Optimization device, optimization method, and program

Abstract

To more efficiently perform optimization calculation of combination of distributed power supplies by using an Ising machine.SOLUTION: An optimization device comprises: a distributed power supply information reception unit which receives distributed power supply information from a distributed power supply; a narrowing-down problem generation unit which generates a quadratic binary optimization problem for narrowing down the distributed power supplies of calculation objects, on the basis of the distributed power supply information; a QUBO problem generation unit which converts the quadratic binary optimization problem into a QUBO problem; an Ising machine execution unit which outputs an Ising machine solution which is a calculation result obtained by narrowing down the distributed power supplies of the calculation objects by solving the QUBO problem by using an Ising machine; a matching determination unit which generates a matching solution which is a result obtained by associating two distributed power supplies among the distributed power supplies of the calculation objects with each other, on the basis of the Ising machine solution; and a matching information transmission unit which transmits the matching solution to the distributed power supply corresponding to the matching solution.SELECTED DRAWING: Figure 1

Description

[Technical Field]

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

[0002] In recent years, many distributed power sources have been introduced, such as renewable energy generators (such as solar power generators) (hereinafter also referred to as "renewable energy"), storage batteries, EVs (electric vehicles), and EV charging facilities (hereinafter also referred to as "charging facilities"). Furthermore, many distributed power sources are sometimes used in combination. For example, the following cases (11) to (13) are considered.

[0003] (11) When operating a virtual power plant (VPP), which integrates and controls multiple distributed power sources to function as a single power plant. (12) When there are multiple EVs and multiple charging facilities, and each EV selects the charging facility to charge. (13) When there are multiple renewable energy sources and multiple storage batteries, and the storage battery to be charged with the electricity generated by each renewable energy source is selected

[0004] When costs are incurred in these cases, there is a need to select the lowest-cost combination. The cost in case (12) is, for example, the power consumption caused by the EV traveling to the charging facility. The cost in case (13) is, for example, the transmission line usage fee (wheeling cost) when transmitting the electricity generated from renewable energy to the storage battery.

[0005] When performing the optimization calculations for the above combinations, if the number of possible combinations is large, conventional algorithms may not be able to obtain low-cost combinations in a practical amount of time. Therefore, it is possible to utilize Ising machines, which can obtain low-cost combinations in a shorter time than conventional algorithms. Summary of the Invention [Problem to be solved by the invention]

[0006] However, with the conventional method using an Ising machine, for example, in the above (12), when there are several thousand EVs and several thousand charging facilities, it may not be possible to obtain a low-cost combination in a practical time.

[0007] Therefore, the problem to be solved by the present invention is to provide an optimization device, an optimization method, and a program that can more efficiently perform optimization calculations for combinations of distributed power sources using an Ising machine. [Means for solving the problem]

[0008] An optimization device according to an embodiment performs matching by matching two or more distributed power sources that receive and transfer power from three or more distributed power sources, and includes a distributed power source information receiving unit that receives distributed power source information from the distributed power sources, a narrowing-down problem generating unit that generates a second-order binary optimization problem for narrowing down the distributed power sources to be calculated based on the distributed power source information, a QUBO problem generating unit that converts the second-order binary optimization problem into a QUBO problem, an Ising machine execution unit that outputs an Ising machine solution that is a calculation result obtained by solving the QUBO problem using an Ising machine to narrow down the distributed power sources to be calculated, a matching determination unit that generates a matching solution that is a result of matching two or more of the distributed power sources to be calculated based on the Ising machine solution, and a matching information transmitting unit that transmits the matching solution to the distributed power source that corresponds to the matching solution. [Brief explanation of the drawings]

[0009] [Figure 1] FIG. 1 is a functional block diagram showing the overall configuration of an optimization device etc. according to the first embodiment. [Figure 2] FIG. 2 is an explanatory diagram of the calculation contents performed by the optimization device of the first embodiment. [Figure 3] FIG. 3 is a flowchart showing the processing performed by the optimization device of the first embodiment. [Figure 4]FIG. 4 is a flowchart showing the details of step S1 in FIG. [Figure 5] FIG. 5 is a flowchart showing a first example of details of step S4 in FIG. [Figure 6] FIG. 6 is a flowchart showing a second example of the details of step S4 in FIG. [Figure 7] FIG. 7 is a functional block diagram showing the overall configuration of an optimization device etc. according to the second embodiment. [Figure 8] FIG. 8 is an explanatory diagram regarding calculation of transportation demand in the second embodiment. [Figure 9] FIG. 9 is a diagram schematically illustrating an overview of the third embodiment. [Figure 10] FIG. 10 is an explanatory diagram of the prior art. DETAILED DESCRIPTION OF THE INVENTION

[0010] Hereinafter, embodiments (first to third embodiments) of the optimization device, optimization method, and program of the present invention will be described with reference to the drawings. Note that in the description of the second embodiment and onward, descriptions of matters similar to those in the previous embodiments will be omitted as appropriate.

