JP4879528B2 - Multi-lateral netting settlement method, system and program - Google Patents

Description

  The present invention relates to a multilateral netting settlement technique for performing netting settlement in which only a difference amount is settled by offsetting payment instructions of a plurality of users in a settlement system used for inter-bank settlement etc., in particular, a combination of payment instructions that can be netted The present invention relates to a technology that is effective when applied to a multilateral netting settlement technology that uses a plurality of strategies for determining the transaction.

  In payment systems used for interbank payments and cash management in corporate groups, reducing payment risk is an important issue.

  Settlement risk is the possibility that a settlement will not take place, resulting in damage. In the fund settlement system, the amount of settlement risk is determined by the time lag period from when the payment instruction is issued until the final settlement is made, and the amount of the transaction amount. It is measured by the total amount, the number of settlements, etc.

  Netting is one of the means for reducing settlement risk in a settlement system. Netting is a method in which payment instructions waiting in a queue are offset between payment system users and only the difference is settled. When it is performed between two users, it is called bilateral netting. When it is performed between three or more users, it is called multilateral netting.

  For example, Patent Document 1 discloses a conventional technique having an effect of reducing a balance of balance, which is a difference in user deposit / withdrawal. According to this method, since the deposit / withdrawal cancellation process is performed when the absolute value of the balance is reduced, it is possible to efficiently reduce the balance of the account.

  For example, Patent Document 2 discloses a conventional technique that has an effect of not increasing a user's debt beyond a threshold in netting. According to this method, it is possible to provide a system in which the debt of the user does not become excessive because the threshold stored in advance is compared with the balance when the netting settlement is made and it is determined whether or not the netting is performed. is there.

  In the netting, when the amount of multi-lateral netting payment is not enough for the user's collateral and payment cannot be made, the prior art for performing payment processing by bilateral netting and transferring funds between users is described in Patent Document 3. ing. This technology also has a function of grouping and managing users according to their creditworthiness, and has the effect of preventing the influence of one user's default on all other users.

  As a conventional technique of an algorithm for obtaining a combination capable of multilateral netting settlement from a waiting payment instruction, there is a method described in Non-Patent Document 1, for example. Here, we confirm the shortage of funds of each user when all the payment instructions on standby are settled, and if there is a shortage of funds, netting the payment instruction from which the user is the payment source from the end of the waiting order It describes how to remove it from the subject and repeat this until there is no shortage of funds.

JP 2000-322475 A Japanese Patent Laid-Open No. 11-3374 US Pat. No. 6,826,545 “Restructuring of Funds Payment System, such as Construction of Large Payment System” March 2004, National Bank Association

  When performing multi-lateral netting, several strategies are conceivable for determining a combination of payment instructions that can be netted from among the payment instructions waiting in the queue, that is, a netting solution. For example, there are a first-in first-out strategy in which payment is made first from a long waiting time, a strategy to maximize the total amount of payment instructions to be settled, and a strategy to maximize the number of payment instructions to be settled. These strategies are all effective strategies for reducing settlement risk. If one of them is adopted, settlement will be executed only based on a specific strategy, and it may happen that some payment instructions are not executed and continue to wait in the queue. For example, when a strategy for maximizing the total amount of payment instructions to be settled is employed, a situation where a small payment instruction waits in a queue for a long time can be considered. In such a case, the settlement risk due to the payment instruction that keeps waiting continues to increase.

  An object of the present invention is to provide a technique capable of solving the above-described problem, preventing a specific payment instruction from being kept waiting in a queue without being settled for a long time, and preventing an increase in settlement risk. It is in.

  The present invention, in a multilateral netting settlement system that settles only the difference by offsetting payment instructions of a plurality of users, calculates a score by evaluating netting solutions derived from a plurality of algorithms by a plurality of evaluation methods. The settlement is made by the netting solution having the highest total score.

  In the multilateral netting settlement system of the present invention, first, a netting solution that is a combination capable of multilateral netting settlement is obtained by a plurality of algorithms. That is, the payment instruction data and balance data that are waiting in the queue are aggregated to calculate the provisional balance, and when the provisional balance is negative, the payment instructions to be excluded are determined by multiple algorithms and are subject to aggregation The netting solution that makes the temporary balance non-negative is obtained for each algorithm.

  Next, for each of the obtained plurality of netting solutions, each solution is scored with a score evaluated by a plurality of evaluation methods. For example, for the plurality of netting solutions obtained, the ratio of the number of payment instructions subject to netting to the total number of payment instructions waiting in queue, or the total amount of payment instructions subject to netting relative to the total amount of payment instructions waiting in queue A score corresponding to the ratio or the like is calculated and stored in the storage device in association with the netting solution.

  Then, after summing up the scored scores for each netting solution and calculating the total score of each netting solution, the total scores are compared, and the netting solution with the highest calculated total score is determined as the solution to be settled. Then, the temporary balance calculated by the netting solution is stored in the balance database to execute settlement by the netting solution.

  According to the present invention, it is possible to prevent a specific payment instruction from continuing to wait in a queue without being settled for a long time, and to prevent an increase in settlement risk.

  Hereinafter, a multi-lateral netting settlement system according to an embodiment capable of further reducing settlement risk by using a plurality of strategies together will be described.

  FIG. 1 is a diagram showing an overall configuration of a multilateral netting settlement system 10 of the present embodiment. As shown in FIG. 1, the multilateral netting settlement system 10 of this embodiment includes a settlement management processing unit 101, a netting solution derivation processing unit 106, a netting solution evaluation processing unit 107, and a netting solution determination processing unit 109. is doing.

  The payment management processing unit 101 manages the entire payment processing of the multilateral netting payment system 10, and the balance after payment by the payment instruction of the netting solution determined by the netting solution determination processing unit 109 is transferred to the balance database 103 (for each user). It is a processing unit that performs settlement by the netting solution by storing.

  The netting solution derivation processing unit 106 is a processing unit that obtains a netting solution indicating a combination of payment instructions capable of multilateral netting settlement using a plurality of algorithms 1061 to 1063 and stores the netting solution in a storage device.

  The netting solution evaluation processing unit 107 scores each of the obtained plurality of netting solutions with scores evaluated by scoring 1071 to 1074 of a plurality of evaluation methods, and associates them with the netting solution in the storage device. A processing unit for storing.

  The netting solution determination processing unit 109 calculates the total score of each netting solution by counting the scored scores for each netting solution, and determines the netting solution having the highest calculated total score as the solution to be settled It is.

  A program for causing the multilateral netting payment system 10 to function as the payment management processing unit 101, the netting solution derivation processing unit 106, the netting solution evaluation processing unit 107, and the netting solution determination processing unit 109 is recorded on a recording medium such as a CD-ROM. Then, after being stored in a magnetic disk or the like, it is loaded into a memory and executed. The recording medium for recording the program may be a recording medium other than the CD-ROM. The program may be used by installing it from the recording medium into the information processing apparatus, or the program may be used by accessing the recording medium through a network.

  The multilateral netting settlement system 10 of this embodiment is connected to a user terminal 104 through a network. The user terminal 104 inputs a payment instruction from the user, and the multilateral netting settlement system 10 stores the data in the queue database (per user) 102. In addition, the multilateral netting settlement system 10 updates the balance database (per user) 103 using the input from the user at the user terminal 104. Here, the balance data in the balance database (per user) 103 is not allowed to take a negative value, that is, is managed not to allow overdraft. When there is a shortage, the user moves funds from account data outside the payment system connected via the network to ensure liquidity.

  An administrator can use the administrator terminal 105 to operate and set the multilateral netting settlement system 10.

  The entire settlement operation by this system is managed by the settlement management processing unit 101. The settlement management processing unit 101 performs multilateral netting settlement every predetermined time. During the execution of multilateral netting settlement, the settlement management processing unit 101 prohibits the updating of the queue database (for each user) 102 and the balance database (for each user) 103 using the user terminal 104 by the user.

