JP2020065407A - Operation planning method, operation planning device, and program - Google Patents

Abstract

To preferably shorten a time for calculating a feasible solution when creating an operation plan for a target device for taking a discrete state.SOLUTION: The operation planning method includes: a problem acquisition step of acquiring an original problem for calculating a plurality of state values including discrete state values indicating discrete states of each target device; a relaxation problem calculation step for calculating a continuous state of each target device by solving a relaxation problem assuming that each target device can assume a continuous state in the original problem; a sub-problem calculation step of calculating a value including at least a part of the discrete state values of the state values by dividing the original problem into multiple sub-problems, and solving the sub-problem using a distance between a solution of the relaxation problem and a solution of the sub-problem; and a slack problem calculation step of calculating a remaining state values of the state values by fixing to a value calculated at the sub-problem calculation step for some of the multiple state values and solving a slack problem using constraints of the original problem.SELECTED DRAWING: Figure 1

Description

  The present invention relates to an operation planning method, an operation planning device and a program.

Conventionally, there is known a technique of creating a start / stop plan of a generator using an iterative method called a Lagrangian relaxation method (for example, refer to Patent Document 1). The Lagrangian relaxation method in Patent Document 1 divides a problem into sub-problems for each target device and calculates a solution. When the solution is calculated by the Lagrangian relaxation method, the number of processing iterations may increase and the calculation time may increase. As a related technique, there is known a control device including a relaxation problem optimization unit that solves a relaxation problem in which discrete constraint conditions are relaxed, and a subproblem optimization unit that solves a subproblem (for example, Patent Document 1). 2).
[Patent Document]
[Patent Document 1] JP 2002-300720 A [Patent Document 2] JP 2013-64245 A

  When creating an operation plan for a target device that takes a discrete state, it is preferable to shorten the time for calculating a feasible solution.

  In order to solve the above-mentioned subject, in the 1st mode of the present invention, an operation plan method which generates an operation plan of a plurality of object devices by a computer is provided. The operation planning method has a problem of obtaining an original problem for calculating a plurality of state values including discrete state values indicating discrete states of respective target devices, which can satisfy a predetermined purpose and constraint conditions. An acquisition stage may be provided. The operation planning method may include a relaxation problem calculation step of solving a relaxation problem assuming that each target device can assume a continuous state in the original problem, and calculating a continuous state of each target device. The operation planning method divides the original problem into a plurality of subproblems, and solves the subproblem by using the distance between the solution of the relaxation problem and the solution of the subproblem. And a sub-problem calculation step for calculating a value including a state value. The operation planning method fixes a part of a plurality of state values to a value calculated in the sub-problem calculation step, and solves the slack problem using the constraint condition of the original problem to A slack problem computation stage may be provided to compute the remaining state values.

  The plurality of state values may include a state value indicating the output amount of the target device. In the sub-problem calculation stage, the state value indicating the output amount of at least a part of the target devices may be calculated. In the slack problem calculation step, the output amount of at least a part of the target devices may be fixed, and the state value indicating the output amount of the remaining target devices may be calculated.

  The target device whose state value is calculated in the slack problem calculation stage may be selected based on at least one characteristic of the output capacity, output efficiency, and responsiveness of the target device.

  The plurality of target devices may include one or more power generation devices and one or more power storage devices. In the partial problem calculation stage, a state value indicating the output amount of the power generation device may be calculated. At the slack problem calculation stage, the output amount of the power generation device may be fixed and the state value indicating the output amount of the power storage device may be calculated.

  The plurality of state values may include a state value indicating the output amount of the target device and a state value indicating the activation state of the target device. In the sub-problem calculation stage, the state value indicating the activation state of the plurality of target devices may be calculated. In the slack problem calculation stage, the state value indicating the activation state of the plurality of target devices may be fixed, and the state value indicating the output amount of the plurality of target devices may be calculated.

  The plurality of state values may include a state value indicating the output amount of the target device and a state value indicating the activation state of the target device. In the first slack problem calculation stage, the output amounts of some target devices may be fixed to the values obtained in the partial problem calculation stage, and the state value indicating the output amount of the remaining target devices may be calculated. . In the first slack problem calculation stage, when there is no state value satisfying the constraint condition, in the second slack problem calculation stage, the state value indicating the activation state of the plurality of target devices is obtained by the partial problem calculation stage. Alternatively, the state value indicating the output amount of the plurality of target devices may be fixed.

  The plurality of state values may include a state value indicating the output amount of the target device and a state value indicating the activation state of the target device. In the first slack problem calculation stage, the state values indicating the activation states of the plurality of target devices are fixed to the values obtained in the partial problem calculation stage, and the state values indicating the output amounts of the plurality of target devices are calculated. You may In the first slack problem calculation stage, if there is no state value that satisfies the constraint, in the second slack problem calculation stage, the output amounts of some target devices are fixed to the values obtained in the partial problem calculation stage. In addition, the state value indicating the output amount of the remaining target device may be calculated.

  In a second aspect of the present invention, a program for causing a computer to execute the operation planning method according to the first aspect is provided.

  According to a third aspect of the present invention, there is provided an operation planning device that generates operation plans for a plurality of target devices. The problem is that the operation planning apparatus obtains an original problem for computing a plurality of state values including discrete state values indicating the discrete states of each target apparatus, which can satisfy predetermined objectives and constraints. An acquisition unit may be provided. The operation planning apparatus may include a relaxation problem calculation unit that solves a relaxation problem in the original problem assuming that each target device can assume a continuous state, and calculates a continuous state of each target device. The operation planning device divides the original problem into a plurality of subproblems, and solves the subproblem by using the distance between the solution of the relaxation problem and the solution of the subproblem, thereby discretizing at least some of the plurality of state values. A partial problem calculator that calculates a value including a state value may be provided. The operation planning apparatus fixes a part of the plurality of state values to a value calculated by the sub-problem calculation unit, and solves the slack problem using the constraint condition of the original problem, A slack problem calculator that calculates the discrete values of the remaining state values may be provided.

