Information processing device, information processing method, and program
The information processing device with multiple processing units and adaptive exchange probabilities enhances the effectiveness of solution searching in combinatorial optimization problems by widening the search space and improving convergence to lower evaluation function values.
Patent Information
- Application Number
- JP2025077749
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2025-05-08
- Publication Date
- 2026-11-19
AI Technical Summary
Existing metaheuristic algorithms like Simulated Annealing (SA) struggle to search a wide range of solution spaces, often getting stuck in local optimal solutions, and Temperature Parallel Simulated Annealing (TPSA) may not effectively converge to lower evaluation function values without appropriate temperature settings.
An information processing device with multiple processing units having different initial temperatures, performing solution exchanges based on evaluation function values and adjusting exchange probabilities, followed by decreasing parameter values to enhance solution search effectiveness.
The proposed method enables more effective solution searching by widening the search space and increasing the likelihood of converging to lower evaluation function values, improving the efficiency of combinatorial optimization problem solving.
Smart Images

Figure 2026181927000001_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to an information processing apparatus, an information processing method, and a program.
Background Art
[0002] In recent years, combinatorial optimization problems in mathematical optimization are problems used in various business scenes, such as planning delivery schedules in the logistics industry or optimizing production schedules in the manufacturing industry. As a solution method for such combinatorial optimization problems, there is a metaheuristic algorithm that can efficiently search for an optimal solution or a good solution close to the optimal solution.
[0003] One of the metaheuristic algorithms is a technique called Simulated Annealing (hereinafter also simply referred to as "SA"). SA is a technique derived from annealing in metallurgical engineering, and is a technique for searching for solutions while gradually transitioning the temperature from a high temperature to a low temperature. However, SA is a technique that makes it difficult to search a wide range of solution spaces, and there is a problem that the solution obtained by the search may fall into a local optimal solution.
[0004] As a technique to solve this problem, Non-Patent Document 1 proposes a technique called Temperature Parallel Simulated Annealing (hereinafter also simply referred to as "TPSA") that searches a wide range of solution spaces. TPSA is a technique in which an initial solution and mutually different initial temperatures are set for each, a plurality of processes that execute SA in parallel according to a constant temperature schedule are constructed, and solutions are exchanged between the processes stochastically at a specific timing.
Prior Art Documents
Non-Patent Documents
[0005]
Non-Patent Document 1
[0006] However, it is desirable that technologies be provided that enable more effective search for solutions. [Means for solving the problem]
[0007] To solve the above problem, according to one aspect of the present invention, an information processing device is provided, comprising: a plurality of processing units, each having parameters with different initial values set and each having an initial solution, and each obtaining a selected solution by searching for a solution based on the initial solution; an exchange unit that performs a solution exchange process relating to the exchange of the selected solution between at least two of the plurality of processing units; and an update unit that decreases the values of the parameters in each of the plurality of processing units after the solution exchange process has been performed.
[0008] The plurality of processing units may perform the search based on evaluation function values corresponding to each of the plurality of candidate solutions that each unit possesses, corresponding to the initial solution.
[0009] The exchange unit may perform the solution exchange process according to the exchange probability.
[0010] The two processing units include a first processing unit and a second processing unit, wherein the selected solution in the first processing unit is the first selected solution, the selected solution in the second processing unit is the second selected solution, the evaluation function value corresponding to the first selected solution is the first selected evaluation function value, and the evaluation function value corresponding to the second selected solution is the second selected evaluation function value, and the exchange unit may set the exchange probability based on the first selected evaluation function value and the second selected evaluation function value.
[0011] The parameter of the first processing unit is the first parameter, and the parameter of the second processing unit is the second parameter, and the initial value of the second parameter is smaller than the initial value of the first parameter. The exchange unit may set the exchange probability to a value less than 1 according to a predetermined rule if the second selection evaluation function value is greater than the first selection evaluation function value.
[0012] The initial value of the first parameter is a positive value, and the predetermined rule may include a rule that reduces the exchange probability as the product of the value of the first parameter and the value of the second parameter becomes smaller.
[0013] The initial value of the first parameter is a positive value, and the predetermined rule may include a rule that increases the exchange probability as the difference between the second selection evaluation function value and the first selection evaluation function value increases.
[0014] The initial value of the second parameter is smaller than the initial value of the first parameter, and the exchange unit may set the exchange probability to 1 if the second selection evaluation function value is smaller than the first selection evaluation function value.
