Optimization apparatus, optimization method, and program

Abstract

To reduce calculation time.SOLUTION: An optimization apparatus comprises: a first calculation unit for calculating a value of a first decision variable that satisfies a first predetermined constraint condition and minimizes the value of a first objective function, where the first decision variable is a variable that represents a combination of a plurality of devices selected from a plurality of types of devices to be selected, the first objective function is a function that calculates the total value of an introduction cost and an operation cost of the selected combination; and a second calculation unit for calculating a value of a second decision variable that minimizes a second objective function as a mathematical optimization problem, where the second decision variable is a variable that represents an operating state of each device included in the combination represented by the first decision variable, the second objective function is a function that formulates a second constraint condition, which is a constraint condition regarding the operation of each device, and calculates the operation costs. The first calculation unit calculates a value of the first objective function by taking the value of the second objective function when the second calculation unit calculates the optimal solution for the second decision variable as the operation cost.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present disclosure relates to an optimization device, an optimization method, and a program. [Background technology]

[0002] Patent Document 1 describes the following combination solution determination system. That is, in the combination solution determination system described in Patent Document 1, a solution candidate generation unit first generates equipment combinations and time series data. The time series data is data such as a target time (24 hours / day x 365 days). The essential series data extraction unit has a function of extracting data necessary for simulation calculation from the time series data, and extracts data (essential series data) necessary for the time series data to verify whether or not the constraint conditions are met. Using this essential series data and equipment combinations, the simulation unit calculates a simulation result (power consumption value) at each time based on the essential series data that is less than the time series data. The evaluation index calculation unit calculates an evaluation value based on the simulation result (power consumption) calculated by the simulation unit. The solution determination unit determines the one with the best evaluation index as a solution from the multiple equipment combinations obtained. According to the combination solution determination system described in Patent Document 1, it is said that the calculation cost required for simulation can be reduced, thereby reducing calculation time and efficiently deriving a solution to a black-box optimization problem. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Patent Publication No. 2021-12479 Summary of the Invention [Problem to be solved by the invention]

[0004] However, although the amount of calculations is reduced in the combination solution determination system described in Patent Document 1, the simulation results (power consumption values) are still calculated by simulation, which poses a problem that the calculation time is likely to be long.

[0005] The present disclosure has been made to solve the above-mentioned problems, and aims to provide an optimization device, an optimization method, and a program that can reduce calculation time. [Means for solving the problem]

[0006] In order to solve the above problem, an optimization device according to the present disclosure includes a first calculation unit that sets a variable representing a combination of multiple devices selected from multiple types of devices to be selected as a first decision variable, sets a function calculating a total value of an introduction cost and an operation cost of the selected combination as a first objective function, and calculates a value of the first decision variable that satisfies a predetermined first constraint condition and minimizes a value of the first objective function; and a second calculation unit that sets a variable representing an operating state of each of the devices included in the combination represented by the first decision variable as a second decision variable, formulates a second constraint condition that is a constraint condition on the operation of each of the devices, sets a function calculating the operation cost as a second objective function, and calculates a value of the second decision variable that minimizes the second objective function as a mathematical optimization problem, and the first calculation unit calculates the value of the first objective function by setting the value of the second objective function at the time when the second calculation unit calculates an optimal solution of the second decision variable as the operation cost.

[0007] The optimization method according to the present disclosure includes a first step of calculating a value of the first decision variable that satisfies a predetermined first constraint condition and minimizes a value of the first objective function, where a variable representing an operating state of each of the devices included in the combination represented by the first decision variable is a second decision variable, a second constraint condition that is a constraint condition on the operation of each of the devices is formulated, a function for calculating the operating cost is a second objective function, and a value of the second decision variable that minimizes the second objective function as a mathematical optimization problem, where in the first step, the value of the second objective function when an optimal solution of the second decision variable is calculated in the second step is set as the operating cost and the value of the first objective function is calculated.

[0008] A program according to the present disclosure is a program for causing a computer to execute the following steps: a first step of calculating a value of the first decision variable that satisfies a predetermined first constraint condition and minimizes a value of the first objective function, where a variable representing an operating state of each of the devices included in the combination represented by the first decision variable is a second decision variable, a second constraint condition that is a constraint condition on the operation of each of the devices is formulated, and a function for calculating the operating cost is a second objective function, where a value of the second decision variable that minimizes the second objective function as a mathematical optimization problem is calculated, where in the first step, the value of the second objective function when an optimal solution for the second decision variable is calculated in the second step is set as the operating cost and the value of the first objective function is calculated. Effect of the Invention

[0009] According to the optimization device, optimization method, and program of the present disclosure, calculation time can be reduced. [Brief description of the drawings]

