Calculation method, calculation device, and program

By adding auxiliary spins and new interaction coefficients to Ising models, the method addresses bit width restrictions, maintaining the ground state and enabling accurate solution determination in combinatorial optimization problems.

JP7578277B2Active Publication Date: 2024-11-06WASEDA UNIV
View PDF 1 Cites 0 Cited by

Patent Information

Application Number
JP2020207989
Authority / Receiving Office
JP · JP
Patent Type
Patents
Current Assignee / Owner
Filing Date
2020-12-15
Publication Date
2024-11-06
Estimated Expiration
2040-12-15

AI Technical Summary

Technical Problem

Existing hardware limitations restrict the bit widths of interaction coefficients and external magnetic field coefficients in Ising models, leading to mutations in the ground state when bit widths are reduced, preventing the attainment of desired solutions.

Method used

A method that adds auxiliary spins to the Ising model and sets new interaction coefficients to expand the model, allowing bit widths to be reduced without altering the ground state, using a processor to execute these calculations.

Benefits of technology

Enables the reduction of bit widths in Ising models without changing the ground state, ensuring accurate determination of solutions in combinatorial optimization problems.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 0007578277000031
    Figure 0007578277000031
  • Figure 0007578277000032
    Figure 0007578277000032
  • Figure 0007578277000033
    Figure 0007578277000033
Patent Text Reader

Abstract

To provide a calculation method, a calculation device, and a program, for reducing a bit width of an Ising model without altering a base state.SOLUTION: In a calculation method for reducing a bit width of an Ising model which is expressed by a plurality of spins (σ1 to σ4), an interaction coefficient among the plurality of spins, an external magnetic field coefficient acting on each of the plurality of spins, a processor adds auxiliary spins (a1 to a6) to an Ising model 10, and sets a new interaction coefficient among at least one spin constituting the Ising model 10 and the auxiliary spins, for computing a new Ising model 80. Accordingly, the bit width of the interaction coefficient and / or the bit width of the external magnetic field coefficient are / is reduced.SELECTED DRAWING: Figure 8
Need to check novelty before this filing date? Find Prior Art

Description

[Technical field]

[0001] The present invention relates to a calculation method, a calculation device, and a program. [Background technology]

[0002] There is a known technique for converting a combinatorial optimization problem, which involves selecting the best combination from among a large number of combinations, into an Ising model and solving it using an annealing machine or an Ising machine (see, for example, Non-Patent Document 1). [Prior art documents] [Non-patent literature]

[0003] [Non-Patent Document 1] K. Takehara, D. Oku, Y. Matsuda, S. Tanaka, and N. Togawa, "A multiple coefficients trial method to solve combinatorial optimization problems for simulated-annealing-based ising machines," in Proc. IEEE 9th International Conference on Consumer Electronics (ICCE-Berlin). ieeexplore.ieee.org, Sep. 2019, pp. 64-69. Summary of the Invention [Problem to be solved by the invention]

[0004] The bit width of the interaction coefficients and external magnetic field coefficients of the Ising model input to the annealing machine or Ising machine is restricted in many hardware, so that arbitrary values ​​cannot be used. If the bit width is reduced by truncating the lower bits using the shift method, the ground state in which the energy of the Ising model is minimized is mutated, and a problem occurs in that the desired solution cannot be obtained.

[0005] The present invention has been made in consideration of the above problems, and has an object to provide a calculation method, a calculation device, and a program for reducing the bit width of an Ising model without changing the ground state. [Means for solving the problem]

[0006] A calculation method according to the present invention is a calculation method for reducing the bit width of an Ising model represented by a plurality of spins, an interaction coefficient between the plurality of spins, and an external magnetic field coefficient acting on each of the plurality of spins, and by adding an auxiliary spin to the Ising model and setting a new interaction coefficient between at least one of the plurality of spins and the auxiliary spin to obtain a new Ising model, the bit width of at least one of the interaction coefficient between the plurality of spins and the external magnetic field coefficient acting on each of the plurality of spins is reduced.

[0007] The computing device of the present invention is a computing device for reducing the bit width of an Ising model represented by a plurality of spins, an interaction coefficient between the plurality of spins, and an external magnetic field coefficient acting on each of the plurality of spins, and includes a processor that adds an auxiliary spin to the Ising model, sets a new interaction coefficient between at least one of the plurality of spins and the auxiliary spin to obtain a new Ising model, and thereby reduces the bit width of at least one of the interaction coefficient between the plurality of spins and the external magnetic field coefficient acting on each of the plurality of spins.