[0015] The information processing device may include an execution control unit that controls the execution of a reprocessing operation a predetermined number of times, which includes: a reacquisition process in which the first processing unit and the second processing unit each perform a search for a solution based on the selected solution after the solution exchange process has been performed to reacquire the selected solution; a resolution exchange process in which the exchange unit resets the exchange probability based on the evaluation function value corresponding to the selected solution reacquired by the first processing unit and the evaluation function value corresponding to the selected solution reacquired by the second processing unit, and performs a solution exchange process again as a solution exchange process according to the reset exchange probability; and a reupdate process in which the update unit again decreases the value of the first parameter and again decreases the value of the second parameter after the solution exchange process has been performed again by the exchange unit.
[0016] The execution control unit may determine the minimum evaluation function value from the evaluation function values corresponding to the selected solution reacquired by the first processing unit and the second processing unit, respectively, after the reprocessing has been performed a predetermined number of times.
[0017] Each of the first and second processing units may obtain as the selected solution a candidate solution from the plurality of candidate solutions that takes the minimum value of the evaluation function.
[0018] The two processing units may be adjacent processing units of the plurality of processing units.
[0019] Furthermore, in order to solve the above problems, according to another aspect of the present invention, there is a computer-based information processing method provided which includes performing a solution exchange process between at least two processing units of a plurality of processing units, each having parameters with different initial values set and each having an initial solution, and each performing a solution search based on the initial solution to obtain a selected solution, and decreasing the values of the parameters held by each of the plurality of processing units after the solution exchange process has been performed.
[0020] Furthermore, in order to solve the above problems, according to another aspect of the present invention, a program is provided that causes a computer to function as a plurality of processing units, each having parameters with different initial values set and each having an initial solution, and obtaining a selected solution by each performing a search for a solution based on the initial solution; an exchange unit that performs a solution exchange process relating to the exchange of the selected solution between at least two of the plurality of processing units; and an update unit that decreases the values of the parameters held by each of the plurality of processing units after the solution exchange process has been performed. [Effects of the Invention]
[0021] As described above, the present invention provides a technology that enables more effective search for solutions. [Brief explanation of the drawing]
[0022] [Figure 1] This is a flowchart showing the calculation algorithm of TPSA. [Figure 2] This is a diagram showing an example of a temperature schedule in which the temperature of each of a plurality of processes is constant. [Figure 3] This is a diagram showing the transition of evaluation function values (energy transition) at each step calculated according to the temperature schedule shown in Figure 2. [Figure 4] This is a diagram showing an example of the functional configuration of the information processing apparatus 1 according to the embodiment of the present invention. [Figure 5] This is a diagram showing a configuration example of the control unit 120 according to the embodiment of the present invention. [Figure 6] This is a diagram showing an example of a temperature schedule in which the value of the temperature parameter for each of a plurality of processes decreases after solution exchange processing. [Figure 7] This is a flowchart showing an operation example of the information processing apparatus 1 according to the embodiment of the present invention. [Figure 8] This is a diagram showing the results of numerical calculation experiments. [Figure 9] This is a diagram showing the hardware configuration of an information processing apparatus 900 as an example of the information processing apparatus 1 according to the embodiment of the present invention. [Detailed Description of the Invention]
[0023] Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. Note that in the present specification and the drawings, components having substantially the same functional configuration are denoted by the same reference numerals, thereby omitting redundant description.
[0024] (0-1. Overview of TPSA) First, before describing the embodiments of the present invention in detail, an overview of TPSA will be explained.
[0025] Figure 1 is a flowchart of the TPSA calculation algorithm. As shown in Figure 1, in TPSA, the problem to be solved and various parameters necessary for the calculation are input (S11). For example, parameters may include temperatures set for multiple processes.
[0026] Next, initialization takes place (S12). During initialization, modeling is performed based on the problem to be solved. For example, the modeling may generate an evaluation function for evaluating the solution. Also, during initialization, initial solutions corresponding to each of the multiple processes are generated.
[0027] Next, solution searches are performed in each of the multiple processes at different temperatures for each process (S13). Subsequently, at specific timings, a solution exchange process is performed between adjacent processes according to the exchange probability corresponding to the temperature (S14).