  Multilateral netting settlement uses payment instruction data stored in a queue database (per user) 102 and balance data stored in a balance database (per user) 103, and uses a netting solution derivation processing unit 106, netting. The solution evaluation processing unit 107 and the netting solution determination processing unit 109 are sequentially executed to identify a netting solution, that is, a combination of payment instructions to be netted, and based on this, a queue database (per user) 102 and a balance database This is done by updating 103 (for each user).

  The netting solution derivation processing unit 106 derives three types of netting solutions using three types of algorithms 1061 to 1063. Details of netting solution derivation in each algorithm will be described later.

  The netting solution evaluation processing unit 107 uses four types of scoring 1071 to 1074 for each of the three types of netting solutions derived by the netting solution derivation processing unit 106, for a total of 12 scores. Calculate Details of the score calculation method in each scoring will be described later.

  The netting solution determination processing unit 109 is stored in the weight database 108 using the four types of scores calculated by the netting solution evaluation processing unit 107 for each of the three types of solutions obtained by the netting solution derivation processing unit 106. Based on the weight data, a weighted total score of each netting solution is calculated. Details of the total score calculation will be described later.

The weight data stored in the weight database 108 can be automatically adjusted by the settlement management processing unit 101 based on the historical data of payment instructions, netting solutions, and scores stored in the historical database 110. Details of this automatic adjustment will be described later.
The above is the overall configuration diagram according to the present embodiment.

  FIG. 45 is a diagram showing a schematic configuration of the multilateral netting settlement system 10 of the present embodiment. As shown in FIG. 45, the multilateral netting settlement system 10 includes hardware resources such as a CPU, a memory, a magnetic disk device, and the like, and software that executes processing of each processing unit and hardware resources cooperate with each other. Performs lateral netting settlement processing.

  Here, the multilateral netting settlement system 10 may be configured by a plurality of processing devices such as a processing device for deriving a netting solution for each algorithm or a processing device for evaluating a netting solution by scoring. good.

  FIG. 13 is a diagram showing a data format of the queue database (per user) 102 of this embodiment. In the queue database (for each user) 102 according to the present embodiment, one table is defined for each user, and FIG. One row of the table represents one payment instruction in which this user is the payment source. In the payee column, the name of the payee user for each waiting payment instruction is stored. In the amount column, the amount of the waiting payment instruction is stored. The input time column stores the input time for each waiting payment instruction.

  FIG. 14 is a diagram showing a data format of the balance database (per user) 103 of the present embodiment. As shown in FIG. 14, in the balance database (for each user) 103 according to the present embodiment, the user name is stored in the user column, and the balance amount of the user is stored in the amount column.

  FIG. 15 is a diagram showing a data format of the weight database 108 of the present embodiment. As shown in FIG. 15, in the weight database 108 of the present embodiment, a weight value is stored for each scoring. Here, scoring (1) to (4) shown in the scoring types in the figure indicate scoring 1071 to 1074, respectively.

  FIG. 16 is a diagram showing a data format of the historical database 110 of this embodiment. As shown in FIG. 16, in the historical database 110 of the present embodiment, one row of the table corresponds to each multilateral netting settlement. The date column stores the date on which the multilateral netting settlement was executed. The times column stores the number of multi-lateral netting settlements on that date. The multilateral netting settlement is specified by both the data stored in the date field and the time field.

  A total score field and four score fields are provided for each of the three types of algorithms, and score data is stored for each netting solution derived by each algorithm. Here, algorithms (1) to (3) in the figure indicate algorithms 1061 to 1063, respectively. Each score column corresponds to the scoring 1071 to 1074 in the order of the numbers, and the score calculated by each scoring is stored in the corresponding score column.

  Among the netting solutions derived by the three types of algorithms, the optimal solution, that is, the netting solution for which multilateral netting settlement has been executed is recorded. In FIG. 16, data when payment is executed by the netting solution derived by the algorithm (3) is stored.

  FIG. 2 is a diagram illustrating a flow of a procedure for executing the multilateral netting settlement according to the present embodiment. The settlement management processing unit 101 of this embodiment measures time (step 2201), determines whether a predetermined time (set to 30 minutes here) has elapsed since the start of the previous multilateral netting settlement (step 2202), and If 30 minutes have not elapsed, the process returns to time measurement (step 2201), and if 30 minutes have elapsed, multilateral netting settlement processing is started (step 2203). At that time, the start time and the number of nettings Are stored in the historical database 110 (step 2801).

  With the start of the multilateral netting payment process, the payment management processing unit 101 notifies the user terminal 104 of each user of the start of the process (step 2101). The settlement management processing unit 101 does not accept updating of queue data and balance data by the user during the execution of the multilateral netting settlement process.

  In response to the start of the multi-lateral netting settlement process, the netting solution derivation processing unit 106 receives the payment instruction data and the balance waiting in the queue from the queue database (for each user) 102 and the balance database (for each user) 103, respectively. Data is read (steps 2301 and 2302), and a program for executing the processing of algorithms 1061 to 1063 is called to derive solutions for multilateral netting settlement independently (steps 2401 to 2403). Details of the solution derivation method in each algorithm will be described later.

  The netting solution evaluation processing unit 107 calls a program for executing the processing of the four types of score calculation methods of scoring 1071 to 1074, and calculates scores for the three types of solutions independently derived in steps 2401 to 2403, respectively. (Steps 2501 to 2512). The solution score derived in step 2401 is calculated in steps 2501 to 2504, the solution score calculated in step 2402 is calculated in steps 2505 to 2508, and the solution score calculated in step 2403 is calculated in steps 2509 to 2512. Each calculated score is stored in the historical database 110 (step 2802). Details of the score calculation method in each scoring will be described later.

  In this way, all the 12 pieces of score data calculated by the four types of scoring for each of the netting solutions derived by the three types of algorithms are transferred to the netting solution determination processing unit 109, and the total score of each solution is calculated (step 2601). ). At that time, the settlement management processing unit 101 reads the score data of each solution derived from each algorithm from the historical database 110 and the data of the solution for which settlement has been executed (step 2803), and performs weight adjustment for total score calculation. (Step 2204), the weight database 108 is updated (Step 2701). Details of the weight adjustment method will be described later.

  The weight database 108 delivers the updated weight data to step 2601 in the netting solution determination processing unit 109 (step 2702). In step 2601, the netting solution determination processing unit 109 calculates the total score of each of the three types of solutions using the score data and the weight data. Each calculated total score of each solution is stored in the total score column for each algorithm in the historical database 110 (step 2804). The netting solution determination processing unit 109 determines the solution having the highest calculated weighted total score as the optimal solution (step 2602). The determined optimal solution is stored in the historical database 110 (step 2805). Details of the total score calculation method will be described later.

  The settlement management processing unit 101 receives the optimal solution data determined in this way from the netting solution determination processing unit 109, and updates the queue database (per user) 102 and the balance database (per user) 103 based on the optimal solution data ( In step 2303 and step 2304, settlement is executed (step 2205). Update information of the queue database (per user) 102 and the balance database (per user) 103 is notified to the user terminal 104 (step 2102).

Upon receiving the end of the multilateral netting settlement process (step 2206), the settlement management processing unit 101 returns to the time measurement in step 2201, and repeats step 2201 and step 2202 until the next multilateral netting settlement starts.
The above is description of the flowchart of the procedure which performs the multilateral netting payment which concerns on this embodiment.

  Next, the netting solution derivation flow of the algorithms 1061 to 1063 of the netting solution derivation processing unit 106 executed in steps 2401 to 2403 in FIG. 2 will be described.