[0008] The program of the present invention is a program for causing a computer to execute a calculation method for reducing the bit width of an Ising model represented by a plurality of spins, an interaction coefficient between the plurality of spins, and an external magnetic field coefficient acting on each of the plurality of spins, and causes the computer to execute a process of reducing the bit width of at least one of the interaction coefficients between the plurality of spins and the external magnetic field coefficient acting on each of the plurality of spins by adding an auxiliary spin to the Ising model and setting a new interaction coefficient between at least one of the plurality of spins and the auxiliary spin to obtain a new Ising model. Effect of the Invention

[0009] According to the present invention, by adding an auxiliary spin to an Ising model and setting a new interaction coefficient between the spin of the Ising model and the added auxiliary spin to obtain a new Ising model, it is possible to reduce the bit width of the Ising model without changing the ground state. [Brief description of the drawings]

[0010] [Figure 1] FIG. 1 is a schematic diagram showing an example in which a shift method is applied to an Ising model. [Diagram 2] FIG. 2 is a block diagram showing the hardware configuration of a computing device according to the present embodiment. [Diagram 3] FIG. 13 is a schematic diagram illustrating a method for expanding an interaction coefficient. [Figure 4] FIG. 11 is a schematic diagram illustrating a method for reducing the bit width of an interaction coefficient. [Diagram 5] FIG. 13 is a schematic diagram showing an example of reducing the bit width of an interaction coefficient. [Figure 6] 11 is a schematic diagram illustrating a method for reducing the bit width of an external magnetic field coefficient. FIG. [Figure 7] FIG. 13 is a schematic diagram showing an example of reducing the bit width of the external magnetic field coefficient. [Figure 8] FIG. 1 is a schematic diagram showing an example in which the calculation method according to the present embodiment is applied to an Ising model. [Figure 9]FIG. 1 is a schematic diagram showing an example in which an integer division problem is represented by an Ising model. [Figure 10] FIG. 1 is a schematic diagram showing an example of a vertex cover problem and its optimal solution. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0011] Hereinafter, an embodiment of the present invention will be described with reference to the drawings. In this embodiment, integers include a sign bit, and even if the sign of an integer is changed, its bit width does not change. Therefore, an n-bit integer is [-(2 n-1 -1), 2 n-1 It refers to any integer in the range [−1].

[0012] <Ising model> The Ising model is a statistical mechanical model defined on an undirected graph G = (V, E). Here, V is a set of vertices, and E is a set of edges. The Ising model is represented by multiple spins, interaction coefficients between the multiple spins, and external magnetic field coefficients acting on each of the multiple spins. Spin σ i is defined on vertex i∈V and takes the values ​​+1, -1 (or up, down). The interaction coefficient J i、j is defined on the edge (i, j) ∈ E, and σ i and σ j The weight of the connection with the external magnetic field coefficient h i is defined on vertex i∈V with spin σ i It represents the force acting on

[0013] The energy function (Hamiltonian) of the Ising model is given by equation (1).

number

[0014] In formula (1), constants that are not dependent on spin are omitted. In this embodiment, J i、j and i Let be an integer.

[0015] The Ising model is stable in the lowest energy state (ground state). It is expected that the optimal solution to a combinatorial optimization problem can be obtained if it can be mapped to the ground state of the Ising model.

[0016] However, as described above, the bit width of the interaction coefficients and external magnetic field coefficients of the Ising model input to the annealing machine or Ising machine is limited in many hardware. Therefore, when interaction coefficients and external magnetic field coefficients with a bit width larger than the bit width that can be loaded into the annealing machine or Ising machine are input, it is necessary to reduce the bit width.

[0017] <Shift method> The shift method is a naive method for reducing the bit width of an Ising model. Figure 1 shows an Ising model 10 defined by four vertices and six edges as an example of an Ising model. In the Ising model 10, J 1、2 =3, J 1、4 =2, J 2、3 =J 2、4 =J 3、4 =J 1、3 =1, h1=3, h2=-2, h3=h4=0. In Figure 1, the arrows at σ1 and σ2 indicate the direction of the external magnetic field. The spin of the ground state of the Ising model10 is (σ1, σ2, σ3, σ4)=(+1, +1, +1, +1).

[0018] Including the sign bit, the bit width of the values ​​"2", "3", and "-2" is 3 bits, and the bit width of the value "1" is 2 bits. In the shift method, the number of bits is reduced by bit shifting. By shifting the 3-bit values ​​"2", "3", and "-2" of the Ising model 10 by 1 bit (right bit shift), the 2-bit values ​​"1", "1", and "-1" are obtained, respectively.

[0019] In this way, when the above-mentioned shift method is applied to the Ising model 10 expressed with a bit width of 3 bits [-3, +3], an Ising model 12 (right side of Fig. 1) expressed with a bit width of 2 bits [-1, +1] is obtained. However, as shown in Fig. 1, there are four types of spin in the ground state of the Ising model 12, and although one of these four types of spin matches the spin in the ground state of the Ising model 10, the ground states of both Ising models do not match completely.