  Note that the above summary of the invention does not enumerate all the necessary features of the invention. Further, a sub-combination of these feature groups can also be an invention.

It is a flow chart which shows an example of the operation planning method concerning one embodiment of the present invention. 6 is a flowchart showing the contents of a slack problem calculation stage in FIG. 1. It is a figure which shows the outline of the solution process of the partial problem in the operation planning method demonstrated in FIG. 6 is a flowchart showing a first example of a slack problem calculation stage in FIG. 1. It is a figure which shows the outline of the operation planning method demonstrated in FIG. It is an example of each solution of the relaxation problem, the partial problem, and the slack problem in the operation planning method described in FIG. It is a flowchart which shows an example of the process which selects the object apparatus used as the calculation object in the slack problem calculation stage. It is a figure which shows an example of the object apparatus which the operation planning apparatus 300 which concerns on one Embodiment of this invention controls. 7 is a flowchart showing a second example of a slack problem calculation stage in FIG. 1. 7 is a flowchart showing a third example of the slack problem calculation stage in FIG. 1. It is a figure which shows the structural example of the operation planning apparatus 300. 1 illustrates an example computer 2200 in which aspects of the present invention may be embodied in whole or in part.

  Hereinafter, the present invention will be described through embodiments of the invention, but the following embodiments do not limit the invention according to the claims. In addition, not all of the combinations of features described in the embodiments are essential to the solving means of the invention.

  FIG. 1 is a flowchart showing an example of an operation planning method according to an embodiment of the present invention. In the operation planning method, a computer generates operation plans for a plurality of target devices. The computer may be a general-purpose computer or a dedicated computer for generating an operation plan.

  The target device is a device that takes a discrete state. For example, the target device is a power supply device, and the discrete states are “start” and “stop”. In this specification, “start” of the device may be represented by “1” as a discrete state value as a state value, and “0” as a discrete state value by “stop”. The target device takes a plurality of state values including a discrete state value indicating a discrete state. The state value may include not only the above-mentioned state value indicating the activation state but also the output amount of the target device. The output amount of the target device is, for example, the value of output power. The target device is not limited to the power supply device. Moreover, the discrete states are not limited to “start” and “stop”. Further, the target device may be a device that takes a discrete state of three or more values.

  The operation plan of the target device is, for example, a plan of the timing of "starting" or "stopping" each of the plurality of power supply devices. As an example, the operation plan of the target device sets the states of the plurality of target devices at each time so as to satisfy the predetermined purpose while satisfying the predetermined constraint condition. In one example, the constraint condition may include a condition caused by the performance of the target device, such as a maximum value or a minimum value of the output power of the power supply device, or may include an artificially agreed condition. The purpose of the operation plan may be the minimization of certain parameters, such as the minimization of power consumption, the minimization of operating costs, etc. The purpose of the operation plan may be the maximization of a predetermined parameter, or may bring the predetermined parameter close to the reference value.

式（１．１）は運転計画の目的を示す目的関数であり、「ｓ．ｔ．」以下に示される式（１．２）から（１．４）は運転計画の制約条件を示す。ａは分割される部分を示しており、例えば、各ノードを示している。ａは、例えば各対象装置の番号を示している。一例において、ｘ_ａおよびｙ_ａは、ａ番目のノード、例えば、ａ番目の対象装置における状態を示す状態値である。本明細書では、ｘ、ｙ等の状態値を、単に「状態ｘ」、「状態ｙ」等のように称する場合がある。ｆ_ａはａ番目のノードの状態ｘ_ａおよびｙ_ａに応じた所定の目的パラメータをノード毎に示す関数である。目的パラメータは、例えば上述したように運転コスト等である。本例の目的関数は、各ノードにおける目的パラメータの総和を最小化することを示している。 First, the computer acquires an original problem for calculating a plurality of state values indicating the states of respective target devices, which can satisfy the above-described purpose and constraint conditions (problem acquisition step S102). The plurality of state values include discrete state values indicating discrete states of the target device. The original problem can be expressed by the following equation as an example.

Expression (1.1) is an objective function indicating the purpose of the operation plan, and Expressions (1.2) to (1.4) shown below “s.t.” indicate the constraint conditions of the operation plan. “A” indicates a divided portion, for example, each node. a indicates the number of each target device, for example. In one example, x _a and y _a are state values indicating the state of the a-th node, for example, the a-th target device. In this specification, state values such as x and y may be simply referred to as “state x”, “state y” and the like. f _a is a function indicating a predetermined object parameters according to the state x _a and y _a the a-th node for each node. The target parameter is, for example, the operating cost as described above. The objective function of this example indicates that the sum of the objective parameters at each node is minimized.

G _a is called an independent constraint. G _a is a predetermined value corresponding to the state of the target device x _a and y _a, a function which defines the relationship between the reference value 0 for each node. In this example, the constraint condition is that the predetermined value is equal to or less than the reference value 0. The independent constraint condition G _a may be an output constraint or a voltage constraint of the power supply at each node. H is a function that defines the relationship between a predetermined value according to the states x _a and y _a of the target device and the reference value 0 at a plurality of or all nodes. For example, H is a constraint condition that spans all variables x and y. Examples of H include a supply and demand balance constraint in which the sum of the output amounts of all target devices is equal to the load, a network constraint, and the like. Note that the constraint condition is not limited to one that defines that the predetermined value is equal to or less than the reference value. As the constraint condition, it is possible to use one that constrains various characteristic values of the target device to predetermined conditions.