  FIG. 3 is a diagram showing a flow of a netting solution derivation procedure common to the three types of algorithms of the present embodiment. In the netting solution derivation procedure of FIG. 3, a method specific to each algorithm is used in determining the payment instruction to be excluded in step 307. Details of this will be described later.

  As shown in FIG. 3, in the netting solution derivation procedure of this embodiment, the payment instruction data 301 and balance data 302 are read, and all the read payment instruction data are specified as netting target payment instruction data (step 303). A temporary balance for each user is calculated from the netting target payment instruction data and the balance data for each user (step 304). The provisional balance is obtained by subtracting the total payment instruction amount of the user from the balance for each user and adding the total payment instruction amount to the user. Then, it is determined whether or not all the temporary balances for each user are non-negative (step 305).

  If the temporary balance for each user is all non-negative, the multilateral netting settlement will be established by updating the balance data for each user to the temporary balance data and deleting the payment instructions subject to netting from each user's queue data. The netting target payment instruction at this time is specified as the solution of the multilateral netting settlement (step 310).

  If there is a user whose provisional balance is not non-negative, the balance data for each user is not allowed to be overwritten, so multi-lateral netting is not established in this netting target payment instruction. Therefore, one specific payment instruction is excluded from the netting target payment instructions. At that time, a payment instruction to be excluded is selected from the payment instructions waiting in the use destination queue with the smallest provisional balance.

  The calculated temporary balance value is compared to identify the usage destination having the smallest temporary balance (step 306), and a method corresponding to each algorithm is selected from the payment instructions waiting in the usage destination queue. The payment instructions to be excluded are identified (step 307). The payment instruction specified in step 307 is excluded and the netting target payment instruction is updated (step 308), and it is determined whether or not the number of netting target payment instructions is 0 (step 309). If it is not 0 (step 310), if it is not 0, the process returns to step 304 to calculate the provisional balance for the updated netting target payment instruction, and this procedure is repeated until the solution is specified.

  The above is the description of the netting solution derivation flow of the algorithms 1061 to 1063 of the netting solution derivation processing unit 106 executed in steps 2401 to 2403 in FIG.

Next, a method for specifying an exclusion target payment instruction specific to each of the aforementioned algorithms will be described below.
The algorithm 1061 uses a first-in first-out strategy. That is, when the payment instruction to be excluded from the payment instructions subject to netting among the payment instructions waiting in the waiting queue with the negative and smallest temporary balance at step 306 in FIG. Use a method that excludes orders. This is a method according to the method exemplified as the conventional technique related to netting solution derivation.

  The algorithm 1062 uses a settlement amount maximization strategy. That is, when specifying a payment instruction to be excluded from the payment instructions subject to netting among the payment instructions waiting in the user queue with the negative and minimum temporary balance at step 306 in FIG. 3, the payment instruction with the smallest amount is selected. Use the method to be excluded. If there are a plurality of payment instructions with the smallest amount, the order of the queue is excluded from the subsequent payment instructions.

The algorithm 1063 uses a settlement number maximization strategy. That is, when the payment instruction to be excluded from the payment instructions subject to netting among the waiting payment instructions in the waiting queue of the user whose negative balance is negative and smallest in step 306 in FIG. Exclude from empty payment instructions. When there are a plurality of payment instructions in which the payee's queue is empty, the order of the queue is excluded from the later payment instructions.
This completes the description of the method for specifying the exclusion target payment instruction specific to each algorithm.

  Next, the scoring 1071 evaluation method in the netting solution evaluation processing unit 107 in FIG. 1 executed in step 2501, step 2505, and step 2509 in FIG. 2 will be described.

  In scoring 1071, the number of payment instructions settled by multilateral netting is evaluated. As the number of settlements increases, the number of payment instructions that remain in the queue continues to decrease, and settlement risk is expected to decrease. In order to evaluate this, the score defined by Equation 1 is calculated.

[Equation 1]
Number of payment instructions to be netted / total number of payment instructions waiting in queue The higher the number of payment instructions that are netted, the higher the score, and the maximum score is 1.0 when all payment instructions in queue are netted . The above is the description of the scoring 1071.

  Next, the scoring 1072 evaluation method in the netting solution evaluation processing unit 107 in FIG. 1 executed in Step 2502, Step 2506, and Step 2510 in FIG. 2 will be described. In scoring 1072, the amount of the payment instruction settled by multilateral netting is evaluated. The larger the total amount to be settled, the smaller the total amount of payment instructions waiting in the queue, and the settlement risk is expected to decrease. In order to evaluate this, the score defined by Equation 2 is calculated.

[Equation 2]
Total amount of payment instructions to be netted / total amount of payment instructions waiting in queue The higher the total amount of payment instructions to be netted, the higher the score, and a maximum score of 1.0 when all payment instructions in queue are netted A perfect score. The above is the description of the scoring 1072.

  Next, the scoring 1073 evaluation method in the netting solution evaluation processing unit 107 in FIG. 1 executed in steps 2503, 2507, and 2511 in FIG. 2 will be described. In scoring 1073, the waiting time of the payment instruction settled by multilateral netting is evaluated. The larger the total waiting time for payment instructions to be settled, the smaller the total waiting time for payment instructions waiting in the queue, and the settlement risk is expected to decrease. In order to evaluate this, the score defined by Equation 3 is calculated.

[Equation 3]
Total payment instruction waiting time to be netted / total payment instruction waiting time waiting in queue The higher the net payment instruction waiting time, the higher the score, the highest when all waiting payment instructions in the queue are netted The score is a perfect score of 1.0. The above is the description of the scoring 1073.

  Next, the scoring 1074 evaluation method in the netting solution evaluation processing unit 107 in FIG. 1 executed in step 2504, step 2508, and step 2512 in FIG. 2 will be described. In scoring 1074, the payment instruction exposure settled by multilateral netting is evaluated. Here, the exposure of each payment instruction is defined as the amount of each payment instruction multiplied by the waiting time of each payment instruction. The larger the total payment instruction exposure that is settled, the smaller the total payment instruction exposure that will remain in the queue, and the settlement risk is expected to decrease. In order to evaluate this, the score defined by Equation 4 is calculated.

[Equation 4]
Total payment instruction exposures to be netted /
Total payment instruction exposures waiting in queue The higher the netted payment instruction exposure, the higher the score, and a maximum score of 1.0 when all waiting payment instructions are netted.

As described above, in this embodiment, the score is obtained by scoring 1071 to 1074 corresponding to the algorithms 1061 to 1063.
In other words, the algorithm 1061 uses the first-in first-out strategy to specify the exclusion target, and thus the payment instruction with a long waiting time is set as the netting target. The scoring 1073 and the scoring 1074 correspondingly wait for the payment instruction to be settled The score is obtained by a formula including time. Further, in the algorithm 1062, a payment instruction with a large amount is set as a netting target by specifying an exclusion target in the settlement amount maximization strategy, and the scoring 1072 and scoring 1074 correspondingly indicate the amount of the payment instruction to be settled. The score is obtained by a mathematical formula that includes it. Further, in the algorithm 1063, by specifying the exclusion target by the settlement number maximization strategy, the payment instruction in which the payee's queue is not empty is set as the netting target, and the scoring 1071 correspondingly corresponds to the number of payment instructions to be settled. The score is obtained by a mathematical formula including.
The above is the description of the scoring 1074.

Next, the total score calculation method executed in step 2601 of FIG. 2 will be described below. For the solutions derived by the algorithm 1061, the scores calculated in the scoring 1071 to 1074 are X ₁₁ , X ₁₂ , X ₁₃ , and X ₁₄ , respectively. Similarly, the scores for the solutions derived by the algorithm 1062 are X ₂₁ , X ₂₂ , X ₂₃ , and X ₂₄ , respectively, and the scores for the solutions derived by the algorithm 1063 are X ₃₁ , X ₃₂ , X ₃₃ , and X ₃₄ , respectively. . The weights set for the scoring 1071 to 1074 are w ₁ , w ₂ , w ₃ , and w ₄ , respectively. The total score X _1t for the solution derived by the algorithm 1061, the total score X _2t for the solution derived by the algorithm 1062, and the total score X _3t for the solution derived by the algorithm 1063 are defined by the following equation (5).