[0020] <Present embodiment> Therefore, in this embodiment, a method for reducing the bit width without changing the ground state of the Ising model is proposed.

[0021] First, the hardware configuration of a computing device 100 that executes the computing method according to this embodiment will be described with reference to FIG.

[0022] As shown in FIG. 2, the computing device 100 includes a processor 102, a memory 104, a storage device 106, an input unit 108, a display 110, and a network interface 112, and these devices are connected via a bus.

[0023] The processor 102 has a central processing unit (CPU) and the like, and executes various processes according to programs stored in the memory 104 and the storage device 106. The calculation method executed by the processor 102 will be described in detail later.

[0024] As the processor 102, instead of a general-purpose computer such as a CPU, a dedicated computer such as an Application Specific Integrated Circuit (ASIC) or a Field Programmable Gate Array (FPGA) for executing the calculation method according to this embodiment may be used.

[0025] The memory 104 includes a read only memory (ROM) and a random access memory (RAM). The ROM stores a boot program such as a BIOS. When the processor 102 reads a program stored in the ROM or a program stored in the storage device 106, the program is loaded into the RAM.

[0026] The storage device 106 has a computer-readable storage medium, and stores a program for executing the calculation method according to the present embodiment, data necessary for executing the program, etc. Examples of the computer-readable storage medium include a hard disk drive (HDD), a solid state drive (SSD), an optical disk, etc.

[0027] The program executed by the processor 102 may be stored in another computer connected via a network, and the processor 102 may read the program from the other computer via the network interface 112.

[0028] The input unit 108 has input devices such as a mouse, a keyboard, etc. The display 110 is configured with a Liquid Crystal Display (LCD) or the like, and displays the results of the processing executed by the processor 102.

[0029] The network interface 112 is an interface for connecting the computing device 100 to a network, such as a Local Area Network (LAN), a Wide Area Network (WAN), and / or the Internet.

[0030] Next, the calculation method executed by the processor 102 will be described with reference to FIGS.

[0031] Before describing a method for reducing the bit width of the interaction coefficients of the Ising model, a method for expanding the interaction coefficients by adding auxiliary spins will be described with reference to FIG.

[0032] As shown in Figure 3, in the Ising model (Model 3-1), there are two spins σ i and σ j is the interaction coefficient J i、j First, we add an auxiliary spin σ x Then, add σ i and σ x and the edge (i, x) between x and σ j Then, add an edge (j, x) between (i, x) and (j, x) and assign a new interaction coefficient J i、j and its absolute value |J i、j In this way, we set σ x Add σ i and σ j J between i、j By expanding this, the extended Ising model (Model 3-2) is obtained.

[0033] If the energy of the Ising model (model 3-1) shown in FIG. 3 is H1, the energy H1 is expressed by the formula (2). In the following, the interaction coefficient J i、j Terms that do not depend on are omitted.

number

[0034] If the energy of the extended Ising model (Model 3-2) shown in Figure 3 is H′1, then J i、j >0, H′1 is expressed by equation (3).

number

[0035] Auxiliary spin σ x takes the value +1 or -1. σ xAssuming that is the spin that minimizes H′1, equation (3) can be expressed as equation (4).

number

[0036] The difference in energy between Model 3-1 and Model 3-2 (the difference between Eq. (2) and Eq. (4)) is |J i、j |Becomes true.

[0037] J i、j <0, H′1 is expressed by equation (5).

number

[0038] σ x Assuming that is the spin that minimizes H′1, equation (5) can be expressed as equation (6).

number

[0039] The difference in energy between Model 3-1 and Model 3-2 (the difference between Eq. (2) and Eq. (6)) is |J i、j |Becomes true.

[0040] In this way, σ x is the spin that minimizes H′1, the energy difference between the original Ising model (Model 3-1) and the extended Ising model (Model 3-2) is always |J i、j Therefore, the energy difference between the ground state of Model 3-1 and the ground state of Model 3-2 is |J i、j |It must be. |J i、j Since the value of | does not depend on the spin state, the spin σ that gives the ground state in Model 3-1 is i , σ j gives the ground state in model 3-2 as well. That is, as shown in Figure 3, if we add an auxiliary spin and an edge to the original Ising model and focus on the original spin, we can obtain the same ground state.

[0041] Next, a method for reducing the bit width of the interaction coefficients will be described with reference to FIG. As shown in Figure 4, in the Ising model (Model 4-1), there are two spins σ i and σ j is the interaction coefficient J i、j It is assumed that the devices are connected in this manner.

[0042] First, J i、j =J′ i、j +J″ i、j So that, J i、j The first interaction coefficient J′ i、j and the second interaction coefficient J″ i、j As shown in Figure 4, J′ i、j and J″ i、j In both cases, σ i and σ j And J′ i、j and J″ i、j One of the two (J′ in Fig. 4) i、j ), the extended Ising model (Model 4-2) is obtained by applying the method for expanding the interaction coefficient shown in Figure 3.

