TWI815453B - Asset allocation optimization system - Google Patents

Abstract

An asset allocation optimization system includes a commodity servo system, an asset allocation servo system electrically connected to the commodity servo system, and a combined optimization servo system electrically connected to the commodity servo system and the asset allocation servo system. The combined optimization servo system performs a combined optimization program on a customer inventory combination of a customer, so that the asset allocation servo system performs an asset matching program on all investment targets to generate a first asset matching combination. The combined optimization servo system then selects at least one of the first asset matching combination to generate non-overlapping N candidate permutations, so that the asset allocation servo system to performs the asset matching program on each candidate permutation to generate N second asset matching combinations and then determines a most suitable asset matching combination.

Description

Asset Adaptation Optimization System

The present invention relates to an optimization system, in particular to an asset matching optimization system for a client's investment portfolio.

The financial management services provided by existing banks, whether through financial managers or computer systems, often result in too fragmented asset allocation when optimizing recommendations for customers' investment portfolios, resulting in the management of individual customers' investments. The increase in the cost of the portfolio will also reduce the willingness of customers to adjust their investment portfolio. In addition, when portfolio optimization recommendations allocate too many products, the transaction amount of a single product will not reach the transaction threshold, increasing the risk that customers will not be able to allocate appropriate products.

Therefore, the purpose of the present invention is to provide an asset adaptation and optimization system that can streamline the number of products and take into account market objectives.

Therefore, the present invention provides an asset adaptation and optimization system, which is suitable for a customer and includes a product servo system, an asset adaptation servo system, and a combination optimization servo system.

The product server system stores multiple investment target information, and each investment target information includes a target name and a target type. The asset matching servo system is electrically connected to the product servo system and is used to execute an asset matching program to generate an asset matching combination according to a target condition. The portfolio optimization servo system is electrically connected to the product server system and the asset matching servo system, and executes a portfolio optimization program for a customer inventory portfolio of the customer.

In the portfolio optimization process, the portfolio optimization server system transmits all the investment targets and the customer inventory combinations, and a corresponding target market portfolio to the asset matching server system to execute the corresponding name of all the targets. The asset matching process that serves as an investment target and uses multiple target proportions of the target market portfolio as distribution targets and meets the target conditions generates a first asset matching portfolio. The portfolio optimization servo system then selects at least one of the subject names of the first asset matching combination to generate N non-overlapping candidate permutations, and then compares each of the candidate permutations with the customer inventory The combination, and the corresponding target market combination are sent to the asset matching server system to execute the corresponding candidate permutation combination as the investment target and use the target proportion of the target market combination as the allocation target and meet the target conditions. The asset matching procedure is performed to obtain N second asset matching combinations, where N is a plural number. The portfolio optimization servo system selects the N second asset matching portfolio that best meets the target condition as the most suitable asset matching portfolio for the customer.

In some implementations, the portfolio optimization server system stores multiple expert market portfolios corresponding to different risk levels, and selects the expert market portfolio corresponding to the risk level of the client as the target market portfolio.

In some implementations, the target condition is that a Sharpe ratio of the first asset matching combination (or the second asset matching combination) is the largest. The Sharpe ratio is equal to the return on an investment group of the first asset matching portfolio (or the second asset matching portfolio) minus a risk-free interest rate, and then divided by the first asset matching portfolio (or the second asset matching portfolio) The standard deviation of an investment group (asset matching portfolio).

In other implementations, the target condition is that the difference between the customer inventory combination and the first asset matching combination (or the second asset matching combination) is minimal. That is, the proportion of the inventory corresponding to each target name of the customer inventory portfolio and the inventory corresponding to each target name of each first asset matching combination (or the second asset matching combination) After corresponding subtraction of the proportions, multiple buying and selling weights are obtained, and then the absolute value of each buying and selling weight is taken and added to obtain the smallest difference among the results.

In other implementations, the target condition is that the first asset matching combination (or the second asset matching combination) is closest to the target market combination. That is, the proportions of the targets corresponding to the target names of the first asset matching portfolio (or the second asset matching portfolio) are respectively the proportions of the targets corresponding to the target names of the corresponding target market portfolio. The Euclidean distance of the target proportion is the smallest.

