, low-accuracy optimized structure information is obtained. By using the aforementioned low-precision optimized structure information as initial structure information and executing the calculation program using the aforementioned reference basis function set, the optimized structure information with the target precision level is obtained. The process of obtaining the calculation result is executed.
[0019] A thirteenth aspect of this disclosure is a computational support program, The computer of the information processing device that executes a calculation program based on the localized wave basis density functional theory, The calculation program is executed at a target accuracy level, and the calculation program is executed at a lower accuracy level. The process of obtaining a basis function includes obtaining a reference basis function set and a low-precision basis function set. When an instruction to execute the calculation program is received regarding the structural information of the target substance, By using the aforementioned structural information as initial structural information and executing the calculation program using the aforementioned low-precision basis function set, low-precision optimized structural information is obtained. By using the aforementioned low-precision optimized structure information as initial structure information and executing the calculation program using the aforementioned reference basis function set, the optimized structure information with the target precision level is obtained. The process of obtaining the calculation result is executed. [Effects of the Invention]
[0020] This disclosure improves the convenience of users who utilize calculation programs based on the localized wave basis density functional method. [Brief explanation of the drawing]
[0021] [Figure 1] Figure 1 shows an example of the system configuration of an information processing system. [Figure 2] Figure 2 is a diagram illustrating the overview of processing in an information processing system. [Figure 3] Figure 3 shows an example of the hardware configuration of an information processing device. [Figure 4]Figure 4 shows an example of the functional configuration of an information processing device and a server device. [Figure 5] Figure 5 shows an example of various types of information. [Figure 6A] Figure 6A is a first sequence diagram showing the processing flow in the information processing system according to the first embodiment. [Figure 6B] Figure 6B is a second sequence diagram showing the processing flow in the information processing system according to the first embodiment. [Figure 7] Figure 7 shows a comparison of calculation speeds. [Figure 8A] Figure 8A is a first sequence diagram showing the processing flow in the information processing system according to the second embodiment. [Figure 8B] Figure 8B is a second sequence diagram showing the processing flow in the information processing system according to the second embodiment. [Figure 8C] Figure 8C is a third sequence diagram showing the processing flow in the information processing system according to the second embodiment. [Modes for carrying out the invention]
[0022] Each embodiment will be described below with reference to the attached drawings. In this specification and the drawings, components having substantially the same functional configuration are denoted by the same reference numerals, and redundant descriptions will be omitted.
[0023] [First Embodiment] <System Configuration of Information Processing System> First, the system configuration of the entire information processing system equipped with the information processing device according to the first embodiment will be described. Figure 1 is a diagram showing an example of the system configuration of the information processing system. As shown in Figure 1, the information processing system 100 has an information processing device 110 and a server device 130. In the information processing system 100, the information processing device 110 and the server device 130 are connected via a network (not shown).
[0024] The server device 130 consists of one or more units. The server device 130 has a calculation program 140 based on the localized wave basis density functional theory installed on it, and when this calculation program is executed, the server device 130 performs calculation processing based on the localized wave basis density functional theory.
[0025] The information processing device 110 has a calculation support program 120 installed. When this calculation support program is executed, the information processing device 110 assists the user 150 when the user 150 has the server device 130 perform calculation processing based on the localized wave basis density functional theory.
[0026] Specifically, the information processing device 110 acquires various information, such as structural information and basis function sets, of the target substance when instructing the server device 130 to perform calculation processing based on the localized wave basis density functional method.
[0027] The information processing device 110 transmits the acquired information along with execution instructions to the server device 130, causing the server device 130 to perform calculation processing based on the localized wave basis density functional theory. The information processing device 110 receives the calculation results obtained from the server device 130 and displays them to the user 150.
[0028] Here, even if the information processing device 110 is instructed by the user 150 to perform calculation processing at a target accuracy level (i.e., high accuracy), • Situations where the calculation process by server device 130 does not converge, or • Situations where the processing time for calculations performed by server device 130 becomes enormous, To avoid this, the calculation process by the server device 130 is controlled (details will be described later). By controlling the calculation process by the server device 130 in this way, the information processing device 110 can avoid the above situation and improve the convenience of the user 150 who uses the calculation program based on the localized wave basis density functional method.
[0029] <Overview of processing in information processing systems> Next, we will describe the overall processing of the information processing system 100 when the information processing device 110 controls the calculation processing performed by the server device 130. Figure 2 is a diagram illustrating the overview of the processing in the information processing system.
[0030] Of these, Figure 2(a) shows an overview of the processing in the information processing system 100 when the information processing device 110 does not control the calculation processing by the server device 130, as a comparative example.
[0031] If user 150 instructs the system to perform high-precision calculation processing, and the information processing device 110 does not control the calculation processing performed by the server device 130, the server device 130 will, as shown in Figure 2(a), First, a highly accurate structural optimization process is performed to optimize the structural information of the target substance. Next, the optimized structural information is subjected to high-precision calculations (for example, calculations of physical properties).