[0043] As described above, if an auxiliary spin that minimizes the energy of the extended Ising model is added to the original Ising model to extend the interaction coefficient, the ground state of the extended Ising model coincides with the ground state of the original Ising model. Therefore, by repeating the process shown in FIG. 4 until the bit width becomes large enough to be loaded onto the processor 102, the bit width of the interaction coefficient of the original Ising model can be reduced without changing the ground state.

[0044] For example, J i、j =2J′ i、j Then, the energy H of Model 4-1 10 is expressed as in Equation (7). In the following, the interaction coefficient J i、j Terms that do not depend on are omitted.

number

[0045] J i、j =2J′ i、j Let J′ be the two identical interaction coefficients. i、j (J′ i、j =J″ i、j ) The energy of model 4-2 is H′ 10 Then, J′ i、j >0, H′ 10 is expressed by equation (8).

number

[0046] σ x is H′ 10 Since this is the spin that minimizes, equation (8) can be expressed as equation (9).

number

[0047] From equations (7) and (9), the spin-independent constant |J′ i、j Excluding |, H 10 =H′ 10 In other words, without changing the ground state, the interaction coefficient J i、j (=2J′ i、j ) can be reduced by 1 bit.

[0048] J′ i、j <0, the energy H′ of model 4-2 10 is expressed by equation (10).

number

[0049] σ x is H′ 10 Since this is the spin that minimizes, equation (10) can be expressed as equation (11).

number

[0050] From equations (7) and (11), the spin-independent constant |J′ i、j Excluding |, H 10 =H′ 10 That is, without changing the ground state, the interaction coefficient J i、j (=2J′ i、j ) can be reduced by 1 bit.

[0051] Next, an example of reducing the bit width of the interaction coefficient will be described with reference to FIG. i and σ j The interaction coefficient J between i、j shows an example of reducing a 5-bit (including sign bit) value to a 3-bit value.

[0052] In the original Ising model (Model 5-1), J i、j Assume that the value of is "14" (5 bits including the sign bit) (step 50). First, decompose "14" in model 5-1 into "3" (3 bits) and "11" (5 bits) (step 52). Then, σ i and σ j For the "3" that connects with, add an auxiliary spin a1, σ i Set the interaction coefficient between and a1 to "3", and between a1 and σ j A new Ising model (model 5-2) is obtained by setting the interaction coefficient between

[0053] Next, J in Model 5-2 i、j The value of "11" is decomposed into "3" (3 bits) and "8" (5 bits) (step 56). i and σ j For the "3" that connects with, add an auxiliary spin a2, σ i Set the interaction coefficient between and a2 to "3", and between a2 and σ j A new Ising model (model 5-3) is obtained by setting the interaction coefficient between

[0054] Next, J of Model 5-3 i、j The value "8" of is decomposed into "3" (3 bits) and "5" (4 bits) (step 60). i and σ j For the "3" that connects with, add an auxiliary spin a3, σ i Set the interaction coefficient between and a3 to "3", and between a3 and σ j A new Ising model (model 5-4) is obtained by setting the interaction coefficient between

[0055] Next, J of Model 5-4 i、j The value "5" of is decomposed into "3" and "2" (both 3 bits) (step 64). Then, σ i and σ j For the "3" that connects with, add auxiliary spin a4, σ i Set the interaction coefficient between and a4 to "3", and between a4 and σ j The interaction coefficient between is also set to "3". This makes it possible to obtain an expanded Ising model (model 5-5) in which all interaction coefficients are expressed in 3 bits or less (step 66).

[0056] Next, a method for reducing the bit width of the external magnetic field coefficients will be described with reference to FIG. As shown in Figure 6, the Ising model (Model 6-1) has a spin σ i External magnetic field coefficient h acting on i It shall have the following.

[0057] h i is the first external magnetic field coefficient h′ i and the second external magnetic field coefficient h x The sum of (h i =h′ i +h x ), first, we add auxiliary spin σ x Add σ i h′ as the external magnetic field coefficient i Set σ x h as the external magnetic field coefficient x Then, set σ iand σ x Add an edge (i, x) between the two, and use the absolute value of the second external magnetic field coefficient |h x In this way, the extended Ising model (Model 6-2) is obtained.

[0058] Next, we consider the conditions for making the ground state of Model 6-2 coincide with that of Model 6-1.