In other implementations, the object type of each investment object information includes a stock, a bond, and an asset. The stock includes a stock and a country and a stock and an industry. The bond includes a bond and an asset. The country and a bond and credit rating, the assets include a customer holding bonds, a customer holding stocks, a customer's other investment positions, and a cash.

In other implementations, the portfolio optimization servo system selects the target proportions corresponding to the target types included in the N second asset matching combinations to be bounded by the corresponding plurality of The one that best meets the target conditions between the final upper bound ratio and multiple final lower bound ratios will be used as the most suitable asset matching portfolio for the customer.

In some implementations, the portfolio optimization servo system is based on the target proportion RA1, an upper bound proportion RA2, and a lower bound proportion RA3 corresponding to each target type of the target market portfolio, and an The final optimal proportion RB, calculate the final upper bound proportion RC1 and the final lower bound proportion RC2 corresponding to the type of the target, where RA2>RA1>RA3, RC1=RB+(RA2-RA1), RC2=RB-(RA1- RA3) .

In some implementations, the portfolio optimization servo system calculates the sum of the inventory proportions of the subject names without transaction restrictions in the customer inventory portfolio to obtain a free proportion sum, and calculates the sum of the inventory proportions according to the target market portfolio. The target proportion RA1 of each target type, and the sum RD1 of all inventory proportions corresponding to the target type and with transaction restrictions in the customer's inventory portfolio, calculate an optimal distance D1 corresponding to the target type, where, D1 =MAX(0,RA1-RD1).

The combined optimization servo system also distributes the free proportion sum according to the ratio of the optimal distance of each target type to the sum of the optimal distances of all target types, and obtains a plurality of corresponding target types respectively. The free proportion is calculated, and then the inventory proportion of each object type of the customer's inventory portfolio is added to the corresponding free proportion to obtain the final optimal proportion corresponding to the object type.

The effect of the present invention is that: the combination optimization servo system generates non-overlapping N candidate permutation combinations based on the first asset matching combination, and then each of them and the customer inventory combination are sent to the asset matching servo system to execute the The asset matching process is to obtain the N second asset matching combinations, and then select the one that best meets the target conditions as the most suitable asset matching combination for the customer. The most suitable asset matching combination that can be achieved is It is most in line with the two purposes of streamlining the number of products and taking into account market goals.

Before the present invention is described in detail, it should be noted that in the following description, similar elements are designated with the same numbering.

Referring to Figure 1, an embodiment of the asset matching and optimization system of the present invention is applicable to a customer and includes a product server system 1, an asset matching server system 2, and a combination optimization server system 3.

The product server system 1 is, for example, one or more servers, and stores a plurality of investment target information. Each investment target information includes an target name and a target type. In this embodiment, the underlying type includes a stock, a bond, and an asset. To differentiate in more detail, the stock includes a stock and a country and a stock and an industry, the bond includes a bond and a country and a bond and a credit rating, and the assets include a bond held by a customer (that is, a bond held by a customer). bonds), one customer's stocks (that is, the stocks the customer already holds), one customer's other investment positions, and one cash.

The asset matching server system 2 is, for example, one or more servers, and is connected to the product server system 1 to form an electrical connection, and is used to execute an asset matching program to generate an asset according to a target condition. Matching combination.

The combination optimization server system 3 is, for example, one or more servers, and is connected to the product server system 1 and the asset matching server system 2 respectively to form an electrical connection, and is to execute a customer inventory combination of the customer. A combination of optimization procedures. For example, the customer's inventory portfolio contains multiple target names, such as A, B, and X, and multiple corresponding target types, such as stocks, stocks, and cash, and multiple corresponding inventory proportions. Such as 0.05, 0.05, 0.9. Then the customer inventory portfolio can be classified as shown in Table 1 below. Target type Inventory proportion stock 0.1 bond 0 Cash 0.9 Table I