[0032] In the case of the processing procedure shown in Figure 2(a), • The structural optimization process does not converge, making it impossible to execute this calculation process, or • Although the structural optimization process converged, the structural information of the target substance was not optimized, and therefore, highly accurate calculation results could not be obtained in this calculation process, or The structural optimization process converged, and the structural information of the target substance was optimized, but the computation process took an enormous amount of time. Such situations could occur.
[0033] On the other hand, Figure 2(b) shows an overview of the processing in the information processing system 100 when the information processing device 110 controls the calculation processing by the server device 130.
[0034] When user 150 instructs the system to perform high-precision calculation processing, and the information processing device 110 controls the calculation processing performed by the server device 130, the server device 130 performs the calculation as shown in Figure 2(b). First, a low-precision structural optimization process is performed to optimize the structural information of the target substance with low precision. Next, the structural information of the target substance is optimized to medium accuracy by performing a medium-accuracy structural optimization process on the low-accuracy optimized structural information. Next, the structural information of the target substance is optimized with high precision by performing a high-precision structural optimization process on the structural information optimized with medium precision. Next, high-precision calculations (for example, calculations of physical properties) are performed on the highly optimized structural information.
[0035] In the case of the processing procedure shown in Figure 2(b), The initial structural optimization process performed is a low-precision structural optimization process. This avoids situations where the computation process requires an enormous amount of time. After completing the low-precision structural optimization process, a medium-precision structural optimization process is performed to obtain medium-precision optimized structural information. Then, a high-precision structural optimization process is performed on the obtained medium-precision optimized structural information. This allows for obtaining high-precision optimized structural information for the target substance while avoiding situations where the calculation process does not converge and the main calculation process cannot be executed. This calculation process is performed on highly accurate optimized structure information. Therefore, highly accurate calculation results can be obtained.
[0036] Furthermore, the "precision" of the calculation process when running the calculation program 140 based on the localized wave basis density functional method depends, for example, on the "number of basis functions included in the basis function set".
[0037] <Hardware configuration of the information processing device> Next, the hardware configuration of the information processing device 110 will be described. Figure 3 shows an example of the hardware configuration of the information processing device. As shown in Figure 3, the information processing device 110 includes a processor 301, memory 302, auxiliary storage device 303, interface device 304, communication device 305, and drive device 306. The hardware components of the information processing device 110 are interconnected via a bus 307.
[0038] The processor 301 has various computing devices such as a CPU (Central Processing Unit) and a GPU (Graphics Processing Unit). The processor 301 executes various programs (for example, computation support programs, etc.) by reading them into the memory 302.
[0039] Memory 302 has main memory devices such as ROM (Read Only Memory) and RAM (Random Access Memory). The processor 301 and memory 302 form a so-called computer, and the computer realizes various functions by the processor 301 executing various programs read into memory 302.
[0040] The auxiliary storage device 303 stores various programs and various data used when those programs are executed by the processor 301. For example, the basis function set storage unit 431 and the structure information storage unit 432, which will be described later, are implemented in the auxiliary storage device 303.
[0041] Interface device 304 is a connection device for connecting an operating device 311 and a display device 312, which are examples of user interface devices. Communication device 305 is a communication device for communicating with server device 130 via a network (not shown).
[0042] The drive device 306 is a device for setting the recording medium 313. The recording medium 313 here includes media for recording information optically, electrically, or magnetically, such as CD-ROMs, flexible disks, and magneto-optical disks. The recording medium 313 may also include semiconductor memory such as ROM and flash memory for recording information electrically.
[0043] The various programs to be installed on the auxiliary storage device 303 are installed, for example, when the distributed recording medium 313 is set in the drive device 306 and the various programs recorded on the recording medium 313 are read by the drive device 306. Alternatively, the various programs to be installed on the auxiliary storage device 303 may be installed when they are downloaded from the network via the communication device 305.
[0044] Note that here, only the hardware configuration of the information processing device 110 is described, and the hardware configuration of the server device 130 is omitted. However, the hardware configuration of the server device 130 is basically the same as that of the information processing device 110.
[0045] <Functional Configuration of Information Processing Equipment and Server Equipment> Next, the functional configuration of the information processing device 110 and the server device 130 will be described. Figure 4 is a diagram showing an example of the functional configuration of the information processing device and the server device. As described above, the information processing device 110 has a calculation support program 120 installed on it, and when the calculation support program 120 is executed, the information processing device 110 functions as an information input unit 410 and a calculation support unit 420.
[0046] The information input unit 410 acquires various information necessary for the server device 130 to perform calculation processing based on the localized wave basis density functional theory, based on instructions from the user 150. The various information necessary for performing calculation processing based on the localized wave basis density functional theory includes, for example, • Structural information of the target substance, • Basis set, This includes, etc.