R in Expression (1.4) indicates that it takes an arbitrary real number value, and Z indicates that it takes a plurality of discrete values. As shown in equation (1.4), the state value x _a is a continuous variable of size n. In one example, the state value x _a may indicate a state of continuous values of n characteristics in one target device, or may indicate the state of each of the n target devices. For example, x _a is a state value indicating power supply output or the like for each time series. Status value y _a is a discrete variable size m. In one example, the state value y _a is a state value indicating the state of discrete binary stop (0) and started (1). Status value y _a may be a state value indicating a stepwise operation, such as the transformer taps. Status value y _a may show the state of the m characteristic in one of the target device, may indicate the status of each of the m target device.

  When solving the original problem, the states x and y of the target device at each time section that satisfy the objective function and constraint conditions can be obtained. However, the original problem represented by the equations (1.1) and (1.2) to (1.4) is a mixed integer nonlinear programming problem. The mixed integer nonlinear programming problem belongs to a class called NP-hard. In the mixed integer nonlinear programming problem, if the number of target devices becomes large, the amount of calculation becomes enormous, and it becomes difficult to obtain a solution.

  In the operation planning method of this example, instead of solving the original problem, the computer solves the relaxation problem assuming that each target device can assume a continuous state in the original problem (relaxation problem calculation step S104). In S104, the computer uses the objective function and the constraint condition in the original problem, and on the assumption that each target device can assume a continuous state under the constraint condition, calculates a continuous state of the target device. The objective function and the constraint conditions in the relaxation problem and the original problem may be the same, except that the state value takes a discrete value or a continuous value.

式（２．４）におけるｙ_ｉハットは、特定の範囲の任意の実数を取り得ることを示している。例えば、ｙ_ｉハットは、０以上１以下の任意の実数を取り得る。本例の緩和問題の目的関数（式（２．１））および制約条件（式（２．２）から（２．４））は、離散的な状態値ｙに代えて、式（２．４）における所定の範囲の任意の値（たとえば、０から１の間の任意の値）を取り得る連続変数である状態値ｙハットを用いること以外は、元問題の目的関数（式（１．１））および制約条件（式（１．２）から（１．４））と同一である。式（２．１）および制約条件（式（２．２）から（２．４））で示される緩和問題は、非線形計画問題である。非線形計画問題は、逐次二次計画法等の手法で高速に解くことができる。たとえば、非線形計画問題である緩和問題は、目的関数が凸性の仮定の下において，解を演算することができる。本明細書では、緩和問題で求めた解（すなわち離散的でなく、所定の範囲の任意の値を取り得る解）を緩和解と称する場合がある。 The relaxation problem can be expressed by the following equation as an example.

The y _i hat in equation (2.4) indicates that any real number in a particular range can be taken. For example, the y _i hat can take any real number between 0 and 1 inclusive. The objective function (Equation (2.1)) and the constraint conditions (Equations (2.2) to (2.4)) of the relaxation problem of this example are replaced by the equation (2.4) instead of the discrete state value y. ), A state variable y hat that is a continuous variable that can take any value within a predetermined range (for example, any value between 0 and 1) is used. )) And constraints (equations (1.2) to (1.4)). The relaxation problem represented by the formula (2.1) and the constraint conditions (formulas (2.2) to (2.4)) is a nonlinear programming problem. The nonlinear programming problem can be solved at high speed by a method such as sequential quadratic programming. For example, a relaxation problem, which is a nonlinear programming problem, can calculate a solution under the assumption that the objective function is convex. In this specification, a solution obtained by the relaxation problem (that is, a solution that is not discrete and can take an arbitrary value within a predetermined range) may be referred to as a relaxation solution.

  Next, the computer determines whether or not there is a feasible solution (that is, states x and y satisfying the constraint condition) in the relaxation problem (first determination step S106). If no feasible solution exists, the process ends. If there is a feasible solution, the computer solves the partial problem obtained by dividing the original problem into a plurality of parts (sub-problem calculation step S108).

  As an example, the partial problem is a problem in which the original problem is divided into a plurality of parts corresponding to a plurality of nodes (target devices). Each sub-problem may include one time section and may include multiple time sections. The number of time sections included in each sub-problem may be the same or different between sub-problems.

  When solving each subproblem, the computer has a relaxed solution of the relaxation problem (ie, state y hat can take any real value and state y hat can take any value between 0 and 1, for example). Hat) and the solution of the subproblem (ie, state x and y, which is a discrete value of 0 or 1). In the sub-problem calculation step S108, the sub-problem is solved using the solution distance to calculate a value including at least some of the discrete state values among the plurality of state values. The solution distance means a measure indicating the distance between two points corresponding to each solution in each state parameter (x and y in this example) space. Types of distance include Euclidean distance and Chebyshev distance. Alternatively, in addition to the distance, an index such as similarity or proximity may be used. As an example, in S108, the solution of the subproblem may be calculated so that the Euclidean distance to the solution of the relaxation problem is minimized.