[0059] If the energy of model 6-1 is H2, H2 is expressed by equation (12). In the following, the external magnetic field coefficient h i Terms that do not depend on are omitted.

number

[0060] If the energy of model 6-2 is H′2, then h x >0, h i >h x Then, H′2 is expressed by equation (13).

number

[0061] σ i Since the value of is +1 or -1, (1+σ i ) is either 2 or 0, and is always a non-negative value. Therefore, σ x When H′2 is at its minimum, equation (13) can be expressed as equation (14).

number

[0062] The difference in energy between Model 6-1 and Model 6-2 (the difference between Eq. (12) and Eq. (14)) is |h x |Becomes true.

[0063] h x <0,h i <h xThen, H′2 is expressed by equation (15).

number

[0064] σ i Since the value of is +1 or -1, (1-σ i ) is either 0 or 2, and is always a non-negative value. Therefore, σ x When =-1, H'2 is minimum and equation (15) can be expressed as equation (16).

number

[0065] The difference in energy between Model 6-1 and Model 6-2 (the difference between Eq. (12) and Eq. (16)) is |h x |Becomes true.

[0066] In this way, σ x Assuming that is the spin that minimizes H′2, the energy difference between the original Ising model (Model 6-1) and the extended Ising model (Model 6-2) is always |h x Therefore, the energy difference between the ground state of Model 6-1 and the ground state of Model 6-2 is |h x It must be |h x Since the value of | does not depend on the spin state, the spin σ that gives the ground state in Model 6-1 is i gives the ground state in model 6-2 as well. That is, as shown in Figure 6, if we add an auxiliary spin and an edge to the original Ising model and focus on the original spin, we can obtain the same ground state.

[0067] By repeating the process shown in FIG. 6 until the bit width becomes one that can be loaded onto the processor 102, the bit width of the external magnetic field coefficients of the original Ising model can be reduced without changing the ground state.

[0068] Next, an example of reducing the bit width of the external magnetic field coefficient will be described with reference to FIG.i External magnetic field coefficient h acting on i shows an example of reducing a 5-bit (including sign bit) value to a 3-bit value.

[0069] In the original Ising model (Model 7-1), h i Assume that the value of is "14" (5 bits including the sign bit) (step 70). i Since the value "14" is expressed as the sum of "3" (3 bits) and "11" (5 bits), we first add an auxiliary spin a1 to model 7-1 and i Set "11" as the external magnetic field coefficient of a1, and set "3" as the external magnetic field coefficient of a2. Then, σ i and a1, and a new interaction coefficient of “3” is set on the added edge to obtain a new Ising model (model 7-2) (step 72).

[0070] Next, h of Model 7-2 i Since the value "11" is expressed as the sum of "3" (3 bits) and "8" (5 bits), we add an auxiliary spin a2 to model 7-2 and obtain σ i Set "8" as the external magnetic field coefficient of a1, and set "3" as the external magnetic field coefficient of a2. Then, σ i and a2, and a new interaction coefficient of “3” is set on the added edge to obtain a new Ising model (model 7-3) (step 74).

[0071] Next, h of Model 7-3 i Since the value "8" is expressed as the sum of "3" (3 bits) and "5" (4 bits), we add an auxiliary spin a3 to model 7-3 and obtain σ i Set "5" as the external magnetic field coefficient of a1, and set "3" as the external magnetic field coefficient of a2. Then, σ i and a3, and a new interaction coefficient of “3” is set on the added edge to obtain a new Ising model (model 7-4) (step 76).

[0072] Next, h of Model 7-4 i Since the value "5" is expressed as the sum of "3" and "2" (both 3 bits), we add an auxiliary spin a4 to model 7-4 and σ i Set "2" as the external magnetic field coefficient of a1, and set "3" as the external magnetic field coefficient of a2. Then, σ i An edge is added between a and a4, and a new interaction coefficient is set to "3" on the added edge. This makes it possible to obtain an extended Ising model (model 7-5) in which all external magnetic field coefficients are expressed in 3 bits or less (step 78).

[0073] The calculation method according to this embodiment makes it possible to formulate the number of auxiliary spins added when reducing the bit width of the original Ising model by 1 bit.

[0074] J of the original Ising model i、j and h i The value of is n bits including the sign bit, and by reducing one bit, it becomes [-(2 n-2 -1), 2 n-2 -1], where n-1 is an arbitrary (n-1)-bit integer (n-1 ≧ 2).

[0075] J i、j The number of auxiliary spins added when is reduced by 1 bit is s i、j Then, s i、j is expressed by equation (17).

number

[0076] h i The number of auxiliary spins added when is reduced by 1 bit is s i Then, s i is expressed by equation (18).

number

[0077] From equations (17) and (18), when an Ising model with a bit width expressed as n bits is reduced by 1 bit using the calculation method according to this embodiment, the total number s of auxiliary spins added is expressed by equation (19).

number

[0078] When the calculation method according to the present embodiment is applied to the Ising model 10 expressed with a bit width of 3 bits ([-3, +3]) shown in FIG. 1 to reduce the bit width by 1 bit, an expanded Ising model 80 (right side of FIG. 8) expressed with a bit width of 2 bits [-1, +1] can be obtained, as shown in FIG. 8.