[0028] Next, it is determined whether the termination condition has been met (S15). If the termination condition has not been met (NO in S15), the process returns to S13. On the other hand, if the termination condition has been met (YES in S15), the process ends. The termination condition is generally the condition that the number of calculations has reached a predetermined threshold.
[0029] The final solution adopted is the process that minimizes the evaluation function value (hereinafter also referred to as "energy") at the end of the last step (S15). The overview of TPSA is as described above.
[0030] (0-2. Areas for improvement in TPSA) In TPSA, it is desirable that the solution search be performed at lower evaluation function values (i.e., lower energy) as the number of calculations for the input problem increases. However, if the temperature in each of the multiple processes is constant, the range of the solution search may not converge to lower evaluation function values unless the maximum temperature or the number of parallel processes is set appropriately.
[0031] Figure 2 shows an example of a temperature schedule in which the temperature of each of the multiple processes remains constant. In Figure 2, the horizontal axis represents the number of calculations (Step), the vertical axis represents the temperature, and the graph shows the relationship between the temperature and the number of calculations for each of the 10 processes. This graph shows an example in which initial temperatures T0 to T9 are set for the 10 processes, and a temperature schedule is set in which the temperature remains constant for each process even as the number of calculations increases.
[0032] Figure 3 shows the transitions (energy transitions) of the evaluation function values at each step, calculated according to the temperature schedule shown in Figure 2. Figure 3 shows the evaluation function values (Energy) at each step of the first process (Process 1), representing the 10 processes.
[0033] In the example shown in Figure 3, the objective is to bring the evaluation function value (Energy) corresponding to the solution closer to the target evaluation function value (Best Known Energy) indicated by the dashed line. However, Figure 3 confirms that in the first process, the evaluation function value (Energy) does not reach the low evaluation function value (low energy). Furthermore, although not shown in Figure 3, calculation results were obtained for the other nine processes as well, showing that the evaluation function values do not reach the low evaluation function value.
[0034] Therefore, embodiments of the present invention primarily propose techniques that enable more effective solution searching. More specifically, embodiments of the present invention propose techniques that enable solution searching with lower-evaluation function values as the number of calculations for an input problem increases.
[0035] The above explains the overview of TPSA and areas for improvement.
[0036] (1. Details of the Embodiment) Next, we will describe the details of embodiments of the present invention.
[0037] (1-1. Functional configuration of information processing equipment) First, an example of the functional configuration of an information processing device according to an embodiment of the present invention will be described. Figure 4 is a diagram showing an example of the functional configuration of an information processing device 1 according to an embodiment of the present invention. The information processing device 1 is implemented by a computer and, as shown in Figure 4, comprises an operation unit 110, a control unit 120, a storage unit 130, a communication unit 140, and a display unit 150.
[0038] (Operation unit 110) The control unit 110 has the function of receiving operations from the user. For example, the control unit 110 may receive input from the user regarding a problem and various parameters. These parameters may include temperature. For example, the control unit 110 may be implemented by a mouse or a touch panel.
[0039] (Control unit 120) The control unit 120 performs various calculation processes. Data necessary for calculations performed by the arithmetic unit is appropriately stored in the storage unit 130.
[0040] For example, the control unit 120 may include a CPU (Central Processing Unit) or other arithmetic unit (processor), and its function may be realized when a program stored in ROM (Read Only Memory) is loaded into RAM (Random Access Memory) by the arithmetic unit and executed. In this case, a computer-readable recording medium on which the program is stored may also be provided. Alternatively, these blocks may be composed of dedicated hardware or a combination of multiple hardware components.
[0041] (Storage unit 130) The memory unit 130 is a memory capable of storing programs and data for operating the control unit 120. The memory unit 130 can also temporarily store various data required during the operation of the control unit 120. For example, the memory unit 130 may be a non-volatile memory.
[0042] (Communications Section 140) The communication unit 140 is configured with a communication interface and communicates with other devices. For example, the communication unit 140 can receive various types of information from other devices. Furthermore, the communication unit 140 can transmit various types of information to other devices.
[0043] (Display section 150) The display unit 150 is implemented by a display and has the function of displaying various information to the user according to the control unit 120. For example, the display unit 150 can display the minimum evaluation function value determined by the control unit 120.
[0044] For example, the display unit 150 may be a liquid crystal display (LCD) device or an OLED (Organic Light Emitting Diode) device. Furthermore, the display unit 150 may be a head-mounted display worn on the user's head or a stationary display.
