CN112035533B - System resource scheduling method and device based on multi-parameter quantization strategy feedback - Google Patents

Abstract

The invention discloses a system resource scheduling method and device based on multi-parameter quantization strategy retest, which comprises the steps of setting a plurality of database combinations according to the idle memory amount of a server, the memory occupied amount of a single-parameter quantization strategy and the memory occupied amount of a single financial database, and retesting multi-parameter quantization strategies according to the plurality of database combinations to obtain the retest running time corresponding to each database combination; the method comprises the steps of searching a plurality of optimal database combinations in a plurality of database combinations and the corresponding backtesting running time of each database combination according to the number of specified backtesting tasks input by a user, determining the number of financial databases to be opened and the number of backtesting tasks served by each financial database according to the optimal database combinations, and backtesting the multi-parameter quantization strategy according to the result, so that the backtesting time of the multi-parameter quantization strategy can be shortened, and the resource cost is reduced.

Description

System resource scheduling method and device based on multi-parameter quantization strategy backtesting

technical field

The present invention relates to the technical field of server resource scheduling, and more particularly, to a system resource scheduling method and device based on multi-parameter quantization strategy backtesting.

Background technique

In recent years, more and more investors have designed corresponding quantitative strategies based on the historical data of stocks to realize the investment analysis of stocks. Existing quantitative strategies are usually written by financial engineers, and the quantitative strategies can be back-tested through historical data to evaluate the performance of quantitative strategies. Specifically, after the user has set some stock indicator combinations, based on the real market data that has occurred within a certain period of time, the simulated trading is performed according to the stock combination obtained by the quantitative strategy, according to the time obtained in the simulated trading process. Data such as profit and drawdown rate in the segment are used to comprehensively evaluate the quantitative strategy.

In general, the quantization strategy involves a large number of parameters, such as the length of the moving average in the single and double moving average quantization strategies, and these parameters will have a greater impact on the performance of the quantization strategy. In order to obtain a higher return on investment, it is necessary to carry out backtesting of quantitative strategies under different parameters and different scenarios (such as different stock targets or different backtesting intervals, etc.). Due to the large number of parameters involved, parallel computing can be used during backtesting to speed up the calculation. However, in the prior art, parallel computing is mainly performed by means of a single database and multiple processes, which imposes a heavy burden on system resources, resulting in an excessively low back-testing efficiency of the quantitative strategy.

After retrieval, the patent "A Cluster-based Quantitative Strategy Backtesting System and its Backtesting Method" (published on 2018.11.06, publication number CN108765149A) disclosed in China when backtesting the quantitative strategy, according to each The load conditions of each computer node are assigned tasks, so that each computer node can be used with maximum efficiency, which can improve the efficiency of backtesting. However, it is obvious that the resource cost is too high to realize the parallel computing backtest by calling multiple computers.

SUMMARY OF THE INVENTION

In order to overcome the above-mentioned defect of high resource cost in the prior art, the present invention provides a system resource scheduling method and device based on multi-parameter quantization strategy back-testing, which can shorten the back-test time of multi-parameter quantization strategy and reduce resources at the same time. cost.

For solving the above-mentioned technical problems, the technical scheme of the present invention is as follows:

A first aspect of the present invention discloses a system resource scheduling method based on multi-parameter quantization strategy backtesting, comprising the following steps:

S1: Back-test the specified single-parameter quantitative strategy according to the preset back-test time period and the preset standard stock target, and record the amount of memory occupied by the single-parameter quantitative strategy during the back-test;

S2: Set up multiple database combinations according to the amount of free memory of the server, the amount of memory occupied by the single-parameter quantization strategy, and the amount of memory occupied by a single financial database; wherein, each of the database combinations includes a certain number of financial databases and each The number of backtesting tasks served by each of the financial databases, and the number of financial databases in each of the database combinations is different;

S3: Back-test the multi-parameter quantification strategy according to the multiple database combinations, the preset back-test time period, and the preset stock target, so as to obtain a back-test corresponding to each of the database combinations operation hours;

S4: According to the specified number of backtesting tasks input by the user for the multi-parameter quantification strategy, search the multiple database combinations and the backtesting running time corresponding to each of the database combinations to obtain multiple optimal databases combination; wherein, the multiple optimal database combinations include the optimal number of financial databases that need to be opened when back-testing the multi-parameter quantitative strategy according to the specified number of back-testing tasks and the services each financial database serves The optimal number of backtesting tasks;

S5: Back-test other stock targets or other back-testing time periods of the multi-parameter quantitative strategy according to the multiple optimal database combinations, where the other stock targets are different from the preset standard stock targets, and the The other back-test time periods are different from the preset back-test time periods, and the other back-test time periods have the same length as the preset back-test time periods.