[0011] (Prior Art) To facilitate understanding of the embodiments, the prior art will first be described again. Fig. 10 is an explanatory diagram of the prior art. Here, two types of distributed power sources are taken as examples: EV taxis (EV1, EV2, ...) and charging facilities (ST1, ST2, ...) that supply power to EVs including EV taxis.

[0012] The system then considers how to find a combination of EV taxis and charging facilities that will minimize the power consumption of EV taxis when they travel to charging facilities while ensuring the necessary amount of charging power based on the EV taxi utilization rate and traffic demand. Traffic demand can represent, for example, the power demand of electric vehicles during the time period under calculation, or the recommended amount of charging power calculated based on the power demand.

[0013] That is, the objective function is as follows so as to minimize the power consumption during driving to the charging facility: Objective function: Power consumption = Σ (power consumption rate x distance) (Electricity consumption rate: kWh of electricity required to travel 1 km)

[0014] The constraints are, for example, as follows: (A) Traffic demand constraints (ensuring the amount of charging energy according to traffic demand) Required power amount ≦ Total power amount after charging (B) Charger number constraint (allocation of chargers without waiting time) Number of EV taxis to be charged ≦ Total number of chargers at charging facilities

[0015] More specifically, the objective function is expressed as, for example, the following equation (1).

number

[0016] The costs incurred when using a combination of an EV taxi and a charging facility may include, for example, the power consumed when the EV taxi travels from its current location to the charging facility, as well as the cost of charging at the charging facility and costs related to road congestion. In addition, incentives for charging may be added as profits.

[0017] To perform the above calculation, a variable is required: the number of EV taxis x the number of charging stations. If we assume that all taxis in a given area in Japan are replaced with EV taxis, for example, there will be 9,339 EV taxis and 2,853 charging stations. Therefore, 9,339 x 2,853 = 26,644,167 variables are required, which is impossible to solve within a practical amount of time using conventional technology (including Ising machines).

[0018] Therefore, the following describes a technology that can more efficiently perform optimization calculations for the combination of distributed power sources using an Ising machine.

[0019] (First embodiment) Next, a first embodiment will be described. Fig. 1 is a functional block diagram showing the overall configuration of an optimization device 1 etc. of the first embodiment. Fig. 2 is an explanatory diagram of the calculation contents by the optimization device 1 of the first embodiment. The optimization device 1 is a device that performs matching to associate two dispersed power sources 2 that transfer power from three or more dispersed power sources 2.

[0020] Here, as in the case of the above-mentioned conventional technology, two types of distributed power sources 2 are taken as examples: EV taxis and charging facilities that supply power to EVs including EV taxis.

[0021] As with the conventional technology described above, we consider a combination of EV taxis and charging facilities that ensures the necessary amount of charging power based on the utilization rate of EV taxis and traffic demand, while minimizing the power consumption of EV taxis when driving to charging facilities. The constraints are also the same as those in the conventional technology described above (details will be described later).

[0022] The objective function differs from that in the case of the above-mentioned conventional technique (Equation (1)) and becomes as follows:

number

[0023] The meaning of each symbol is as follows: min,cost: Minimization of total power consumption, N: Number of EV taxis, ST: Number of charging facilities, C i,j : Power consumption when EV taxi i travels to charging facility j X EVi : "1" if EV taxi i is charging, "0" if not X STj : "1" if charging at charging facility j, "0" if not

[0024] In addition to the power consumption, other costs and benefits may be taken into account as costs. For example, costs related to road congestion may be added, or incentives for charging may be subtracted.

[0025] As with conventional technology, when calculating using equation (2) assuming there are 9,339 EV taxis and 2,853 charging stations, the number of variables is 9,339 + 2,853 = 12,192. The calculation to minimize this objective function is then performed using an Ising machine. In other words, first, the EV taxis and charging stations to be charged are determined (narrowed down), but a one-to-one correspondence between EV taxis and charging stations is not determined.