[0045] (1-2. Detailed functions of the control unit) Next, the detailed functions of the control unit 120 according to an embodiment of the present invention will be described. Figure 5 is a diagram showing an example of the configuration of the control unit 120 according to an embodiment of the present invention. As shown in Figure 5, the control unit 120 comprises processing units 121-0 to 121-9, an exchange unit 122, an update unit 123, and an execution control unit 124.
[0046] The control unit 120 generates an evaluation function for evaluating the solution based on modeling of the problem to be solved. The evaluation function may be a common evaluation function for processing units 121-0 to 121-9. The control unit 120 also generates an initial solution corresponding to each of the processing units 121-0 to 121-9. The initial solutions corresponding to each of the processing units 121-0 to 121-9 may be generated randomly or according to a predetermined rule.
[0047] (Processing step 121) Each of the processing units 121-0 to 121-9 can correspond to a process. That is, the control unit 120 has started 10 processes. However, the number of processes can be multiple and is not limited to 10. When explaining content that can apply to all of the processing units 121-0 to 121-9, the processing units 121-0 to 121-9 may be referred to simply as processing unit 121 without distinction.
[0048] Each of the processing units 121-0 to 121-9 has a temperature parameter, and each of these temperature parameters is set to a different initial temperature. In the following explanation, it is assumed that the values of the temperature parameters of each of the processing units 121-0 to 121-9 are positive values and that a greater-than / less-than relationship is maintained. Note that the temperature parameter is an example of a parameter. The initial temperature is the initial value of the parameter.
[0049] The temperature parameter of processing unit 121-k (where k is an integer satisfying 0 ≤ k ≤ 9) is set to an initial temperature Tk. For example, the temperature parameter of processing unit 121-1 is set to an initial temperature T1, the temperature parameter of processing unit 121-2 is set to an initial temperature T2, and the temperature parameter of processing unit 121-9 is set to an initial temperature T9.
[0050] The order of initial temperatures T0 to T9 is given as T0 > T1 > T2 > T3 > T4 > T5 > T6 > T7 > T8 > T9. When the initial temperatures T0 to T9 are arranged in a line in order of magnitude, two processing units 121 whose initial temperatures are adjacent to each other are also called adjacent processing units 121. For example, processing unit 121-0 with initial temperature T0 and processing unit 121-1 with initial temperature T1 are adjacent processing units 121. Also, processing unit 121-1 with initial temperature T1 and processing unit 121-2 with initial temperature T2 are adjacent processing units 121.
[0051] Furthermore, as described above, the relative magnitudes of the temperature parameter values of each of the processing units 121-0 to 121-9 are maintained. Therefore, even if the temperature parameter values of processing units 121-0 to 121-9 change from the initial temperatures T0 to T9, the temperature parameters of two adjacent processing units 121 remain adjacent to each other.
[0052] Each of the processing units 121-0 to 121-9 has an initial solution generated by initialization. Each of the processing units 121-0 to 121-9 obtains a selected solution by searching for a solution based on the initial solution it possesses.
[0053] More specifically, processing units 121-0 to 121-9 each generate multiple candidate solutions corresponding to their respective initial solutions, and calculate the output from the evaluation function when these generated candidate solutions are input to the evaluation function as the evaluation function value corresponding to each of the multiple candidate solutions. Then, each of the processing units 121-0 to 121-9 searches for a solution based on the evaluation function value corresponding to each of the multiple candidate solutions.
[0054] Processing units 121-0 to 121-9 may select a candidate solution from among several candidate solutions corresponding to the initial solution, the one that minimizes the evaluation function value.
[0055] (Exchange part 122) The exchange unit 122 performs a solution exchange process between at least two processing units 121 of processing units 121-0 to 121-9. In this process, the solution exchange process is performed between two adjacent processing units 121. Specifically, the solution exchange process is performed between two adjacent processing units 121 according to the exchange probability. The exchange probability is the probability that solutions will be exchanged. The solution exchange process can mean determining whether to perform a solution exchange according to the exchange probability and then performing processing based on the determination result. Therefore, the solution exchange process can include both cases where solutions are exchanged and cases where they are not.
[0056] In the following explanation, among two adjacent processing units 121, the processing unit 121 with a higher initial temperature parameter will also be referred to as the "high-temperature processing unit R(high)". On the other hand, among two adjacent processing units 121, the processing unit 121 with a lower initial temperature parameter will also be referred to as the "low-temperature processing unit R(low)". The high-temperature processing unit R(high) may correspond to the first processing unit. On the other hand, the low-temperature processing unit R(low) may correspond to the second processing unit.