[0010] [Figure 1] FIG. 1 is a block diagram showing an example configuration of an optimization device according to an embodiment of the present disclosure. [Diagram 2] FIG. 2 is a schematic diagram for explaining an example of input data according to an embodiment of the present disclosure. [Diagram 3] FIG. 2 is a schematic diagram for explaining an example of input data according to an embodiment of the present disclosure. [Figure 4] FIG. 2 is a schematic diagram for explaining an example of input data according to an embodiment of the present disclosure. [Diagram 5] FIG. 2 is a schematic diagram for explaining an example of input data according to an embodiment of the present disclosure. [Figure 6] FIG. 4 is a schematic diagram for explaining a first calculation unit according to an embodiment of the present disclosure. [Figure 7] FIG. 11 is a schematic diagram for explaining a backup machine constraint according to an embodiment of the present disclosure. [Figure 8] FIG. 11 is a schematic diagram for explaining a backup machine constraint according to an embodiment of the present disclosure. [Figure 9] FIG. 11 is a schematic diagram for explaining a backup machine constraint according to an embodiment of the present disclosure. [Figure 10] 6 is a schematic diagram for explaining a second calculator according to an embodiment of the present disclosure. FIG. [Figure 11] FIG. 13 is a schematic diagram for explaining a calculation example of power consumption according to an embodiment of the present disclosure. [Figure 12] FIG. 13 is a schematic diagram for explaining a calculation example of power consumption according to an embodiment of the present disclosure. [Figure 13] FIG. 2 is a schematic diagram for explaining an example of the operation of the optimization device according to an embodiment of the present disclosure. [Figure 14] 1 is a flowchart illustrating an example of the operation of an optimization device according to an embodiment of the present disclosure. [Figure 15] FIG. 13 is a schematic diagram for explaining a modified example of an optimization device according to an embodiment of the present disclosure. [Figure 16] FIG. 13 is a schematic diagram for explaining a modified example of an optimization device according to an embodiment of the present disclosure. [Figure 17] FIG. 1 is a schematic block diagram illustrating a configuration of a computer according to at least one embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0011] Hereinafter, an optimization device, an optimization method, and a program according to an embodiment of the present disclosure will be described with reference to Figs. 1 to 17. In each figure, the same or corresponding configurations are designated by the same reference numerals, and the description is omitted as appropriate. In addition, the optimization device according to an embodiment of the present disclosure can be used, for example, in the design stage, when designing, examining, analyzing, etc., an optimal combination of multiple devices to be introduced, for example, into a specific facility or facility from among multiple types of devices to be selected (selected). In the following example, as a combination of multiple types of devices, a combination of multiple types of heat source machines that operate by consuming electricity as energy is illustrated, but the devices and combinations of devices are not limited to heat source machines and combinations of heat source machines.

[0012] (term) First, the main terms used in the embodiments of the present disclosure (or the terms used to explain the main terms) will be described. A "mathematical optimization problem" (or a "mathematical programming problem") is a problem of mathematically finding an optimal one for an objective from among candidates that satisfy conditions, or a problem of finding a decision variable that maximizes or minimizes an objective function after setting a constraint equation that represents a constraint condition. The "elements" of a mathematical optimization problem include a "decision variable," an "objective function," and a "constraint condition." A "decision variable" is a variable that represents a number, amount, etc. to be determined in a mathematical optimization problem (hereinafter, a decision variable is also simply referred to as a variable). An "objective function" is a function that represents a measure of whether a decision variable is good or bad for an objective, a function with a decision variable as an independent variable, and a function that is to be maximized (or minimized) in a mathematical optimization problem. Hereinafter, the value of the objective function for a certain decision variable is referred to as an objective function value or an evaluation function value. A "constraint condition" is a condition that a candidate decision variable must satisfy, or a condition that gives a range that the decision variable can take. Note that mathematical optimization problems may not include constraints, or the constraints may not be expressed by constraint equations. An "optimal solution" is the value of the decision variables that achieves the best objective function value among all feasible solutions. A "feasible solution" is the value of the decision variables that satisfies all of the constraint equations. An "optimal value" is the objective function value of the optimal solution. "Optimization" refers to finding the optimal solution (or quasi-optimal solution) of a mathematical optimization problem.

[0013] A "linear programming problem" is an optimization problem in which the objective function and constraints are all expressed as linear expressions (first-order expressions), and is a mathematical programming problem in which the objective function is a linear function and all constraint conditions are expressed as equations or inequalities of linear functions. A "mixed-integer programming problem" (or "mixed-integer linear programming problem") is a mathematical programming problem in which the objective function and constraints are expressed as linear expressions (linear expressions), and the decision variables include integer and real values. "Formulation" means to express relationships and conditions using mathematical expressions.

[0014] "Black-Box Optimization (also known as BBO)" is an optimization problem in which the objective function in the optimization problem can be obtained for a certain decision variable, but analytical information (such as gradient or Hesse matrix) or properties (such as convexity or inter-variable dependency) of the objective function cannot be obtained. Solutions to black-box optimization algorithms include direct search methods such as genetic algorithms, and sequential approximation optimization that generates a surrogate model based on a probabilistic model such as the Bayesian optimization algorithm.