本例の部分問題では、それぞれノードａについて、式（３．２）から（３．４）を満たす状態ｘ´_ａ、ｙ´_ａを演算する。部分問題における状態ｘ´_ａは、式（３．４）に示すように任意の実数の値を取る。また状態ｙ´_ａは、緩和問題で求めた緩和解ｙ_ａハット（すなわち、０から１の間の任意の値を取り得る）に所定の偏差Δを加算したものであり、０または１の２値のいずれかの値を取る。偏差Δは、状態ｙ´_ａと緩和解ｙ_ａハットとの差分（すなわち距離）を示している。 The sub-problem can be expressed by the following equation as an example. Here, variables other than the subproblems x ′ _{a and} y ′ _a , that is, x ₁ hat, y ₁ hat _{, ...} x _a−1 hat, ya ₋₁ hat, x _{a + 1} hat, y _{a + 1} hat, _... · · _{x a} hat, _{y a} hat are treated as fixed variables using relaxation solutions. In this example, the objective function in the subproblem minimizes _a given objective parameter f _a at each node.

The partial problem of the present embodiment, the nodes a respective state _x'a satisfying the equation (3.2) and (3.4), calculates the y _'a. State in the subproblem _x'a takes a value of any real number as shown in equation (3.4). The state y _'a is relaxed problem obtained relaxation solutions y _a hat (i.e., may take any value between 0 and 1) in are those obtained by adding a predetermined difference delta, 0 or 1 of 2 Takes one of the values. Deviation Δ shows a state y _'a mitigation solution y _a hat and the difference (i.e. distance).

In the objective function and constraints of subproblem, using state y _'a in place of the state y _a in the objective function and constraints of the original problem. State y _'a have a difference of the deviation Δ with respect to alleviating the solution y _a hat. Purpose In this example, as the objective function of subproblems, as shown in equation (3.1), to the objective parameters defined by the function f _a, the sum of the distance parameter d corresponding to the deviation Δ minimize Use a function. Note that d is a scale indicating the magnitude of the deviation Δ, and may be arbitrarily set by the user. d may be a proximity index that evaluates the distance (proximity) between the y hat and the feasible modified solution y ′ _a . d may be a square norm or the like as the simplest index.

  By solving such a subproblem, the original problem can be divided into multiple subproblems, and each subproblem can be solved using the distance between the relaxation solution of the relaxation problem and the discrete value solution of the subproblem. . In the sub-problem, the original problem is divided into a plurality of problems, so that the size of one problem is reduced. Therefore, even if the discrete state y is included, the calculation can be performed at high speed. The calculation of the subproblem can be executed at high speed by using, for example, a branch and bound method.

Then, by introducing a distance parameter (d in this example) into the objective function to solve the subproblem, the solution of the subproblem is prevented from becoming a suboptimal that is optimized considering only one subproblem. be able to. In particular, the mitigation solution y _a hat, a distance between the solutions y _'a subproblem by setting the objective function to minimize, reasonable solutions near the exact solution can be obtained at high speed.

In the example shown from equation (3.1) to (3.4), but has been used relaxation solutions _{y a} hat, may be used relaxation solution _{x a} hat. In this case, x ′ _a = x _a hat + Δ. May also be used both relaxation solution x _a hat and y _a hat.

  After the solution has been obtained for one subproblem, it is determined whether or not the solution has been obtained for a predetermined number of subproblems (second determination step S110). If the solution has not been obtained for a predetermined number of subproblems, the processing from S108 to S110 is repeated for the next subproblem. However, the partial problems a = 1, ... A may be solved in parallel. When the solution is calculated for a predetermined number of subproblems, the computer solves the slack problem (slack problem calculation step S112).

  FIG. 2 is a flowchart showing the contents of the slack problem calculation stage in FIG. The computer fixes a part of the plurality of state values to the value calculated in the partial problem calculation step S108 (S202). Then, the computer solves the slack problem using the constraint condition of the original problem, and calculates the remaining state value among the plurality of state values (S204).

The slack problem can be expressed by the following equation as an example. In the present _{example, x's,} except for the _{y's, x'1, y'1} , x'2, y'2, ···, x's-1, y's-1, x's + 1, _{y's + 1, ··· x'a} , y'1, may be secured as a part of _x'a is a plurality of state values.

In the objective function and constraints of slack problem, use state y _'s in place of the state y _a in the objective function and constraints of the original problem. The state y ′ _s has a difference of deviation Δ with respect to the relaxation solution y _s hat. In this example, an objective function that minimizes the objective parameter defined by the function f _s is used as the objective function of the slack problem, as shown in equation (4.1).

  In the solution method of the slack problem, a part of the plurality of state values is fixed to the value calculated in the sub-problem calculation stage, and the remaining state values are calculated. The computer may solve all subproblems in the subproblem calculation step S108, and then resolve at least some of the subproblems as slack problems in the slack problem calculation step (S112).

  Then, except for the state value to be solved as the slack problem, the computer fixes the plurality of state values to the values calculated in the sub-problem calculation step S108, and then solves the slack problem using the constraint condition of the original problem, The remaining state value of the plurality of state values is calculated. Therefore, the occurrence of a state in which there is no state value satisfying the constraint is suppressed as compared with the case where the slack problem calculation stage is not included. That is, it is possible to adjust the solution at the slack problem calculation stage so that a feasible solution can be obtained. Therefore, the time for calculating a feasible solution can be shortened.

FIG. 3 is a diagram showing an outline of solution processing of a partial problem in the operation planning method described in FIG. The horizontal axis in FIG. 3 represents the state value y ₁ and the vertical axis represents the state value y ₂ . For the state values y ₁ and y ₂ , discrete values that the target device can actually take are plotted by a plurality of diamond-shaped marks 204. A straight line 202 indicates a boundary line between the state values y ₁ and y _{2 that} satisfy the constraint condition. An area hatched with a broken line outside the straight line 202 is an area that does not satisfy the constraint condition, and an area that is not hatched is an area that satisfies the constraint condition. In FIG. 3, a straight line 202 in a certain time section is shown.

