Optimization device and optimization method

The optimization device improves convergence to optimal solutions by determining energy changes and limiting value transitions within permissible ranges, addressing the challenge of complex evaluation functions with 1-hot constraints in combinatorial optimization.

JP7590642B2Active Publication Date: 2024-11-27FUJITSU LTD
View PDF 4 Cites 0 Cited by

Patent Information

Application Number
JP2020199761
Authority / Receiving Office
JP · JP
Patent Type
Patents
Current Assignee / Owner
Filing Date
2020-12-01
Publication Date
2024-11-27
Estimated Expiration
2040-12-01

AI Technical Summary

Technical Problem

Conventional optimization devices face challenges in converging to optimal solutions due to complex potential shapes of evaluation functions with many local maximums and local minimums, especially when dealing with combinatorial optimization problems that include 1-hot constraints.

Method used

An optimization device and method that determine the amount of energy change when the value of a state variable changes, using a search unit to find optimal solutions while adhering to 1-hot constraints, and a transition allowable range determining unit to limit value changes within permissible ranges, thereby improving convergence.

Benefits of technology

The solution enhances convergence to optimal solutions by limiting value changes within permissible ranges, reducing the need for additional constraint terms, and maintaining constraint satisfaction without increasing the number of state variables, thus improving the efficiency of combinatorial optimization.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 0007590642000017
    Figure 0007590642000017
  • Figure 0007590642000018
    Figure 0007590642000018
  • Figure 0007590642000019
    Figure 0007590642000019
Patent Text Reader

Abstract

To improve convergence to an optimum solution.SOLUTION: On the basis of a change amount of energy when a value of one of x1 to xN included in an evaluation function changes and on the basis of a result of determination processing of determining the propriety of the change of the value, a searching unit 11 searches for an optimum solution that minimizes energy by repeating processing of changing a value of one of x1 to xN. A transition permissible range determination unit 12 determines an upper limit or a lower limit of a second identification number of a second state variable for which a change from 0 to 1 is permitted in a second state variable group out of a plurality of state variable groups (groups gD1 to gD3) which is included in x1 to xN and in each of which one state variable is 1 and the rest state variables are 0, based on a first identification number of a first state variable in which the value is 1 in a first state variable group.SELECTED DRAWING: Figure 1
Need to check novelty before this filing date? Find Prior Art

Description

[Technical field]

[0001] The present invention relates to an optimization device and an optimization method. [Background technology]

[0002] As an optimization device that can calculate large-scale combinatorial optimization problems, which von Neumann-type computers are not good at, there is an Ising device (also called a Boltzmann machine) that uses an Ising-type evaluation function (also called an energy function, etc.).

[0003] In calculations using an Ising device, the problem to be calculated is replaced with an Ising model, which is a model that represents the behavior of spin in a magnetic material. Then, using Markov chain Monte Carlo methods such as simulated annealing and the replica exchange method (also called the exchange Monte Carlo method), a search is performed for a state in which the value of an Ising model-type evaluation function (corresponding to the energy of the Ising model) is minimized. The state is represented by the values ​​of multiple state variables included in the evaluation function.

[0004] Conventionally, there is an optimization device that searches for a state with the minimum energy by performing the Markov chain Monte Carlo method using a digital circuit. Conventional optimization devices calculate the amount of change in energy by assuming that only one state variable changes at a time, and decide whether or not to allow the change in the state variable based on the amount of change plus a noise value corresponding to temperature. Changes in the value of the state variable that increase the energy are also allowed with a certain probability, and the probability decreases as the temperature decreases.

[0005] Incidentally, some optimization problems have a constraint (1-hot constraint) that the number of state variables that have a value of 1 among the state variables included in the evaluation function is only one. In the 1-hot constraint, each state variable appears only once in the set of constraint equations, and N 2When state variables are arranged in a matrix with N rows and N columns, the sum of the values ​​of the state variables in each row and column is 1. Hereinafter, the former 1-hot constraint will be called the 1-Way 1-hot constraint, and the latter 1-hot constraint will be called the 2-Way 1-hot constraint. For example, traffic flow optimization problems and bin packing problems have 1-Way 1-hot constraints. For example, traveling salesman problems, vehicle dispatching problems, and quadratic assignment problems have 2-Way 1-hot constraints.

[0006] Conventionally, methods have been proposed for solving vehicle dispatch planning problems using genetic algorithms and quantum computing (see, for example, Patent Documents 1 and 2). [Prior art documents] [Patent documents]

[0007] [Patent Document 1] JP 2003-285930 A [Patent Document 2] JP 2003-114132 A Summary of the Invention [Problem to be solved by the invention]

[0008] Some combinatorial optimization problems include many constraint conditions, and the evaluation function for such combinatorial optimization problems includes constraint terms corresponding to each constraint condition. Evaluation functions that include many constraint terms have a complex potential shape that includes many maximum and minimum values, which can lead to a problem of poor convergence to the optimal solution.

[0009] In one aspect, the present invention aims to provide an optimization device and an optimization method capable of improving convergence to an optimal solution. [Means for solving the problem]

[0010] In one embodiment, there is provided an optimization device comprising: a search unit that searches for an optimal solution that minimizes the energy by repeating a determination process of determining whether or not a value can be changed based on an amount of change in energy when a value of any of a plurality of state variables included in an evaluation function representing the energy of an Ising model that models a combinatorial optimization problem is changed, and a process of changing the value based on a result of the determination process; and a transition allowable range determination unit that determines an upper limit or lower limit of a second identification number of a second state variable that is allowed to change from the second value in a second state variable group among the plurality of state variable groups, based on a first identification number of a first state variable that has the first value in a first state variable group among the plurality of state variable groups, each of which has one state variable having a first value and the remaining state variables having a second value.

[0011] Also, in one embodiment, an optimization method is provided. Effect of the Invention

[0012] In one aspect, the present invention can improve convergence to an optimal solution. [Brief description of the drawings]

[0013] [Figure 1] FIG. 1 illustrates an example of an optimization device according to a first embodiment. [Diagram 2] FIG. 11 is a diagram illustrating an example of a conversion result. [Diagram 3] 13A and 13B are diagrams illustrating examples of cases where constraint conditions are satisfied and not satisfied. [Figure 4] 13A and 13B are diagrams illustrating another example of determining the range in which a value change is permitted. [Diagram 5] FIG. 11 illustrates an example of an optimization device according to a second embodiment. [Figure 6] FIG. 13 illustrates an example of a transition allowable range determining unit; [Figure 7] FIG. 13 is a diagram illustrating an example of calculation of a boundary value. [Figure 8] 11 illustrates an example of a transition permission / prohibition signal output unit and a storage unit; FIG. [Figure 9] FIG. 13 is a diagram illustrating an example of a transition possible / prohibited bit generating circuit. [Figure 10] FIG. 13 illustrates an example of a ΔE calculation unit. [Figure 11] FIG. 13 is a diagram showing an example of storing Dt and ri,t. [Figure 12] FIG. 13 is a diagram illustrating an example of Case 1. [Figure 13] FIG. 13 is a diagram illustrating an example of storage of ri,t before and after a state transition in Case 1. [Figure 14] FIG. 13 is a diagram illustrating an example of Case 2. [Figure 15] FIG. 13 is a diagram illustrating an example of Case 3. [Figure 16] FIG. 13 is a diagram illustrating an example of storage of ri,t before and after a state transition in Case 3. [Figure 17] FIG. 13 is a diagram illustrating another example of the ΔE calculation unit. [Figure 18] 13 is a flowchart showing an example of the overall operation flow of the optimization device. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0014] Hereinafter, an embodiment of the invention will be described with reference to the drawings. (First embodiment) FIG. 1 illustrates an example of an optimization device according to a first embodiment.

[0015] The optimization device 10 searches for an optimal solution that minimizes the energy of an Ising model that models a combinatorial optimization problem. The energy of the Ising model is defined, for example, by an evaluation function (E(x)) as shown in the following equation (1).