In the portfolio optimization process, the portfolio optimization server system 3 transmits the subject names of all the investment subject information, the client inventory combination, and a corresponding target market portfolio to the asset matching server system for execution. The asset matching program corresponding to all the target names as investment targets and using multiple target proportions of the target market portfolio as allocation targets and meeting the target conditions generates a first asset matching portfolio. In more detail, the portfolio optimization servo system 3 also stores a plurality of expert market portfolios corresponding to different risk levels, and selects the expert market portfolio corresponding to the risk level of the customer as the target market portfolio. For example, the risk levels range from low to high respectively from risk 1 to 4. The expert market portfolio corresponding to risk 2 is as shown in Table 2 below, and the values are determined according to different risk levels after expert discussions, for example. For example, if the target type is a stock, the target proportion should be between the lower bound proportion of 0.2 and the upper bound proportion of 0.4. Target type target proportion Lower bound proportion Upper bound proportion stock 0.3 0.2 0.4 bond 0.7 0.6 0.8 Cash 0 0 1 Table II

In this embodiment, the target condition is that the Sharpe ratio of the first asset matching combination is the largest, or the difference between the customer inventory combination and the first asset matching combination is the smallest, or the first asset matching combination is the smallest. The asset matching portfolio is closest to the target market portfolio.

The Sharpe ratio is equal to the return rate of an investment group of the first asset matching portfolio minus a risk-free interest rate, and then divided by the standard deviation of an investment group of the first asset matching portfolio. The risk-free interest rate is a preset known parameter. For example, the first asset matching group includes three target names W1, W2, and W3, and the corresponding three target proportions are 0.6, 0.3, and 0.1 , whose historical returns are 1.7%, 1.6%, and 1.5% respectively, then the return rate of this investment group is 0.6*1.7%+0.3*1.6%+0.1*1.5%=1.65%, and the standard deviation of this investment group is equal to ([ 0.6 0.3 0.1]*risk matrix*[0.6 0.3 0.1]^(T)) ^(1/2)=0.5597. The risk matrix is a 3*3 matrix and corresponds to the three target names W1 and W2. , and the price variation (Variance) or covariance (Covariance) between , and W3.

The minimum difference between the customer inventory portfolio and the first asset adaptation portfolio is: the inventory proportion corresponding to each subject name of the customer inventory portfolio and the inventory proportion corresponding to each subject name of the first asset adaptation portfolio. After correspondingly subtracting the corresponding target proportions, multiple buying and selling weights are obtained, and then the absolute value of each buying and selling weight is taken and then added to obtain the smallest difference among the results. For example, the customer's inventory portfolio includes the subject name A, the corresponding subject type is stocks, and the corresponding inventory proportion is 0.1; and the subject name is B, the corresponding subject type is bonds, and the corresponding stock ratio is 0.1. The inventory ratio is 0.9. Combination 1 and combination 2 are shown in Table 3. The difference corresponding to combination 1 is 0.1+0.1=0.2, and the difference corresponding to combination 2 is 0.1+0.9+0.8=1.8. Therefore, the difference corresponding to combination 1 is the smallest. Subject name Target type target proportion Buy and sell weight Absolute value of buy and sell weight Combination 1 A stock 0.2 +0.1 0.1 B bond 0.8 -0.1 0.1 Combination 2 A stock 0.2 +0.1 0.1 B bond 0 -0.9 0.9 C bond 0.8 +0.8 0.8 Table 3

The target condition is that the first asset matching combination is closest to the target market combination: the target proportions corresponding to the target names of the first asset matching combination are respectively the same as the corresponding target market combination. The Euclidean distance of the proportion of such targets corresponding to the names of the targets is the smallest. For example, the target market portfolio includes the target name A, the corresponding target type is stocks, and the corresponding target proportion is 0.3; and the target name is B, the corresponding target type is bonds, and the corresponding target proportion is The proportion is 0.7. Then the Euclidean distance of combination 1 in Table 3 is equal to , the Euclidean distance of combination 2 is equal to , therefore, combination 1 has the smallest Euclidean distance and is closest to the target market combination.

Then, in the combination optimization process, the combination optimization servo system 3 selects at least one of the target names of the first asset matching combination to generate N non-overlapping candidate permutation combinations. For example, all the investment target information has 26 target names from A to Z, and the result of the first asset matching combination is that the target names A to D correspond to different proportions. is zero, and the proportions corresponding to the underlying names E~Z are all zero, then the number of the underlying names included in the first asset matching portfolio is 4 (that is, the proportions are not The number of the underlying names is zero), and the corresponding underlying types are, for example, stocks, stocks, bonds, bonds. Then the subject names of each of the N candidate permutations are ABCD, ABC, ABD, ACD, BCD, AB, AC, AD, BC, BD, CD, A, B, C, and D respectively. In other words, the N candidate permutations include all permutations of some or all of the underlying names of the first asset matching combination.