[0047] As shown in Figure 4, the information processing device 110 has a basis function set storage unit 431. The information input unit 410 accepts the user 150's selection of a basis function set by presenting the user 150 with a list of combinations of basis function sets stored in the basis function set storage unit 431. The information input unit 410 reads the basis function set selected by the user 150 from the basis function set storage unit 431 and notifies the calculation support unit 420. The basis function set selected by the user 150 includes: • A set of low-precision basis functions used when performing structural optimization processing at a lower precision level than the target precision level (i.e., low precision), • A set of medium-precision basis functions used when performing an operation at a precision level lower than the target precision level but higher than the low-precision basis function set. • The reference basis set used when performing the structural optimization process and the main calculation process at the target accuracy level (i.e., high accuracy), This includes the number of basis functions included in each basis function set. Low-precision basis set < Medium-precision basis set < Reference basis set, They are in a relationship.
[0048] As shown in Figure 4, the information processing device 110 has a structural information storage unit 432, and the information input unit 410 presents the user 150 with a list of structural information stored in the structural information storage unit 432. This allows the information processing device 110 to receive a selection of structural information for a target substance from the user 150. The information input unit 410 reads the structural information for the target substance selected by the user 150 from the structural information storage unit 432 and notifies the calculation support unit 420. The structural information read from the structural information storage unit 432 predefines the lattice shape, atomic arrangement, and other details of the target substance.
[0049] Furthermore, the method for obtaining the basis function set and structural information by the information input unit 410 is arbitrary and not limited to the method described above.
[0050] The calculation support unit 420 further includes a basis function acquisition unit 421, a structure information acquisition unit 422, and a calculation result acquisition unit 423.
[0051] When the basis function acquisition unit 421 receives notification of a low-precision basis function set, a medium-precision basis function set, and a reference basis function set from the information input unit 410, it transmits the low-precision basis function set to the server device 130 as a basis function set for executing low-precision structural optimization processing.
[0052] When the low-precision structural optimization processing by the server device 130 is completed, the basis function acquisition unit 421 transmits the medium-precision basis function set notified by the information input unit 410 to the server device 130 as a basis function set for executing the medium-precision structural optimization processing.
[0053] When the server device 130 completes the medium-precision structural optimization processing, the basis function acquisition unit 421 transmits the reference basis function set notified by the information input unit 410 to the server device 130 as a basis function set for executing high-precision structural optimization processing.
[0054] When the structural information acquisition unit 422 receives structural information of the target substance from the information input unit 410, it transmits the structural information to the server device 130 as initial structural information for executing a low-precision structural optimization process.
[0055] When the low-precision structural optimization processing by the server device 130 is completed, the structural information acquisition unit 422 acquires low-precision optimized structural information from the server device 130. The structural information acquisition unit 422 transmits the acquired low-precision optimized structural information to the server device 130 as initial structural information for executing medium-precision structural optimization processing.
[0056] When the server device 130 completes the medium-precision structural optimization processing, the structural information acquisition unit 422 acquires medium-precision optimized structural information from the server device 130. The structural information acquisition unit 422 then transmits the acquired medium-precision optimized structural information to the server device 130 as initial structural information for executing high-precision structural optimization processing.
[0057] When the calculation result acquisition unit 423 receives an instruction from the user 150 to perform a low-precision structural optimization process, it transmits the instruction to the server device 130. As a result, the calculation result acquisition unit 423 receives low-precision optimized structural information from the server device 130 and notifies the structural information acquisition unit 422.
[0058] When the calculation result acquisition unit 423 receives an instruction from the user 150 to perform a medium-precision structural optimization process, it transmits the execution instruction to the server device 130. As a result, the calculation result acquisition unit 423 receives medium-precision optimized structural information from the server device 130 and notifies the structural information acquisition unit 422.
[0059] When the calculation result acquisition unit 423 receives an instruction from the user 150 to perform a high-precision structural optimization process, it transmits the instruction to the server device 130. As a result, the calculation result acquisition unit 423 receives high-precision optimized structural information and energy calculation results from the server device 130 and displays them to the user 150.
[0060] When the calculation result acquisition unit 423 receives an instruction from the user 150 to execute a high-precision calculation process (for example, a calculation process for material properties), it transmits the execution instruction to the server device 130. As a result, the calculation result acquisition unit 423 receives the high-precision calculation results (for example, material properties) from the server device 130 and displays the received high-precision calculation results (for example, material properties) to the user 150.
[0061] As described above, the server device 130 has a calculation program 140 based on the localized wave basis density functional theory installed. When this calculation program 140 based on the localized wave basis density functional theory is executed, the server device 130 functions as a calculation unit 440.
[0062] As shown in Figure 4, the calculation unit 440 includes a structure optimization unit 441, an energy calculation unit 442, and a physical property calculation unit 443.
[0063] The structure optimization unit 441 receives a set of low-precision basis functions from the basis function acquisition unit 421 to be used when executing a low-precision structure optimization process. The structure optimization unit 441 receives structural information of the target substance as initial structural information from the structural information acquisition unit 422. The structure optimization unit 441 receives an instruction to execute a low-precision structure optimization process from the calculation result acquisition unit 423. As a result, the structure optimization unit 441 performs a low-precision structure optimization process using the conjugate gradient method on the structural information of the target substance and transmits the low-precision optimized structure information to the information processing device 110.