  Also, the position corresponding to the relaxation solution is plotted by a round mark 206. The relaxation solution is the solution that best fits the objective function within the region that satisfies the constraint condition. On the other hand, in the subproblem, of the discrete solutions (marks 204) in the region that satisfy the constraint condition, the solution (mark 208) having the smallest deviation Δ from the relaxation solution is calculated. By such processing, a solution that does not necessarily match the exact solution obtained from the original problem but has a close distance to the relaxation solution and high validity can be found at high speed.

  FIG. 4 is a flowchart showing a first example of the slack problem calculation stage in FIG. The plurality of state values may include not only the state value indicating the activation state of the target device but also the output amount of the target device. Then, in the sub-problem calculation step S108 of FIG. 1, not only the state value indicating the activation state of at least a part of the target devices but also the output amount of the target devices is calculated.

  As shown in FIG. 4, the state value indicating the output amount and the state value indicating the activation state for some of the target devices among the plurality of target devices are fixed. The state value indicating the output amount and the state value indicating the activation state for some target devices are fixed to the values obtained by solving the subproblem using the distance between the solution of the relaxation problem and the solution of the subproblem. (S302). For example, when there is a target device (A), a target device (B), a target device (C), and a target device (D), among them, the target device (A), the target device (B), and the target device (D). ), The state value indicating the output amount and the state value indicating the starting state are fixed to the values obtained by solving the subproblem using the solution distance.

  Next, the computer calculates a state value indicating the output amount and a state value indicating the activation state for the remaining target devices among the plurality of target devices (S304). According to the above example, the computer calculates the state value indicating the output amount and the state value indicating the activation state for the remaining target device (C). The slack problem calculation process is executed after the sub-problem calculation process to fix each state value of the sub-problems other than the problem to be solved as a slack problem and to calculate each non-fixed state value.

  FIG. 5 is a diagram showing an outline of the operation planning method described in FIG. First, the computer solves the relaxation problem in the original problem assuming that each target device can assume a continuous state, and calculates the continuous state of each target device. Then, the computer solves the partial problem with the power supplies G1 and G2 as the target devices. In particular, the computer uses the distance between the solution of the relaxation problem and the solution of the subproblem to fix the state value indicating the output amount and the state value indicating the activation state for the power source G1 and the power source G2. The computer then uses the power supply G3 as the remaining target device to solve the slack problem. As for the partial problem, the slack problem may be solved again for the power source 3 after solving the partial problem by solving all the target devices, that is, the power source G1, the power source G2, and the power source G3. On the other hand, the sub-problem calculation process may be omitted for the power supply 3 to be solved as the slack problem.

FIG. 6 is an example of each solution of the relaxation problem, the partial problem, and the slack problem in the operation planning method described in FIG. The state values indicating the output amounts of the power source G1, the power source G2, and the power source G3 are x1, x2, and x3, respectively. First, the relaxation solutions x1 hat, x2 hat, and x3 hat are calculated under the restriction state in the relaxation problem. Then, the solution x ′ ₁ , the solution x ′ ₂ , and the solution x ′ ₃ of the partial problem are calculated by solving the partial problem using the distance between the relaxation solution and the solution of the partial problem. However, as shown in FIG. 6, the solution obtained in the sub-problem calculation stage may not satisfy the limiting conditions. For example, the sum of solutions of each subproblem may exceed the load. In this case, no feasible solution is obtained. Therefore, by solving the slack problem for some of the plurality of subproblems, the feasible solution is processed.

  A process of selecting a target device to be a calculation target in the slack problem calculation stage from a plurality of target devices will be described. FIG. 7 is a flowchart showing an example of a process of selecting a target device as a calculation target in the slack problem calculation stage. The operation planning apparatus acquires characteristic information about a plurality of target apparatuses and stores the acquired characteristic information in the database. The characteristic information may include at least one of the output capacity, output efficiency, and responsiveness of the target device. Each time a new target device is provided, the new target device may transmit the characteristic information to the operation planning device via the communication network.

  When the operation plan is executed (S404: YES), the computer functioning as the operation plan device refers to the database and outputs based on at least one characteristic of the output capacity, output efficiency, and responsiveness of the target device. Then, the target device to be the calculation target in the slack problem calculation stage is selected (S406). Specifically, a target device whose output capacity of the target device is larger than a predetermined value, particularly a target device having the maximum output capacity may be selected. By selecting the target device with the maximum output capacity, the adjustment range is expanded in the slack problem calculation stage.

  Further, a target device whose output efficiency is lower than a predetermined value, particularly a target device whose output efficiency is the lowest may be selected. The output efficiency is the output efficiency with respect to the consumed energy. As shown in FIG. 6, the target device to be operated in the slack problem operation stage may be the solution farthest from the optimal solution. Therefore, by setting the target device having low output efficiency as the calculation target in the slack problem calculation stage, it is possible to reduce the influence caused by the possibility that the solution is far from the optimum solution.

  Further, a target device having a responsiveness higher than a predetermined value, particularly a target device having the fastest responsiveness among a plurality of target devices may be selected as a target device to be a calculation target in the slack problem calculation stage. The responsiveness may be the following speed with respect to the control command. Note that the processing similar to that in FIG. 1 may be executed after S406 in FIG.