[0057] In the following explanation, the temperature parameter of the high-temperature side processing unit R(high) will also be referred to as "temperature T(high)". Conversely, the temperature parameter of the low-temperature side processing unit R(low) will also be referred to as "temperature T(low)". Temperature T(high) may correspond to the first parameter. Conversely, temperature T(low) may correspond to the second parameter.
[0058] In the following explanation, the selected solution in the high-temperature side processing unit R(high) will also be referred to as "selected solution X(high)". Conversely, the selected solution in the low-temperature side processing unit R(low) will also be referred to as "selected solution X(low)". Selected solution X(high) may correspond to the first selected solution. Conversely, selected solution X(low) may correspond to the second selected solution.
[0059] In the following explanation, the evaluation function value corresponding to the selected solution X(high) will also be referred to as the "evaluation function value E(high)". Conversely, the evaluation function value corresponding to the selected solution X(low) will also be referred to as the "evaluation function value E(low)". The evaluation function value E(high) may correspond to the first selected evaluation function value. Conversely, the evaluation function value E(low) may correspond to the second selected evaluation function value.
[0060] An example of a method for setting the exchange probability is shown below. Specifically, the exchange unit 122 sets the exchange probability based on the evaluation function value E(high) and the evaluation function value E(low). In particular, the exchange unit 122 sets the exchange probability to 1 if the evaluation function value E(low) is smaller than the evaluation function value E(high).
[0061] On the other hand, if the evaluation function value E(low) is greater than the evaluation function value E(high), the exchange unit 122 sets the exchange probability to a value less than 1 according to a predetermined rule. For example, the predetermined rule may include a rule that the smaller the product of the temperature T(high) value and the temperature T(low) value, the smaller the exchange probability. The predetermined rule may also include a rule that the larger the difference between the evaluation function value E(low) and the evaluation function value E(high), the larger the exchange probability.
[0062] As an example, if the exchange probability is P, the difference in the values of the temperature parameter is ΔT = T(high) - T(low), and the difference in the evaluation function values is ΔE = E(high) - E(low), then the given rule can be expressed by the following equation (1).
[0063]
number
[0064] The exchange unit 122 performs a resolution exchange process between the high-temperature side processing unit R(high) and the low-temperature side processing unit R(low) according to the set exchange probability P.
[0065] (Updated part 123) The update unit 123 decreases the temperature parameter values of each of the processing units 121-0 to 121-9 after the solution exchange process has been performed. As the temperature parameter values decrease each time the solution exchange process is performed, the search in the solution space becomes wider, which makes it easier for the evaluation function value (Energy) to converge to a lower evaluation function value (lower energy), and thus the search for solutions becomes more effective.
[0066] Specifically, the update unit 123 may monotonically decrease the values of the temperature parameters of each of the processing units 121-0 to 121-9 after the solution exchange process has been performed, as the number of calculations increases. The number of calculations corresponds to the number of steps and may increase each time the solution exchange process is performed. For example, the update unit 123 may monotonically decrease the values of the temperature parameters of each of the processing units 121-0 to 121-9 after the solution exchange process has been performed, exponentially as the number of calculations increases, as shown in equation (2) below.
[0067]
number
[0068] In equation (2), T new This is the value of the temperature parameter after the decrease. initial is the initial temperature. α is the decay rate, and can take any value between 0 and α ≤ 1. α is mainly assumed to be a constant, but it can also be variable. i is the current number of calculations (current number of steps). step is the maximum number of calculations (maximum number of steps). int is a function that takes an integer argument.
[0069] Figure 6 shows an example of a temperature schedule in which the value of each temperature parameter of multiple processes decreases after the solution exchange process. In Figure 6, the horizontal axis is the number of calculations (Step), the vertical axis is the temperature, and the graph shows the relationship between the value of the temperature parameter and the number of calculations. As can be seen from this graph, in the embodiment of the present invention, a temperature schedule is set in which the value of the temperature parameter decreases from the initial temperature T0 to T9 as the number of calculations increases.