[0064] The structure optimization unit 441 receives a set of medium-precision basis functions from the basis function acquisition unit 421 to be used when executing a medium-precision structure optimization process. The structure optimization unit 441 receives low-precision optimized structure information as initial structure information from the structure information acquisition unit 422. The structure optimization unit 441 receives an instruction to execute a medium-precision structure optimization process from the calculation result acquisition unit 423. As a result, the structure optimization unit 441 executes a medium-precision structure optimization process using the conjugate gradient method on the low-precision optimized structure information and transmits the medium-precision optimized structure information to the information processing device 110.
[0065] The structural optimization unit 441 receives a set of reference basis functions from the basis function acquisition unit 421 to be used when executing high-precision structural optimization processing. The structural optimization unit 441 receives medium-precision optimized structural information as initial structural information from the structural information acquisition unit 422. The structural optimization unit 441 receives an instruction to execute high-precision structural optimization processing from the calculation result acquisition unit 423. As a result, the structural optimization unit 441 executes high-precision structural optimization processing using the conjugate gradient method on the medium-precision optimized structural information and transmits the high-precision optimized structural information and energy calculation results to the information processing device 110.
[0066] The energy calculation unit 442 performs energy calculations based on the localized wave basis density functional method when the structural optimization unit 441 performs low-precision structural optimization processing, medium-precision structural optimization processing, and high-precision structural optimization processing.
[0067] The physical property calculation unit 443 receives an instruction from the basis function acquisition unit 421 to execute a high-precision main calculation process (physical property calculation process) and receives high-precision optimized structure information from the structure optimization unit 441. Based on this, the physical property calculation unit 443 executes the high-precision physical property calculation process and transmits the physical property calculation results to the information processing device 110.
[0068] <Details of the basis function set storage unit and structural information storage unit> Next, we will describe the details of the basis function sets (low-precision basis function set, medium-precision basis function set, reference basis function set) stored in the basis function set storage unit 431 and the details of the structural information stored in the structural information storage unit 432. Figure 5 shows an example of the various types of information.
[0069] Figure 5(a) shows an example of a basis function set stored in the basis function set storage unit 431. As shown in Figure 5(a), the basis function set 510 includes the following information items: "ID", "low-precision basis function set", "medium-precision basis function set", and "reference basis function set". "ID" stores an identifier for identifying combinations of multiple basis function sets (low-precision basis function set, medium-precision basis function set, and reference basis function set) with different precision levels. "Low-precision basis function set" stores the low-precision basis function set. "Medium-precision basis function set" stores the medium-precision basis function set. "Reference basis function set" stores the reference basis function set.
[0070] Figure 5(b) shows an example of structural information stored in the structural information storage unit 432. As shown in Figure 5(b), the structural information 520 includes the following information items: "ID", "Substance Name", and "Structural Information". "ID" stores an identifier for identifying the structural information. "Substance Name" stores the name of the substance identified by the structural information. "Structural Information" stores the structural information (lattice shape, atomic arrangement, etc.) of the substance having the corresponding substance name.
[0071] <Processing flow in an information processing system> Next, the processing flow in the information processing system 100 will be described. Figures 6A and 6B are the first and second sequence diagrams showing the processing flow in the information processing system according to the first embodiment. In Figure 6A, steps S601 to S606 and S621 and S622 represent the steps of the low-precision structural optimization process. Steps S607 to S611 and S623 and S624 represent the steps of the medium-precision structural optimization process. In Figure 6B, steps S612 to S615 and S625 and S626 represent the steps of the high-precision structural optimization process. Steps S616 and S627 to S629 represent the steps of the high-precision main calculation process.
[0072] In step S601 of Figure 6A, the information input unit 410 receives input of structural information of the target substance from the user 150 and notifies the structural information acquisition unit 422 of the received structural information.
[0073] In step S602, the information input unit 410 receives input of a low-precision basis function set from the user 150 and notifies the basis function acquisition unit 421 of the low-precision basis function set that it has received.
[0074] In step S603, the basis function acquisition unit 421 transmits a low-precision basis function set to the calculation unit 440.
[0075] In step S604, the structural information acquisition unit 422 transmits the structural information notified by the information input unit 410 in step S601 to the calculation unit 440 as initial structural information.
[0076] In step S605, the calculation result acquisition unit 423 receives an instruction from the user 150 to execute a low-precision structural optimization process and transmits it to the calculation unit 440.
[0077] In step S621, the structural optimization unit 441 and the energy calculation unit 442 of the calculation unit 440 optimize the initial structural information and generate low-precision optimized structural information by performing energy calculations under a low-precision basis function set.
[0078] In step S622, the structural optimization unit 441 transmits low-precision optimized structural information to the calculation support unit 420. The calculation result acquisition unit 423 of the calculation support unit 420 acquires the low-precision optimized structural information.
[0079] In step S606, the calculation result acquisition unit 423 notifies the structural information acquisition unit 422 of the low-precision optimized structure information.
[0080] In step S607, the information input unit 410 receives input of a medium-precision basis function set from the user 150 and notifies the basis function acquisition unit 421 of the medium-precision basis function set that it has received.