[Equation 5]
X _1t = w ₁ × X ₁₁ + w ₂ × X ₁₂ + w ₃ × X ₁₃ + w ₄ × X ₁₄
X _2t = w ₁ × X ₂₁ + w ₂ × X ₂₂ + w ₃ × X ₂₃ + w ₄ × X ₂₄
X _3t = w ₁ × X ₃₁ + w ₂ × X ₃₂ + w ₃ × X ₃₃ + w ₄ × X ₃₄
By setting the total weight to 1.0, each total score becomes a perfect score of 1.0. If the weights are all equal to 0.25, there are four types of scoring: evaluation of the number of payment instructions for scoring, evaluation of the amount of payment instructions for scoring, evaluation of waiting time for payment instructions for scoring, and payment for scoring The evaluation of instruction exposure will be evaluated equally. By setting the weight value to be large by specific scoring, an evaluation is made with an emphasis on scoring evaluation in which the weight value is increased by the total score at this time. For example, as an extreme example, if w ₁ is 1.0 and all others are 0, X _t = X ₁ , and the evaluation based on the total score is made only by evaluating the number of payment instructions for scoring. A total score is calculated for each of the three types of solutions derived by the three types of algorithms, and it is determined which algorithm is used as the optimum solution depending on the magnitude. The above is the description of the total score calculation method executed in step 2601 of FIG.

  Next, the weight adjustment executed in step 2204 of FIG. 2 will be described below. The weight adjustment is based on the average score of multi-lateral netting settlement in the period preset by the administrator of the score calculated by each scoring in the netting solution by the algorithm adopted in the previous netting stored in the historical database 110. Made. The scoring weight having the lowest average score is increased by a value preset by the administrator, and the weights of the other algorithms are evenly decreased so that the total weight becomes 1.

  A specific example of the weight adjustment procedure will be described using the past data illustrated in FIG. Assume that in the execution of the third multilateral netting settlement on July 1, 2005, the administrator uses the average of the last two multilateral netting settlements. According to the optimal solution column of FIG. 16, the solution by the algorithm (3) which is the algorithm 1063 is adopted and executed as the optimal solution for the netting settlement immediately before.

  In the example of FIG. 16, the score by scoring of the solution by the algorithm 1063, which is the immediately preceding optimal solution, is 0.5 for the previous multilateral netting settlement score (1) for the scoring 1071. Is 0.7, the score (2) of the previous multi-lateral netting settlement is 0.6, the score (2) of the previous two times is 0.5, and the score of the previous multi-lateral netting settlement is 0.5 for the scoring 1073 (3) is 0.6, the previous score (3) is 0.4, and for scoring 1074, the previous multilateral netting settlement score (4) is 0.8, and the previous score (4) is 0.6. The average score of each scoring in the last two multilateral netting settlements is 0.6 for scoring 1071. Ring 1072 is 0.55, scoring 1073 is 0.5, scoring 1074 0.7, has an average score minimized by scoring 1073.

In response, for example, when the weight adjustment range is set to 0.09 by the administrator, the weight of the scoring 1073 is increased by 0.09, and the weights of the scoring 1071, the scoring 1072, and the scoring 1074 are equalized to 0.03. The weight is adjusted by decreasing each time. As described above, in the present embodiment, the score is obtained by scoring 1071 to 1074 corresponding to the algorithms 1061 to 1063, and thus the algorithm for deriving the optimal solution is changed by adjusting the weight for the scoring. Can do.
The above is the description of the weight adjustment executed in step 2204 in FIG.

  FIG. 4 is a diagram showing a screen of the user terminal 104 of the present embodiment. The user terminal 104 is prepared for each user, and the screen of FIG. 4 is also prepared for each user. Using this screen, the user can enter payment instructions. When a payee is entered in the payee column 401, a payment amount is entered in the amount column 402, and the execute button 403 is pressed, the multilateral netting settlement system 10 accepts the payment instruction input and receives the queue database (user The queue data of the user 102 is updated, and the queue data status of the user is displayed in the waiting queue payment instruction list 404.

  Using the screen shown in FIG. 4, the user can deposit and withdraw balances. The current balance column 405 displays the current balance of the user. By inputting the deposit amount in the deposit column 406 and pressing the deposit execution button 407, the user can deposit the balance. Further, the user can withdraw the balance by inputting the withdrawal amount in the withdrawal column 408 and pressing the withdrawal execution button 409. The multilateral netting settlement system 10 accepts the input of the deposit amount or withdrawal amount, and updates the balance data in the balance database (for each user) 103. The above is the description of the screen of the user terminal 104 in FIG.

  FIG. 5 is a diagram showing a screen of the administrator terminal 105 of this embodiment. The administrator uses this screen to set up the system. In FIG. 5, the current setting is displayed as an initial state, and the setting is updated by inputting the setting in each input cell and pressing the update button 509. When the input information is restored, a cancel button 510 is pressed. The multilateral netting settlement system 10 receives the input of those settings and updates the setting data in the memory and the magnetic disk device.

  In the multilateral netting settlement execution interval column 501, an interval for performing multilateral netting settlement is input. FIG. 5 shows a state where 30 minutes is set. The weight value column 502 displays the current weight data value. The radio button 503 is selected when automatic weight adjustment is performed, and the radio button 504 is selected when the weight is manually adjusted. When the radio button 504 is selected, the weight data can be manually input in the weight value column 502, and the weight data is updated by pressing the update button 509 after inputting the weight of each scoring in the weight value column 502. The

  When the radio button 503 is selected, automatic weight adjustment is performed based on the setting data input from the automatic adjustment execution interval field 505 to the used historical data period field 508. In an automatic adjustment execution interval column 505, an interval for executing automatic weight adjustment is set in units of business days and netting times. FIG. 5 shows an example in which settings for automatic weight adjustment are input every business day, every time, that is, every business day. In the automatic adjustment width column 506, the width of automatic weight adjustment is set. FIG. 5 shows an example in which 0.09 is input.

In the used historical data period columns 507 and 508, the period of historical data used when executing weight adjustment is set. The year before column of the use historical data period column 507 sets how many years ago data is used on the basis of the executed multilateral netting settlement. The month before column of the use historical data period column 507 sets how many months ago the data is used based on the multilateral netting settlement to be executed. In the column before the business day of the use historical data period column 507, how many business days before the data is used is set based on the executed multilateral netting settlement. Similarly, the year before column of the use historical data period column 508 sets how many years ago data is used on the basis of the executed multilateral netting settlement. The month before column of the use historical data period column 508 sets how many months ago data is used on the basis of the executed multilateral netting settlement. In the usage historical data period column 508, the number of business days before the data is set based on the multi-lateral netting settlement to be executed.
FIG. 5 shows an example of multi-lateral netting settlement to be executed and setting using data from the previous two times to the previous business day of the same month in the same year. The above is the description of the screen of the administrator terminal 105 in FIG.

  FIG. 6 is a diagram showing a notification screen for starting the multilateral netting settlement notified to the user terminal in step 2101 of FIG. 2 of the present embodiment. The multilateral netting settlement system 10 of the present embodiment continues to display the message screen of FIG. 6 on the front of the screen of FIG. 5 during execution of multilateral netting settlement. During this time, the input on the screen of FIG. 5 is not accepted.