[0079] In the augmented Ising model 80 in FIG. 8, edges drawn with solid lines and arrows drawn with solid lines respectively indicate that the interaction coefficient and the external magnetic field coefficient are +1, and arrows drawn with dotted lines indicate that the external magnetic field coefficient is -1.

[0080] From equations (17) to (19), the total number s of auxiliary spins added to the Ising model 10 is 6. In fact, to reduce h1 (=3) by 1 bit, auxiliary spins a1 and a2 are added, and J 1、2 In order to reduce h2 (=3) by 1 bit, auxiliary spins a3 and a4 are added, and in order to reduce h2 (=-2) by 1 bit, auxiliary spin a5 is added. 1、4 In order to reduce (=2) by one bit, an auxiliary spin a6 is added, and the extended Ising model 80 is obtained.

[0081] In the expanded Ising model, J 1、2 =J 1、3 =J 1、4 =J 2、3 =J 2、4 =J 3、4 =+1, h1=+1, h2=-1, h3=h4=0. This shows that the spin of the ground state of the expanded Ising model 80, excluding the auxiliary spins, is (σ1, σ2, σ3, σ4)=(+1, +1, +1, +1), which is consistent with the spin of the ground state of the original Ising model 10.

[0082] Next, an example will be described in which the calculation method according to this embodiment is applied to a random Ising model and two combinatorial optimization problems (an integer partition problem and a vertex cover problem).

[0083] <Random Ising model> The following random Ising model is a connected graph in which vertices and edges are randomly generated, the number of vertices is in the range of 5 to 20, and the edge density is set to approximately 1.0. In addition, the interaction coefficients and external magnetic field coefficients are randomly generated to have a uniform distribution of a predetermined bit width, and the bit width is in the range of 5 bits to 11 bits.

[0084] Table 1 shows an example in which the calculation method according to this embodiment is applied to each Ising model generated randomly in this way. [Table 1]

[0085] In Table 1, #1 to #11 in the "Name" column represent graphs of different Ising models. In the "Name" column, |V| represents the number of vertices, and |E| represents the number of edges. The "Bit-Width" column represents the bit width of the original Ising model, the "#Reduction bits" column represents the number of bits reduced by applying the calculation method according to this embodiment to the original Ising model, and the "#Spins" column represents the total number of spins of the expanded Ising model obtained by reducing the bit width of the original Ising model. When the number of reduced bits is "0", the total number of spins is the same as the number of spins of the original Ising model. A blank in the "#Spins" column means that the bit width cannot be reduced any further. The configuration of each table described later is the same as Table 1.

[0086] In Table 1, for example, the Ising model #1 has 5 vertices, 10 edges, and a bit width of 5 bits. When the bit width is reduced by 1 bit by applying the calculation method according to this embodiment to the Ising model #1, the total number of spins becomes 15, when the bit width is reduced by 2 bits, the total number of spins becomes 34, and when the bit width is reduced by 3 bits, the total number of spins becomes 109. It can be seen from Table 1 that the total number of spins increases as the bit width of the Ising models #1 to #11 is reduced.

[0087] <Integer division problem> The Number Partition Problem (NPP) is a problem for a set of N natural numbers, M = {m1, m2, ..., m N Given a set of 1, M is divided into two subsets, M1 and M2, and the problem is to determine whether the sum of the natural numbers in M1 is equal to the sum of the natural numbers in M2. NPP is known to be an NP-complete problem.

[0088] For example, given a set M = {1, 2, 4, 7}, if M is divided into subsets M1 = {1, 2, 4} and M2 = {7}, the sum of the natural numbers in M1 will be equal to the sum of the natural numbers in M2.

[0089] NPP can be expressed by the Ising model. N (=|M|) spins σ i (1≦i≦N) and define the energy function (Hamiltonian) as H NPP Then, the NPP can be formulated as an Ising model as shown in equations (20) and (21).

number

number

[0090] As shown in equation (21), σ i (1≦i≦N) is m i When assigned to M1, it becomes +1, and m iWhen H is assigned to M2, it becomes -1. NPP The minimum value of is zero. H NPP When is zero, the sum of the natural numbers belonging to M1 is equal to the sum of the natural numbers belonging to M2.

[0091] Given the above set M = {1, 2, 4, 7}, an example of NPP expressed as an Ising model is shown in Figure 9. The interaction coefficient J i、j From equation (20), 2m i m j Spin σ i (1≦i≦N) is m i As mentioned above, the set M can be divided into M1 = {1, 2, 4} and M2 = {7}. As shown in FIG. 9, when (σ1, σ2, σ3, σ4) = (+1, +1, +1, -1) or (σ1, σ2, σ3, σ4) = (-1, -1, -1, +1), H NPP takes the minimum value.