[0015] (Configuration of the optimization device) Fig. 1 is a block diagram showing a configuration example of an optimization device according to an embodiment of the present disclosure. The optimization device 1 shown in Fig. 1 can be configured using a computer such as a personal computer, and includes an input unit 11, a first calculation unit 12, a second calculation unit 13, and an output unit 14 as a functional configuration configured by a combination of hardware such as a computer and peripheral devices of the computer, and software such as a program executed by the computer.

[0016] The input unit 11 inputs input data to be input to the first calculation unit 12 and the second calculation unit 13, setting data related to processing by the first calculation unit 12 and the second calculation unit 13, etc., according to, for example, a user's instruction operation. FIGS. 2 to 5 are schematic diagrams for explaining examples of input data input by the input unit 11 according to the embodiment of the present disclosure. The input data 21 shown in FIG. 2 is an example of data representing the relationship between the model, the equipment cost (introduction cost), and the cooling capacity. The input data 22 shown in FIG. 3 is an example of data representing a pattern of load, cooling water, and air temperature conditions. The horizontal axis is the time step, and the vertical axis is the required load and temperature. The time step is 288 steps for 12 months, each hour. The input data 22 represents the expected values ​​of the required load (solid line), dry bulb temperature (chain line), and wet bulb temperature (dashed line) for each time step. The input data 23 shown in FIG. 4 is an example of required load data. The horizontal axis is time, and the vertical axis is the required load. The input data 22 represents an estimated value of the required load for each month in one-hour units. The input data 24 shown in Fig. 5 is an example of data representing equipment characteristics. The input data 24 shown in Fig. 5 represents the corresponding relationship between the load rate, the cooling water inlet temperature, and the power consumption. The input unit 11 inputs, for example, the input data 21 to 24 shown in Figs. 2 to 5, and stores them in a predetermined storage device.

[0017] The first calculation unit 12 calculates the value of the first decision variable that satisfies a first constraint condition and minimizes the value of the first objective function, using a variable representing a combination of multiple devices selected from multiple types of devices to be selected as a first decision variable and a function that calculates the total value of the introduction cost and the operation cost of the selected combination as a first objective function. At this time, the first calculation unit 12 calculates the value of the first objective function by using the value of the second objective function when the second calculation unit 13 calculates the optimal solution of the second decision variable as the operation cost. Note that the terms "first" and "second" are terms for distinguishing between decision variables, constraint conditions, and objective functions with different contents, and there is no difference in the meaning of the above-mentioned terms. Note that the introduction cost is the cost (expense) required for introducing the device, and includes the cost of the device to be introduced and the cost required for installing the device. Note that the operation cost is the cost required for operating the device over a predetermined period of time, and includes, for example, the annual load (output) and the cost of the device power consumption corresponding to the temperature pattern. Note that in this embodiment, the load may be read as output.

[0018] The second calculation unit 13 sets a variable representing the operating state of each device included in the combination represented by the first decision variable as a second decision variable, formulates a second constraint condition which is a constraint condition on the operation of each device, sets a function for calculating the operating cost as a second objective function, and calculates the value of the second decision variable that minimizes the second objective function as a mathematical optimization problem.

[0019] The output unit 14 outputs the calculation results of the first calculation unit 12, the calculation results of the second calculation unit 13, etc. to a predetermined output device such as a display device.

[0020] (First calculation section) FIG. 6 is a schematic diagram for explaining the first calculation unit 12 according to an embodiment of the present disclosure. In this embodiment, the first calculation unit 12 solves the optimization of the equipment combination (equipment configuration) as an arrangement planning problem 101 using a black box optimization algorithm, for example, as shown in FIG. 6. In the example shown in FIG. 6, the input / output information 102 has the operation cost calculated by the second calculation unit 13 as input information and the equipment configuration calculated by the first calculation unit 12 as output information. The first decision variable 103 in the arrangement planning problem 101 represents, for example, the number of devices to be installed for each model. The first decision variable 103 represents, for example, information such as 3 devices a, 0 devices b, and 2 devices c, as shown in an example 106.

[0021] In this case, if there are N_j (the character after "_" indicates a subscript) types of models to be combined (selected), the first decision variable 103 can be defined as the variable (decision variable) ω in formula (1). Note that the bold ω in the formula represents a vector (or array).

[0022]

number

[0023] Here, each element ω_j (j=0, ..., N_j) of the decision variable ω represents the number of installed devices j (devices of type j). The element ω_j can be treated as an integer variable of 0 or more, as shown in equation (2). In equation (2), "Z" is a set of integers of 0 or more.

[0024]

number

[0025] The first objective function 104 can be the sum of the introduction cost and the operation cost of each device in the combination of devices represented by the first decision variable 103. In the layout planning problem 101, the first decision variable 103 is optimized so that the first objective function 104 is minimized.

[0026] The first objective function 104 can be defined as, for example, a function F(ω) of formula (3). In formula (3), the variable C_j represents the introduction cost (installation cost) per device j. The function E(ω) represents the total operation cost of each device in a predetermined period when multiple devices are introduced in a combination represented by the decision variable ω. The value of the function E(ω) is calculated by the second calculation unit 13.