[0026] After that, the one-to-one correspondence between EV taxis and charging facilities is determined through formulation, focusing only on the narrowed-down EV taxis and charging facilities (details will be provided later).For example, if the number of EV taxis is narrowed down to 100 and the number of charging facilities is narrowed down to 3000, the number of variables will be reduced significantly, reducing the problem scale.

[0027] The process after narrowing down is performed by, for example, the optimization solver 17, but the following rule-based process may also be used. This is expected to increase the speed. - Matching is performed between EV taxis and charging facilities in order of lowest power consumption (if the power consumption is the same, priority is given to the EV taxi with the lowest SOC (State of Charge)). Charging facilities that exceed the charger number constraint are excluded from matching.

[0028] Next, the optimization device 1 will be described with reference to Fig. 1. Note that in Fig. 1, arrows indicate the main flow of information, and information may flow through parts without arrows.

[0029] The optimization device 1 includes an Ising machine 11, a processing unit 12, a storage unit 13, a distributed power supply information receiving unit 14, an input unit 15, an output unit 16, and an optimization solver 17. A distributed power supply 2 is connected to the optimization device 1. Note that the optimization solver 17 is not an essential component.

[0030] The Ising machine 11 is realized by, for example, a quantum annealing device or a simulated annealing device. However, the Ising machine 11 is not limited to these and may be another device or algorithm for solving a QUBO (Quadratic Unconstrained Binary Optimization) problem (a QUBO-type problem). Furthermore, the Ising machine 11 may be implemented as a device separate from the processing unit 12 and connected to the processing unit 12, or may be implemented as the same device as the processing unit 12.

[0031] The processing unit 12 is, for example, a computer including one or more processors. The storage unit 13 is, for example, a random access memory (RAM), a read only memory (ROM), a flash memory (such as a solid state drive (SSD)), or a hard disk drive (HDD). The storage unit 13 may be implemented as a device separate from the processing unit 12 and connected to the processing unit 12, or may be implemented as the same device as the processing unit 12, for example, as a main storage device or cache memory.

[0032] The distributed power source information receiver 14 is connected to the distributed power source 2 and the storage unit 13, and receives distributed power source information (described in detail below) from the dispersed power source 2. The input unit 15 is connected to the processing unit 12, and includes, for example, some or all of various keys, buttons, dial switches, a mouse, a touch panel, etc. The input unit 15 may also be a connection unit that is electrically connected to an external device.

[0033] The output unit 16 may be a connection unit that is connected to the processing unit 12 and electrically connected to an external device, or may be an LCD (Liquid Crystal Display) or an organic EL (Electro-luminescence) display device.

[0034] The processing unit 12 includes, as functional units, a refinement problem generation unit 121, a QUBO problem creation unit 122, an Ising machine execution unit 123, a matching determination unit 124, a matching information transmission unit 125, and a data management unit 126. Each functional unit in the processing unit 12 is realized, for example, by a processor such as a CPU (Central Processing Unit) executing a program (software). Furthermore, some or all of these functional units may be realized by hardware (including circuitry) such as an LSI (Large Scale Integration), an ASIC (Application Specific Integrated Circuit), an FPGA (Field-Programmable Gate Array), or a GPU (Graphics Processing Unit), or may be realized by a combination of software and hardware. The matching information transmission unit 125 is connected to the distributed power source 2.

[0035] The storage unit 13 stores distributed power supply information 130, cost information 131, a second-order binary optimization problem 132, a QUBO problem 133, an Ising machine solution 134, and a matching solution 135.

[0036] The distributed power source information 130 includes distributed power source information received from the distributed power source information receiver 14 (for example, charger usage status information in charging facilities) and location information of the distributed power sources 2 (EV taxis, charging facilities).

[0037] The cost information 131 is information such as various costs and profits related to charging an EV taxi at a charging facility. The cost information 131 is, for example, information such as the power consumption caused by the EV taxi traveling to the charging facility, costs related to road congestion, and incentive profits from charging. The cost information 131 is input via the input unit 15 and the data management unit 126, and is used when the refinement problem generation unit 121 creates a quadratic binary optimization problem 132.

[0038] The quadratic binary optimization problem 132 is created and saved in the refinement problem generator 121 .