[0081] In step S608, the basis function acquisition unit 421 transmits the medium-precision basis function set to the calculation unit 440.
[0082] In step S609, the structural information acquisition unit 422 transmits the low-precision optimized structural information notified by the calculation result acquisition unit 423 in step S606 to the calculation unit 440 as initial structural information.
[0083] In step S610, the calculation result acquisition unit 423 receives an instruction from the user 150 to execute a medium-precision structural optimization process and transmits it to the calculation unit 440.
[0084] In step S623, the structural optimization unit 441 and the energy calculation unit 442 of the calculation unit 440 optimize the initial structural information and generate optimized structural information with medium accuracy by performing energy calculations under a set of medium-precision basis functions.
[0085] In step S624, the structural optimization unit 441 transmits moderately accurate optimized structural information to the calculation support unit 420. The calculation result acquisition unit 423 of the calculation support unit 420 acquires the moderately accurate optimized structural information.
[0086] In step S611, the calculation result acquisition unit 423 notifies the structural information acquisition unit 422 of the optimized structure information with moderate accuracy.
[0087] In step S612 of Figure 6B, the information input unit 410 receives input of a reference basis function set from the user 150 and notifies the basis function acquisition unit 421 of the received reference basis function set.
[0088] In step S613, the basis function acquisition unit 421 transmits the reference basis function set to the calculation unit 440.
[0089] In step S614, the structural information acquisition unit 422 transmits the medium-precision optimized structural information notified by the calculation result acquisition unit 423 in step S611 to the calculation unit 440 as initial structural information.
[0090] In step S615, the calculation result acquisition unit 423 receives an instruction from the user 150 to execute a high-precision structural optimization process and transmits it to the calculation unit 440.
[0091] In step S625, the structural optimization unit 441 and the energy calculation unit 442 of the calculation unit 440 optimize the initial structural information and generate highly accurate optimized structural information by performing energy calculations under a set of reference basis functions.
[0092] In step S626, the structural optimization unit 441 transmits highly accurate optimized structural information and energy calculation results to the calculation support unit 420. The calculation result acquisition unit 423 of the calculation support unit 420 acquires highly accurate optimized structural information and energy calculation results and displays them to the user 150.
[0093] In step S616, the calculation result acquisition unit 423 receives an instruction from the user 150 to execute a physical property calculation process for highly accurate optimized structural information and transmits it to the calculation unit 440.
[0094] In step S627, the structural optimization unit 441 of the calculation unit 440 notifies the physical property calculation unit 443 of highly accurate optimized structural information to be used in the physical property calculation process.
[0095] In step S628, the physical property calculation unit 443 of the calculation unit 440 calculates physical property values for highly accurate optimized structural information.
[0096] In step S629, the physical property calculation unit 443 transmits the physical property calculation results to the calculation support unit 420. The calculation result acquisition unit 423 of the calculation support unit 420 acquires the physical property calculation results and displays them to the user 150.
[0097] <Calculation example> Next, we will explain a comparison between the calculation time when the calculation process by the server device 130 is not controlled (Figure 2(a)) and the calculation time when the calculation process by the server device 130 is controlled (Figure 2(b)). Figure 7 shows a comparison of calculation times.
[0098] In Figure 7(a), the horizontal axis represents the types of molecules of the target substances used for comparison, and the vertical axis represents the calculation time. For all types of molecules, the calculation time when the calculation process was controlled (Figure 2(b)) was shorter than the calculation time when the calculation process was not controlled (Figure 2(a)).
[0099] Figure 7(b) shows box plots representing the computation time for each type of molecule when the computation process is not controlled (Figure 2(a)) and box plots representing the computation time for each type of molecule when the computation process is controlled (Figure 2(b)).
[0100] As shown in Figure 7(b), the median computation time for each type of molecule when the computation process was not controlled (Figure 2(a)) was 1.46 times the median computation time for each type of molecule when the computation process was controlled (Figure 2(b)).
[0101] <Summary> As is clear from the above explanation, the information processing device 110 according to the first embodiment, which executes a calculation program based on the localized wave basis density functional method, • Obtain a set of reference basis functions that will allow the calculation program to run at the target level of accuracy. • Obtain a set of low-precision basis functions that will run the calculation program at a lower precision level than the target precision level. Obtain a set of medium-precision basis functions that allows the calculation program to run at a precision lower than the target precision level, but higher than the low-precision basis function set. When an instruction is received to run a calculation program using the structural information of the target substance, the calculation program is run using the structural information of the target substance as initial structural information and a low-precision basis set to obtain low-precision optimized structural information. Next, using the low-precision optimized structure information as initial structure information, the calculation program is executed using a medium-precision basis function set to obtain the medium-precision optimized structure information. Next, using the medium-precision optimized structure information as initial structure information, the calculation program is executed using a reference basis function set to obtain optimized structure information at the target precision level.