[0016]

number

[0017] The first term on the right-hand side is the sum of the values ​​of the two state variables and the weighting coefficients for all combinations of two state variables that can be selected from all state variables included in the evaluation function, without omissions or duplications. x i is the i-th state variable. x j is the jth state variable. W ij is a weighting coefficient indicating the weight (for example, the strength of the connection) between the i-th state variable and the j-th state variable. ii = 0. Also, W ij =W ji (That is, the coefficient matrix based on the weighting factors is often a symmetric matrix.)

[0018] The second term on the right-hand side is the sum of the products of the bias coefficients of all state variables and the values ​​of the state variables. i denotes the bias coefficient for the i-th state variable. c is a constant.

[0019] For example, a spin of -1 in the Ising model corresponds to a state variable value of 0. A spin of +1 in the Ising model corresponds to a state variable value of 1. For this reason, a state variable can also be called a bit that takes on a value of 0 or 1.

[0020] The combination of state variable values ​​that minimizes the value of equation (1) is the solution to the problem (optimal solution). The optimization device 10 includes a search unit 11 and a transition allowable range determination unit 12 . The search unit 11 performs a determination process to determine whether or not a value can be changed based on the amount of change in energy when the value of any of the multiple state variables included in the evaluation function described above is changed. Then, based on the result of the determination process, the search unit 11 performs a process (update process) to change the value of any of the multiple state variables. The search unit 11 repeats these processes to search for an optimal solution that minimizes energy.

[0021] In the following example, the search unit 11 performs the update process while satisfying the 1-hot constraint (in the following example, the 1-way 1-hot constraint). Note that the search unit 11 may perform the update process while satisfying the 2-way 1-hot constraint.

[0022] The search unit 11 determines whether or not to update the values ​​of any two state variables in a group of state variables (hereinafter referred to as a group) that should satisfy the 1-hot constraint, based on the amount of change in energy calculated using the weighting coefficients included in the formula (1), for example. In order to satisfy the 1-hot constraint, when a state variable having a value of 1 among the state variables included in the group is updated to 0, only one state variable having a value of 0 is updated to 1.

[0023] By the way, in formula (1), x i The value of 1-x changes i So, x i The increase is Δx i =(1-x i )-x i =1-2x i The change in energy due to the change in this value (ΔE i ) is expressed by the following equation (2).

[0024]

number

[0025] In formula (2), x i When changes from 1 to 0, Δx i becomes -1, and x i When changes from 0 to 1, Δx i is 1. Note that h i is called the local field, and Δx i Depending on h i Multiplying by the sign (+1 or -1) is ΔE i It is.

[0026] x j When h changes from 0 to 1 iThe change in is Δh i (j) =+W ij , x j When h changes from 1 to 0 i The change in is Δh i (j) =-W ij Similarly, x i When x changes j About h j The change in is Δh j (i) =Δx i W ij , which can be expressed as:

[0027] Therefore, x i ,x j The change in energy when both change (ΔE ij ) can be expressed by the following equation (3).

[0028]

number

[0029] As mentioned above, to transition from one state that satisfies a 1-hot constraint to another state that satisfies the constraint, the values ​​of two state variables are changed. x i As x changes from 1 to 0 j The change in energy when changes from 0 to 1 is called ΔE j Then, Δx i =-1, Δx j = 1, so from equation (3), ΔE j can be expressed by the following equation (4).

[0030]

number

[0031] In addition, in formula (1), W ij x i x j x i ,x jIf either of these is 0, it does not contribute to the energy, so W in equation (4) ij is not necessary. The search unit 11 calculates the amount of change in energy as described above for each state variable whose value is 0 among the state variables of each group that should satisfy the 1-hot constraint. Then, the search unit 11 calculates the amount of change in energy ΔE j Based on this, the search unit 11 uses the simulated annealing method, the replica exchange method, or the like to determine whether or not to allow a change in the state variables that causes a change in the energy. The search unit 11 preferentially accepts changes in the state variables that reduce E(x) in equation (1), but also probabilistically allows changes that increase E(x). However, ΔE j If is a very large positive value, the probability that the change will be tolerated becomes very small.

[0032] When the search unit 11 determines a state variable in a group whose value is to be updated from 0 to 1, it updates the value of that state variable from 0 to 1, and also updates the value of the state variable in that group whose current value is 1 from 1 to 0.

[0033] The search unit 11 searches all the current state variables (x1 to x N ) is stored in a storage unit 11a. The search unit 11 is realized by an electronic circuit such as an ASIC (Application Specific Integrated Circuit) or an FPGA (Field Programmable Gate Array). The storage unit 11a is an electronic circuit such as an SRAM (Static Random Access Memory) or a register. The search unit 11 may be realized by software processing in which a processor such as a CPU (Central Processing Unit) or a GPU (Graphics Processing Unit) executes a program.

[0034] The transition allowable range determination unit 12 determines x1 to x Nand determines a range in which a value change is permitted for a plurality of groups each having one state variable having a first value and the remaining state variables having a second value. In the following example, the transition allowable range determination unit 12 will be described as determining a range in which a value change is permitted for a plurality of groups that should satisfy the 1-way 1-hot constraint as described above. That is, the first value is 1 and the second value is 0, but the present invention is not limited to this and the first value may be 0 and the second value may be 1.

[0035] The transition allowable range determination unit 12 determines, based on the identification number of a state variable having a value of 1 in one group among the above multiple groups, an upper or lower limit for the identification number of a state variable in another group that is allowed to change in value from 0 to 1.

[0036] In this way, certain constraints can be satisfied without adding additional constraint terms, as will be described later. The transition allowable range determination unit 12 is realized by, for example, an electronic circuit such as an ASIC, an FPGA, etc. Note that the transition allowable range determination unit 12 may be realized by software processing in which a processor such as a CPU or a GPU executes a program.

[0037] In the following, we will explain the calculation of a routing problem with multiple nodes, such as the Capacitated Vehicle Routing Problem (CVRP). In such problems, for example, state variables are used that indicate whether a certain vehicle is at a certain node at a certain time with 0 or 1, but the number of state variables increases as the number of nodes and vehicles increases.

[0038] CVRP is a problem of determining a customer visiting sequence (route) that minimizes the total travel distance, etc., based on various input variables, when a transport vehicle waiting at a specific facility called a depot (starting point) delivers (or collects) demand to a customer location (hereafter referred to as a node) and returns to the depot.

[0039] The constraints of CVRP are as follows, A to G. (Constraint A) For all routes, the total demand of the nodes excluding the depots on the route is within the maximum load capacity of one transport vehicle (hereinafter referred to as a truck).

[0040] (Constraint B) At all times, a truck passes through (visits) only one location (one node) on one route at the same time. (Constraint C) All nodes except the depots are passed through by a truck only once.

[0041] (Constraint D) Except for the depot, when a truck passes through a node at a certain time (>0) on a certain route, the truck passes through some node at one time before on that route.

[0042] (Constraint E) Except for the depot, if a truck passes through a node at a certain time (<M (time to visit all nodes)) on a certain route, the truck passes through some node at one time after on that route.

[0043] (Constraint F) On all routes, a truck passes through the depot twice on the same route. (Constraint G) On all routes, at the start time and end time, the truck does not pass through any place other than the depot.

[0044] The optimization device 10 of the first embodiment uses, for example, state variables as shown in FIG. 1. The squares marked with "1" indicate state variables with a value of 1, and the squares without a value indicate state variables with a value of 0. FIG. 1 shows an example of state variables when calculating a CVRP with 4 trucks and 13 nodes. Although there may be multiple depots, it is assumed to be one hereinafter.