  FIG. 7 is a diagram showing a notification screen of multi-lateral netting settlement completion and data update notified to the user terminal in step 2102 of FIG. 2 of the present embodiment. The multilateral netting settlement system 10 of the present embodiment switches the message screen of FIG. 6 to the message screen of FIG. 7 upon completion of settlement, and when the press of the OK button on the message screen of FIG. 7 is accepted, the screen of FIG. Returning to the screen of FIG. 5, the user can confirm the contents of the executed settlement, and can input a new payment instruction and update the balance. This completes the explanation of the message screens of FIGS.

  In the following, an example in which the specified multilateral netting solution changes by the three types of algorithms described above will be described, and an example in which the score calculated by the four types of scoring described above will change for each solution. To do. The state immediately before the multi-lateral netting settlement for six users of user A, user B, user C, user D, user E, and user F is shown below.

  FIG. 17 is a diagram illustrating an example of a payment waiting instruction for each user according to the present embodiment. FIG. 17 shows the payment instructions waiting in each user's queue, and the queues of user C, user E, and user F are empty.

  Although the input time is stored in the waiting time column of FIG. 17 as shown in FIG. 14 in the original database structure, the input time is subtracted from the multilateral netting settlement start time here for easy explanation. The waiting time in minutes.

  FIG. 8 is a schematic diagram showing the waiting instruction in queue waiting shown in FIG. 17 of the multilateral netting settlement system 10 of the present embodiment. In FIG. 8, the payment instruction between each user waiting in the queue is represented by an arrow and its amount. For example, the arrow from A to B indicates the payment instruction for the amount 250 from user A to user B. Represents.

  FIG. 18 is a diagram showing the balance of each user according to the present embodiment. FIG. 18 shows that the balance of user A is “250” and the balance of other users is “0”.

  First, a case where a netting solution is obtained based on the algorithm 1061 in the multilateral netting settlement system 10 of the present embodiment will be described.

  FIG. 19 is a diagram showing the transition of the temporary balance according to the algorithm 1061 of this embodiment. FIG. 19 shows the transition of the provisional balance when the netting solution is obtained by the algorithm 1061 from the states shown in FIGS. 17, 18, and 8.

  FIG. 20 is a diagram illustrating a change in the waiting instruction for waiting in a queue according to the present embodiment. FIG. 20 shows changes in the netting solution when the netting solution is obtained by the algorithm 1061 from the states shown in FIGS. 17, 18, and 8.

  When the netting target payment instructions are excluded in order to obtain a solution according to the algorithm 1061, the payment instructions erased by the disappearing lines shown in (1) to (3) of FIG. 20 are excluded in the order of the numbers in the figure. The FIG. 19 shows the transition of the temporary balance every time one payment instruction is excluded. The user whose provisional balance is negative and minimum in the first stage is User B, and the payment instruction to be excluded from the netting target is selected from User B's queued payment instruction.

  Since the algorithm 1061 excludes the first-in first-out strategy, that is, the payment instruction at the end of the queue as described above, the payment instruction of (1) in FIG. 20 is excluded. As a result, the temporary balance changes to the value shown in the “B → D deleted” column of FIG. At this time, the usage destination with the negative and minimum temporary balance is the usage destination B, and the payment instruction shown in (2) of FIG. As a result, the provisional balance changes to the value shown in the “B → C deleted” column of FIG. At this time, the usage destination with the negative and minimum provisional balance is the usage destination D, and the payment instruction shown in FIG. As a result, the temporary balance changes to the value shown in the column “D → E deleted” in FIG. 19, the temporary balances of all the users are non-negative, and if the balance is updated to this temporary balance value, Multi-lateral netting is established by the payment instructions that are not excluded, that is, two payment instructions that are not deleted by the erasing line in FIG. The payment instruction for this solution is schematically shown in FIG.

  FIG. 9 is a payment instruction showing a solution by the algorithm 1061 of the multilateral netting settlement system 10 of the present embodiment. In FIG. 9, a dotted line indicates a payment instruction excluded from the netting target by this solution, and a solid line indicates a payment instruction that is a netting target. FIG. 21 shows score values for each of the four types of scoring for this solution.

FIG. 21 is a diagram showing a solution score by the algorithm 1061 of this embodiment. The score (1) in FIG. 21 is the result of calculating 2/5 based on the calculation formula shown in Equation 1. The score (2) in FIG. 21 is the result of calculating (250 + 150) / (250 + 150 + 200 + 150 + 40) based on the calculation formula shown in Equation 2. The score (3) in FIG. 21 is the result of calculating (20 + 19) / (20 + 19 + 10 + 5 + 4) based on the calculation formula shown in Equation 3. The score (4) in FIG. 21 is (20 × 250 + 19 × 150) / (250 × 20 + 150 × 19 + 200 × 10 + 150 × 5 + 40 × 4) based on the calculation formula shown in Equation 4. Is the result of calculating.
The above is an explanation of a case where a solution is obtained based on the algorithm 1061.

  Next, a case where a solution is obtained based on the algorithm 1062 in the multilateral netting settlement system 10 of the present embodiment will be described.

  FIG. 22 is a diagram showing the transition of the temporary balance according to the algorithm 1062 of the present embodiment. FIG. 22 shows the transition of the provisional balance when the netting solution is obtained by the algorithm 1062 from the states shown in FIGS. 17, 18, and 8.

  FIG. 23 is a diagram showing a change in the waiting instruction for waiting in the queue according to the present embodiment. FIG. 23 shows changes in the netting solution when the netting solution is obtained by the algorithm 1062 from the states shown in FIGS. 17, 18, and 8.

  When the netting target payment instructions are excluded in order to obtain a solution according to the algorithm 1062, the payment instructions erased by the disappearing lines shown in (1) to (3) of FIG. 23 are excluded in this order. FIG. 22 shows the transition of the provisional balance every time one payment instruction is excluded. The user whose provisional balance is negative and minimum in the first stage is User B, and the payment instruction to be excluded from the netting target is selected from User B's queued payment instruction.

  In the algorithm 1062, as described above, the payment amount maximization strategy, that is, the payment instruction with the smallest amount in the payment instruction waiting in the queue (if there are a plurality of corresponding payment instructions, the payment in the order of the queue later) (Instruction) is excluded, so the payment instruction of (1) in FIG. 23 is excluded. As a result, the provisional balance changes to the value shown in the column “B → D deleted” in FIG. At this time, the usage destination with the negative and minimum temporary balance is the usage destination B, and the payment instruction (2) in FIG. As a result, the temporary balance changes to the value shown in the “B → F erase” column in FIG. At this time, the usage destination having the negative and minimum provisional balance is the usage destination D, and the payment instruction shown in FIG. As a result, the temporary balance changes to the value shown in the column “D → E deleted” in FIG. 22, the temporary balances of all the users are non-negative, and if the balance is updated to this temporary balance value, Multilateral netting is established by the payment instructions that are not excluded, that is, by two payment instructions that are not deleted by the erasing line in FIG. The payment instructions for this solution are schematically shown in FIG.

  FIG. 10 is a payment instruction showing a solution by the algorithm 1062 of the multilateral netting settlement system 10 of the present embodiment. In FIG. 10, a dotted line indicates a payment instruction excluded from the netting target by this solution, and a solid line indicates a payment instruction that is a netting target. FIG. 24 shows score values obtained by the four types of scoring for this solution.

FIG. 24 is a diagram showing a solution score by the algorithm 1062 of this embodiment. The score (1) in FIG. 24 is the result of calculating 2/5 based on the calculation formula shown in Equation 1. The score (2) in FIG. 24 is the result of calculating (250 + 200) / (250 + 150 + 200 + 150 + 40) based on the calculation formula shown in Equation 2. The score (3) in FIG. 24 is the result of calculating (20 + 10) / (20 + 19 + 10 + 5 + 4) based on the calculation formula shown in Equation 3. The score (4) in FIG. 24 is based on the calculation formula shown in Equation 4, (20 × 250 + 10 × 200) / (250 × 20 + 150 × 19 + 200 × 10 + 150 × 5 + 40 × 4) Is the result of calculating.
The above is a description of a case where a solution is obtained based on the algorithm 1062.