[0027]

number

[0028] The first constraint condition 105 in the allocation planning problem 101 may include, for example, a backup machine constraint, a maximum number of installed units constraint, a required load constraint, and the like. The backup machine constraint is a constraint that the combination of equipment is operable even if one of the models of one or more installed units fails. In this case, the backup machine constraint is a constraint condition that even if at least one device included in the selected combination fails, operation can be continued by other devices. Here, a specific example of the backup machine constraint will be described with reference to Figs. 7 to 9. Figs. 7 to 9 are schematic diagrams for explaining the backup machine constraint according to the embodiment of the present disclosure. Fig. 7 shows an example of a required load. The maximum load shown in Fig. 7 is 5,068 kW. Fig. 8 shows an example of an equipment configuration in which the first constraint condition 105 does not include the backup machine constraint, as a comparative example. In the example shown in Fig. 8, the equipment configuration includes only one device a with a maximum load (maximum output) of 5,415 kW. Also, Fig. 9 shows an example of an equipment configuration in which the first constraint condition 105 includes the backup machine constraint. In the example shown in Figure 9, the equipment configuration includes two units of equipment b with a maximum load (maximum output) of 3,235 kW and one unit of equipment c with a maximum load (maximum output) of 3,235 kW. The equipment configuration shown in Figure 9 satisfies the backup unit constraint, and can meet the required load (maximum load is 5,068 kW) even if one of the units fails. The backup unit constraint is a nonlinear constraint that is difficult to linearize. In contrast, when a black-box optimization algorithm is used as shown in Figure 6, the backup unit constraint can be easily taken into account. Note that the maximum number of units to be installed constraint is a constraint on the maximum number of units of equipment to be installed, and the required load constraint is a constraint on the total required load.

[0029] As shown in FIG. 6, the first calculation unit 12 uses, for example, a black-box optimization algorithm to determine a variable representing a combination of multiple devices selected from multiple types of devices to be selected as first decision variables 103, a function for calculating the total value of the introduction cost and the operation cost of the selected combination as first objective function 104, and calculates a value of the first decision variables 103 that satisfies a predetermined first constraint condition 105 and minimizes the value of the first objective function 104.

[0030] (Second calculation section) FIG. 10 is a schematic diagram for explaining the second calculation unit 13 according to the embodiment of the present disclosure. In this embodiment, the second calculation unit 13 solves the optimization of an operation plan (load sharing among devices, operation stop timing, etc.) as an operation plan problem 201 as a mathematical optimization problem, as shown in FIG. 10, for example. In the example shown in FIG. 10, the input / output information 202 has the device configuration (first decision variable 103) calculated by the first calculation unit 12 as input information, and the optimal value of the second objective function 204 calculated by the second calculation unit 13 as output information. The second decision variable 203 in the operation plan problem 201 includes a plurality of variables representing the number of operating units, the load factor, the chilled water / cooling water flow rate, the number of started / stopped units, etc.

[0031] In the operation planning problem 201, the second decision variable 203, which is a variable that determines the operation method of the equipment, can be expressed by a variable that determines the start / stop of the equipment until time τ and a variable that determines the load of the equipment. In this embodiment, the equipment is classified into a main engine and an auxiliary machine (or an auxiliary device) that is used in conjunction with the main engine, and the second constraint condition 205 applied to the main engine and the auxiliary device can be made different. The auxiliary machine (auxiliary device) is a pump connected to the main engine, a cooling tower, etc. In this example, a case where A auxiliary machines (auxiliary devices) are linked to one main engine is considered. The start / stop can be treated as an integer variable representing the start and stop of the equipment, and the load of the equipment can be treated as a continuous variable equal to or greater than 0.

[0032] The second decision variable 203 can be defined as, for example, the variables (decision variables) of equations (4) to (7). Equation (4) shows a variable representing start / stop (0 or 1) of main engine j at time t. Equation (5) shows a variable representing the load (0 to 100%) of main engine j at time t. Equation (6) shows a variable representing start / stop (0 or 1) of auxiliary engine a at time t. Equation (7) shows a variable representing the load (0 to 100%) of auxiliary engine a at time t. In addition, in equations (5) and (7), "R" is a set of real numbers greater than or equal to 0.

[0033]

number

[0034]

number

[0035]

number

[0036]

number

[0037] It is also assumed that constraint conditions (second constraint conditions 205) to be satisfied at each time for the operation of the equipment are determined in advance. Representative second constraint conditions 205 include, for example, that the installed and operating equipment exhibits the required performance, and that the availability rate of the equipment is within an available range. The second constraint conditions 205 shown in Fig. 10 include a time series constraint, a required load constraint, a load rate upper and lower limit constraint, and a flow rate upper and lower limit constraint. The time series constraint is a constraint related to a time series, such as a constraint that a device (e.g., a heat source machine) must continue to operate for a certain period of time after startup.

[0038] The second constraint condition 205 can be expressed as the following formula (8) using the variables related to the operation shown in formulas (4) to (7).