[0045] The horizontal axis, 1 to 13, represent the node numbers, and D0 to D4 represent the depots. The vertical axis represents the time t. As mentioned above, there is only one depot, but D0 to D4 are used to distinguish the time when the four trucks depart from the depot and the time when they return. That is, the depot from which the first truck departs is represented by D0, the depot from which the first truck returns and the depot from which the second truck departs are represented by D1, the depot from which the second truck returns and the depot from which the third truck departs are represented by D2. In addition, the depot from which the third truck returns and the depot from which the fourth truck departs are represented by D3, and the depot from which the fourth truck returns is represented by D4.

[0046] For this reason, the number of state variables should be 18 x 18, but if the first truck departs from the depot at time t = 0 and the fourth truck returns to the depot at the final time t = 17, the state variables in the rows for t = 0, 17 and in the columns D0, D4 may not be necessary. Therefore, the number of state variables indicating whether any of the four trucks is in any of the 13 nodes or depots at time t may be 16 x 16 = 256, within box 15 in Figure 1.

[0047] In other words, the number of state variables to be included in the evaluation function can be the square of the number of nodes excluding depots + the number of trucks -1. In order to satisfy constraints B, C, and F, the 16×16 state variables have a 2-way 1-hot constraint in which the sum of the values ​​of the state variables in each row and column is 1.

[0048] Therefore, the groups of state variables in each of the columns D1 to D3 are groups that should satisfy the 1-way 1-hot constraint. Also, the groups of state variables in each of the columns D1 to D3 in the frame 15 can be used to calculate the sum of the loads of the four trucks on each route (total demand value of the nodes).

[0049] Hereinafter, the groups of state variables in each of the columns D1 to D3 in the box 15 are referred to as groups gD1, gD2, and gD3. In addition, in the box 15, the state variables (x1 to x 208 ) to x t,n In addition, the 16 state variables (x 209 ~x 224 ) to y D1,1 ~y D1,16 , 16 state variables of group gD2 (x 225 ~x 240 ) to y D2,1 ~y D2,16 , 16 state variables of group gD3 (x 241 ~x 256 ) to y D3,1 ~y D3,16 It is written as follows.

[0050] In order to calculate the sum of the load amounts of the four trucks on each route, the state variables of the groups gD1 to gD3 expressed as above are converted by the search unit 11, for example, based on the following equation (5).

[0051]

number

[0052] In addition, in equation (5), k = 1, 2, 3, and i = 1, 2, 3, ..., 16. FIG. 2 is a diagram showing an example of a conversion result. In the example in Figure 2, y D1,5 ,y D2,8 ,y D3,14 Therefore, from equation (5), the converted value is z D1,1 ~z D1,5 ,z D2,1 ~z D2,8 ,z D3,1 ~z D3,14 is set to 1 and the rest are set to 0.

[0053] The sum of the loads of the four trucks on each route (D TOT1 ~D TOT4) can be calculated using the converted value, for example, by the following formula (6).

[0054]

number

[0055] In equation (6), D t is the sum of the demands at time t and is expressed by the following equation (7).

[0056]

number

[0057] In equation (7), D n is the demand of the node with node number n. The demand of the depot is 0. By the way, the above D TOT1 ~D TOT4 To enable the formulation of the above, the constraint is that the order in which the state variables of groups gD1 to gD3 become 1 is the order of groups gD1, gD2, and gD3. In other words, it is required that the first truck, which visits any of the multiple nodes at a time before the second truck, returns to the depot at a time before the second truck. Furthermore, it is required that the second truck, which visits any of the multiple nodes at a time before the third truck, returns to the depot at a time before the third truck. If this constraint is not met, it is a constraint violation (constraint violation 1).

[0058] In the following example, the case where the state variable of group gD2 becomes 1 immediately after the state variable of group gD1 becomes 1, and the case where the state variable of group gD3 becomes 1 immediately after the state variable of group gD2 becomes 1 are also considered to be constraint violations (constraint violation 2). In other words, it is required that the second truck does not return to the depot immediately after the time when the first truck returns to the depot. Furthermore, it is required that the third truck does not return to the depot immediately after the time when the second truck returns to the depot.

[0059] Furthermore, the cases where the state variables of groups gD1 to gD3 are 1 just after the truck first leaves the depot and just before the truck finally arrives at the depot are also considered to be constraint violations (constraint violation 3). Constraint violations 2 and 3 mean that none of the four trucks visits any node. However, depending on the problem setting, it does not necessarily have to be considered a constraint violation.

[0060] FIG. 3 is a diagram showing examples of cases where the constraint conditions are satisfied and not satisfied. In Fig. 3, the example marked "OK" does not violate any of the constraints listed above. The examples "NG1" and "NG2" fall under constraint violation 1. The example "NG3" falls under constraint violation 2, and the example "NG4" falls under constraint violation 3.

[0061] If a constraint term is added to the evaluation function in order to prevent violation of the constraint conditions as described above, the number of constraint terms increases, which may deteriorate the convergence to the optimal solution. Therefore, based on the identification number of a state variable having a value of 1 in one of the groups gD1 to gD3, the transition allowable range determination unit 12 determines the upper or lower limit of the identification number of a state variable that is allowed to change in value from 0 to 1 in another group.

[0062] The transition allowable range determination unit 12 includes a storage unit 12a. The storage unit 12a stores identification numbers for identifying state variables of groups gD1, gD2, and gD3. Furthermore, the storage unit 12a stores identification numbers for identifying state variables (hot bits) hD1, hD2, and hD3 whose values ​​are 1 in the groups gD1, gD2, and gD3. The storage unit 12a is, for example, an electronic circuit such as an SRAM or a register.

[0063] In the example of Figure 1, the identification number of the state variable of group gD1 is x 209 ~x 224 The identification numbers 209 to 224 of the state variables of group gD1 are stored. 225 ~x 240 The identification numbers of the state variables of group gD3 are 225 to 240, and the identification numbers of the state variables of group gD4 are x 241 ~x 256 The hot bit hD1 has an identification number of 241 to 256 stored therein. 213 The identification number of the hot bit hD2 is 213, and the identification number of the hot bit hD2 is x 232 The identification number of the hot bit hD3 is 232, and the identification number of the hot bit hD3 is x 254 The identification number 254 is stored in the memory.

[0064] FIG. 1 shows an example of the process performed by the transition allowable range determining unit 12. In FIG. In order to avoid the above constraint violation 3, the lower limit of the identification number of a state variable in group gD1 that is allowed to change in value from 0 to 1 is 209+1=210, which is the first identification number of a state variable in group gD1.

[0065] The upper limit of the identification numbers of state variables that are allowed to change in value from 0 to 1 in group gD1 is determined based on the identification number of hot bit hD2 to avoid the above-mentioned constraint violations 1 and 2. As in the example of Fig. 1, when the number of state variables included in each of groups gD1 to gD3 (group size) is 16, the identification number of hot bit hD2 -16-2 becomes the upper limit of the identification numbers of state variables that are allowed to change in value from 0 to 1 in group gD1. When the identification number of hot bit hD2 is 232, the upper limit is 232-16-2=214.

[0066] The lower limit of the identification numbers of state variables that are allowed to change their values ​​from 0 to 1 in group gD2 is determined based on the identification number of hot bit hD1 to avoid the above constraint violations 1 and 2. When the group size=16, the identification number of hot bit hD1+16+2 becomes the lower limit of the identification numbers of state variables that are allowed to change their values ​​from 0 to 1 in group gD2. When the identification number of hot bit hD1 is 213, the lower limit is 213+16+2=231.

[0067] The upper limit of the identification numbers of state variables that are allowed to change their values ​​from 0 to 1 in group gD2 is determined based on the identification number of hot bit hD3 to avoid the above constraint violations 1 and 2. When the group size=16, the identification number of hot bit hD3-16-2 becomes the upper limit of the identification numbers of state variables that are allowed to change their values ​​from 0 to 1 in group gD2. When the identification number of hot bit hD3 is 254, the upper limit is 254-16-2=236.