  Next, a case where a netting solution is obtained based on the algorithm 1063 in the multilateral netting settlement system 10 of the present embodiment will be described.

  FIG. 25 is a diagram showing the transition of the temporary balance according to the algorithm 1063 of this embodiment. FIG. 25 shows the transition of the provisional balance when the netting solution is obtained by the algorithm 1063 from the states shown in FIGS. 17, 18, and 8.

  FIG. 26 is a diagram showing a change of the waiting instruction in queue waiting according to the present embodiment. FIG. 26 shows a change in the netting solution when the netting solution is obtained by the algorithm 1063 from the states shown in FIGS. 17, 18, and 8.

  When the netting target payment instructions are excluded in order to obtain a solution according to the algorithm 1063, the payment instructions erased by the disappearing lines shown in (1) and (2) of FIG. 26 are excluded in this order. FIG. 25 shows the transition of the provisional balance every time one payment instruction is excluded. The user whose provisional balance is negative and minimum in the first stage is User B, and the payment instruction to be excluded from the netting target is selected from User B's queued payment instruction.

  In the algorithm 1063, as described above, the number of settlement maximization strategy, that is, a payment instruction in which the payee's queue is empty among the payment instructions waiting in the queue (in the case where there are a plurality of payment instructions, the order of the queue is later) Payment instruction) is excluded, the payment instruction (1) in FIG. 26 is excluded. As a result, the provisional balance changes to the value shown in the column “B → C deleted” in FIG. At this time, the usage destination with the negative and minimum temporary balance is the usage destination B, and the payment instruction shown in (2) of FIG. As a result, the temporary balance changes to the value shown in the column “B → F is deleted” in FIG. 25, the temporary balances of all users are non-negative, and if the balance is updated to this temporary balance value, Multi-lateral netting is established by the payment instructions that are not excluded, that is, the three payment instructions that are not deleted by the disappearing line in FIG. The payment instructions for this solution are schematically shown in FIG.

  FIG. 11 is a payment instruction showing a solution by the algorithm 1063 of the multilateral netting settlement system 10 of the present embodiment. In FIG. 11, a dotted line indicates a payment instruction excluded from the netting target by this solution, and a solid line indicates a payment instruction that is a netting target. FIG. 27 shows score values obtained by the four types of scoring for this solution.

FIG. 27 is a diagram showing a solution score by the algorithm 1062 of this embodiment. The score (1) in FIG. 27 is the result of calculating 3/5 based on the calculation formula shown in Equation 1. The score (2) in FIG. 27 is the result of calculating (250 + 150 + 40) / (250 + 150 + 200 + 150 + 40) based on the calculation formula shown in Equation 2. The score (3) in FIG. 27 is the result of calculating (20 + 5 + 4) / (20 + 19 + 10 + 5 + 4) based on the calculation formula shown in Equation 3. The score (4) in FIG. 27 is based on the calculation formula shown in Equation 4, and (20 × 250 + 5 × 150 + 4 × 40) / (250 × 20 + 150 × 19 + 200 × 10 + 150 × 5 + This is the result of calculating 40 × 4).
The above is a description of a case where a solution is obtained based on the algorithm 1063.

  21, 24, and 27, the highest score (1) by scoring 1071 is the solution by algorithm 1063, and the highest score (2) by scoring 1072 is by the solution by algorithm 1062. It can be seen that the highest score (3) by scoring 1073 and the highest score (4) by scoring 1074 are solutions by the algorithm 1061. In this way, even in the same payment instruction and balance situation, the solution derived by the algorithm used is different, and the score by each scoring in each solution is also different. It can be seen that the method of selecting a solution that is considered effective for reduction is effective.

  The above is an explanation of an example in which the specified multilateral netting solution is changed by the three types of algorithms described above, and an example in which the score calculated by the four types of scoring described above is changed for each solution. .

Next, the effect of the weight adjustment executed in step 2204 in FIG. 2 will be described below with an example.
It is assumed that the payment instruction shown in FIG. 28 is input immediately after the start of use of the settlement system, and the balance at that time is in the state shown in FIG.

FIG. 28 is a diagram showing an example of payment instruction input according to the present embodiment. FIG. 29 is a diagram showing the balance state of the present embodiment. The “pay → receipt” column in FIG. 28 represents the user names of the payment source and the reception destination. For example, A → B indicates a payment instruction paid from the user A to the user B. The amount column shows the amount of each payment instruction. FIG. 29 shows the balance status for each user.
The queue at this time is in the state shown in FIG. This situation is schematically shown in FIG.

  FIG. 30 is a diagram illustrating a queue state when the payment instruction of FIG. 28 of the present embodiment is input. FIG. 30 shows the queue state of each user when the payment instruction shown in FIG. 28 is input.

  FIG. 12 is a schematic diagram showing the payment instruction of FIG. 30 of the multilateral netting settlement system 10 of the present embodiment. In FIG. 12, the payment instruction between each user waiting in the queue of FIG. 30 is represented by an arrow and its amount. For example, the arrow from A to B indicates the amount 100 from user A to user B. Represents the payment instruction.

  For simplicity, only two types of algorithms will be used in the following, the algorithm 1061 settlement number maximization strategy and the algorithm 1062 settlement amount maximization strategy, and scoring is performed by scoring the number of scoring 1071s. Consider a case where only two types of those that evaluate the amount of 1072 are used.

  Assuming that the initial weight data has the value shown in FIG. 31, when the first multilateral netting settlement is executed, the transition of the temporary balance by the algorithm 1061 is shown in FIG. 32 and the transition of the temporary balance by the algorithm 1062 is 33, the algorithm 1061 has a score shown in FIG. 34, and the algorithm 1062 has a score shown in FIG.

  FIG. 31 is a diagram showing initial weight data of the present embodiment. FIG. 31 shows that the initial weight data of scoring 1071 and 1072 is 0.5, and score (1) and score (2) indicate scoring 1071 and scoring 1072, respectively.

  FIG. 32 is a diagram showing the transition of the temporary balance according to the algorithm 1061 of the present embodiment. FIG. 32 shows the transition of the temporary balance when the netting solution is obtained by the algorithm 1061 from the states shown in FIGS. 29, 30 and 12.

  FIG. 33 is a diagram showing the transition of the temporary balance according to the algorithm 1062 of the present embodiment. FIG. 33 shows the transition of the provisional balance when the netting solution is obtained by the algorithm 1062 from the states shown in FIGS. 29, 30, and 12.

  FIG. 34 is a diagram showing a score of the algorithm 1061 of the present embodiment. The scores (1) and (2) in FIG. 34 represent the results calculated based on the formulas (1) and (2) when the netting solution is obtained by the algorithm 1061 as shown in FIG.

  FIG. 35 is a diagram showing a score of the algorithm 1062 of this embodiment. The scores (1) and (2) in FIG. 35 represent the results calculated based on the equations (1) and (2) when the netting solution is obtained by the algorithm 1062 as shown in FIG.

  FIG. 36 is a diagram showing the total score of each algorithm of the present embodiment. In FIG. 36, the total score for each algorithm when the score of each algorithm is the value of FIG. 34 and FIG. 35 is represented.

  FIG. 37 is a diagram showing the balance update status of this embodiment. FIG. 37 shows the balance status when a settlement is made by a solution by the algorithm 1061.

  FIG. 38 is a diagram showing the status of the queue of this embodiment. FIG. 38 shows the state of the queue when the settlement is made by the solution by the algorithm 1061.