[0039]

number

[0040] The function g representing the second constraint 205 takes a value of 0 when each element satisfies all the constraints at time t, and takes a value of -1 when even one constraint is not satisfied. Therefore, when each element satisfies all the constraints at each time, the function g takes a value of 0 or more.

[0041] Next, an example of formulating the second constraint condition 205 will be described. In this embodiment, the operation method is formulated as a mathematical program. The annual heat load and temperature of the facility are given in advance, and an operation method that minimizes power consumption is obtained by optimization based on that information.

[0042] The required load constraint is a constraint condition that requires the required load to be satisfied. The required load is determined at each time, and the required load at time t is D_t. The total load of each device at that time must match or be greater than the required load D_t. The constraint equation is written in the following equation (9).

[0043]

number

[0044] The upper and lower load factor constraints are constraints on the range of the load. For equipment j, the minimum and maximum loads are assumed to be ρ_j^low, ρ_j^up (the character after "^" indicates a superscript). In this case, the load ρ_(j,t) of equipment j can be limited to within a range by considering the following constraint equation.

[0045]

number

[0046] The following constraints can be set for the auxiliaries: The following constraints can be considered for the load ρ_(j,t) of the main engine j, and the corresponding start / stop η_(a,j,t) and load θ_(a,j,t) of the auxiliary a.

[0047] A constraint that starts the related auxiliary only when the main engine j is running ⇒ The start and stop of the main engine and the start and stop of the auxiliary can be constrained to match.

[0048]

number

[0049] · Constraint to change the load on the auxiliary engine depending on the load on the main engine j ⇒ If the relationship can be described linearly, it can be expressed as follows using the coefficient c.

[0050]

number

[0051] Next, a description will be given of the second objective function 204. The second objective function 204 (function E(ω)) is shown as equation (13).

[0052]

number

[0053] In formula (13), f_j is a function that calculates the power consumption of the main unit j, h_(a,j) is a function that calculates the power consumption of the auxiliary unit a of the main unit j, and k is a coefficient that converts the power consumption into an electricity fee. The second objective function 204 is the total value of the power consumption costs of the main unit and its auxiliary units. This power consumption value becomes the function E(ω) in optimizing the equipment configuration in the first calculation unit 12.

[0054] A function f_j for calculating the power consumption of a main unit j is usually a nonlinear and time-varying system. Since it is difficult to express the system itself in a mathematical formula, the power consumption is calculated using a simulation in the background art. In contrast, in this embodiment, the function f_j is approximately expressed in a mathematical formula. Specifically, the function f_j is a function for calculating the power consumption from the load (load factor) of the device and the surrounding outdoor air temperature. The outdoor air temperature is given in advance as an optimization condition. The relationship between the air temperature, the load factor, and the power consumption is measured discretely and given in advance (FIG. 11). FIG. 11 is a schematic diagram for explaining an example of calculation of power consumption according to an embodiment of the present disclosure.

[0055] The function f_j is used in the optimization calculation by interpolating the discrete values ​​of the temperature, load factor, and power consumption, treating them as a continuous function. In this embodiment, in order to improve the calculation speed, linear interpolation is performed to calculate as a mixed integer linear problem, but it may be supplemented by quadratic approximation and formulated as a mixed integer quadratic programming problem. Although the calculation cost tends to be high, it is also possible to perform nonlinear approximation to make it a mixed integer nonlinear problem.

[0056] These problems formulated by the equations (4) to (13) and solved by the second calculation unit 13 can be treated as mixed integer mathematical programming problems. In particular, when the functions g, f_j, and h_(a,j) are all described in linear form, the optimization problem for determining an operation method can be formulated as a mixed integer linear optimization problem.

[0057] As shown in FIG. 10, the second calculation unit 13 sets a variable representing the operating state of each device included in the combination represented by the first decision variable 103 (FIG. 6) as a second decision variable 203, formulates a second constraint condition 205 which is a constraint condition on the operation of each device, sets a function for calculating the operating cost as a second objective function 204, and calculates the value of the second decision variable 203 that minimizes the second objective function 204 as a mathematical optimization problem.

[0058] 12 is a schematic diagram for explaining a calculation example of power consumption according to an embodiment of the present disclosure. For example, as shown in FIG. 12, the second calculation unit 13 can determine that the operation of only device B is optimal for the required load RL1, and that the operation of a combination of devices A and B is optimal for the required load RL2.

[0059] (Isolation of the problem between the first and second calculation parts) In this embodiment, as shown in FIG. 13, for example, by dividing the optimization problem into two, an arrangement planning problem (e.g., a black-box optimization algorithm using a solution method based on a genetic algorithm) 101 and an operation planning problem (e.g., a mathematical optimization problem solved by linear programming) 201, it becomes possible to quickly solve an equipment configuration that minimizes the sum of the power consumption cost corresponding to the operation that minimizes the power consumption of the required load and the introduction cost of the equipment while considering the constraints of the nonlinear equipment configuration. Note that the evaluation function value 204d (optimum value of the second objective function 204) used in the arrangement planning problem 101 is calculated when solving the operation planning problem 201. In addition, it becomes possible to obtain a strictly optimal solution by solving the operation method by mathematical programming. Therefore, it becomes possible to quickly solve the equipment configuration optimization problem that considers the required load, which is a large-scale mixed integer problem, by dividing the problem and performing optimization by a method suitable for each problem. FIG. 13 is a schematic diagram for explaining an example of the operation of the optimization device according to an embodiment of the present disclosure.