[0068] The lower limit of the identification numbers of state variables that are allowed to change their values ​​from 0 to 1 in group gD3 is determined based on the identification number of hot bit hD2 to avoid the above constraint violations 1 and 2. When group size=16, the identification number of hot bit hD2+16+2 becomes the lower limit of the identification numbers of state variables that are allowed to change their values ​​from 0 to 1 in the state variable group of column D3. When the identification number of hot bit hD2 is 232, the lower limit is 232+16+2=250.

[0069] In order to avoid the above constraint violation 3, the upper limit of the identification number of a state variable in group gD3 that is allowed to change in value from 0 to 1 is 256-1=255, which is the last identification number of the state variable in group gD3.

[0070] Based on the upper or lower limit determined as above, the transition allowable range determination unit 12 outputs a signal (a signal indicating transition prohibition or transition allowance) indicating whether or not a change in value from 0 to 1 is allowed for each state variable of groups gD1 to gD3. In the example of FIG. 1, the signal indicating transition prohibition is 1 and the signal indicating transition allowance is 0.

[0071] For example, for group gD2, x is allowed to change from 0 to 1. 231 ~x 236 x already has a value of 1 232 For all other cases, a signal indicating transition permission is output. 232 , x 225 ~x 230 , x 237 ~x 240 For the above, a signal indicating a transition prohibition is output.

[0072] For state variables for which a signal indicating a transition prohibition has been output, the search unit 11 uses a predetermined large positive value as the amount of change in energy when the value of the state variable changes. This prevents the value of the state variable from being allowed to change, and can prevent the occurrence of the constraint violations 1 to 3.

[0073] FIG. 4 is a diagram showing another example of how to determine the range in which a value change is permitted. In the example of Fig. 4, in groups gD1 to gD3, state variables that are allowed to change in value from 0 to 1 are state variables with identification numbers of +1 and -1 relative to the identification number of the state variable whose value is 1. However, as in the above example, the upper and lower limits of the identification numbers of state variables that are allowed to change are determined so as not to fall under the above constraint violations 1 to 3.

[0074] For example, as in the example of Figure 4, in group gD1, when the value of a state variable with an identification number of +1 relative to the identification number of a state variable with a value of 1 changes to 1, this falls under constraint violation 2 in relation to the state variable in group gD2 with a value of 1. Therefore, in group gD1, only the state variable with an identification number of -1 is allowed to change relative to the identification number of the state variable with a value of 1.

[0075] In group gD2, when the value of a state variable with an identification number of -1 changes to 1 relative to the identification number of a state variable with a value of 1, this falls under constraint violation 2 in relation to the state variable with a value of 1 in group gD1. Therefore, in group gD2, only changes in state variables with an identification number of +1 relative to the identification number of a state variable with a value of 1 are permitted.

[0076] In group gD3, if the value of a state variable with an identification number of +1 relative to the identification number of a state variable with a value of 1 changes to 1, this corresponds to constraint violation 3. Therefore, in group gD3, only changes to the state variable with an identification number of -1 relative to the identification number of a state variable with a value of 1 are allowed.

[0077] By further limiting the range of allowable value changes in this manner, it is possible to simplify the hardware configuration of the search unit 11, which determines whether or not to allow a change in the value of a state variable based on the amount of change in energy when the change occurs.

[0078] In the example of Figure 4, the state variables that are allowed to change in value from 0 to 1 are state variables with identification numbers of +1 and -1 relative to the identification number of the state variable with a value of 1, but the identification numbers are not limited to +1 and -1 and can be set as appropriate.

[0079] As described above, according to the optimization device 10 of the first embodiment, in a certain group, based on the identification number of a state variable whose value is 1, an upper or lower limit of the identification number of a state variable whose value is allowed to change from 0 to 1 is determined in another group. This makes it possible to restrict the range of state variables whose values ​​are allowed to change next according to the current state, and makes it possible to search for a solution that satisfies the constraint conditions without increasing the number of constraint terms, thereby improving the convergence to an optimal solution.

[0080] In addition, as shown in Figure 1, the number of state variables to be included in the evaluation function can be the square of the number of nodes excluding depots + the number of trucks - 1, so the number of state variables can be kept low even if the number of nodes or the number of trucks increases.

[0081] Furthermore, in the optimization device 10, for the truck number -1, groups gD1 to gD3 (groups of the truck number -1) are provided that indicate whether the trucks have returned to the depot at each time, and the sum of the load amounts of each truck along each route can be calculated using these groups. This makes it possible to calculate a constraint term for preventing the sum of the load amounts from exceeding the maximum load amount (an example of calculating the constraint term will be described later). This eliminates the need to prepare multiple patterns of combinations of the load amounts of each truck that satisfy the maximum load amount and search for an optimal solution for each pattern, and thus makes it possible to suppress an increase in the number of times the optimal solution search process is executed.

[0082] The method of determining the upper or lower limit of the identification number of a state variable whose value is allowed to change from 0 to 1 as described above is not limited to routing problems of multiple nodes such as CVRP, but can also be applied to other combinatorial optimization problems.

[0083] (Second embodiment) FIG. 5 illustrates an example of an optimization device according to the second embodiment. The optimization device 20 of the second embodiment includes a search unit 21 and a transition allowable range determination unit 22 .

[0084] The search unit 21 performs a determination process to determine whether or not a value can be changed based on the amount of change in energy when the value of any of the multiple state variables included in the evaluation function is changed. Then, based on the result of the determination process, the search unit 21 searches for an optimal solution that minimizes energy by repeating an update process that changes the value of any of the multiple state variables.

[0085] The search unit 21 performs an update process while satisfying the 1-hot constraint (in the following example, the 2-way 1-hot constraint). Note that the search unit 21 may perform an update process while satisfying the 1-way 1-hot constraint, like the search unit 11 in the first embodiment.

[0086] The 16 × 16 state variables (arranged in 16 rows and 16 columns) set up as shown in Figure 1 above to calculate CVRP have a 2-way 1-hot constraint. When excluding the search for states other than the states that satisfy the 2-way 1-hot constraint, the values ​​of four state variables are changed in one state transition.

[0087] In addition, when the 2-way 1-hot constraint is satisfied, if one of the state variables with a value of 0 is selected as a candidate for updating, the other three state variables are determined. j When x is the candidate for update, j Among the state variables in the same row and column as x, the state variable with a value of 1 i ,x l is a candidate for updating. i The same column as x l x with value 0, which is in the same row as k are candidates for updating.

[0088] The change in energy of the Ising model that occurs when the values ​​of these four state variables are changed is called ΔE j Then, ΔE j can be expressed as the following equation (8).

[0089]

number

[0090] Also, x i ,x j ,x k ,x l The change in the local field due to the change in (Δh m (m=1,2,…,N) can be expressed as the following equation (9).

[0091]

number

[0092] The search unit 21 includes a ΔE calculation unit 21a, a selection circuit 21b, an identification number calculation unit 21c, an update unit 21d, and a control unit 21e. The ΔE calculation unit 21a calculates the amount of change in energy (ΔE1 to ΔE N ) is calculated. In order to perform the update process while satisfying the 1-hot constraint, the ΔE calculation unit 21a calculates the amount of change in energy as shown in equation (8).

[0093] However, for state variables for which a signal indicating transition prohibition has been output from transition allowable range determination unit 22 or state variables whose current value is 1, ΔE calculation unit 21a outputs a predetermined large positive value as the amount of change in energy when the value of the state variable changes. The predetermined large positive value is, for example, the maximum positive value that optimization device 20 can output. For state variables whose current value is 1, notification is given by N control signals EN output by update unit 21d. For example, the control signal EN corresponding to a state variable whose current value is 1 is 1, and the control signal EN corresponding to a state variable whose current value is 0 is 0.

[0094] The ΔE calculation unit 21a is realized, for example, using a register that holds weighting coefficients and local fields, a selector that selects the weighting coefficients used in the calculations of equations (8) and (9), and an adder / subtractor that performs the calculations of equations (8) and (9).