[0070] (Execution Control Unit 124) The execution control unit 124 controls the reprocessing, which includes a reacquisition process to reacquire the selected solution, a re-exchange process which is a second solution exchange process, and a re-update process which decreases the temperature value again, to be executed a predetermined number of times (hereinafter also referred to as the "number of re-executions"). Here, the number of re-executions does not have to be limited. For example, the number of re-executions can be one or more. Note that the number of re-executions is the number of times the solution exchange process is performed again, so it is one less than the number of solution exchange processes that may correspond to the number of calculations.
[0071] Specifically, the reacquisition process is a process in which processing units 121-0 to 121-9 each perform a search for a solution based on the selected solution after the solution exchange process has been performed, thereby reacquisition of the selected solution.
[0072] The re-exchange process is a process in which the exchange unit 122 resets the exchange probability based on the evaluation function value corresponding to the selected solution reacquired by the high-temperature side processing unit R(high) and the evaluation function value corresponding to the selected solution reacquired by the low-temperature side processing unit R(low), and performs a solution exchange process again according to the reset exchange probability.
[0073] The re-update process is a process in which the update unit 123 decreases the values of each temperature parameter of the processing units 121-0 to 121-9 again after the replacement unit 122 has performed the de-replacement process again.
[0074] After the reprocessing has been performed a certain number of times, the execution control unit 124 determines the minimum evaluation function value (lowest energy) from the evaluation function values corresponding to the selected solution reacquired by each of the processing units 121-0 to 121-9. The minimum evaluation function value may be displayed by the display unit 150, transmitted to another device by the communication unit 140, or stored in the storage unit 130.
[0075] The detailed functions of the control unit 120 according to the embodiment of the present invention have been described above.
[0076] (1-3. Operation of Information Processing Device 1) Next, with reference to Figure 7 (and Figures 1 to 6 as appropriate), an example of the operation of the information processing device 1 according to an embodiment of the present invention will be described.
[0077] Figure 7 is a flowchart showing an example of the operation of the information processing device 1 according to an embodiment of the present invention. As shown in Figure 7, in the embodiment of the present invention, the problem to be solved and various parameters necessary for calculation are input (S21). The input parameters include the initial temperature set for the temperature parameters corresponding to the processing units 121-0 to 121-9.
[0078] Next, initialization takes place (S22). During initialization, modeling is performed based on the problem to be solved. For example, an evaluation function for evaluating the solution may be generated through modeling. Also, during initialization, initial solutions corresponding to each of the processing units 121-0 to 121-9 are generated.
[0079] Next, each of the processing units 121-0 to 121-9 performs a solution search using the temperature parameter value (S23). Subsequently, at specific timings, a solution exchange process is performed between two adjacent processing units 121 according to the exchange probability corresponding to the temperature parameter value (S24).
[0080] Next, the values of the temperature parameters of the processing units 121-0 to 121-9 are reduced (S25). This reduction in temperature parameters makes the solution search more effective.
[0081] Next, it is determined whether the termination condition has been met (S26). If the termination condition has not been met (NO in S26), the process returns to S23. On the other hand, if the termination condition has been met (YES in S26), the process ends. The termination condition is generally the number of calculations reaching a predetermined threshold.
[0082] The final solution adopted is the one in processing unit 121 that minimizes the evaluation function value (energy) at the time the last step (S26) is completed.
[0083] The above describes an example of the operation of the information processing device 1 according to an embodiment of the present invention.
[0084] (1-4. Effects) As described above, according to the embodiment of the present invention, the value of the temperature parameter decreases each time the solution exchange process is performed. With this configuration, the search in the solution space is performed over a wider area, so the evaluation function value (Energy) is more likely to converge to a lower evaluation function value (lower energy), and the effect of more effectively searching for solutions can be enjoyed.
[0085] To verify these effects, the inventors of this application conducted numerical calculation experiments using an algorithm according to the flowchart shown in Figure 7 and a temperature schedule (Figure 6) shown in equation (2). The parameters used to calculate the temperature schedule are as follows: T0 is the initial temperature set for the temperature parameter of the processing unit 121-0. α is the decay rate of the temperature parameter value. The number of parallelisms is the number of processing units 121.
[0086] T0 = 1,000,000 α = 0.8 Step = 10000 Number of parallels=10 Deviation rate=0.75(Tk+1=0.75×Tk)
[0087] Figure 8 shows the results of a numerical calculation experiment. In Figure 8, the evaluation function value (Energy) at each step of processing unit 121-0 (Process 1) is shown, representing processing units 121-0 to 121-9.