[0060] (Example of the optimization device in action) Fig. 14 is a flowchart showing an example of the operation of the optimization device according to the embodiment of the present disclosure. In the example of the operation shown in Fig. 14, first, the input unit 11 inputs data to be used in the optimization process (step S1). Next, the first calculation unit 12 sets an initial value to the first decision variable 103 (step S2). Next, the second calculation unit 13 calculates the second decision variable 203 that minimizes the second objective function 204 as a mathematical optimization problem (step S4). Next, the first calculation unit 12 calculates the value of the first decision variable 103 that minimizes the first objective function 104 only once (step S4). Next, the first calculation unit 12 determines whether the termination conditions (whether the upper limit of the number of calculation repetitions has been reached, whether the first objective function 104 has become equal to or lower than a predetermined threshold, etc.) are satisfied (step S5). If the conditions are not satisfied (step S5: NO), the process returns to step S3. If the conditions are satisfied (step S5: YES), the output unit 14 outputs the calculation results such as the optimal solution and the optimal value, and the process shown in FIG. 14 is terminated.

[0061] (Effects of this embodiment) In this embodiment, by using black-box optimization for the equipment configuration, it is possible to take into account nonlinear constraints on the placement of equipment. In addition, by applying mathematical programming to the operation method, it is possible to quickly obtain an exact solution. In addition, it is not dependent on simulation time. In addition, by dividing the problem into parts, it is possible to quickly obtain a solution even for large-scale problems.

[0062] That is, according to this embodiment, it is possible to quickly solve a large-scale mixed integer problem, which is an equipment configuration optimization problem that takes into account the required load, by dividing the problem into parts and performing optimization using a method suitable for each part.

[0063] In addition, since black-box optimization, which can take nonlinear constraints into account, is used for optimizing the equipment configuration, it is possible to take nonlinear constraints into account.

[0064] Let us consider a method of solving equipment configuration optimization and operation method optimization as one mathematical optimization problem. When nonlinear backup machine constraints are taken into account, the problem to be handled becomes a mixed integer nonlinear optimization problem. As the scale of the problem increases, the mixed integer nonlinear optimization requires longer calculation times or becomes more difficult to solve. In this embodiment, the problem regarding a large-scale operation method can be solved as a mixed integer linear problem, while considering nonlinear constraints only for the equipment configuration. Therefore, optimization that considers nonlinear constraints while maintaining the overall calculation speed is possible.

[0065] The backup machine constraint given as an example will be described. The backup machine constraint assumes the failure of a single device. In this embodiment, since black-box optimization is used for the problem of determining the device configuration, the backup machine constraint can be taken into consideration.

[0066] Furthermore, according to this embodiment, restrictions on the operation of devices and auxiliaries can be taken into consideration.

[0067] (Modification 1 of the above embodiment) As a variation of the above embodiment, learning can be performed using past optimization results, and an optimal combination of equipment and an operation method can be derived from the required load and other conditions. In black box optimization and mathematical optimization, optimization calculations are started based on an initial solution, and the closer this initial solution is to the optimal solution, the faster the calculations can be completed. Therefore, by using the solution obtained by the learning model as the initial solution, it is possible to complete the optimization calculations quickly.

[0068] In the above embodiment, since optimization is performed by dividing the problem into the device configuration and the operation, it is possible to learn each problem individually. First, consider learning the device configuration. At this time, there is no need to specifically obtain the operation method, and learning is performed to output only the device configuration. Specifically, learning is performed to output the device configuration from device information such as the required load and the maximum and minimum load settings of the device. Next, learning is performed regarding the operation method. In this learning, the device operation method is learned from the device configuration, the required load, and other conditions. Since the operation method is learned with the device configuration given, it is possible to calculate a reasonable operation method. In this way, by learning the optimization result and using it as an initial solution, it is possible to perform efficient optimization calculation.

[0069] According to the first modification, the learning result is used to use an initial solution close to the optimal solution, thereby making it possible to reduce the calculation time. Note that the learning result may be applied to only one of the optimization of the combination of devices and the optimization of the operation method.

[0070] (Modification 2 of the above embodiment) 15 and 16 are schematic diagrams for explaining a modified example of an optimization device according to an embodiment of the present disclosure. In the above embodiment, as shown in FIG. 15, the optimization device 1 according to the above embodiment is configured to perform optimization of an operation method for each equipment configuration that appears as a calculation process by mathematical optimization inside the black box optimization of the equipment configuration. As shown in FIG. 15, in the optimization device 1 according to the above embodiment, in the black box optimization 101, for example, a solution search 401 is performed in a search space 402 of a genetic algorithm, and in the mathematical optimization (linear programming) 201, a solution search is performed as a subproblem in a search space 403 of the mathematical optimization, and as a result of the search, an optimal solution is calculated in the space 404 of the original problem, and at that time, an evaluation function value 204d, which is an optimal value of the second objective function, is calculated. Then, the search 401 is performed again using the evaluation function value 204d.