[0102] In this way, by configuring the server device to control the structure optimization process, in the case of the information processing device 110 according to the first embodiment, The initial structural optimization process performed is a low-precision structural optimization process. This avoids situations where the computation process takes an enormous amount of time. After the low-precision structural optimization process is completed, a medium-precision structural optimization process is executed to obtain medium-precision optimized structural information, and then a high-precision structural optimization process is performed on the obtained medium-precision optimized structural information. This allows for obtaining high-precision optimized structural information for the target substance while avoiding situations where the calculation process does not converge and the main calculation process cannot be executed. This calculation process is performed on highly accurate optimized structure information. Therefore, highly accurate calculation results can be obtained.
[0103] As a result, the convenience for users utilizing calculation programs based on localized wave basis density functional theory can be improved.
[0104] [Second Embodiment] In the first embodiment described above, the user 150 is configured to input execution instructions for low-precision structural optimization processing, medium-precision structural optimization processing, high-precision structural optimization processing, and high-precision main calculation processing. In addition, the user 150 is configured to input a low-precision basis function set, a medium-precision basis function set, and a reference basis function set.
[0105] However, the information processing device 110 may be configured so that the user 150 only inputs execution instructions for high-precision main calculation processing as execution instructions. Specifically, the calculation support unit 420 of the information processing device 110 may be configured to automatically generate execution instructions for low-precision structural optimization processing, medium-precision structural optimization processing, and high-precision structural optimization processing, and send them to the server device 130.
[0106] Furthermore, the information processing device 110 may be configured to accept only a reference basis function set as input by the user 150. Specifically, the calculation support unit 420 of the information processing device 110 may be configured to automatically read the low-precision basis function set and the medium-precision basis function set and transmit them to the server device 130. The second embodiment will now be described, focusing on the differences from the first embodiment described above.
[0107] <Processing flow in an information processing system> First, the processing flow in the information processing system 100 according to the second embodiment will be described. Figures 8A to 8C are the first to third sequence diagrams showing the processing flow in the information processing system according to the second embodiment. The differences from Figures 6A to 6B described in the first embodiment above are steps S801 to S803 and S811 to S813.
[0108] In step S801 of Figure 8A, the information input unit 410 receives input of a reference basis function set from the user 150 and notifies the basis function acquisition unit 421 of the received reference basis function set.
[0109] In step S802, when the basis function acquisition unit 421 receives notification of a reference basis function set from the information input unit 410, it refers to the basis function set storage unit 431. The basis function acquisition unit 421 reads out the low-precision basis function set corresponding to the notified reference basis function set.
[0110] As a result, in step S603, the basis function acquisition unit 421 transmits the read low-precision basis function set to the calculation unit 440.
[0111] In step S803, the basis function acquisition unit 421 generates an instruction to execute a low-precision structural optimization process and notifies the calculation result acquisition unit 423.
[0112] As a result, in step S605, the calculation result acquisition unit 423 sends an instruction to the calculation unit 440 to execute a low-precision structural optimization process.
[0113] In step S811 of Figure 8B, the basis function acquisition unit 421 reads out a medium-precision basis function set corresponding to the reference basis function set notified in step S801.
[0114] As a result, in step S608, the basis function acquisition unit 421 transmits the read-out medium-precision basis function set to the calculation unit 440.
[0115] In step S812, the basis function acquisition unit 421 generates an instruction to execute a medium-precision structural optimization process and notifies the calculation result acquisition unit 423.
[0116] As a result, in step S610, the calculation result acquisition unit 423 sends an instruction to the calculation unit 440 to execute a medium-precision structural optimization process.
[0117] In step S813, the basis function acquisition unit 421 generates an instruction to execute a high-precision structural optimization process and notifies the calculation result acquisition unit 423.
[0118] As a result, in step S615, the calculation result acquisition unit 423 sends an instruction to the calculation unit 440 to execute a high-precision structural optimization process.
[0119] <Summary> As is clear from the above description, the information processing device 110 according to the second embodiment is When a reference basis function set is input by user 150, the low-precision basis function set is read and sent to the server device 130, and an execution instruction for low-precision structural optimization processing is generated and sent to the server device 130. When low-precision optimized structure information is obtained from the server device 130, the medium-precision basis function set is read and sent to the server device 130, and an execution instruction for the medium-precision structure optimization process is generated and sent to the server device 130. When medium-precision optimized structure information is obtained from the server device 130, a reference basis function set is sent to the server device 130, and an execution instruction for high-precision structure optimization processing is generated and sent to the server device 130.
[0120] As a result, user 150 only needs to input an execution instruction for the high-precision main calculation process as an execution instruction, and only needs to input a reference basis function set as a basis function set. In other words, according to the information processing device 110 of the second embodiment, even when the same information input and execution instructions are performed as in the conventional, The initial structural optimization process performed is a low-precision structural optimization process. This avoids situations where the computation process takes an enormous amount of time. After the low-precision structural optimization process is completed, a medium-precision structural optimization process is executed to obtain medium-precision optimized structural information, and then a high-precision structural optimization process is performed on the obtained medium-precision optimized structural information. This allows for obtaining high-precision optimized structural information for the target substance while avoiding situations where the calculation process does not converge and the main calculation process cannot be executed. This calculation process is performed on highly accurate optimized structure information. Therefore, highly accurate calculation results can be obtained.