[0039] The QUBO questions 133 are created and saved in the QUBO question creation unit 122.

[0040] The Ising machine solution 134 is calculated by executing the Ising machine 11 in the Ising machine execution unit 123 and is stored.

[0041] A matching solution 135 is determined in the matching determiner 124 and stored.

[0042] The narrowing-down problem generator 121 generates a second-order binary optimization problem for narrowing down the distributed power sources 2 to be calculated based on the distributed power source information 130, cost information 131, and the like, and saves the generated problem as a second-order binary optimization problem 132.

[0043] The quadratic binary optimization problem 132 includes, for example, an objective function that aims to minimize the total cost or maximize the total profit incurred when using a combination of an electric vehicle and charging facility using location information of the charging facility and the electric vehicle, a binary variable representing whether or not the electric vehicle (EV taxi) is being charged (whether or not to charge), and a binary variable representing whether or not the charging facility is being used (whether or not to use), a constraint that the total charging power of the electric vehicle satisfies a specified transportation demand, and a constraint regarding the number of chargers per charging facility.

[0044] Furthermore, the narrowing problem generation unit 121 may generate the quadratic binary optimization problem 132 using a predetermined relaxed constraint condition that relaxes the constraint condition regarding the number of chargers per charging facility (i.e., a condition that allows EV taxis to wait for a certain amount of time at the charging facility for charging).

[0045] The QUBO problem creation unit 122 converts the quadratic binary optimization problem 132 into a QUBO problem and saves it as a QUBO problem 133.

[0046] The Ising machine execution unit 123 solves the QUBO problem 133 using the Ising machine 11, outputs an Ising machine solution that is a calculation result obtained by narrowing down the distributed power sources 2 to be calculated, and saves it as an Ising machine solution 134.

[0047] The matching determination unit 124 generates a matching solution that is a result of associating two by two of the distributed power sources 2 to be calculated based on the Ising machine solution 134, and stores the matching solution as a matching solution 135. The matching determination unit 124 generates the matching solution using, for example, the optimization solver 17.

[0048] The matching information transmission unit 125 transmits the matching solution 135 to the distributed power source 2 corresponding to the matching solution 135. In the example of the destination distributed power source 2 being an EV taxi and a charging facility, the EV taxi is required and the charging facility is optional. The EV taxi drives to the charging facility indicated in the received matching solution to charge.

[0049] The data management unit 126 manages input and output data for the processing unit 12 .

[0050] Next, we will explain the formulation in detail. The objective function is formulated as shown in the above formula (2). In that case, the constraints on traffic demand are formulated as shown in the following formula (3).

number

[0051] The meanings of symbols other than those explained in equation (2) are as follows: Demand: Amount of charging (amount of charging energy) that corresponds to traffic demand CAP i :EV i Battery capacity [kWh] SOC i :EV i Remaining battery capacity [kWh]

[0052] Furthermore, the constraint on the number of chargers is formulated as shown in the following equation (4).

number

[0053] The meaning of each symbol is as follows: CNUM j : Number of chargers at charging facility j

[0054] Also, C i,j (Power consumption for driving to the charging facility) is calculated, for example, using the following formula: C i,j =EM i D i,j EM i :EV i Reciprocal of electricity cost [kWh / km] D i,j : Distance between EVi and charging facility j [km]

[0055] Other constraints may be used, such as the requirement that the power consumption of the entire charging facility be equal to or less than the contracted power consumption.

[0056] Furthermore, although the above is assumed to be for a single time frame, it may be formulated to create a charging plan for multiple future times.

[0057] Demand can be manually entered by the operator of the EV taxi business, or it can be calculated by estimating with high accuracy the charging demand according to the required driving distance from traffic statistics information using machine learning techniques, etc.

[0058] Next, the formulation when the QUBO problem creation unit 122 converts the quadratic binary optimization problem 132 into the QUBO problem 133 is as shown in the following equation (5).

number

[0059] The first term is the original objective function. The second term is a penalty term corresponding to the constraint on traffic demand. W1 is the weight for that penalty term. The third term is a penalty term corresponding to the constraint on the number of chargers. W2 is the weight for that penalty term. y 1,n , y 2,n is a binary variable used to express the slack variable when transforming inequality constraints into equality constraints. Here, instead of eliminating the constraint, a penalty term corresponding to the constraint is inserted.