The portfolio optimization servo system 3 then sends each candidate permutation combination, the customer inventory combination, and the corresponding target market combination to the asset matching server system 2 to implement the corresponding candidate permutation combination as the investment target and use the The target proportions of the target market portfolio are the allocation target and the asset matching process meets the target conditions, and corresponding N second asset matching combinations are obtained, where N is a plural number.

The target condition is that the Sharpe ratio of the second asset matching combination is the largest, or the difference between the customer inventory portfolio and the second asset matching combination is the smallest, or the second asset matching combination is closest to the target market combination. , that is, the same one corresponding to the aforementioned target condition.

The portfolio optimization servo system 3 also keeps the inventory proportions of all subject names with transaction restrictions (that is, the transaction restriction is "yes") in the customer's inventory portfolio unchanged, and calculates the rest of the customer's inventory portfolio (that is, the transaction restriction is "yes"). The sum of the inventory proportions of the subject names for which the transaction limit is "None" is obtained to obtain a free proportion sum. For example, the target market combination is shown in the aforementioned Table 2, and the customer inventory combination is shown in the following Table 4. Among them, the customer inventory portfolio also includes a transaction restriction corresponding to each subject name. The transaction restriction is, for example, "yes" and "no", and the reasons for the transaction restriction include at least the following: (1) Due to market perception And limit the holding range of the underlying types, such as stocks, bonds, the United States (i.e., countries), Europe (i.e., regions), finance (i.e., industries), technology (i.e., industries), etc.; (2) Restrictions on selling due to business logic ( Such as IPO lock-up period) or purchase restrictions (such as commodity risk level); (3) Restrictions caused by customers' or professionals' market views or investment preferences. This sum of free proportions is, for example, 0.05+0.15+0.15+0.2=0.55. Subject name Inventory proportion Target type Transaction restrictions A 0.2 stock have B 0.25 stock have C 0.05 stock without D 0.15 bond without E 0.15 bond without F 0.2 bond without Table 4

The combination optimization servo system 3 then calculates a plurality of optimal distances respectively corresponding to the target types of the target market combination. That is to say, the optimal distance corresponding to any one of the target types is equal to Max (0, corresponding to the target type The target proportion - the sum of all the inventory proportions in the customer's inventory portfolio that correspond to the target type and the transaction limit is "yes"), for example: the optimal distance for stocks is Max(0,0.3-0.45)=0; The optimal distance for bonds is Max(0,0.7-0)=0.7; the optimal distance for cash is Max(0,0-0)=0.

The combined optimization servo system 3 then distributes the sum of the free proportions according to the ratio of each optimum distance to the sum of the optimum distances to obtain a plurality of free proportions. For example: the free ratio allocated to stocks is 0.55*(0/0.7)=0; the free ratio allocated to bonds is 0.55* (0.7/0.7)=0.55; the free ratio allocated to cash is 0.55*(0 /0.7)=0. The portfolio optimization servo system 3 then calculates a plurality of final optimal proportions corresponding to the target types of the customer's inventory portfolio, that is, the inventory proportions plus the corresponding free proportions. For example: the final optimal ratio corresponding to stocks is 0.45+0= 0.45; the final optimal ratio corresponding to bonds is 0+0.55=055; the final optimal ratio corresponding to cash is 0+0=0.

The portfolio optimization servo system 3 then calculates a plurality of corresponding final upper bounds based on the target proportions, the upper bound proportions, the lower bound proportions, and the final optimal proportions corresponding to the target market portfolio. Scale and multiple final lower bound scales. As shown in List 5 below. Target type final optimum ratio final lower bound ratio final upper bound ratio stock 0.45 Max(0,0.45-0.1)= 0.35 Min(1,0.45+0.1)= 0.55 bond 0.55 Max(0,0.55-0.1)= 0.45 Min(1,0.55+0.1)= 0.65 Cash 0 Max(0,0-0)=0 Min(1,0+1)=1 Table 5

For example, if the target type is a stock, the final upper bound proportion is equal to the final optimal proportion plus the corresponding upper bound proportion minus the target proportion (as shown in Table 2), that is, 0.55=0.45 + (0.4-0.3 ), the final lower bound proportion of the underlying type of bond is equal to the final optimal proportion minus the corresponding target proportion plus the lower bound proportion (as shown in Table 2), that is, 0.45=0.55-0.7+0.6=0.55-(0.7 -0.6).