Preferably, step S4 includes the following steps:

S4.1: compress the multiple database combinations and the backtest running time corresponding to each of the database combinations in a speedup ratio grouping manner to obtain a speedup ratio queue;

S4.2: According to the specified number of backtesting tasks input by the user for the multi-parameter quantization strategy, search in the speedup ratio queue to obtain multiple optimal database combinations.

Preferably, step S4.1 includes:

S4.1.1: Calculate the speedup ratio corresponding to each of the database combinations according to the backtest running time corresponding to each of the database combinations to obtain multiple speedup ratios;

S4.1.2: Group the multiple speedup ratios to form a speedup ratio queue.

Preferably, step S4.1.2 includes the following steps:

S4.1.2.1: According to the maximum acceleration ratio, the minimum acceleration ratio and the preset number of groups among the multiple acceleration ratios, calculate the interval Δspeedup of each group by the following formula:

In the formula, T represents the preset number of groups, speedup _max represents the maximum speedup ratio, and speedup _min represents the minimum speedup ratio;

S4.1.2.2: Group multiple speedup ratios according to each group interval by the following formula to obtain multiple speedup ratio groups S _i :

S _i ={s|[speedup _min +(i-1)*Δspeedup]≤s≤[speedup _min +i*Δspeedup],s∈speedup},1≤i≤T

Wherein, each acceleration ratio group includes multiple acceleration ratios, a database combination corresponding to each acceleration ratio, and a backtesting running time corresponding to the database combination;

S4.1.2.3: Determine the speedup ratio with the largest number of backtesting tasks in each speedup ratio group as the representative parameter of the speedup ratio group to form a set speedup:

Among them, S represents the speedup ratio, S _i represents the speedup ratio group, nf represents the number of financial databases, and ps represents the number of backtesting tasks served by each financial database;

S4.1.2.4: According to a plurality of speedup ratio groups and the representative parameters of each speedup ratio group, a speedup ratio queue is formed.

Preferably, the multi-parameter quantization strategy is obtained by setting the moving average length of the single-parameter quantization strategy as a variable parameter.

Preferably, the single parameter quantization strategy is specifically a single moving average strategy.

Preferably, in step S2, according to the amount of free memory of the server, the amount of memory occupied by the single-parameter quantization strategy, and the amount of memory occupied by a single financial database, multiple database combinations are set to satisfy the following inequality:

Ms*ps*nf+Mf*nf<M

Wherein, M represents the amount of free memory of the server, Ms represents the amount of memory occupied by the single-parameter quantization strategy, and Mf represents the amount of memory occupied by a single financial database.

Preferably, the multiple database combinations are specifically an ordered pair (nf, ps), and the number N of the multiple database combinations is represented by the following formula:

N=nf*ps

Among them, nf represents the number of financial databases, and ps represents the number of backtesting tasks served by each financial database.

Preferably, the multiple optimal database combinations are specifically one or more optimal ordered pairs (nf _optimal , ps _optimal ), and the number of optimal database combinations included in each optimal ordered pair N is expressed by _the following formula:

N _excellent = nf _excellent * ps _excellent

Among them, nf is the _optimal number of financial databases, ps is the _optimal number of _backtesting tasks served by each financial database; and N is ≤ N.

A second aspect of the present invention discloses a system resource scheduling device based on multi-parameter quantization strategy backtesting, comprising:

The standard testing unit is used to back-test the specified single-parameter quantitative strategy according to the preset back-testing time period and the preset standard stock target, and record the amount of memory occupied by the single-parameter quantitative strategy during the back-testing process;

A setting unit, configured to set a plurality of database combinations according to the amount of free memory of the server, the amount of memory occupied by the single-parameter quantification strategy, and the amount of memory occupied by a single financial database; wherein, each of the database combinations includes a certain amount of financial The database and the number of backtesting tasks served by each of the financial databases, and the number of financial databases in each of the database combinations is different;

A back-testing unit for back-testing the multi-parameter quantification strategy according to the multiple database combinations, the preset back-test time period and the preset stock target, so as to obtain each of the databases The backtest running time corresponding to the combination;

The optimization unit is configured to search the multiple database combinations and the backtesting running time corresponding to each of the database combinations according to the specified number of backtesting tasks input by the user for the multi-parameter quantification strategy, so as to obtain multiple database combinations. optimal database combinations; wherein, the multiple optimal database combinations include the optimal number of financial databases that need to be opened when back-testing the multi-parameter quantitative strategy according to the specified number of back-testing tasks and each of the The optimal number of backtesting tasks served by the financial database;

A backtest execution unit, configured to perform backtests on other stock targets or other backtest time periods of the multi-parameter quantitative strategy according to the multiple optimal database combinations, the other stock targets and the preset standard stocks The target is different, the other backtest time period is different from the preset backtest time period, and the other backtest time period and the preset backtest time period have the same length.