[0060] Next, the formulation when the matching determination unit 124 generates a matching solution that is the result of matching two of the distributed power sources 2 to be calculated based on the Ising machine solution 134 is as shown in the following equations (6) to (9).

number

[0061] Other constraints may be used, such as the time required for charging, the travel time of EV taxis to charging facilities, and ensuring a certain EV utilization rate. In this case, the EV utilization rate becomes a variable that extends in the time direction because it changes over time. In addition, the amount of charging energy for each EV taxi may be introduced as a continuous variable, and the cost for the amount of charging energy may be simultaneously optimized.

[0062] The optimization solver 17 is a device or software capable of solving the quadratic binary optimization problem 132. The optimization solver 17 may be implemented in the same device as the processing unit 12, or may be implemented as a computer different from the processing unit 12. The optimization solver 17 can also be realized as software that implements, for example, a branch and bound method. The optimization solver 17 solves the quadratic binary optimization problem 132 for the narrowed-down EV taxis and charging facilities.

[0063] Next, the processing by the optimization device 1 will be described with reference to Fig. 3. Fig. 3 is a flowchart showing the processing by the optimization device 1 of the first embodiment.

[0064] In step S1, the narrowing-down problem generator 121 generates a second-order binary optimization problem 132 used to narrow down the number of distributed power sources 2 to be calculated, based on the distributed power source information 130, cost information 131, and the like.

[0065] Next, in step S2, the QUBO problem creation unit 122 transforms the quadratic binary optimization problem 132 into a QUBO problem 133.

[0066] Next, in step S3, the Ising machine execution unit 123 uses the Ising machine 11 to solve the QUBO problem 133, thereby generating an Ising machine solution 134 that is a calculation result obtained by narrowing down the distributed power sources 2 to be calculated.

[0067] Next, in step S4, the matching determination unit 124 generates a matching solution 135, which is a result of associating two by two of the narrowed-down distributed power sources 2 to be calculated, based on the Ising machine solution 134.

[0068] Next, in step S5, the matching information transmitter 125 transmits the matching solution 135 to the dispersed power source 2 corresponding to the matching solution 135.

[0069] Fig. 4 is a flowchart showing the details of step S1 in Fig. 3. First, in step S101, the refinement problem generator 121 refers to the distributed power source information 130 and the like to acquire the status of the distributed power source 2 (EV taxi, charging facility) (their locations, the usage status of the charging facility, etc.).

[0070] Next, in step S102, the refinement problem generator 121 calculates the cost ("C" in the above formula (2)). i,j ") is calculated.

[0071] Next, in step S103, the refinement problem generator 121 calculates a quadratic objective function (for example, the above-mentioned formula (2)).

[0072] Next, in step S104, the narrowing-down problem generator 121 calculates constraint conditions (for example, the above-mentioned formulas (3) and (4)).

[0073] Next, Fig. 5 is a flowchart showing a first example of the details of step S4 in Fig. 3. In the processing of Fig. 5, the matching determination unit 124 generates an optimization problem including an objective function and constraint conditions related to the narrowed-down distributed power sources 2 based on the information of the distributed power sources 2 narrowed down by the Ising machine execution unit 123, and generates a matching solution by solving the optimization problem. This will be described in detail below.

[0074] First, in step S401, the matching determination unit 124 calculates a first-order objective function (for example, the above-mentioned formula (6)) for the narrowed-down dispersed power sources 2 (EV taxis and charging facilities).

[0075] Next, in step S402, the matching determination unit 124 calculates constraint conditions (for example, the above-mentioned formulas (7) to (9)).

[0076] Next, in step S403, the matching determination unit 124 solves the optimization problem using the optimization solver 17. The optimization solver 17 may be a general-purpose solver that implements, for example, a branch and bound method, or may implement a metaheuristic method such as a genetic algorithm or tabu search.

[0077] Next, Fig. 6 is a flowchart showing a second example of the details of step S4 in Fig. 3. In this second example, the optimization solver 17 is not used. In the process of Fig. 6, the matching determination unit 124 generates a matching solution based on the information of the dispersed power sources 2 narrowed down by the Ising machine execution unit 123, and on the costs incurred when the dispersed power sources 2 are used in combination. At the same time, the matching determination unit 124 repeatedly performs a process of excluding from the matching targets any dispersed power sources 2 for which a constraint violation has occurred, thereby generating a matching solution. This will be described in detail below.