The portfolio optimization servo system selects the target proportions corresponding to the target types included in the N second asset matching portfolios to be respectively bounded by the corresponding final upper bound proportions and the final lower bound proportions. time, and the one that best meets the target conditions will be used as the most suitable asset matching portfolio for the customer.

For example, the target condition is that when the Sharpe ratio of the first asset matching combination (or the second asset matching combination) is the largest, one of the asset matching combinations will The target proportions of stocks, bonds, and cash in the portfolio and another of the asset-matching portfolios are 0.25, 0.55, and 0.2, and 0.4, 0.55, and 0.05 respectively, and the Sharp of one of the asset-matching portfolios The ratio is the highest, and the Sharpe ratio of the other asset allocation combination is the second highest. However, since 0.25 is not between 0.35 and 0.55 (as shown in Table 5), the portfolio optimization servo system 3 selects the other asset allocation combination. The suitable combination is used as the most suitable asset matching combination for the customer.

In addition, it should be noted that in this embodiment, the subject type of the expert market portfolio is stocks, bonds, and cash as examples. In other embodiments, the subject type of the expert market portfolio can also be It is at least one of the stock and the country, the stock and the industry, the bond and the country, the bond and the credit rating, the bond held by the customer, the stock held by the customer, other investment positions of the customer, and the cash. combination.

To sum up, the combination optimization servo system 3 generates non-overlapping N candidate permutation combinations based on the first asset adaptation combination, and then sends each of the N candidate permutation combinations and the customer inventory combination to the asset adaptation servo system 2 for execution. The asset matching process is to obtain the N second asset matching combinations, and then select the one that is between the corresponding final upper bound ratios and the final lower bound ratios and that best meets the target conditions as The most suitable asset matching combination for this customer can achieve the following two advantages: (1) Streamline the average number of bought and sold products to reduce the number of remaining inventory products after executing sell transactions during the portfolio optimization process, or execute The number of new products added after the purchase transaction is too high, resulting in the customer having too many products in stock, in order to reduce the cost of financial management staff to manage the customer's investment portfolio; (2) By reducing the number of products bought and sold, the allocation amount of the same product is increased, reducing the number of customers Since the amount cannot be allocated because the amount does not reach the transaction threshold, the purpose of the present invention can indeed be achieved.

However, the above are only examples of the present invention. They cannot be used to limit the scope of the present invention. All simple equivalent changes and modifications made based on the patent scope of the present invention and the contents of the patent specification are still within the scope of the present invention. within the scope covered by the patent of this invention.

1: Product servo system 2: Asset adaptation servo system 3: Combination optimization servo system

Other features and effects of the present invention will be clearly presented in the embodiments with reference to the drawings, in which: FIG. 1 is a block diagram illustrating an embodiment of the asset adaptation and optimization system of the present invention.

1: Product servo system

2: Asset adaptation servo system

3: Combination optimization servo system

Claims (9)