[0071] On the other hand, in the second modification, as shown in Fig. 16, the calculation method for optimizing the equipment configuration is to calculate multiple equipment configurations at once, and to proceed with the search for black-box optimization from the multiple results. With this method, it becomes possible to execute the optimization calculation of each operation method for multiple equipment configurations calculated simultaneously in parallel.

[0072] According to this modification, the calculation speed can be improved by parallelizing the calculation.

[0073] (Action and effect) The optimization device of the above configuration includes a first calculation unit 12 that calculates a value of the first decision variable that satisfies a predetermined first constraint condition and minimizes the value of the first objective function, a variable that represents the operation state of each device included in the combination represented by the first decision variable as a second decision variable, a second constraint condition that is a constraint condition regarding the operation of each device is formulated, a function that calculates the operation cost is the second objective function, and a second calculation unit 13 that calculates a value of the second decision variable that minimizes the second objective function as a mathematical optimization problem. The first calculation unit 12 calculates the value of the first objective function by taking the value of the second objective function when the second calculation unit 13 calculates the optimal solution of the second decision variable as the operation cost. According to the optimization device, optimization method, and program disclosed herein, simulation time is not consumed, so that calculation time can be reduced.

[0074] (Other embodiments) Although the embodiments of the present disclosure have been described in detail above with reference to the drawings, the specific configuration is not limited to this embodiment, and design changes and the like that do not depart from the gist of the present disclosure are also included.

[0075] Computer Configuration FIG. 17 is a schematic block diagram illustrating a configuration of a computer according to at least one embodiment. The computer 90 comprises a processor 91, a main memory 92, a storage 93, and an interface 94. The optimization device 1 described above is implemented in a computer 90. The operations of the above-mentioned processing units are stored in the form of a program in a storage 93. A processor 91 reads the program from the storage 93, loads it in a main memory 92, and executes the above-mentioned processing in accordance with the program. The processor 91 also secures storage areas in the main memory 92 corresponding to the above-mentioned storage units in accordance with the program.

[0076] The program may be for realizing a part of the functions to be performed by the computer 90. For example, the program may be for realizing the functions by combining with other programs already stored in the storage or by combining with other programs implemented in other devices. In another embodiment, the computer may include a custom LSI (Large Scale Integrated Circuit) such as a PLD (Programmable Logic Device) in addition to or instead of the above configuration. Examples of PLDs include PAL (Programmable Array Logic), GAL (Generic Array Logic), CPLD (Complex Programmable Logic Device), and FPGA (Field Programmable Gate Array). In this case, some or all of the functions to be realized by the processor may be realized by the integrated circuit.

[0077] Examples of the storage 93 include a hard disk drive (HDD), a solid state drive (SSD), a magnetic disk, a magneto-optical disk, a compact disc read only memory (CD-ROM), a digital versatile disc read only memory (DVD-ROM), and a semiconductor memory. The storage 93 may be an internal medium directly connected to the bus of the computer 90, or an external medium connected to the computer 90 via an interface 94 or a communication line. In addition, when the program is distributed to the computer 90 via a communication line, the computer 90 that receives the program may load the program into the main memory 92 and execute the above-mentioned process. In at least one embodiment, the storage 93 is a non-transitory tangible storage medium.

[0078] <Additional Notes> The optimization device 1 according to the above embodiment or its modified example can be understood, for example, as follows.

[0079] (1) An optimization device 1 according to a first aspect includes a first calculation unit 12 that sets a variable representing a combination of a plurality of devices selected from a plurality of types of devices to be selected as a first decision variable, sets a function for calculating a total value of an introduction cost and an operation cost of the selected combination as a first objective function, and calculates a value of the first decision variable that satisfies a predetermined first constraint condition and minimizes a value of the first objective function; and a second calculation unit 13 that sets a variable representing an operating state of each of the devices included in the combination represented by the first decision variable as a second decision variable, formulates a second constraint condition that is a constraint condition on the operation of each of the devices, sets a function for calculating the operation cost as a second objective function, and calculates a value of the second decision variable that minimizes the second objective function as a mathematical optimization problem, and the first calculation unit 12 calculates the value of the first objective function by setting a value of the second objective function at the time when the second calculation unit 13 calculates an optimal solution of the second decision variable as the operation cost.

[0080] (2) The optimization device 1 of a second aspect is the optimization device 1 of (1), wherein the first constraint condition includes a constraint condition that the operation can be continued even if at least one of the devices included in the selected combination fails.