[0078] First, in step S411, the matching determination unit 124 determines whether there is a distributed power source 2 to be matched, and if the answer is Yes, the process proceeds to step S412, and if the answer is No, the process ends. For example, if the traffic demand constraint (charging energy amount constraint) is not satisfied, the process determines Yes and continues matching.

[0079] In step S412, the matching determination unit 124 matches the narrowed-down distributed power sources 2 in ascending order of cost. For example, among combinations of EV taxis and charging facilities, matching is performed in ascending order of power consumption. In this case, if the power consumption is the same, for example, the EV taxi with the lowest SOC is given priority.

[0080] Next, in step S413, the matching determination unit 124 determines whether or not there is a matching target that deviates from the constraint conditions, and if Yes, the process proceeds to step S414, and if No, the process returns to step S411.

[0081] In step S414, the matching determination unit 124 excludes the target dispersed power source 2 in which a constraint violation (deviation from the constraint conditions) has occurred from the matching candidates. For example, in the case of a charging facility, the dispersed power source 2 is excluded from the matching candidates when the charger number constraint is exceeded.

[0082] In this way, according to the optimization device 1 of the first embodiment, first, the distributed power sources 2 to be calculated are narrowed down by solving the QUBO problem using an Ising machine, and then matching is performed for the narrowed down distributed power sources 2, thereby making it possible to more efficiently perform optimization calculations for combinations of distributed power sources 2. For example, even in the case of an order of magnitude of numbers such as 9,339 EV taxis and 2,853 charging facilities, as explained in the above-mentioned prior art, it is possible to obtain a low-cost combination in a practical amount of time.

[0083] (Second embodiment) Next, a second embodiment will be described. Fig. 7 is a functional block diagram showing the overall configuration of an optimization device 1 etc. according to the second embodiment. Compared to the optimization device 1 in Fig. 1, the optimization device 1 in Fig. 7 additionally includes a future constant prediction unit 127.

[0084] The future constant prediction unit 127 predicts the values ​​at future times of constants in the quadratic binary optimization problem 132. Specifically, for example, the future constant prediction unit 127 predicts the values ​​at future times of constants based on the values ​​of the constants at the current time and the distributed power source information 130 received from the distributed power source information receiving unit 14.

[0085] Here, the constant in the quadratic binary optimization problem 132 is, for example, Demand (amount of charging corresponding to traffic demand) in the above-mentioned equation (3).

[0086] 8 is an explanatory diagram regarding the calculation of traffic demand in the third embodiment. (a) to (c) are graphs showing traffic statistical information used to estimate a target time. (a) is a graph showing the relationship between traffic demand at a target time (vertical axis) and traffic demand at the immediately preceding time (horizontal axis). (b) is a graph showing the relationship between traffic demand at a target time (vertical axis) and the operating rate of electric vehicles at the target time (horizontal axis). (c) is a graph showing the relationship between traffic demand at a target time (vertical axis) and precipitation at the target time (horizontal axis).

[0087] Then, the narrowing problem generation unit 121 models the relationship between the traffic demand at the target time and other data (traffic demand at the immediately preceding time, the operating rate of electric vehicles at the target time, and the amount of precipitation at the target time) using, for example, multiple regression analysis, and uses the model to estimate and use the traffic demand at the target time.

[0088] (d) to (f) are graphs showing traffic statistical information used to estimate future times. (d) is a graph showing the relationship between future traffic demand (vertical axis) and current traffic demand (horizontal axis). (e) is a graph showing the relationship between future traffic demand (vertical axis) and the rate of change in electric vehicle operation rate up to the present time (horizontal axis). (f) is a graph showing the relationship between future traffic demand (vertical axis) and weather forecast values ​​of precipitation (horizontal axis).

[0089] Then, the narrowing problem generation unit 121 models the relationship between the traffic demand at a future time and other data (current traffic demand, the rate of change in the operation rate of electric vehicles up to the current time, and weather forecast values ​​for precipitation) using, for example, multiple regression analysis, and uses the model to estimate and use the traffic demand at a future time.

[0090] In this way, according to the second embodiment, by predicting the value of the constant at a future time and using it to generate narrowed-down problems, it is possible to generate narrowed-down problems with higher accuracy.

[0091] The constants to be predicted in the quadratic binary optimization problem 132 are not limited to traffic demand. Other constants include the above-mentioned cost "C i,j" and the SOC of an EV taxi are also possible, and their values ​​at future times may be predicted and used in the same way.