[0095] The selection circuit 21b selects thermal excitation energy and ΔE1 to ΔE N Based on the magnitude relationship between ΔE1 and ΔE2, an identification number=j that identifies one of the state variables whose value is 0 and that is allowed to change from 0 to 1, among those included in each group, is output. The thermal excitation energy is determined based on a random number and a parameter T that represents the temperature input from the control unit 21e, and is sometimes called a noise value. Note that the thermal excitation energy and ΔE1 to ΔE2 are N Depending on the magnitude relationship with j, there may be cases where no change is allowed for any of the state variables whose value is 0. In the following, it is assumed that the selection circuit 21b outputs, together with the identification number=j, a flag f indicating whether or not a change in the value of the state variable with the identification number=j is allowed. For example, if the value of the flag f is 1, this indicates that a change in the value of the state variable is allowed, and if the value of the flag f is 0, this indicates that a change is not allowed.

[0096] The identification number calculation unit 21c has, for example, a register that stores the identification number of the group to which each state variable belongs and the identification number of the state variable in each group whose value is 1. Based on the identification number = j output by the selection circuit 21b, the identification number calculation unit 21c calculates the other three identification numbers = i, k, l.

[0097] For example, i,l is x j Among the state variables in the same row and column as x, the state variable with a value of 1 i ,x l k is the identification number of x i The same column as x l The same row as x k The identification number of the device. Note that k can be calculated by k=i+lj.

[0098] In the following, it is assumed that the identification number calculation unit 21c also outputs the identification number=j and flag f supplied from the selection circuit 21b. The identification numbers=i, j, k, l are supplied to the ΔE calculation unit 21a and are used when updating the local field used to calculate the amount of change in energy based on equation (9). If the flag f indicates that no change is permitted, the identification number calculation unit 21c sets the identification numbers=i, j, k, l to invalid values ​​(for example, 0), for example.

[0099] The identification number calculation unit 21c may be, for example, a processor such as a CPU or GPU that performs the above-mentioned processing based on the identification number stored in a register, or may be realized using various logic circuits.

[0100] The update unit 21d updates N state variables (x1 to x N The update unit 21d has a storage unit 21d1 that holds the values ​​of i, l, j, k, and the like. The storage unit 21d1 is realized by using electronic circuits such as a register and an SRAM, for example. When the flag f indicates that the change is permitted, the update unit 21d updates the values ​​of the state variables of the identification numbers=i, l output by the identification number calculation unit 21c from 1 to 0, and updates the values ​​of the state variables of the identification numbers=j, k from 0 to 1.

[0101] The update unit 21d updates the energy based on the amount of change in the energy corresponding to the change in the state variables of the identification numbers=i, j, k, l. The storage unit 21d1 holds the minimum energy at each update point and the state when the minimum energy was obtained (the state at the time of the minimum energy). The update unit 21d also receives the above-mentioned control signal EN and x1 to x N is supplied to the ΔE calculation unit 21a.

[0102] The update unit 21d can be realized using an addition circuit for updating the energy, a comparator for comparing the updated energy with the previous minimum energy, and various logic circuits for inverting the values ​​of state variables of identification numbers = i, j, k, l from 0 to 1 or from 1 to 0.

[0103] The control unit 21e performs an initial setting process for the optimization device 20. As the initial setting process, the setting of weight coefficients for calculating the formula (4), the setting of initial values ​​of local fields and state variables, the setting of identification numbers of groups to which each state variable belongs, etc. are performed. The initial values ​​of the state variables are set so as to satisfy the 1-hot constraint in each group. Furthermore, for groups for which the transition allowable range determination unit 22 described later determines the range in which a value change is allowed, a hot bit is set so as not to fall under the constraint violations 1 to 3 described above, for example.

[0104] Furthermore, the control unit 21e decreases the value of T in accordance with an externally specified temperature schedule, for example, every time the update process for updating the state of the Ising model is repeated a predetermined number of times.

[0105] Furthermore, the control unit 21e repeats the update process a predetermined number of times, and then the state (x1 to x N ) and output it to the outside as, for example, a solution to the optimization problem. Note that the control unit 21e may obtain and output the minimum energy or the state at the time of the minimum energy stored in the storage unit 21d1 after the update process has been repeated a predetermined number of times. The control unit 21e may output the obtained various information to a display device (not shown) for display, or may transmit it to an external information processing device.

[0106] The control unit 21e can be realized by, for example, an electronic circuit such as an ASIC or an FPGA. The control unit 21e may be a processor such as a CPU or a GPU. In this case, the processor performs the above-mentioned processing by executing a program stored in a memory (not shown).

[0107] The permissible transition range determining unit 22 determines, for a plurality of groups, a permissible range of value change so as not to violate a predetermined constraint condition. FIG. 6 illustrates an example of the transition allowable range determining unit.

[0108] The transition allowable range determination unit 22 includes a storage unit 22a, a hot bit update unit 22b, a boundary value calculation unit 22c, and a transition permission / prohibition signal output unit 22d. The storage unit 22a stores, for example, values ​​and identification numbers of N state variables, identification numbers of groups to which each state variable belongs, identification numbers for identifying state variables of a plurality of groups for which a transition tolerance range is determined, and group sizes (the number of state variables belonging to each group). Furthermore, the storage unit 22a stores identification numbers of hot bits in each group for which a transition tolerance range is determined. These identification numbers and group sizes are stored in the storage unit 22a in the above-mentioned initial setting process by the control unit 21e. Furthermore, the storage unit 22a stores boundary values ​​(upper and lower limits) of identification numbers of state variables that allow a change in value from 0 to 1 in each group for which a transition tolerance range is determined, calculated by the boundary value calculation unit 22c. The storage unit 22a is, for example, an electronic circuit such as an SRAM or a register.

[0109] The hot bit update unit 22b updates the identification number (j, k) and flag f output by the identification number calculation unit 21c, and the state (x1 to x N For example, the hot bit update unit 22b updates the identification number of the hot bit stored in the storage unit 22a based on j or k. j or x k When j or k becomes 1, j or k becomes the identification number of the hot bit of that group. Such a hot bit update unit 22b can be realized by various logic circuits.

[0110] The boundary value calculation unit 22c calculates boundary values ​​(upper and lower limits) of the identification numbers of the state variables that are allowed to change in value from 0 to 1 in each group, based on the identification numbers of the state variables and hot bits of each group for which the allowable transition range is determined, and the group size.

[0111] FIG. 7 is a diagram illustrating an example of calculation of the boundary value. FIG. 7 shows an example of calculating boundary values ​​so as not to fall under the constraint violation conditions 1 to 3 described above.

[0112] In group gD1, the lower limit Min is the first identification number of the state variable of group gD1 + 1, and the upper limit Max is the identification number of the hot bit of group gD2 minus 18. As mentioned above, 18 is the group size (16) + 2.

[0113] In group gD2, the lower limit Min is the value obtained by adding 18 to the identification number of the hot bit in group gD1, and the upper limit Max is the value obtained by subtracting 18 from the identification number of the hot bit in group gD3.

[0114] In group gD3, the lower limit Min is a value obtained by adding 18 to the identification number of the hot bit in group gD2, and the upper limit Max is a value obtained by subtracting 1 from the last identification number of the state variable in group gD3.

[0115] The boundary value calculation unit 22c that performs such processing is realized by using various logic circuits such as an adder / subtractor circuit. The transition possibility signal output unit 22d outputs a signal (transition possibility signal) indicating whether or not a change in value is possible for each state variable belonging to the above-mentioned multiple groups, based on the boundary values ​​calculated by the boundary value calculation unit 22c and stored in the storage unit 22a.

[0116] FIG. 8 is a diagram illustrating an example of the transition permission signal output unit and the storage unit. FIG. 8 shows register groups 22a1, 22a2, 22a3, 22a4, and 22a5 that store some of the information stored in the memory unit 22a.