[0088] In the example shown in Figure 8, as in the example shown in Figure 3, the objective is to bring the evaluation function value (Energy) corresponding to the solution closer to the target evaluation function value (Best Known Energy) shown by the dashed line. Unlike the example shown in Figure 3, in the example shown in Figure 8, it can be confirmed that the evaluation function value (Energy) converges to a lower evaluation function value (lower energy) in processing unit 121-0. Furthermore, although not shown in Figure 8, calculation results were obtained in processing units 121-1 to 121-9 that similarly showed the evaluation function value converging to a lower evaluation function value.
[0089] The effects of the information processing device 1 according to the embodiment of the present invention have been described above.
[0090] (2. Hardware Configuration Example) Next, an example of the hardware configuration of the information processing device 1 according to the present invention will be described.
[0091] In the following, an example of the hardware configuration of the information processing device 900 will be described as an example of the hardware configuration of the information processing device 1 according to an embodiment of the present invention. Note that the example of the hardware configuration of the information processing device 900 described below is merely one example of the hardware configuration of the information processing device 1. Therefore, the hardware configuration of the information processing device 1 may be modified by removing unnecessary components from the hardware configuration of the information processing device 900 described below, or new components may be added.
[0092] Figure 9 shows the hardware configuration of an information processing device 900 as an example of an information processing device 1 according to an embodiment of the present invention. The information processing device 900 includes a CPU (Central Processing Unit) 901, a ROM (Read Only Memory) 902, a RAM (Random Access Memory) 903, a host bus 904, a bridge 905, an external bus 906, an interface 907, an input device 908, an output device 909, a storage device 910, and a communication device 911.
[0093] The CPU 901 functions as both an arithmetic processing unit and a control unit, controlling the overall operation of the information processing unit 900 according to various programs. The CPU 901 may also be a microprocessor. The ROM 902 stores programs and arithmetic parameters used by the CPU 901. The RAM 903 temporarily stores programs used in the execution of the CPU 901 and parameters that change as needed during its execution. These are interconnected by a host bus 904, which consists of a CPU bus and other components.
[0094] The host bus 904 is connected to an external bus 906, such as a PCI (Peripheral Component Interconnect / Interface) bus, via a bridge 905. It is not always necessary to configure the host bus 904, bridge 905, and external bus 906 separately; these functions may be implemented on a single bus.
[0095] The input device 908 consists of input means for the user to input information, such as a mouse, keyboard, touch panel, buttons, microphone, switches, and levers, and an input control circuit that generates input signals based on the user's input and outputs them to the CPU 901. The user operating the information processing device 900 can input various types of data to the information processing device 900 or instruct it to perform processing operations by operating this input device 908.
[0096] The output device 909 includes, for example, display devices such as CRT (Cathode Ray Tube) display devices, liquid crystal display (LCD) devices, OLED (Organic Light Emitting Diode) devices, lamps, and audio output devices such as speakers.
[0097] The storage device 910 is a device for storing data. The storage device 910 may include a storage medium, a recording device for recording data on the storage medium, a reading device for reading data from the storage medium, and a deletion device for deleting data recorded on the storage medium. The storage device 910 is composed of, for example, an HDD (Hard Disk Drive). This storage device 910 drives the hard disk and stores programs executed by the CPU 901 and various data.
[0098] The communication device 911 is a communication interface composed of, for example, a communication device for connecting to a network. The communication device 911 may support either wireless or wired communication.
[0099] The hardware configuration example of the information processing device 1 according to the embodiment of the present invention has been described above.
[0100] (3. Summary) Although preferred embodiments of the present invention have been described in detail above with reference to the attached drawings, the present invention is not limited to these examples. It is clear to any person with ordinary skill in the art to which the present invention belongs that various modifications or alterations can be conceived within the scope of the technical idea described in the claims, and these are also understood to fall within the technical scope of the present invention.