[0092] Table 2 shows an example in which the calculation method according to this embodiment is applied to the Ising model of an NPP. [Table 2]

[0093] #12 to #23 in the "Name" column of Table 2 each represent a different set M. Here, N natural numbers (N mod 3=0 or 2) in set M are generated based on f(i,j) defined in equation (22), which represents the Fibonacci sequence. When set M is generated based on equation (22), NPP always has a solution.

number

[0094] #13, #15, #18, and #20 to #23 in Table 2 represent sets generated based on f(i,j) in formula (22). Set #14 corresponds to the Ising model 90 with a bit width of 7 shown in FIG. 9. For example, when the calculation method according to this embodiment is applied to the Ising model 90 to reduce the bit width by 1 bit, the total number of spins becomes 5, and when the bit width is reduced by 5 bits, the total number of spins becomes 124. In this way, it can be seen that the total number of spins increases when the bit width is reduced by the calculation method according to this embodiment in the NPP as well.

[0095] <Vertex coverage problem> A vertex cover of an undirected graph G = (V, E) is when at least one vertex of each edge in E is included in a subset of vertices V' ⊆ V. The Vertex Cover Problem (hereafter referred to as VCP) is to find a minimum-size V', and is known to be an NP-hard problem.

[0096] An example of a VCP is shown in Fig. 10. The graph 1000 shown in Fig. 10 has eight vertices (1 to 8) and 16 edges. The minimum size V' of the graph 1000 is a set of five vertices, and as shown in Fig. 10, there are a subset V'1 consisting of vertices 1, 3, 4, 6, and 7, and a subset V'2 consisting of vertices 2, 4, 5, 6, and 8. That is, at least one vertex of each edge of the graph 1000 is included in V'1 or V'2.

[0097] VCP can also be expressed by the Ising model. N (=|V|) spins σ i (1≦i≦N) and define the energy function (Hamiltonian) as H VCP Then, VCP can be formulated as an Ising model as shown in equations (23) and (24).

number

number

[0098] In formula (23), α VCP and β VCP is a positive weight parameter. As shown in equation (24), σ i (1≦i≦N) is +1 if vertex i belongs to V', and is -1 if vertex i does not belong to V'.

[0099] The first and second terms of Eq. (23) are the penalty function and the objective function, respectively. If at least one vertex of each edge of E belongs to V′, the first term of Eq. (23) is zero. If V′ is the minimum size, then H VCP is minimized, and the optimal solution can be obtained. VCP ≦α VCP By setting the above, we can obtain solutions that satisfy the constraints, called feasible solutions.

[0100] Table 3 shows an example in which the calculation method according to this embodiment is applied to the Ising model of the VCP. [Table 3]

[0101] The "Name" column in Table 3 shows benchmark graphs selected from the AG-Monien Graph Collection at the following site: “AG-Monien - SuiteSparse matrix collection,” https: / / sparse.tamu.edu / AG-Monien. In addition, for all graphs in Table 3, α VCP and β VCP Both values ​​are set to 1.

[0102] Regarding each graph shown in Table 3, it can be seen that when the calculation method according to this embodiment is applied to reduce the bit width by 1 bit, the total number of spins increases.

[0103] Note that the annealing machine or Ising machine has a predetermined number of spins, and can perform processing as long as the number of spins is equal to or less than the predetermined number of spins. In fact, in Tables 1 to 3, when the total number of spins is equal to or less than the predetermined number of spins possessed by the processor 102, the processor 102 can correctly obtain the ground state.

[0104] Next, for comparison with the calculation method according to this embodiment, the results obtained by applying the naive method (shift method) to the random Ising model, NPP, and VCP are shown in Tables 4, 5, and 6, respectively.

[0105] [Table 4]

[0106] [Table 5]

[0107] [Table 6]

[0108] The contents of the "Name" columns in Tables 4 to 6 are the same as those in Tables 1 to 3, respectively. In addition, in the "a" / "b" columns of "#Ground states by naive method / #Original ground states matched" in Tables 4 to 6, "a" represents the number of ground states of the Ising model whose bit width has been reduced by the shift method, and "b" represents the number of ground states that match the ground states of the original Ising model among the ground states of the Ising model after the shift method has been applied.

[0109] Table 4 (random Ising model) shows that the ground state of each Ising model whose bit width has been reduced to 3 bits by the shift method is almost identical to the ground state of the original Ising model (see the underlined data in Table 4). On the other hand, the ground state of each Ising model whose bit width has been reduced to 2 bits by the shift method is almost identical to the ground state of the original Ising model (see the double underlined data in Table 4).