[0092] (Third embodiment) Next, a third embodiment will be described. Fig. 9 is a diagram schematically illustrating an overview of the third embodiment. In the above-described first and second embodiments, two types of distributed power sources 2, namely, EV taxis and charging facilities, have been described as examples. In the third embodiment, the distributed power sources 2 are of two types: renewable energy generators (RE1, RE2, ...) and storage batteries (SB1, SB2, ...).

[0093] In this case, the second-order binary optimization problem 132 includes an objective function that aims to minimize the total value of costs incurred when using a combination of a renewable energy generator and a storage battery (for example, the transmission line usage fee (wheeling cost) when transmitting generated power from renewable energy to the storage battery via transmission line T) or to maximize the total value of profits, using a binary variable that indicates whether or not a renewable energy generator is used and a binary variable that indicates whether or not a storage battery is used.

[0094] Specifically, if you want to absorb (store) surplus electricity generated by renewable energy in a storage battery, the present invention can be used to determine the combination of renewable energy and storage battery that minimizes the total value of transmission line usage fees.

[0095] Under the current system, wheeling costs are a flat rate, but if the system design changes in the future and wheeling costs change depending on distance, time of day, system congestion, etc., the effect will be greater.

[0096] Other anticipated costs are listed below. - Individually contracted rates when peer-to-peer electricity trading is realized. Costs for virtual energy storage services. For example, the cost of using a storage battery located in a remote location, which varies depending on the type of battery.

[0097] As described above, according to the third embodiment, the optimization calculation of the combination of renewable energy and storage batteries can also be performed more efficiently using an Ising machine, as in the first and second embodiments.

[0098] The program executed by the optimization device 1 of this embodiment is provided as a file in an installable or executable format recorded on a computer-readable recording medium such as a CD (Compact Disc)-ROM, a flexible disk (FD), a CD-R, or a DVD (Digital Versatile Disk).

[0099] The program may also be configured to be stored on a computer connected to a network such as the Internet and provided by being downloaded via the network, or to be provided or distributed via a network such as the Internet, or to be provided by being pre-installed in a ROM or the like.

[0100] The program has a modular structure that includes each of the functional components described above, and in actual hardware, the CPU reads the program from the storage medium and executes it, loading each functional component into the main memory device and generating it on the main memory device.

[0101] Although several embodiments of the present invention have been described, these embodiments are presented as examples and are not intended to limit the scope of the invention. These novel embodiments can be embodied in various other forms, and various omissions, substitutions, and modifications can be made without departing from the spirit of the invention. These embodiments and their modifications are included within the scope and spirit of the invention, and are also included in the scope of the invention and its equivalents as defined in the claims.

[0102] For example, additional data, such as weather data, may be used in generating the optimization problem.

[0103] Furthermore, in the objective function in the first embodiment, the charging cost may be variable. For example, the objective function may include not only the power consumption during driving but also the charging cost. In this case, the charging cost may be determined, for example, by an aggregator or based on the nodal price (additional power supply cost incurred for additional power demand). [Explanation of symbols]

[0104] 1...Optimization device, 2...Distributed power source, 11...Ising machine, 12...Processing unit, 13...Memory unit, 14...Distributed power source information receiving unit, 15...Input unit, 16...Output unit, 17...Optimization solver, 121...Narrowing problem generation unit, 122...QUBO problem creation unit, 123...Ising machine execution unit, 124...Matching determination unit, 125...Matching information transmission unit, 126...Data management unit, 127...Future constant prediction unit, 131...Cost information, 132...Second-order binary optimization problem, 133...QUBO problem, 134...Ising machine solution, 135...Matching solution

Claims

1. 1. An optimization device that performs matching to associate two or more distributed power sources that transfer power from three or more distributed power sources, a distributed power source information receiving unit that receives distributed power source information from the distributed power source; a narrowing-down problem generator that generates a second-order binary optimization problem for narrowing down the distributed power sources to be calculated based on the distributed power source information; a QUBO problem generator that converts the quadratic binary optimization problem into a QUBO (Quadratic Unconstrained Binary Optimization) problem; an Ising machine execution unit that solves the QUBO problem using an Ising machine and outputs an Ising machine solution that is a calculation result in which the distributed power sources to be calculated are narrowed down; a matching determination unit that generates a matching solution that is a result of matching two or more of the distributed power sources to be calculated based on the Ising machine solution; and a matching information transmitting unit that transmits the matching solution to the distributed power source that corresponds to the matching solution.