  FIG. 8 is a diagram showing an example of a target device controlled by the operation planning device 300 according to one embodiment of the present invention. The plurality of target devices in this example are a plurality of power sources. Specifically, the plurality of target devices (304, 306) include one or more power generation devices 304 and one or more power storage devices 306. The power generation device 304 and the power storage device 306 supply electric power to the external device 302. The external device 302 may be a server.

  The operation planning apparatus 300 calculates a state value indicating the output amounts of the plurality of power generation devices 304 in the partial problem calculation step S108. The operation planning apparatus 300 may fix the output amount of the power generation device 304 and calculate a state value indicating the output amount of the power storage device 306 in the slack problem calculation step S112. When there are a plurality of power storage devices 306, one power storage device 306 may be selected as a target device to be calculated in the slack problem calculation stage. In this case, the output amounts of the power storage devices other than the power storage device 306 to be calculated in the slack problem calculation stage may be fixed in the slack problem calculation stage. The responsiveness of the power storage device 306 is faster than the responsiveness of the power generation device 304.

  As described above, in FIGS. 4 to 8, in a plurality of nodes, that is, a plurality of target devices, after fixing the output amount and the activation state of the target device other than the device to be the calculation target in the slack problem calculation stage, At the slack problem calculation stage, the state value indicating the output amount of the target device to be calculated and the state value indicating the starting state were calculated. However, the present invention is not limited to the process of selecting the calculation target in the slack problem calculation stage according to the node as described above.

FIG. 9 is a flowchart showing a second example of the slack problem calculation stage in FIG. The state value of the number includes a state value indicating the output amount of the target device and a state value indicating the activation state of the target device. Then, in the sub-problem calculation step S108 of FIG. 1, at least the state value y indicating the activation state of the plurality of target devices is calculated. In particular, at least the state value y indicating each activation state in all target devices is calculated. The formula, fix to the calculated value y _a bar in subproblems operation stages for the state value y _a which indicates the activation state of the plurality of target devices, calculates the output quantity x _a of each target device.

As shown in FIG. 9, for fixing the state value y _a bar indicating the activation state of the plurality of target devices. The status value y _a bar indicating the activation state for all of the target device may be fixed. The state value indicating the activation state of the plurality of target devices is fixed to a value obtained by solving the partial problem using the distance between the solution of the relaxation problem and the solution of the partial problem (step S502). For example, when there is a target device (A), a target device (B), a target device (C), and a target device (D), all the target devices (A), target devices (B), and target devices (C) among them. ), and the target device (D), a state value y _a bar indicating the activated state, is fixed to the value obtained by solving the subproblem using the distance of the solution.

Next, the remaining state value among the plurality of state values, that is, the state value indicating the output amount of the target device is calculated (S504). According to the above example, the output amount x _a of each target device is calculated for all of the plurality of target devices (A), target devices (B), target devices (C), and target devices (D). In the slack problem calculation process, a plurality of state values other than the state value indicating the output amount solved as the slack problem may be fixed, and the solution may be calculated using the state value indicating the output amount solved as the slack problem.

  As described above, as described with reference to FIGS. 4 to 8, the process of selecting the calculation target in the slack problem calculation stage according to the node and the slack according to the types of the plurality of state values as described with reference to FIG. There is a process of selecting a calculation target in the problem calculation stage. The possibility of obtaining a feasible solution may differ depending on such two types of processing. Therefore, when one of the processes is executed and the feasible solution of the slack problem is not obtained, the other process may be executed.

  FIG. 10 is a flowchart showing a third example of the slack problem calculation stage in FIG. The state value of the number includes a state value indicating the output amount of the target device and a state value indicating the activation state of the target device. First, the first slack problem solution process is executed (S602), and if there is a feasible solution (S604: YES), the slack problem solution process is completed, and the process directly returns. On the other hand, if there is no feasible solution, the second slack problem solution process is executed (step S606).

  The first slack problem solving process S602 may be the processes of S302 and S304 of FIG. 4, and the second slack problem solving process S606 may be the processes of S502 and S504 of FIG. In this case, in the first slack problem calculation step (S602), the output amounts of some target devices are fixed to the values obtained in the partial problem calculation stage, and the output amounts of the remaining target devices are indicated. The value is calculated (S302 and S304 in FIG. 4). Then, in the first slack problem calculation stage, if there is no state value that satisfies the constraint condition (S604: NO), in the second slack problem calculation stage (S606), a state value indicating the activation state of the plurality of target devices. Is fixed to the value obtained in the sub-problem calculation stage, and the state value indicating the output amount of the plurality of target devices is calculated (S502 and S504 in FIG. 9).

  On the other hand, the first slack problem solving process S602 may be the processes of S502 and S504 of FIG. 9, and the second slack problem solving process S606 may be the processes of S302 and S304 of FIG. In this case, in the first slack problem calculation step (S602), the state values indicating the activation states of the plurality of target devices are fixed to the values obtained in the partial problem calculation step, and the output amounts of the plurality of target devices are set. The state value shown is calculated (S502 and S504 of FIG. 9). Then, in the first slack problem calculation stage, when there is no state value that satisfies the constraint condition (S604: NO), in the second slack problem calculation stage (S606), the output amounts of some target devices are partially It is fixed to the value obtained in the problem calculation stage, and the state value indicating the output amount of the remaining target device is calculated (S302 and S304 in FIG. 4).

  According to such processing, even when the state value that satisfies the constraint cannot be obtained by one type of slack problem operation, the state value that satisfies the constraint can be obtained by another type of slack problem operation. There are cases. Therefore, it becomes easy to obtain a feasible solution.