[0081] (3) The optimization device 1 of a third aspect is the optimization device 1 of (1) or (2), wherein the second decision variable includes a variable representing the start or stop of the equipment and a variable representing the output of the equipment, the multiple types of equipment include a main engine and an auxiliary equipment used in conjunction with the main engine, and the second constraint condition includes a constraint condition of starting the auxiliary equipment used in conjunction with the main engine only when the main engine is started, and / or a constraint condition of changing the output of the auxiliary equipment used in conjunction with the main engine in accordance with a change in the output of the main engine.

[0082] (4) The optimization device 1 of a fourth aspect is the optimization device 1 of (1) to (3), wherein the first calculation unit sets an initial solution based on a past optimal solution of the first decision variables when predetermined requirement conditions for the combination of equipment match within a predetermined range, and / or the second calculation unit sets an initial solution based on a past optimal solution of the second decision variables when the combination of equipment matches within a predetermined range and predetermined requirement conditions for the combination match within a predetermined range.

[0083] (5) The optimization device 1 of a fifth aspect is the optimization device 1 of (1) to (4), wherein the first calculation unit calculates a plurality of values ​​of the first decision variables in parallel using a black-box optimization algorithm, and the second calculation unit calculates a plurality of values ​​of the second decision variables in parallel corresponding to the plurality of first decision variables calculated in parallel. [Explanation of symbols]

[0084] 1. Optimization device 11...Input section 12...First calculation section 13...Second calculation section 14...Output section

Claims

1. a first calculation unit that calculates a value of the first decision variable that satisfies a first predetermined constraint condition and minimizes a value of the first objective function, the first decision variable being a variable that represents a combination of a plurality of devices selected from a plurality of types of devices to be selected, and a function that calculates a total value of an introduction cost and an operation cost of the selected combination being a first objective function; a second calculation unit that sets a variable representing an operating state of each of the devices included in the combination represented by the first decision variable as a second decision variable, formulates a second constraint condition which is a constraint condition on the operation of each of the devices, sets a function for calculating the operating cost as a second objective function, and calculates a value of the second decision variable that minimizes the second objective function as a mathematical optimization problem; Equipped with The first calculation unit calculates a value of the first objective function by setting a value of the second objective function when the second calculation unit calculates an optimal solution of the second decision variable as the operating cost. Optimizer.

2. The first constraint condition includes a constraint condition that the operation can be continued even if at least one of the devices included in the selected combination fails. The optimization device according to claim 1 .

3. The second decision variables include a variable representing start or stop of the device and a variable representing an output of the device, The plurality of types of equipment include a main engine and an auxiliary engine that is used in conjunction with the main engine, The second constraint condition includes a constraint condition that the auxiliary machine used in conjunction with the main engine is started only when the main engine is started, and / or a constraint condition that the output of the auxiliary machine used in conjunction with the main engine is changed according to a change in the output of the main engine. The optimization device according to claim 2 .

4. The first calculation unit sets an initial solution based on a past optimal solution of the first decision variable when a predetermined required condition for the combination of devices is met within a predetermined range; and / or The second calculation unit sets an initial solution based on a past optimal solution of the second decision variable when the combination of the devices matches within a predetermined range and a predetermined required condition for the combination matches within a predetermined range. The optimization device according to claim 3 .

5. The first calculation unit calculates a plurality of values ​​of the first decision variable in parallel using a black-box optimization algorithm; The second calculation unit calculates, in parallel, a plurality of values ​​of the second decision variable corresponding to the plurality of first decision variables calculated in parallel.

5. The optimization device according to claim 4.

6. a first step of calculating a value of the first decision variable that satisfies a first predetermined constraint condition and minimizes a value of the first objective function, the first decision variable being a variable that represents a combination of a plurality of devices selected from a plurality of types of devices to be selected, and the first objective function being a function that calculates a total value of an introduction cost and an operation cost of the selected combination; a second step of setting a variable representing an operating state of each of the devices included in the combination represented by the first decision variable as a second decision variable, formulating a second constraint condition which is a constraint condition on the operation of each of the devices, setting a function for calculating the operating cost as a second objective function, and calculating a value of the second decision variable which minimizes the second objective function as a mathematical optimization problem; Including, In the first step, a value of the first objective function is calculated by setting a value of the second objective function when an optimal solution of the second decision variable is calculated in the second step as the operating cost. Optimization methods.

7. a first step of calculating a value of the first decision variable that satisfies a first predetermined constraint condition and minimizes a value of the first objective function, the first decision variable being a variable that represents a combination of a plurality of devices selected from a plurality of types of devices to be selected, and the first objective function being a function that calculates a total value of an introduction cost and an operation cost of the selected combination; a second step of setting a variable representing an operating state of each of the devices included in the combination represented by the first decision variable as a second decision variable, formulating a second constraint condition which is a constraint condition on the operation of each of the devices, setting a function for calculating the operating cost as a second objective function, and calculating a value of the second decision variable which minimizes the second objective function as a mathematical optimization problem; A program for causing a computer to execute the following: In the first step, a value of the first objective function is calculated by setting a value of the second objective function when an optimal solution of the second decision variable is calculated in the second step as the operating cost. program.

Images