2. The distributed power sources are of two types: electric vehicles and charging facilities that supply power to the electric vehicles; 2. The optimization device according to claim 1, wherein the second-order binary optimization problem includes an objective function that aims to minimize the total cost or maximize the total profit incurred when the electric vehicle and the charging facility are used in combination, using location information of the charging facility and the electric vehicle, a binary variable that indicates whether the electric vehicle is being charged, and a binary variable that indicates whether the charging facility is being used, a constraint that the total value of charging power of the electric vehicles satisfies a predetermined transportation demand, and a constraint related to the number of chargers for each charging facility.

3. The distributed power sources are of two types: a renewable energy generator and a storage battery; 2. The optimization device according to claim 1, wherein the second-order binary optimization problem includes an objective function that aims to minimize the total cost or maximize the total profit incurred when the renewable energy generator and the storage battery are used in combination, using a binary variable that represents whether the renewable energy generator is used or not and a binary variable that represents whether the storage battery is used or not.

4. The traffic demand represents the power demand of the electric vehicle during a calculation time period, or a recommended amount of charging power calculated based on the power demand; The optimization device according to claim 2 , wherein the narrowing-down problem generator estimates the power demand based on accumulated past traffic statistics information.

5. The optimization device according to claim 2 , wherein the narrowing-down problem generator generates the quadratic binary optimization problem using a predetermined relaxed constraint condition obtained by relaxing a constraint condition related to the number of chargers for each charging facility.

6. a future constant prediction unit that predicts a value of a constant in the quadratic binary optimization problem at a future time; 2. The optimization device according to claim 1, wherein the future constant prediction unit predicts a value of the constant at a future time based on a value of the constant at a current time and the distributed power supply information received from the distributed power supply information receiving unit.

7. 3. The optimization device according to claim 2, wherein the cost incurred when using the electric vehicle in combination with the charging facility is the power consumption incurred when the electric vehicle travels from its current location to the charging facility and / or the charging cost at the charging facility.

8. 2. The optimization device according to claim 1, wherein the matching determination unit generates an optimization problem including an objective function and constraint conditions related to the narrowed-down distributed power sources based on information about the distributed power sources narrowed down by the Ising machine execution unit, and generates the matching solution by solving the optimization problem.

9. 2. The optimization device according to claim 1, wherein the matching determination unit generates the matching solution based on the information of the distributed power sources narrowed down by the Ising machine execution unit, and on costs incurred when the combined distributed power sources are used, by repeatedly performing a process of excluding from matching targets any distributed power sources for which a constraint violation has occurred.

10. An optimization method using an optimization device that performs matching to associate two or more distributed power sources that transfer power from three or more distributed power sources, the method comprising: a distributed power source information receiving step of receiving distributed power source information from the distributed power source; a narrowing-down problem generation step of generating a second-order binary optimization problem for narrowing down the distributed power sources to be calculated based on the distributed power source information; A QUBO problem generation step of converting the quadratic binary optimization problem into a QUBO problem; an Ising machine execution step of solving the QUBO problem using an Ising machine to output an Ising machine solution that is a calculation result in which the distributed power sources to be calculated are narrowed down; a matching determination step of generating a matching solution that is a result of matching two or more of the distributed power sources to be calculated based on the Ising machine solution; a matching information transmission step of transmitting the matching solution to the distributed power source corresponding to the matching solution.

11. a computer that is an optimization device that performs matching to associate two or more distributed power sources that transfer power from three or more distributed power sources; a distributed power source information receiving unit that receives distributed power source information from the distributed power source; a narrowing-down problem generator that generates a second-order binary optimization problem for narrowing down the distributed power sources to be calculated based on the distributed power source information; a QUBO problem generator that converts the quadratic binary optimization problem into a QUBO problem; an Ising machine execution unit that solves the QUBO problem using an Ising machine and outputs an Ising machine solution that is a calculation result in which the distributed power sources to be calculated are narrowed down; a matching determination unit that generates a matching solution that is a result of matching two or more of the distributed power sources to be calculated based on the Ising machine solution; and a program for causing the matching solution to function as a matching information transmitting unit that transmits the matching solution to the distributed power source corresponding to the matching solution;

Images