  FIG. 11 is a diagram showing a configuration example of the operation planning apparatus 300. The operation planning apparatus 300 includes a problem acquisition unit 1102, a relaxation problem calculation unit 1104, a partial problem calculation unit 1106, and a slack problem calculation unit 1108. The relaxation problem calculator 1104, the partial problem calculator 1106, and the slack problem calculator 1108 may be a common calculator 1110.

  The problem acquisition unit 1102 executes the problem acquisition step S102 shown in FIG. The relaxation problem calculation unit 1104 executes the relaxation problem calculation step S104 and the first determination step S106 shown in FIG. The sub-problem calculation unit 1106 executes the processes up to S1108 and S1110 shown in FIG. With such a configuration, the operation planning method described in FIGS. 1 to 10 can be executed.

  FIG. 12 illustrates an example computer 2200 in which aspects of the present invention may be embodied in whole or in part. The program installed in the computer 2200 can cause the computer 2200 to function as an operation associated with an apparatus according to an embodiment of the present invention or one or more sections of the apparatus, or the operation or the one or more sections. Sections may be executed, and / or computer 2200 may execute methods according to embodiments of the invention or steps of such methods. Such programs may be executed by CPU 2212 to cause computer 2200 to perform certain operations associated with some or all of the blocks in the flowcharts and block diagrams described herein.

  The computer 2200 according to this embodiment includes a CPU 2212, a RAM 2214, a graphic controller 2216, and a display device 2218, which are interconnected by a host controller 2210. Computer 2200 also includes input / output units such as communication interface 2222, hard disk drive 2224, DVD-ROM drive 2226, and IC card drive, which are connected to host controller 2210 via input / output controller 2220. There is. The computer also includes legacy input / output units such as ROM 2230 and keyboard 2242, which are connected to input / output controller 2220 via input / output chip 2240.

  The CPU 2212 operates according to the programs stored in the ROM 2230 and the RAM 2214, thereby controlling each unit. The graphics controller 2216 obtains image data generated by the CPU 2212 in a frame buffer or the like provided in the RAM 2214 or itself, and causes the image data to be displayed on the display device 2218.

  The communication interface 2222 communicates with other electronic devices via the network. The hard disk drive 2224 stores programs and data used by the CPU 2212 in the computer 2200. The DVD-ROM drive 2226 reads a program or data from the DVD-ROM 2201 and provides the hard disk drive 2224 with the program or data via the RAM 2214. The IC card drive reads programs and data from the IC card and / or writes programs and data to the IC card.

  The ROM 2230 stores therein a boot program or the like executed by the computer 2200 at the time of activation, and / or a program depending on the hardware of the computer 2200. The input / output chip 2240 may also connect various input / output units to the input / output controller 2220 via parallel ports, serial ports, keyboard ports, mouse ports, etc.

  The program is provided by a computer-readable medium such as a DVD-ROM 2201 or an IC card. The program is read from the computer-readable medium, installed in the hard disk drive 2224, the RAM 2214, or the ROM 2230, which is also an example of the computer-readable medium, and executed by the CPU 2212. The information processing described in these programs is read by the computer 2200 and brings about the cooperation between the programs and the various types of hardware resources described above. An apparatus or method may be configured by implementing the manipulation or processing of information according to the use of the computer 2200.

  For example, when communication is executed between the computer 2200 and an external device, the CPU 2212 executes the communication program loaded in the RAM 2214, and based on the processing described in the communication program, executes the communication processing with respect to the communication interface 2222. You may order. The communication interface 2222 reads the transmission data stored in the transmission buffer processing area provided in the recording medium such as the RAM 2214, the hard disk drive 2224, the DVD-ROM 2201, or the IC card under the control of the CPU 2212, and the read transmission is performed. The data is transmitted to the network, or the received data received from the network is written in a reception buffer processing area or the like provided on the recording medium.

  Further, the CPU 2212 allows the RAM 2214 to read all or necessary portions of files or databases stored in an external recording medium such as a hard disk drive 2224, a DVD-ROM drive 2226 (DVD-ROM 2201), and an IC card. Various types of processing may be performed on the data in RAM 2214. The CPU 2212 then writes back the processed data to the external recording medium.

  Various types of information such as various types of programs, data, tables, and databases may be stored on the recording medium and processed. The CPU 2212 is described elsewhere in this disclosure for the data read from the RAM 2214, and performs various types of operations, information processing, conditional judgment, conditional branching, unconditional branching, and information retrieval specified by the instruction sequence of the program. Various types of processing may be performed, including / replacement, etc., and the result is written back to RAM 2214. Further, the CPU 2212 may search for information in files, databases, etc. in the recording medium. For example, when a plurality of entries each having the attribute value of the first attribute associated with the attribute value of the second attribute are stored in the recording medium, the CPU 2212 specifies the attribute value of the first attribute. That is, the entry that matches the condition is searched from the plurality of entries, the attribute value of the second attribute stored in the entry is read, and the entry is associated with the first attribute that satisfies the predetermined condition. The attribute value of the obtained second attribute may be acquired.

  The programs or software modules described above may be stored on a computer-readable medium on or near computer 2200. Further, a recording medium such as a hard disk or a RAM provided in a server system connected to a dedicated communication network or the Internet can be used as a computer readable medium, and thereby the program is provided to the computer 2200 via the network. To do.