  From FIG. 31, FIG. 34, FIG. 35, and Equation 5, the total score of each algorithm is as shown in FIG. 36, the solution by the algorithm 1061 having a large total score is selected as the optimum solution, and the netting settlement is executed. The balance of each user is as shown in FIG. 37, and the queue is as shown in FIG.

  FIG. 39 is a diagram showing an input state of a payment instruction after the netting settlement according to the present embodiment. FIG. 39 shows a payment instruction input after the netting settlement.

  After the netting settlement and before the second netting, there is an input of the payment instruction shown in FIG. 39, there is a deposit of “70” in the balance of user A, and “70” is issued from the balance of user B. Suppose there was money.

  That is, consider the case where the same situation as in FIG. 29 and the queue in FIG. 30 is created again. Thereafter, weight adjustment is performed at the time of the second netting. As for the weight adjustment setting, it is assumed that the historical data to be used is one netting settlement immediately before, the width for performing the weight adjustment is 0.1, and the frequency for performing the weight adjustment is set to be the netting settlement for each time. . At this time, the algorithm executed in the last netting is the algorithm 1061, and the score (2) of the scoring 1072 is low as shown in FIG. Therefore, the weight is adjusted as shown in FIG. 40 by adjusting the weight by the method described above.

  FIG. 40 is a diagram showing the status of weight data according to this embodiment. FIG. 40 shows that the weight data of scoring 1071 and 1072 are changed to 0.4 and 0.5, respectively.

  Since the balance and payment instructions are the same as before the first netting, the scores of each algorithm are the same as the values shown in FIGS. 34 and 35, and the transition of the temporary balance is the same as in FIGS. 33 and 34. It is. Since the weight is updated to the value shown in FIG. 40, the total score of each algorithm is as shown in FIG.

  FIG. 41 is a diagram showing the total score of the multilateral netting settlement system 10 of the present embodiment. At this time, as in the first netting, the solution by the algorithm 1061 having a large total score is selected as the optimal solution, the netting settlement is executed, and the balance of each user is updated to the same value as in FIG. The queue is updated as in FIG.

  Thereafter, even before the third netting, there is an input of the payment instruction shown in FIG. 39 as before the second netting, and there is a deposit of “70” in the balance of user A, It is assumed that “70” has been withdrawn from the balance. Therefore, consider a case where the balance is created in FIG. 29 and the queue is created in the same situation as shown in FIG. Thereafter, the same weight adjustment as that before the second netting is performed before the third netting, and the weights are updated as shown in FIG.

  FIG. 42 is a diagram illustrating the status of weight data according to the present embodiment. FIG. 42 shows that the weight data of scoring 1071 and 1072 has been changed to 0.3 and 0.7, respectively. As a result, the total score is calculated as shown in FIG. 43 as in the second netting.

  FIG. 43 is a diagram showing the total score of the multilateral netting settlement system 10 of the present embodiment. As shown in FIG. 43, in the third netting, unlike the first and second times, the algorithm having the largest total score is the algorithm 1062, and as a result of the multilateral netting settlement based on the solution by the algorithm 1062, the queue The state is updated as shown in FIG.

  FIG. 44 is a diagram showing a queue status of the multilateral netting settlement system 10 of the present embodiment. Comparing FIG. 38 and FIG. 44, the payment instruction from user B to user E and the payment instruction from user E to user A, which were not settled in the first and second netting settlement, are weighted. As a result of the adjustment, it can be seen that the settlement was made by the third netting. If weight adjustment is not performed, the same balance and payment instruction status are repeated in this example, so the same status as the first netting is repeated, and the payment instruction from user B to user E and user E forever. The payment instruction from user A to user A will not be settled, and it will not be possible to prevent an increase in settlement risk due to the remaining of these two payment instructions. That is, the increase in settlement risk can be prevented by adjusting the weight.

It should be noted that the usage period of historical data used for weight adjustment is relatively large, for example, one month, and the payment instruction is input by making the frequency relatively small, for example, once a month. When there is a pattern such as a long-term trend or cycle, an effective settlement strategy can be adopted by this method.
The above is an example of the effect of weight adjustment.

  As described above, in this embodiment, each of netting solutions derived by a plurality of algorithms corresponding to a plurality of strategies is derived using scores calculated for each solution by a plurality of evaluation methods and preset weights. The solution with the highest total score is identified as the optimal solution, and payment can be made using this optimal solution, so that a specific payment instruction is not settled for a long time but waits in the queue It is possible to provide a settlement system that can prevent an increase in settlement risk by continuing to do so. In addition, by automatically adjusting the weight used in calculating the above total score based on historical data, an optimal solution is derived according to the past payment instruction situation, for example, a payment instruction that has been waiting in a queue for a long time Can be provided.

  As described above, according to the multilateral netting settlement system of this embodiment, a netting solution derived from a plurality of algorithms is evaluated by a plurality of evaluation methods, a score is calculated, and settlement is performed using a netting solution having the highest total score. Therefore, it is possible to prevent a specific payment instruction from waiting in a queue without being settled for a long time, and to prevent an increase in settlement risk.

It is a figure which shows the whole structure of the multilateral netting payment system of this embodiment. It is a figure which shows the flow of the procedure which performs the multilateral netting settlement of this embodiment. It is a figure which shows the flow of the netting solution derivation | leading-out procedure common in each of three types of algorithms of this embodiment. It is a figure which shows the screen of the user terminal 104 of this embodiment. It is a figure which shows the screen of the administrator terminal 105 of this embodiment. It is a figure which shows the notification screen of the start of the multilateral netting payment notified to a user terminal by step 2101 of FIG. 2 of this embodiment. It is a figure which shows the notification screen of the completion of the multilateral netting settlement notified to a user terminal by step 2102 of FIG. 2 of this embodiment, and a data update. It is a schematic diagram which shows the payment instruction in queue waiting shown in FIG. 17 of the multilateral netting settlement system of this embodiment. It is a payment instruction | indication which shows the solution by the algorithm 1061 of the multilateral netting payment system of this embodiment. It is a payment instruction | indication which shows the solution by the algorithm 1062 of the multilateral netting payment system of this embodiment. It is a payment instruction | indication which shows the solution by the algorithm 1063 of the multilateral netting payment system of this embodiment. It is a schematic diagram which shows the payment instruction | indication of FIG. 30 of the multilateral netting payment system of this embodiment. It is a figure which shows the data format of the queue database (per user) 102 of this embodiment. It is a figure which shows the data format of the balance database (per user) 103 of this embodiment. It is a figure which shows the data format of the weight database 108 of this embodiment. It is a figure which shows the data format of the historical database 110 of this embodiment. It is a figure which shows the example of the payment instruction in queue waiting of each user of this embodiment. It is a figure which shows the balance of each user of this embodiment. It is a figure which shows transition of the temporary balance by the algorithm 1061 of this embodiment. It is a figure which shows the change of the payment instruction in queue waiting of this embodiment. It is a figure which shows the score of the solution by the algorithm 1061 of this embodiment. It is a figure which shows transition of the temporary balance by the algorithm 1062 of this embodiment. It is a figure which shows the change of the payment instruction in queue waiting of this embodiment. It is a figure which shows the score of the solution by the algorithm 1062 of this embodiment. It is a figure which shows transition of the temporary balance by the algorithm 1063 of this embodiment. It is a figure which shows the change of the payment instruction in queue waiting of this embodiment. It is a figure which shows the score of the solution by the algorithm 1062 of this embodiment. It is a figure which shows the example of input of the payment instruction | indication of this embodiment. It is a figure which shows the state of the balance of this embodiment. It is a figure which shows the state of the queue at the time of the input of the payment instruction | indication of FIG. 28 of this embodiment. It is a figure which shows the initial weight data of this embodiment. It is a figure which shows transition of the temporary balance by the algorithm 1061 of this embodiment. It is a figure which shows transition of the temporary balance by the algorithm 1062 of this embodiment. It is a figure which shows the score of the algorithm 1061 of this embodiment. It is a figure which shows the score of the algorithm 1062 of this embodiment. It is a figure which shows the total score of each algorithm of this embodiment. It is a figure which shows the update condition of the balance of this embodiment. It is a figure which shows the condition of the queue of this embodiment. It is a figure which shows the input condition of the payment instruction | indication after the said netting settlement of this embodiment. It is a figure which shows the condition of the weight data of this embodiment. It is a figure which shows the total score of the multilateral netting payment system of this embodiment. It is a figure which shows the condition of the weight data of this embodiment. It is a figure which shows the total score of the multilateral netting payment system of this embodiment. It is a figure which shows the queue condition of the multilateral netting payment system of this embodiment. It is a figure which shows schematic structure of the multilateral netting settlement system 10 of this embodiment.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 ... Multilateral netting payment system, 102 ... Queue database (per user), 103 ... Balance database (per user), 104 ... User terminal, 105 ... Administrator terminal, 108 ... Weight database, 110 ... Historical database 1061-1063 ... algorithm, 1071-1074 ... scoring, 101 ... settlement management processing unit, 106 ... netting solution derivation processing unit, 107 ... netting solution evaluation processing unit, 109 ... netting solution determination processing unit, 301 ... payment instruction data , 302 ... balance data, 401 ... payee column, 402 ... amount column, 403 ... execute button, 404 ... list of payment instructions in queue waiting, 405 ... current balance column, 406 ... deposit column, 407 ... deposit execution button, 408 ... withdrawal column, 409 ... withdrawal execution button, 501 ... Vertical netting settlement execution interval field, 502 ... Weight value field, 503 and 504 ... Radio button, 505 ... Automatic adjustment execution interval field, 506 ... Automatic adjustment width field, 507 and 508 ... Usage historical data period field, 509 ... Update button 510 ... Cancel button.