[0121] As a result, according to the second embodiment, the convenience for users utilizing a calculation program based on the localized wave basis density functional method can be further improved.
[0122] [Third Embodiment] In the first and second embodiments described above, the basis function set storage unit 431 is configured to pre-store combinations of low-precision basis function sets, medium-precision basis function sets, and reference basis function sets. Furthermore, in the second embodiment described above, when a reference basis function set is input, the corresponding low-precision reference basis function set and medium-precision basis function set are read from the basis function set storage unit 431.
[0123] However, the method for obtaining a low-precision base function set and a medium-precision base function set when a base function set is input is not limited to this. For example, the low-precision base function set and the medium-precision base function set may be configured to be generated based on the base function set. Specifically, a medium-precision base function set may be generated by removing some of the basis functions included in the base function set, and a low-precision base function set may be generated by removing some of the generated medium-precision base function set.
[0124] [Fourth Embodiment] In each of the above embodiments, the process of optimizing the structural information of the target substance is performed in three stages (low-precision structural optimization, medium-precision structural optimization, and high-precision structural optimization).
[0125] However, the process of optimizing the structural information of the target substance may be configured to be executed in two stages (for example, a low-precision structural optimization process and a high-precision structural optimization process). In this case, the basis function acquisition unit 421 acquires a low-precision basis function set and a reference basis function set. The calculation support unit 420 uses the low-precision optimized structural information as initial structural information and the reference basis function set to execute a calculation program based on localized wave basis density functional theory to acquire optimized structural information at the target precision level.
[0126] Alternatively, the process of optimizing the structural information of the target substance may be configured to be executed in, for example, N steps (where N is an integer greater than or equal to 2). In this case, the basis function acquisition unit 421 acquires the Nth set of basis functions, which executes a calculation program based on localized wave basis function density functional theory at an Nth precision level higher than the N-1th precision level (where N is an integer greater than or equal to 2). The calculation support unit 420 uses the N-1th precision level of optimized structural information as initial structural information and executes a calculation program based on localized wave basis function density functional theory using the Nth set of basis functions to acquire the Nth precision level of optimized structural information.
[0127] [Fifth Embodiment] In the embodiments described above, the number of basis functions included in the basis function set was changed according to the accuracy level. However, the objects to be changed according to the accuracy level are not limited to these. For example, in low-accuracy structural optimization processes, a configuration may be adopted to terminate structural optimization processes that are not converging early. Specifically, • Let C1 be the number of structural optimization steps when running a calculation program using a low-precision basis set, with the structural information of the target substance as the initial structural information. • When executing a calculation program using a reference basis set, with low-precision optimized structure information as the initial structure information, the number of structure optimization steps is C2(C1 <C2)とする、 By doing so (that is, by making C1 smaller than C2), the structure optimization process can be terminated early. Alternatively, • The structural information of the target substance is used as the initial structural information, and the convergence criterion value when running the calculation program using a low-precision basis set is defined as Th1. • When executing a calculation program using low-precision optimized structure information as initial structure information and a reference basis function set, the convergence criterion is defined as Th2 (Th1 > Th2). By doing so (that is, by making Th1 larger than Th2), the structure optimization process can be terminated early.
[0128] [Sixth Embodiment] In each of the above embodiments, it was explained that structural information is pre-stored in the structural information storage unit 432. However, structural information is, for example, • Commercial molecular modeling software such as GaussView, or • Avogadro, an open-source molecular modeling software. It may be generated by the following methods. Specifically, structural information may be generated in these molecular modeling software programs by generating three-dimensional coordinate information of the atoms constituting the molecule in response to user input operations.
[0129] Furthermore, structural information may be generated, for example, by Molecular Mechanics (MM) using classical force fields. Alternatively, it may be generated by semi-empirical methods including classical force fields, or by machine learning methods using neural network potentials. As for semi-empirical methods, ·AM1 (Austin Model 1) method, ·PM3 (Parameterized Model number 3) method, ·PM6 (Parameterized Model number 6) method, ·PM7 (Parameterized Model number 7) method, ·MINDO (Modified Intermediate Neglect of Diatomic Overlap) method, ·MNDO (Modified Neglect of Diatomic Overlap) method, Hückel method, • Extended Hückel method, Examples include semi-empirical molecular orbital methods. Other machine learning techniques besides neural network potentials include, for example, Gaussian approximation potentials.
[0130] Although specific examples of calculation programs based on the localized wave basis density functional method were not mentioned in the above embodiments, examples of such calculation programs include Gaussian and GAMES.
[0131] It should be noted that the present invention is not limited to the configurations shown in the above embodiments, including combinations with other elements. These aspects can be modified without departing from the spirit of the present invention and can be appropriately determined according to their application.