An asset matching optimization system is suitable for a customer and includes: a product server system that stores multiple investment target information, and each investment target information includes a target name and a target type; an asset matching server system that electronically Connected to the product server system and used to execute an asset matching program to generate an asset matching combination according to a target condition; and a combination optimization server system electrically connected to the product server system and the asset matching server system, and A customer inventory portfolio of the client executes a portfolio optimization program. In the portfolio optimization program, the portfolio optimization server system transmits all the investment targets and the client inventory portfolio, and a corresponding target market portfolio to the asset allocation The adaptation server system generates a first asset allocation program by executing the asset matching program corresponding to all the target names as investment targets and using multiple target proportions of the target market portfolio as allocation targets and meeting the target conditions. The combination optimization servo system then selects at least one of the subject names of the first asset matching combination to generate N non-overlapping candidate permutations and combinations, and then compares each candidate permutation and combination with The customer inventory portfolio and the corresponding target market portfolio are sent to the asset matching server system to execute the corresponding candidate permutation combination as the investment target and use the target proportion of the target market portfolio as the allocation target and comply with the target. The asset matching program meets the conditions, and N second asset matching combinations are obtained, N is a plural number, and the combination optimization servo system selects one of the N second asset matching combinations. Among them, the one that best meets the target conditions will be used as the most suitable asset matching combination for the customer. The asset matching optimization system as described in claim 1, wherein the portfolio optimization server system stores a plurality of expert market portfolios corresponding to different risk levels, and selects the expert market portfolio corresponding to the risk level of the client as the target Market mix. The asset matching optimization system as described in claim 2, wherein the target condition is that a Sharpe ratio of the first asset matching combination is the largest, and the Sharpe ratio is equal to a Sharpe ratio of the first asset matching combination. After subtracting a risk-free interest rate from the group return rate, it is divided by the standard deviation of an investment group of the first asset matching portfolio, or the target condition is that the Sharpe ratio of the second asset matching portfolio is the largest, and the Sharpe ratio It is equal to the return rate of the investment group of the second asset matching portfolio minus the risk-free interest rate, and then divided by the standard deviation of the investment group of the second asset matching portfolio. The asset adaptation optimization system as described in claim 2, wherein the target condition is that the difference between the customer inventory portfolio and the first asset adaptation portfolio is the smallest, that is, the customer inventory portfolio corresponds to each target name. After the inventory proportion is subtracted from the inventory proportion corresponding to each target name of each first asset matching combination, a plurality of buying and selling weights are obtained, and then the absolute value of each buying and selling weight is obtained. The smallest difference among the summed results, or the target condition is the smallest difference between the customer inventory portfolio and the second asset matching combination, that is, the inventory proportion corresponding to each target name of the customer inventory portfolio. The ratio is subtracted from the proportion of the inventory corresponding to each subject name of each second asset matching combination. After that, the buying and selling weights are obtained, and then the absolute value of each buying and selling weight is obtained and then the sum of the results has the smallest difference. The asset matching optimization system as described in claim 2, wherein the target condition is that the first asset matching combination is closest to the target market combination, that is, the target names of the first asset matching combination correspond to The Euclidean distance between the target proportions and the target proportions corresponding to the target names of the corresponding target market portfolio is the smallest, or the target condition is that the second asset matching portfolio is closest to the target market The combination is the smallest Euclidean distance between the target proportions corresponding to the target names of the second asset matching combination and the target proportions corresponding to the target names of the corresponding target market combination. . The asset matching optimization system as described in claim 2, wherein the object type of each investment object information includes a stock, a bond, and an asset, and the stock includes a stock and a country and a stock and an industry, and the Bonds include a bond and a country and a bond and a credit rating. The assets include a customer holding bonds, a customer holding stocks, a customer's other investment positions, and a cash. The asset matching optimization system as described in claim 2, wherein the portfolio optimization servo system selects the target proportions corresponding to the target types included in the N second asset matching combinations. Among the corresponding multiple final upper bound ratios and multiple final lower bound ratios, the one that best meets the target conditions will be used as the most suitable asset matching portfolio for the customer. The asset adaptation optimization system as described in claim 7, wherein the portfolio optimization servo system is based on the target proportion RA1, an upper bound proportion RA2, and a lower bound corresponding to each target type of the target market portfolio. Proportion RA3, and a final optimal proportion RB, calculate the final upper bound proportion RC1 and the final lower bound proportion RC2 corresponding to the target type, where, RA2>RA1>RA3, RC1=RB+(RA2-RA1), RC2=RB-( RA1-RA3). The asset adaptation optimization system as described in request item 8, wherein the portfolio optimization servo system calculates the sum of the inventory proportions of the subject names without transaction restrictions in the customer inventory portfolio to obtain a free proportion sum, and Based on the target proportion RA1 of each target type in the target market portfolio and the sum RD1 of all inventory proportions corresponding to the target type and with transaction restrictions in the customer inventory portfolio, calculate the most appropriate target type corresponding to the target market. Distance D1, where D1=MAX(0,RA1-RD1), the combined optimization servo system also calculates the sum of free proportions according to the optimal distance of each object type and the sum of the optimal distances of all object types. Allocate the proportion to obtain multiple free proportions corresponding to the corresponding target types, and then calculate the inventory proportion of each target type in the customer's inventory portfolio plus the corresponding free proportion to obtain the corresponding target type. of the final optimum ratio.

Images