[0117] The register group 22a1 includes x1 to x N The identification numbers (1 to N) of the registers 22a2 are stored in the register group 22a1. N The identification number of the group to which each of these belongs (g1 to g N ) is stored in the register group 22a3, and x1 to x NThe register group 22a4 stores the values ​​of x1 to x N The lower limits (Min1 to Min N ) is stored in the register group 22a5, and x1 to x N The upper limit (Max1 to Max N ) is stored.

[0118] In addition, Min1~Min N and Max1~Max N The jth Min j and Max j is the state variable (x j ) belongs to matches the identification number of the group for which the transition allowable range is determined, stored in the storage unit 22a. Therefore, for groups other than the group for which the transition allowable range is determined, the above lower limit and upper limit remain at their initial values.

[0119] The transition possibility signal output unit 22d has transition possibility bit generation circuits 22d1, 22d2, ..., 22dN. The transition possibility bit generation circuits 22d1 to 22dN are N The value of Min1~Min N And Max1~Max N Based on x1~x N For each of these, transition bits p1, p2, ..., p N This generates and outputs N transition enable / disable bits p1 to p N The transition permission / prohibition signal based on the above is output from the transition permission / prohibition signal output unit 22d.

[0120] FIG. 9 is a diagram illustrating an example of a transition possible / prohibited bit generating circuit. Fig. 9 shows an example of a circuit of the i-th transition possible / prohibited bit generation circuit 22di among the transition possible / prohibited bit generation circuits 22d1 to 22dN shown in Fig. 8. The other transition possible / prohibited bit generation circuits can also be realized by a similar circuit configuration.

[0121] The transition permission bit generation circuit 22di includes comparison circuits 30, 31 and a NAND (negative logical product) circuit 32. The comparison circuit 30 outputs the comparison result between i the aforementioned lower limit Min in the group to which x belongs i and the identification number = i. When Min i ≤ i, the comparison circuit 30 outputs 1; when Min i > i, the comparison circuit 30 outputs 0.

[0122] The comparison circuit 31 outputs the comparison result between i the aforementioned upper limit Max in the group to which x belongs i and the identification number = i. When Max i ≥ i, the comparison circuit 31 outputs 1; when Max i < i, the comparison circuit 31 outputs 0.

[0123] The NAND circuit 32 outputs the negative logical product of the output signals of the comparison circuits 30, 31 and the value obtained by inverting the value of x i (when the value of x i is 0, it is 1; when it is 1, it is 0). When all three inputs of the NAND circuit 32 are 1, as the transition permission bit p i for x i , the NAND circuit 32 outputs 0, indicating that the change of the value of x from 0 to 1 is allowed. When any one of the three inputs of the NAND circuit 32 is 0, as the transition permission bit p i for x i , the NAND circuit 32 outputs 1, indicating that the change of the value of x from 0 to 1 is prohibited.

[0124] By using the transition permission range determination unit 22 as described above, for a plurality of groups, it is possible to determine the range in which the change of the value is allowed so as not to violate the predetermined constraint conditions. For groups different from the groups for which the transition permission range is determined, for example, the initial value of the aforementioned lower limit may be set to 1 and the initial value of the aforementioned upper limit may be set to N. In that case, for the state variables belonging to that different group, 0, which is the transition permission bit indicating that the change of the value from 0 to 1 is allowed, will be output.

[0125] (Calculation method for energy change considering CVRP constraints, part 1) As mentioned above, CVRP has a constraint that the total demand value of all nodes on the route, excluding depots, must be within the maximum load capacity of one truck. For example, if the sum of the loads on each route by four trucks (D TOT1 ~D TOT4 ) can be calculated using the above equation (6), but these are quadratic expressions for the state variables. Therefore, D TOT1 ~D TOT4 can be transformed as shown in the following equation (10).

[0126]

number

[0127] In equation (10), x is (x1, x2, …, x N ) matrices, and each of V1 to V4 is an N×N matrix. If the maximum load is Q, the above constraints are D TOT1 ~D TOT4 This can be expressed as ≦Q.

[0128] An evaluation function including such a quadratic inequality constraint as a constraint term can be expressed by the following equation (11).

[0129]

number

[0130] In equation (11), C is a cost term and represents the total distance traveled by the four trucks. The cost term C can be expressed by the following equation (12).

[0131]

number

[0132] In equation (12), W is the W ijThis is an N × N matrix by The change in the value of the evaluation function (change in energy) ΔE can be expressed by the following equation (13).

[0133]

number

[0134] In order to calculate the above-mentioned ΔE as shown in equation (13) in the ΔE calculation unit 21a as shown in FIG. 5, ΔC and ΔP1 to ΔP4 are calculated in parallel and added together.

[0135] FIG. 10 is a diagram illustrating an example of the ΔE calculation unit. The ΔE calculation unit 40 includes storage units 41a, 41b1 to 41b4, a ΔC calculation circuit 42a, and a ΔD TOT The device includes calculation circuits 42b1 to 42b4 and a ΔE output circuit 43.

[0136] The storage unit 41a stores the above-mentioned W. The storage units 41b1 to 41b4 store the above-mentioned V1 to V4. The storage units 41a, 41b1 to 41b4 can be realized using electronic circuits such as registers and SRAMs.

[0137] The ΔC calculation circuit 42a calculates x1~x N For each of these, ΔE j The amount of change in energy (ΔC) corresponding to x1 to x2 is calculated. The identification numbers (i, j, k, l) supplied to the ΔC calculation circuit 42a are used to select the elements of W to be used in the calculation. N Let x1~x be the local field of (8) (e.g., held in a register). N The energy change is then propagated to a circuit that calculates the change in energy for either

[0138] However, the ΔC calculation circuit 42a calculates the state variables whose corresponding control signals EN are 1 or the corresponding transition enable / disable bits p1 to p NFor state variables with a value of 1, a large positive value is output as the change amount.

[0139] ΔD TOT The calculation circuits 42b1 to 42b4 calculate x1 to x N For each of these, ΔE j The change in energy corresponding to (ΔD TOT1 ~ΔD TOT4 ) where ΔD TOT The calculation circuits 42b1 to 42b4 differ from the ΔC calculation circuit 42a in that they use V1 to V4 instead of W. For example, ΔD TOT The calculation circuit 42b1 calculates x1 to x N The change in energy for each of these, ΔD TOT1 Calculate.

[0140] ΔD TOT The identification numbers (=i, j, k, l) supplied to the calculation circuits 42b1 to 42b4 are used to select elements of V1 to V4 to be used in the calculation. N The local field of equation (8) is expressed as x1~x N The energy change is then propagated to a circuit that calculates the change in energy for either

[0141] In addition, ΔD TOT For the calculation circuits 42b1 to 42b4, the state variables whose corresponding control signals EN are set to 1 or the corresponding transition enable / disable bits p1 to p N For state variables with a value of 1, a large positive value may be output as the amount of change.

[0142] The ΔE output circuit 43 calculates x1 to x N The change in energy for each of these is ΔE1 to ΔE N Calculate and output. (Calculation method for energy change considering CVRP constraints, part 2) In the second calculation method, the optimization device 20 calculates the total demand D at each time t. tThe optimization device 20 calculates and stores a variable r i,t Memorize.

[0143] Figure 11 shows the D t andr i,t FIG. 13 is a diagram showing an example of storage of the above. Figure 11 shows the D for i = 1 to 4 and t = 1 to 16. t andr i,t An example of memory is shown. i,t is 1 if the truck passes through route i at time t, and 0 if it does not.

[0144] x1~x N Among these, there are three cases: Case 1, in which the order in which nodes are visited within the same route changes depending on the identification number of the state variable whose value changes; Case 2, in which the nodes visited between different routes are swapped; and Case 3, in which the order in which depots and nodes are visited is swapped.

[0145] FIG. 12 is a diagram illustrating an example of Case 1. Figure 12 shows an example in which the order of visiting nodes with node numbers 2 and 4 is swapped as the values ​​of four state variables (16 × 16 state variables in box 15) change when calculating a CVRP with four trucks and 13 nodes, as in Figure 1.