[0101] For example, as described above, after the solution exchange process is performed, the update unit 123 monotonically decreases the values of the temperature parameters of each of the processing units 121-0 to 121-9 as the number of calculations increases. At this time, the update unit 123 may monotonically decrease the values of the temperature parameters using an exponential function as described above, but it may also monotonically decrease the values of the temperature parameters using other monotonically decreasing functions. For example, the update unit 123 may monotonically decrease the values of the temperature parameters using a linear function. [Explanation of Symbols]
[0102] 1. Information Processing Device 110 Operation section 120 Control Unit 121 Processing Unit 122 Exchange section 123 Update Department 124 Execution Control Unit 130 Storage section 140 Communications Department 150 Display section
Claims
1. Multiple processing units, each having parameters with different initial values and an initial solution, and each performing a solution search based on the initial solution to obtain a selected solution, An exchange unit that performs solution exchange processing related to the exchange of the selected solution between at least two of the plurality of processing units, An update unit that reduces the value of the parameter in each of the plurality of processing units after the aforementioned solution exchange process has been performed, An information processing device equipped with the following features.
2. The plurality of processing units each perform the search based on the evaluation function value corresponding to each of the plurality of candidate solutions that correspond to the initial solution they each possess. The information processing apparatus according to claim 1.
3. The exchange unit performs the solution exchange process according to the exchange probability. The information processing apparatus according to claim 2.
4. The two processing units include a first processing unit and a second processing unit, The selected solution in the first processing unit is the first selected solution, The selected solution in the second processing unit is the second selected solution, The evaluation function value corresponding to the first selection solution is the first selection evaluation function value, The evaluation function value corresponding to the second choice solution is the second choice evaluation function value, The exchange unit sets the exchange probability based on the first selection evaluation function value and the second selection evaluation function value. The information processing apparatus according to claim 3.
5. The parameters of the first processing unit are the first parameters, The parameters of the second processing unit are the second parameters, The initial value of the second parameter is smaller than the initial value of the first parameter. The exchange unit sets the exchange probability to a value less than 1 according to a predetermined rule if the second selection evaluation function value is greater than the first selection evaluation function value. The information processing apparatus according to claim 4.
6. The initial value of the first parameter is a positive value, The aforementioned predetermined rule includes a rule that reduces the exchange probability as the product of the value of the first parameter and the value of the second parameter becomes smaller. The information processing apparatus according to claim 5.
7. The initial value of the first parameter is a positive value, The aforementioned predetermined rule includes a rule that increases the exchange probability as the difference between the second selection evaluation function value and the first selection evaluation function value increases. The information processing apparatus according to claim 5.
8. The initial value of the second parameter is smaller than the initial value of the first parameter. The exchange unit sets the exchange probability to 1 if the second selection evaluation function value is smaller than the first selection evaluation function value. The information processing apparatus according to claim 5.
9. The aforementioned information processing device is The first processing unit and the second processing unit each perform a search for a solution based on the selected solution after the solution exchange process has been performed, thereby performing a reacquisition process to reacquire the selected solution. The exchange unit resets the exchange probability based on the evaluation function value corresponding to the selected solution reacquired by the first processing unit and the evaluation function value corresponding to the selected solution reacquired by the second processing unit, and performs a re-solution exchange process as a second solution exchange process according to the reset exchange probability. The update unit performs a re-update process in which, after the exchange unit has performed the re-exchange process, it further decreases the value of the first parameter and further decreases the value of the second parameter. The system includes an execution control unit that controls the execution of reprocessing, including the reprocessing process, a predetermined number of times. The information processing apparatus according to any one of claims 5 to 8.
10. The execution control unit, After the reprocessing is performed the predetermined number of times, the minimum evaluation function value is determined from the evaluation function values corresponding to the selected solution reacquired by the first processing unit and the second processing unit, respectively. The information processing apparatus according to claim 9.
11. Each of the first processing unit and the second processing unit obtains a candidate solution from the plurality of candidate solutions that takes the minimum value of the evaluation function as the selected solution. The information processing apparatus according to claim 10.
12. The two processing units are adjacent processing units of the plurality of processing units. The information processing apparatus according to claim 1.
13. A solution exchange process is performed between at least two of a plurality of processing units, each having parameters with different initial values set, each having an initial solution, and each performing a solution search based on the initial solution to obtain a selected solution, regarding the exchange of the selected solution. After the aforementioned solution exchange process is performed, the value of the parameter possessed by each of the plurality of processing units is reduced, A computer-based information processing method, including [a specific example].
14. Computers, Multiple processing units, each having parameters with different initial values and an initial solution, and each performing a solution search based on the initial solution to obtain a selected solution, An exchange unit that performs solution exchange processing related to the exchange of the selected solution between at least two of the plurality of processing units, An update unit that reduces the value of the parameter in each of the plurality of processing units after the aforementioned solution exchange process has been performed, A program that makes it function as such.