[0132] This application claims priority based on Japanese Patent Application No. 2024-112744, filed on 12 July 2024, which is incorporated herein by reference to the entire contents of the said Japanese Patent Application. [Explanation of Symbols]
[0133] 100: Information Processing Systems 110: Information Processing Device 120: Calculation support program 130: Server device 140: Calculation program based on localized wave basis density functional theory 410: Information Input Section 420:Calculation support department 421: Basis function acquisition section 422: Structure information acquisition unit 423: Calculation result acquisition part 431: Base function set storage 432: Structural Information Storage Unit 440: Calculation section 441: Structural Optimization Department 442: Energy Calculation Unit 443: Physical property calculation department 510: Basis set 520 :Structural information
Claims
1. An information processing device that executes a calculation program based on the localized wave basis density functional method, A basis function acquisition unit that acquires the following: a reference basis function set for executing the calculation program at a target precision level; a first basis function set for executing at a precision level lower than the target precision level; and an Nth basis function set for executing at an N-th precision level lower than the target precision level and higher than the (N-1)th precision level (where N is an integer greater than or equal to 2), wherein the Nth basis function set is generated stepwise by increasing the precision level by one step at a time. When an instruction to execute the calculation program is received regarding the structural information of the target substance, Using the aforementioned structural information as initial structural information, the calculation program is executed using the first basis function set to obtain optimized structural information of a first accuracy level. By using the N-1 precision level optimization structure information as initial structure information and executing the calculation program using the N basis function set, the Nth precision level optimization structure information is obtained. By using the aforementioned N-level optimization structure information as initial structure information and executing the calculation program using the aforementioned reference basis function set, the target-level optimization structure information is obtained. Calculation result acquisition unit and An information processing device having
2. The number of basis functions included in the first basis function set is less than the number of basis functions included in the reference basis function set. The information processing apparatus according to claim 1.
3. It has a storage unit that stores combinations of basis function sets with different levels of precision. The information processing apparatus according to claim 1.
4. Each basis function set obtained by the basis function acquisition unit is a basis function set selected by the user from among a plurality of basis function sets included in the combination. The information processing apparatus according to claim 3.
5. When the basis function acquisition unit acquires a reference basis function set selected by the user from among a plurality of basis function sets included in the combination, it acquires other basis function sets included in the combination that correspond to the reference basis function set. The information processing apparatus according to claim 4.
6. When the basis function acquisition unit acquires a reference basis function set input by the user, it generates a basis function set with a lower precision level than the reference basis function set based on the reference basis function set. The information processing apparatus according to claim 1.
7. The convergence criterion value when executing the calculation program using the first basis set with the aforementioned structural information as initial structural information is greater than the convergence criterion value when executing the calculation program using the reference basis set with the aforementioned N precision level optimized structural information as initial structural information. The information processing apparatus according to claim 1.
8. The structural information of the aforementioned target substance is generated using one of the following methods: molecular mechanics, semi-empirical methods, or machine learning methods. The information processing apparatus according to claim 1.
9. The number of structural optimization steps when executing the calculation program using the first basis set with the aforementioned structural information as initial structural information is smaller than the number of structural optimization steps when executing the calculation program using the reference basis set with the aforementioned optimized structural information of accuracy level N as initial structural information. The information processing apparatus according to claim 1.
10. The computer of the information processing device that executes a calculation program based on the localized wave basis density functional theory, A basis function acquisition step, which involves obtaining a set of base functions for the calculation program, which includes a reference base function set for execution at a target precision level, a low precision base function set for execution at a precision level lower than the target precision level, and an Nth base function set for execution at a precision level lower than the target precision level and higher than the (N-1)th precision level (where N is an integer greater than or equal to 2), wherein the Nth base function set is generated stepwise by increasing the precision level by one step at a time. When an instruction to execute the calculation program is received regarding the structural information of the target substance, Using the aforementioned structural information as initial structural information, the calculation program is executed using the first basis function set to obtain optimized structural information of a first accuracy level. By using the N-1 precision level optimization structure information as initial structure information and executing the calculation program using the N basis function set, the Nth precision level optimization structure information is obtained. By using the aforementioned N-level optimization structure information as initial structure information and executing the calculation program using the aforementioned reference basis function set, the target-level optimization structure information is obtained. Calculation result acquisition process and An information processing method that performs the following.
11. The computer of the information processing device that executes a calculation program based on the localized wave basis density functional theory, A basis function acquisition step, which involves obtaining a set of base functions for the calculation program, which includes a reference base function set for execution at a target precision level, a low precision base function set for execution at a precision level lower than the target precision level, and an Nth base function set for execution at a precision level lower than the target precision level and higher than the (N-1)th precision level (where N is an integer greater than or equal to 2), wherein the Nth base function set is generated stepwise by increasing the precision level by one step at a time. When an instruction to execute the calculation program is received regarding the structural information of the target substance, Using the aforementioned structural information as initial structural information, the calculation program is executed using the first basis function set to obtain optimized structural information of a first accuracy level. By using the N-1 precision level optimization structure information as initial structure information and executing the calculation program using the N basis function set, the Nth precision level optimization structure information is obtained. By using the aforementioned N-level optimization structure information as initial structure information and executing the calculation program using the aforementioned reference basis function set, the target-level optimization structure information is obtained. Calculation result acquisition process and A computational support program to perform the necessary actions.