[0146] In case 1, the total demand at each time t is D t In the example of Figure 12, D2 and D3 are swapped, but the sum of the loads on each route (D TOT1 ~D TOT4 ) does not change. Therefore, there is no change in the constraint terms in the evaluation function shown in equation (11).

[0147] Figure 13 shows the state transition of r i,t FIG. 13 is a diagram showing an example of storage of the above. In case 1, as shown in Figure 13, before and after the state transition (changes in four state variables (four-bit transitions)), r i,t does not change.

[0148] FIG. 14 is a diagram illustrating an example of Case 2. In FIG. In the example of Fig. 14, in the first route, the node that the truck visits at time t = 2 is changed from node number = 2 to node number = 3. In addition, in the second route, the node that the truck visits at time t = 3 is changed from node number = 3 to node number = 2. In other words, the nodes visited between the different routes are swapped.

[0149] In case 2, the total demand at each time t is D t Of these, D2 and D6 will be swapped, and D TOT1 and D TOT2 changes. Therefore, the constraint term changes. However, r before and after the state transition in Case 2 i,t is the same as in FIG. 13 and does not change before and after the transition.

[0150] FIG. 15 is a diagram illustrating an example of Case 3. In the example of Figure 15, in the first route, the truck returns to the depot at time t = 3 instead of visiting node number = 4. In the second route, the truck visits node number = 4 instead of being at the depot at time t = 5.

[0151] In case 3, the total demand at each time t is D t Of these, D3 and D5 will be swapped, and D TOT1 and D TOT2 changes. This causes a change in the constraint term. Figure 16 shows the state transition of r i,t FIG. 13 is a diagram showing an example of storage of the above.

[0152] In case 3, the number of visited nodes in the two routes changes before and after the state transition. Therefore, r i,t As shown in FIG. 16, the value of the sigma-based ...

[0153] The optimization device 20 is D t andr i,t Based on D TOT1 ~D TOT4 Calculate ΔD while considering the above cases 1 to 3. TOT1 ~ΔD TOT4 As mentioned above, in the state transition of Case 1, D TOT1 ~D TOT4 On the other hand, in the state transitions of cases 2 and 3, D TOT1 ~D TOT4 A change occurs in any two of the above.

[0154] In the example shown in Figure 14 (Case 2), D before the state transition TOT1 D TOT1 = D1 + D2 + D3 + D4 + D5, and D before the state transition TOT2 D TOT2 =D6+D7+D8. D after state transition TOT1 D TOT1 = D1 + D'2 + D3 + D4 + D5, and D after the state transition TOT2 D TOT2 =D'6+D7+D8. Note that D'2=-D'6.

[0155] The change in the sum of the loads before and after the state transition is ΔD TOT1 =ΔD2=D'2-D2, which can be expressed as the following equation (14).

[0156]

number

[0157] In equation (14), Δr 1,i is the state transition time r before and after the state transition at time t=i. 1,i This represents the change in the current value, which is 0 in the example of FIG. On the other hand, ΔD TOT2 =-ΔD TOT1 This can be expressed as:

[0158] In the example shown in Figure 15 (Case 3), D before the state transition TOT1 D TOT1= D1 + D2 + D3 + D4 + D5, and D before the state transition TOT2 D TOT2 =D6+D7+D8. D after state transition TOT1 D TOT1 = D1 + D2 + D'3, and D after the state transition TOT2 D TOT2 =D4+D'5+D6+D7+D8.

[0159] The change in the sum of the loads before and after the state transition is ΔD TOT1 =ΔD3-(D4+D5)=ΔD3+(D4Δr 1,4 +D5Δr 1,5 ), which can be expressed as the following equation (15).

[0160]

number

[0161] In equation (15), Δr 1,i is the state transition time r before and after the state transition at time t=i. 1,i represents the change in, which is 1 in case 3. On the other hand, ΔD TOT2 =-ΔD TOT1 This can be expressed as:

[0162] Considering the above example, the change in the sum of the loads before and after the state transition for a route p is ΔD TOTp can be calculated by the following procedure. Step 1: The optimization device 20 generates identification numbers=i, j, k, and l of the four state variables of transition candidates.

[0163] Step 2: The optimization device 20 calculates a route p that causes a change in the sum of the load amounts based on the identification numbers=i, j, k, l. Step 3: The optimization device 20 finds an interval [s1, s2] of the identification numbers in which a change in the sum of the load amounts occurs for the route p.

[0164] Step 4: The optimization device 20 calculates ΔD t, Δr i,t Calculate. Step 5: The optimization device 20 calculates ΔD based on the following equation (16): TOTp Calculate.

[0165]

number

[0166] In order to execute the above-mentioned steps 1 to 5 in the optimization device 20, for example, a ΔE calculation unit as described below can be used. Fig. 17 is a diagram showing another example of the ΔE calculation unit, in which the same elements as those shown in Fig. 10 are denoted by the same reference numerals.

[0167] The ΔE calculation unit 50 includes a storage unit 51, an update circuit 52, a transition candidate generation circuit 53, a route p calculation circuit 54, and a ΔD TOT It has a calculation circuit 55 and a ΔE output circuit 56. The storage unit 51 stores the D t andr i,t The storage unit 51 can be realized by using electronic circuits such as a register and an SRAM.

[0168] The update circuit 52 is N Based on the value of t andr i,t Update. The transition candidate generating circuit 53 is N Based on the values ​​of the transition candidate bits p1 to p2, the transition candidate generation circuit 53 generates the identification numbers = i, j, k, l of the four state variables of the transition candidate. For example, the transition candidate generation circuit 53 selects the identification number = j from the identification numbers of the state variables whose current value is 0, and repeats the process of generating other identification numbers = i, k, l so as to satisfy the 2-way 1-hot constraint, the number of times corresponding to the number of state variables whose value is 0. However, the transition candidate generation circuit 53 uses the transition possibility bits p1 to p N The identification numbers of state variables whose are equal to 1 are not adopted as identification numbers = i, j, k, or l.

[0169] By further limiting the range in which the value change is permitted as in FIG. 4 described above, it is possible to simplify the hardware configuration of the transition candidate generating circuit 53 and the circuits that perform the subsequent processes. The route p calculation circuit 54 calculates a route p that causes a change in the sum of the load amounts based on the identification numbers=i, j, k, l generated by the transition candidate generation circuit 53. For example, this process is performed the same number of times as the number of pairs of identification numbers=i, j, k, l generated by the transition candidate generation circuit 53.

[0170] ΔD TOT The calculation circuit 55 determines the interval [s1, s2] of the identification number in which the sum of the loads changes with respect to the route p. For example, in the case of the example of FIG. 14 (case 2), s1=s2=2, and in the case of the example of FIG. 15 (case 3), s1=3, s2=5. Then, ΔD TOT The calculation circuit 55 is D t andr i,t Based on this, ΔD for the interval [s1, s2] t , Δr i,t Calculate ΔD as shown in equation (16) TOTp Calculate.

[0171] The ΔE output circuit 56 is t andr i,t Based on D TOTp Then, the ΔE output circuit 56 calculates x1 to x N For each of ΔC, ΔP for all routes p where the sum of the loads changes p By adding ΔE1 to ΔE N Calculate and output ΔP p =max(D TOTp +ΔD TOTp -Q,0)-max(D TOTp -Q,0).

[0172] (Example of Overall Operation of Optimization Device 20) FIG. 18 is a flowchart showing an example of the overall operational flow of the optimization device. In the following, the case where the simulated annealing method is used will be described as an example, but the present invention is not limited to this, and a method such as the replica exchange method can also be used.

[0173] First, under the control of the control unit 21e, an initial setting process is performed (step S1). In the initial setting process, the control unit 21e sets weight coefficients for calculating the formula (4), sets initial values ​​of local fields and state variables, sets identification numbers of groups to which each state variable belongs, and so on. When the first calculation method of ΔE described above is used, the control unit 21e sets V1 to V4, and when the second calculation method of ΔE is used, sets D t andr i,t Set the initial value of .