[0110] From Table 5 (NPP), it can be seen that the number of basis states of each Ising model whose bit width has been reduced by the shift method hardly matches the number of basis states of the original Ising model. For example, if the bit width of the Ising model corresponding to #14 (Ising model 90 shown in FIG. 9) is reduced by 3 bits or more, the optimal solution of NPP cannot be obtained. Also, in the problems shown in #12 to #19 in Table 5, it can be seen that the number of basis states of each Ising model whose bit width has been reduced to 2 bits by the shift method is significantly larger than the number of basis states that match the original Ising model (see the underlined data in Table 5). In this case, the optimal solution of NPP can hardly be obtained.

[0111] From Table 6 (VCP), for graphs with less than 15 vertices (se 1, cage 1, cage 2 in Table 6), the ground state of the Ising model with 1 bit reduced by the shift method matches the ground state of the original Ising model. However, when the size of the graph increases, the number of ground states of the Ising model with 1 bit reduced by the shift method does not match the number of ground states of the original Ising model. In particular, for bfly 1 (|V|=24, |E|=96) in Table 6, simply reducing the bit width of the original Ising model by 1 bit makes it impossible to obtain the optimal solution of VCP at all.

[0112] According to this embodiment, an auxiliary spin is added to the Ising model, and a new Ising model is obtained by setting a new interaction coefficient between the spin of the Ising model and the auxiliary spin, thereby dispersing the magnitude of the coefficients (interaction coefficient, external magnetic field coefficient) of the Ising model. This makes it possible to reduce the bit width of the Ising model without changing the ground state. In this way, in this embodiment, even if an Ising model with a bit width larger than the bit width that can be mounted on the hardware is input, the ground state of the Ising model can be correctly obtained.

[0113] The present invention is not limited to the above-described embodiment, and various modifications are possible without departing from the spirit of the present invention. Other embodiments and modifications made by those skilled in the art are also included in the present invention. [Explanation of symbols]

[0114] 100 Computing equipment 102 processors 104 Memory 106 Storage device 108 Input section 110 Display 112 Network Interface

Claims

1. A calculation method executed by a processor for reducing a bit width of an Ising model represented by a plurality of spins, an interaction coefficient between the plurality of spins, and an external magnetic field coefficient acting on each of the plurality of spins, the method comprising: Adding auxiliary spins to the Ising model, A calculation method for reducing a bit width of at least one of the interaction coefficients between the plurality of spins and the auxiliary spin by determining a new Ising model by setting a new interaction coefficient between the plurality of spins and the auxiliary spin, the bit width of at least one of the interaction coefficients between the plurality of spins and the external magnetic field coefficient acting on each of the plurality of spins.

2. The computational method according to claim 1 , wherein the auxiliary spin is a spin that minimizes the energy of the new Ising model.

3. When reducing the bit width of an interaction coefficient between two spins among the plurality of spins, If the interaction coefficient between the two spins is expressed as the sum of the first interaction coefficient and the second interaction coefficient, dividing an interaction coefficient between the two spins into the first interaction coefficient between the two spins and the second interaction coefficient between the two spins; 3. The calculation method according to claim 1, further comprising the step of: setting the first interaction coefficient between one of the two spins and the auxiliary spin; and setting an absolute value of the first interaction coefficient between the other of the two spins and the auxiliary spin, thereby expanding the first interaction coefficient between the two spins.

4. When reducing the bit width of an external magnetic field coefficient acting on one of the plurality of spins, If the external magnetic field coefficient acting on one spin is expressed as the sum of a first external magnetic field coefficient and a second external magnetic field coefficient, setting the first external magnetic field coefficient for the one spin; setting the second external magnetic field coefficient for the auxiliary spins; The calculation method according to claim 1 , further comprising setting an absolute value of the second external magnetic field coefficient as the new interaction coefficient between the one spin and the auxiliary spin.

5. A calculation device for reducing a bit width of an Ising model represented by a plurality of spins, an interaction coefficient between the plurality of spins, and an external magnetic field coefficient acting on each of the plurality of spins, Adding auxiliary spins to the Ising model, A computing device comprising a processor that reduces a bit width of at least one of the interaction coefficients between the plurality of spins and the auxiliary spin by setting a new interaction coefficient between the at least one spin among the plurality of spins and the auxiliary spin to obtain a new Ising model.

6. A program for causing a computer to execute a calculation method for reducing a bit width of an Ising model represented by a plurality of spins, an interaction coefficient between the plurality of spins, and an external magnetic field coefficient acting on each of the plurality of spins, Adding auxiliary spins to the Ising model, A program for causing a computer to execute a process of reducing a bit width of at least one of the interaction coefficients between the plurality of spins and the auxiliary spin by obtaining a new Ising model by setting a new interaction coefficient between at least one of the plurality of spins and the auxiliary spin.

Citation Information

Patent Citations

  • Performing fault tree analysis on quantum computers

    US20190026645A1