  Although the present invention has been described using the embodiments, the technical scope of the present invention is not limited to the scope described in the above embodiments. It is apparent to those skilled in the art that various modifications and improvements can be added to the above-described embodiment. It is apparent from the scope of the claims that the embodiments added with such modifications or improvements can be included in the technical scope of the present invention.

  The execution order of each process such as operation, procedure, step, and step in the device, system, program, and method shown in the claims, the specification, and the drawings is, in particular, “before” or “prior to”. It should be noted that the output of the previous process can be realized in any order unless the output of the previous process is used in the subsequent process. Even if the operation flow in the claims, the specification, and the drawings is described by using “first,” “next,” and the like for convenience, it means that it is essential to carry out in this order. Not a thing.

  202 ... Straight line, 204, 206, 208 ... Mark, 300 ... Operation planning device, 302 ... External device, 304 ... Power generation device, 306 ... Power storage device, 1102 ... Problem Acquisition unit, 1104 ... Relaxation problem calculation unit, 1106 ... Partial problem calculation unit, 1108 ... Slack problem calculation unit, 1110 ... Calculation unit, 2200 ... Computer, 2201 ... DVD-ROM 2210 ... Host controller, 2212 ... CPU, 2214 ... AM, 2216 ... Graphic controller, 2218 ... Display device, 2220 ... Input / output controller, 2222 ... Communication interface, 2224 ... Hard disk drive, 2226 ... DVD-ROM drive, 2230 ... ROM, 240 ... input / output chip, 2242 ... keyboard

Claims (9)

An operation planning method for generating an operation plan of a plurality of target devices by a computer,
A problem acquisition step of acquiring an original problem for calculating a plurality of state values including discrete state values indicating discrete states of respective target devices, which can satisfy a predetermined purpose and constraint conditions;
A relaxation problem calculation step of solving a relaxation problem assuming that each target device can assume a continuous state in the original problem, and calculating a continuous state of each target device,
By dividing the original problem into a plurality of subproblems and solving the subproblem by using the distance between the solution of the relaxation problem and the solution of the subproblem, at least a part of the plurality of state values of discrete states A sub-problem calculation stage that calculates a value including a value,
Of some of the plurality of state values, a part of the plurality of state values is fixed to a value calculated in the sub-problem calculation step, and a slack problem using the constraint condition of the original problem is solved. And a slack problem calculation step of calculating the remaining state value. The plurality of state values include a state value indicating an output amount of the target device,
In the partial problem calculation step, a state value indicating the output amount of at least a part of the target device is calculated,
The operation planning method according to claim 1, wherein in the slack problem calculating step, the output amount of at least a part of the target devices is fixed, and the state value indicating the output amount of the remaining target devices is calculated. The operation planning method according to claim 2, wherein the target device for which a state value is calculated in the slack problem calculation step is selected based on at least one characteristic of the output capacity, output efficiency, and responsiveness of the target device. The plurality of target devices include one or more power generation devices and one or more power storage devices,
In the partial problem calculation step, a state value indicating the output amount of the power generator is calculated,
The operation planning method according to claim 2, wherein, in the slack problem calculation step, the output amount of the power generation device is fixed and a state value indicating the output amount of the power storage device is calculated. The plurality of state values include a state value indicating an output amount of the target device and a state value indicating an activation state of the target device,
In the sub-problem calculation step, calculating a state value indicating the activation state of the plurality of target devices,
The operation planning method according to claim 1, wherein, in the slack problem calculating step, a state value indicating an activation state of the plurality of target devices is fixed and a state value indicating an output amount of the plurality of target devices is calculated. The plurality of state values include a state value indicating an output amount of the target device and a state value indicating an activation state of the target device,
In the first slack problem calculation step, the output amount for some of the target devices is fixed to the value obtained in the partial problem calculation step, and the state value indicating the output amount of the remaining target devices is calculated. Then
In the first slack problem calculation step, if the state value that satisfies the constraint condition does not exist,
In the second slack problem calculation stage, the state value indicating the activation state of the plurality of target devices is fixed to the value obtained in the partial problem calculation stage, and the state value indicating the output amount of the plurality of target devices is fixed. The operation planning method according to claim 1. The plurality of state values include a state value indicating an output amount of the target device and a state value indicating an activation state of the target device,
In the first slack problem calculation stage, the state value indicating the activation state of the plurality of target devices is fixed to the value obtained in the partial problem calculation stage, and the state value indicating the output amount of the plurality of target devices is fixed. Is calculated,
In the first slack problem calculation step, if the state value that satisfies the constraint condition does not exist,
In the second slack problem calculation step, the output amount for some of the target devices is fixed to the value obtained in the partial problem calculation step, and the state value indicating the output amount of the remaining target devices is calculated. The operation planning method according to claim 1.   A program for causing a computer to execute the operation planning method according to any one of claims 1 to 7. An operation planning device for generating operation plans for a plurality of target devices,
A problem acquisition unit that can satisfy a predetermined purpose and constraint conditions, and that acquires an original problem for calculating a plurality of state values including a discrete state value indicating a discrete state of each target device,
A relaxation problem calculation unit that solves a relaxation problem assuming that each target device can take a continuous state in the original problem, and calculates a continuous state of each target device,
By dividing the original problem into a plurality of subproblems and solving the subproblem by using the distance between the solution of the relaxation problem and the solution of the subproblem, at least a part of the plurality of state values of discrete states A sub-problem calculation part that calculates a value including a value,
A part of the plurality of state values is fixed to a value calculated by the sub-problem calculation unit, and a slack problem using the constraint condition of the original problem is solved to obtain one of the plurality of state values. And a slack problem calculator that calculates the discrete values of the remaining state values.

Images