Claims (8)

In a multi-lateral netting settlement method using a multi -lateral netting settlement system that offsets payment instructions of multiple users and settles only the difference amount,
The multilateral netting settlement system includes a netting solution derivation processing unit, a netting solution evaluation processing unit, a netting solution determination processing unit, and a settlement management processing unit.
The netting solution derivation processing unit obtains a netting solution that is a combination of payment instructions capable of multilateral netting settlement from payment instructions waiting in a queue from a queue database for each user and a balance database for each user. The payment instruction data and balance data that are waiting in the queue are read, the algorithm that excludes the last payment instruction in the queue, the algorithm that excludes the payment instruction with the smallest amount, and the payee's queue Obtaining and storing in a storage device by a plurality of algorithms including two or more of the algorithms excluded from an empty payment instruction ;
The netting solution evaluation processing unit determines, for each of the obtained plurality of netting solutions, an evaluation method based on the number of payment instructions subject to netting / the total number of payment instructions waiting in queue, the total amount of payment instructions subject to netting / Valuation method based on the total amount of payment instructions in queue waiting, total payment instruction waiting time for netting / evaluation method based on total waiting time for all payment instructions in queue waiting, total payment instruction exposure for netting / waiting queue Scoring with a score evaluated by a plurality of evaluation methods including two or more of the evaluation methods based on the total payment instruction exposure of all, and storing in a storage device in association with the netting solution;
The netting solution determination processing unit calculates the total score of each netting solution by totaling the scored scores for each netting solution, and determines the netting solution having the highest calculated total score as a solution to be settled; ,
The settlement management processing unit includes a step of performing settlement by the netting solution by storing a balance after payment according to the payment instruction of the determined netting solution in a storage device. 2. The multi-lateral netting settlement method according to claim 1, wherein the netting solution evaluation processing unit obtains the score by a mathematical expression including a waiting time of a payment instruction to be settled. The netting solution evaluation processing unit calculates the total score by weighting the score using weight data set in advance for each of the evaluation methods. Multi-lateral netting payment method. The said netting solution evaluation process part changes the value of the weight data set for every said evaluation method according to the past average score, The Claim 1 thru | or 3 characterized by the above-mentioned. Multi-lateral netting payment method. The netting solution derivation processing unit calculates a temporary balance by aggregating payment instruction data and balance data that are waiting in a queue, and when the temporary balance is negative, payment that is excluded by the plurality of algorithms 5. The multi-value according to claim 1, wherein a netting solution in which the temporary balance is non-negative is obtained for each algorithm by determining an instruction and excluding it from the target of aggregation. Lateral netting payment method. The settlement management processing unit performs settlement using the netting solution by storing the temporary balance calculated when obtaining the netting solution in a balance database for the netting solution determined as the solution for the settlement. The multilateral netting settlement method according to any one of claims 1 to 5. In a multi-lateral netting payment system that offsets payment instructions of multiple users and settles only the difference,
Payment instructions in queue waiting from the queue database for each user and the balance database for each user, which are combinations of payment instructions that can be settled by multilateral netting from the payment instructions in the queue. Reads data and balance data, excludes the last payment instruction in the queue, excludes the payment instruction with the smallest amount, and excludes it from the payment instructions whose payee queue is empty A netting solution derivation processing unit that obtains a plurality of algorithms including two or more algorithms and stores them in a storage device;
For each of the obtained plurality of netting solutions, each solution is evaluated based on the number of payment instructions to be netted / the total number of payment instructions in queue, the total amount of payment instructions to be netted / all payment instructions in queue. Valuation method based on total amount, total payment instruction waiting time subject to netting / evaluation method based on total waiting time for all payment instructions waiting in queue, totaling payment instruction exposure subject to netting / evaluation based on total payment instruction exposures waiting in queue A netting solution evaluation processing unit for scoring with scores evaluated by a plurality of evaluation methods including two or more of the methods and storing the score in a storage device in association with the netting solution;
Netting solution for each netting solution to calculate the total score of each netting solution, the netting solution determination processing unit for determining the highest netting solution of the calculated total score as a solution to be settled,
A multi-lateral netting settlement system, comprising: a settlement management processing unit that performs settlement based on the netting solution by storing a balance after payment according to the payment instruction of the determined netting solution in a storage device. In a program for causing a computer to execute a multilateral netting settlement method using a multilateral netting settlement system that offsets payment instructions of multiple users and settles only the difference amount,
The multilateral netting settlement system includes a netting solution derivation processing unit, a netting solution evaluation processing unit, a netting solution determination processing unit, and a settlement management processing unit.
The netting solution derivation processing unit obtains a netting solution that is a combination of payment instructions capable of multilateral netting settlement from payment instructions waiting in a queue from a queue database for each user and a balance database for each user. The payment instruction data and balance data that are waiting in the queue are read, the algorithm that excludes the last payment instruction in the queue, the algorithm that excludes the payment instruction with the smallest amount, and the payee's queue Obtaining and storing in a storage device by a plurality of algorithms including two or more of the algorithms excluded from an empty payment instruction ;
The netting solution evaluation processing unit determines, for each of the obtained plurality of netting solutions, an evaluation method based on the number of payment instructions subject to netting / the total number of payment instructions waiting in queue, the total amount of payment instructions subject to netting / Valuation method based on the total amount of payment instructions in queue waiting, total payment instruction waiting time for netting / evaluation method based on total waiting time for all payment instructions in queue waiting, total payment instruction exposure for netting / waiting queue Scoring with a score evaluated by a plurality of evaluation methods including two or more of the evaluation methods based on the total payment instruction exposure of all, and storing in a storage device in association with the netting solution;
The netting solution determination processing unit calculates the total score of each netting solution by totaling the scored scores for each netting solution, and determines the netting solution having the highest calculated total score as a solution to be settled; ,
A program for causing the computer to execute a step of performing settlement by the netting solution by storing the balance after payment by the payment instruction of the determined netting solution in a storage device by the settlement management processing unit .

Images