[0174] The initial values ​​of the state variables are set so as to satisfy the 1-hot constraint in each group. Furthermore, for the group for which the transition allowable range determination unit 22 determines the range in which the value change is allowed, a hot bit is set so as not to fall under the constraint violations 1 to 3 described above.

[0175] Furthermore, the control unit 21e sets the initial value of T based on a predetermined temperature change schedule, the number of times the update process is repeated, and so on. Thereafter, the allowable transition range determining unit 22 determines a range in which a change in value is allowed (allowable transition range) (step S2).

[0176] Then, the ΔE calculation units 21a, 40, and 50 calculate ΔE1 to ΔE2 when a state that satisfies the 1-hot constraint transitions from a state that satisfies the 1-hot constraint to another state that satisfies the 1-hot constraint through a state transition with a Hamming distance of 4 in each group that satisfies the 1-hot constraint. N is calculated (step S3).

[0177] However, for a transition prohibition bit that is a state variable for which a signal indicating a transition prohibition has been output from the transition allowable range determination unit 22 or a state variable whose current value is 1, the ΔE calculation units 21a, 40, and 50 output a predetermined large positive value as the amount of change in energy.

[0178] After the process of step S3, the selection circuit 21b selects ΔE1 to ΔE N Based on this, an identification number=j is selected (step S4). The process in step S4 is ΔE1 to ΔE N Based on x1~x N For example, the selection circuit 21b selects thermal excitation energy generated based on T and a uniform random number, and ΔE1 to ΔE N , selects the amount of change in energy that is smaller than the thermal excitation energy, and selects the identification number j corresponding to that amount of change. If there are multiple amounts of change in energy that are smaller than the thermal excitation energy, the selection circuit 21b selects one of the amounts of change in energy, for example, according to a predetermined rule or randomly. If there is no amount of change in energy that is smaller than the thermal excitation energy, no change occurs in any of the state variables, but the selection circuit 21b may prompt the occurrence of a state transition by adding an offset value to the thermal excitation energy, for example.

[0179] After the process of step S4, the identification number calculation unit 21c further calculates the identification numbers=i, k, l from the selected identification number=j (step S5). Thereafter, an update process is performed (step S6). In the process of step S6, the ΔE calculation units 21a, 40, and 50 update the local fields, and the update unit 21d updates the four state variables stored in the storage unit 21d1. When the ΔE calculation unit 50 is used, the update circuit 52 updates the D t andr i,t will be updated.

[0180] Furthermore, in the process of step S6, the hot bit update unit 22b of the transition allowable range determination unit 22 updates the hot bit identification number stored in the storage unit 22a. The control unit 21e determines whether the number of repetitions of the processes from steps S2 to S6 has reached a predetermined number N1 (step S7). If the number of repetitions has not reached the predetermined number N1, the processes from step S2 are repeated.

[0181] When the number of repetitions reaches the predetermined number N1, the control unit 21e determines whether or not the number of changes in T (the number of temperature changes) reaches a predetermined number N2 (step S8). If the number of temperature changes has not reached the predetermined number N2, the control unit 21e changes T (lowers the temperature) (step S9). The method of changing the predetermined number of times N1, N2, and the value of T (such as how much to reduce the value at one time) is determined based on a predetermined temperature change schedule, etc. After the process of step S9, the process from step S2 is repeated.

[0182] When the number of temperature changes reaches the predetermined number N2, the control unit 21e outputs, for example, the values ​​of all state variables stored in the storage unit 21d1 at that time as the calculation results (step S10), and ends the process. Note that the control unit 21e may calculate the energy based on the values ​​of all state variables every time a state transition occurs, sequentially update the values ​​of all state variables that provide the minimum energy, and output the values ​​of all state variables at the time when the number of temperature changes reaches the predetermined number N2 as the solution.

[0183] The order of the above processes is not limited to the above example, and may be changed as appropriate. As described above, the optimization device 20 of the second embodiment can provide the same effects as the optimization device 10 of the first embodiment. That is, the transition allowable range determination unit 22 determines the upper or lower limit of the identification number of a state variable that is allowed to change in value from 0 to 1 in a certain group, based on the identification number of a state variable whose value is 1 in another group. This makes it possible to restrict the range of the state variable whose value is allowed to change next according to the current state, making it possible to search for a solution that satisfies the constraint conditions without increasing the number of constraint terms, and improving the convergence to an optimal solution.

[0184] While one aspect of the optimization device and optimization method of the present invention has been described above based on the embodiment, these are merely examples and the present invention is not limited to the above description. [Explanation of symbols]

[0185] 10 Optimization Device 11 Exploration Department 11a,12a Storage section 12 Transition tolerance determination section 15 slots

Claims

1. a search unit that searches for an optimal solution that minimizes the energy by repeating a determination process of determining whether or not a value can be changed based on an amount of change in energy when a value of any one of a plurality of state variables included in an evaluation function that represents the energy of an Ising model that models a combinatorial optimization problem is changed, and a process of changing the value based on a result of the determination process; a transition tolerance range determination unit that determines an upper limit or a lower limit of a second identification number of a second state variable that is allowed to change from the second value in a second state variable group among the plurality of state variable groups, based on a first identification number of a first state variable that has the first value in a first state variable group among a plurality of state variable groups, each of which has one state variable having a first value and the remaining state variables having a second value; An optimization device having the above configuration.

2. The search unit includes an energy change amount calculation unit that calculates an amount of change in the energy, 2. The optimization device according to claim 1, wherein the energy change amount calculation unit outputs a predetermined positive value as the amount of change when a state variable in the second state variable group, other than the second state variable, which is allowed to change from the second value, changes from the second value.

3. 3. The optimization device according to claim 1, wherein when the combinatorial optimization problem is a routing problem involving a plurality of nodes, the number of the plurality of state variables is the square of a value obtained by adding the number of the plurality of nodes excluding a starting point to the number of the plurality of transport vehicles minus 1.

4. 4. The optimization device according to claim 3, wherein a number of the plurality of state variable groups is the number of vehicles minus 1, and each of the plurality of state variable groups includes state variables equal to a value obtained by adding the number of vehicles minus 1 to the number of the plurality of nodes excluding the starting point, and indicating whether any of the number of vehicles minus 1 has returned to the starting point at each time.

5. 5. The optimization device according to claim 4, wherein the transition allowable range determination unit determines the upper limit or the lower limit such that a second vehicle among the number-1 of vehicles, which visits any of the plurality of nodes at an earlier time than a first vehicle, returns to the starting point at an earlier time than the first vehicle.

6. 6. The optimization device according to claim 5, wherein the transition allowable range determination unit determines the upper limit or the lower limit so that the first transport vehicle does not return to the starting point at a time immediately after the time that the second transport vehicle returns to the starting point.

7. a search unit searches for an optimal solution that minimizes the energy by repeating a determination process of determining whether or not a value can be changed based on an amount of change in energy when a value of any one of a plurality of state variables included in an evaluation function that represents the energy of an Ising model that models a combinatorial optimization problem is changed, and a process of changing the value based on a result of the determination process; a transition tolerance range determination unit that determines an upper limit or a lower limit of a second identification number of a second state variable that is allowed to change from the second value in a second state variable group among the plurality of state variable groups, based on a first identification number of a first state variable having the first value in a first state variable group among a plurality of state variable groups, each of which has one state variable having a first value and the remaining state variables having a second value; Optimization methods.

Citation Information

Patent Citations

  • Method and apparatus for searching for driving route using quantum calculation, driving route search program and record medium recording the program

    JP2003114132A

  • Transportation schedule making method and its system

    JP2003285930A

  • Optimization device and method for controlling optimization device

    JP2020064535A

  • Optimization device and method for controlling optimization device

    JP2020173661A