Compared with the prior art, the beneficial effects of the technical solution of the present invention are as follows: the present invention discloses a method and device for scheduling system resources based on multi-parameter quantization strategy back-testing, which can be used according to the amount of free memory of the server and the occupation of the single-parameter quantization strategy. The amount of memory and the amount of memory occupied by a single financial database, set up multiple database combinations, and perform backtesting on the multi-parameter quantitative strategy according to the multiple database combinations to obtain the backtesting running time corresponding to each database combination; based on this, when the user inputs Specify the number of backtesting tasks When you want to backtest other stock targets or other backtesting time periods of the multi-parameter quantitative strategy, you can perform backtesting in multiple database combinations and the backtesting running time corresponding to each database combination according to the specified number of backtesting tasks. According to the multiple optimal database combinations, you can determine how many financial databases need to be opened and how many backtest tasks each financial database serves, and use the results to return the multi-parameter quantification strategy. It can shorten the back-test time of multi-parameter quantization strategies, speed up the back-test efficiency of multi-parameter quantization strategies, and reduce resource costs.

Description of drawings

1 is a flowchart of a system resource scheduling method based on multi-parameter quantization strategy backtesting in Embodiment 1.

FIG. 2 is a schematic diagram of a system resource scheduling apparatus based on multi-parameter quantization strategy backtesting in Embodiment 3. FIG.

Among them: 201, standard test unit; 202, setting unit; 203, backtest test unit; 204, optimization unit; 205, backtest execution unit.

Detailed ways

The accompanying drawings are for illustrative purposes only, and should not be construed as limitations on this patent;

The technical solutions of the present invention will be further described below with reference to the accompanying drawings and embodiments.

Example 1

As shown in FIG. 1 , this embodiment provides a system resource scheduling method based on multi-parameter quantization strategy backtesting, including the following steps:

S1: According to the preset backtest time period and the preset standard stock target, backtest the specified single-parameter quantitative strategy, and record the amount of memory occupied by the single-parameter quantitative strategy during the backtesting process.

Optionally, the preset standard stock target is a stock target serving as a reference standard, which can be preset by the developer.

Among them, the single-parameter quantitative strategy refers to a quantitative strategy with a fixed moving average length, and the number of parameters is one, which may be a single moving average or a double moving average strategy.

Optionally, the occupancy information of the single-parameter quantization strategy in the backtesting process can be acquired, which includes the occupied memory amount Ms and the occupied CPU core number. For example, the amount of occupied memory Ms is 12KB, and the number of occupied CPU cores is 1 physical core.

S2: According to the amount of free memory of the server, the amount of memory occupied by the single-parameter quantification strategy, and the amount of memory occupied by a single financial database, multiple database combinations are set; wherein, each database combination includes a certain number of financial databases and each financial database. The number of backtesting tasks for the service, the number of financial databases in each database combination is different.

Optionally, before step S2, a tool or command can be used to obtain the hardware information of the server, and the hardware information includes basic CPU information and memory information; wherein, the basic information of the CPU can be the CPU frequency, the number of CPU cores, etc.; the memory information includes free memory. The amount of free memory, memory frequency, etc.; wherein, the amount of free memory may specifically be the number of bytes of free memory, hereinafter referred to as M.

And, before step S2, a financial database may also be opened in the server, and the amount of memory Mf occupied by the financial database may be obtained. Optionally, the financial database can be a redis database.

Optionally, the combination of multiple databases is specifically an ordered pair (nf, ps), where nf represents the number of financial databases, and ps represents the number of backtesting tasks served by each financial database. The selection range of nf and ps is limited by the amount of free memory M of the server, the amount of memory occupied by the single-parameter quantization strategy Ms, and the amount of memory occupied by a single financial database Mf, that is, the following inequality needs to be satisfied:

Ms*ps*nf+Mf*nf<M (1)

S3: Back-test the multi-parameter quantitative strategy according to multiple database combinations, preset back-test time periods and preset stock targets to obtain the back-test running time corresponding to each database combination;

It should be noted that the single-parameter quantization strategy refers to a quantization strategy with a fixed moving average length. When a user intends to find the single-parameter quantization strategy with the best return from several single-parameter quantization strategies with different moving average lengths, the single-parameter quantization strategy can be used. The average length of the parameter quantization strategy is set as a variable parameter, which constitutes a multi-parameter quantization strategy. The multi-parameter quantization strategy may include multiple single-parameter quantization strategies, and the average length of each single-parameter quantization strategy is different. That is to say, the multi-parameter quantization strategy can be obtained by setting the moving average length of the single-parameter quantization strategy in step S1 as a variable parameter.

Among them, the number of parameters of the multi-parameter quantization strategy is greater than one.

S4: According to the specified number of backtesting tasks input by the user for the multi-parameter quantitative strategy, search through multiple database combinations and the backtesting running time corresponding to each database combination to obtain multiple optimal database combinations;

The multiple optimal database combinations include the optimal number of financial databases that need to be opened when backtesting the multi-parameter quantitative strategy according to the specified number of backtesting tasks, and the optimal number of backtesting tasks served by each financial database.

Optionally, the multiple database combinations are specifically an ordered pair (nf, ps), and the number N of the multiple database combinations is calculated by N=nf*ps.

Further, the multiple optimal database combinations may specifically be one or more optimal ordered pairs (nf _optimal , ps _optimal ), and the number of optimal database combinations included in each _optimal ordered pair N optimally passes through N _excellent = nf _excellent * ps _excellent calculated. Among them, nf is the _optimal number of financial databases, ps is the _optimal number of _backtesting tasks served by each financial database; and N is ≤ N. That is, the total number of backtesting tasks served by financial databases in each optimal database combination is less than or equal to the total number of backtesting tasks served by financial databases in multiple database combinations.

Specifically, when the user inputs the specified number of backtesting tasks n to backtest the multi-parameter quantification strategy, the user can search through multiple database combinations and the backtesting running time corresponding to each database combination according to the specified number of backtesting tasks n , to obtain multiple optimal database combinations.

However, due to the large search space for searching in multiple database combinations and the backtesting running time corresponding to each database combination, the optimization process takes more time. In order to reduce the time spent in the optimization process, optionally, step S4 may include the following steps:

S4.1: Compress multiple database combinations and the backtest running time corresponding to each database combination in a speedup ratio grouping manner to obtain a speedup ratio queue;

S4.2: According to the specified number of backtesting tasks input by the user for the multi-parameter quantization strategy, search in the acceleration ratio queue to obtain multiple optimal database combinations.

Specifically, a search can be performed in the speedup queue to obtain multiple optimal database combinations.

Further optionally, step S4.1 may include the following steps:

S4.1.1: According to the backtest running time corresponding to each database combination, calculate the speedup ratio corresponding to each database combination to obtain multiple speedup ratios.

It should be noted that the speedup ratio is used to characterize the ratio of the running time of the same backtest task in a single-processor system and a parallel-processor system, and is used to measure the performance and effect of parallel systems or program parallelization. Wherein, the single processor system refers to the situation where there is only one financial database and the financial database only serves one backtesting task, the database combination in the present invention belongs to the parallel processor system (that is, one or more financial databases and each financial database service one or more backtesting tasks).

Specifically, the speedup ratio corresponding to each database combination (nf, ps) can be calculated by the following formula:

Among them, N represents the number of operating parameters, N=nf*ps, nf represents the number of financial databases, ps represents the number of backtesting tasks served by each financial database; t(1,1) represents the backtesting tasks in a single-processor system. The running time of the test, t(nf, ps) represents the running time of the backtest corresponding to the database combination.

S4.1.2: Group multiple speedup ratios to form a speedup ratio queue.

Further optionally, step S4.1.2 may include the following steps:

S4.1.2.1: According to the maximum acceleration ratio, the minimum acceleration ratio and the preset number of groups among the multiple acceleration ratios, calculate the interval Δspeedup of each group by the following formula (3):

Among them, T represents the preset number of groups, speedup _max represents the maximum speedup ratio, and speedup _min represents the minimum speedup ratio.

S4.1.2.2: Group multiple speedup ratios according to the following formula (4) according to each group interval to obtain multiple speedup ratio groups S _i :

S _i ={s|[speedup _min +(i-1)*Δspeedup]≤s≤[speedup _min +i*Δspeedup],s∈speedup},1≤i≤T(4)

Wherein, each speedup ratio group includes multiple speedup ratios, a database combination corresponding to each speedup ratio, and a backtesting running time corresponding to the database combination.

It should be noted that when grouping, it is necessary to record the database combination (nf, ps) corresponding to each speedup ratio and the backtest running time t (nf, ps) corresponding to the database combination (nf, ps).

The preset number of groups T can be adjusted according to actual needs. The larger the T, the finer the granularity of the search space, the closer the search results are to the optimal, and the longer the search time. The smaller T is, the coarser the granularity of the search space is, the less the search results are close to the optimal, and the shorter the search time is.

S4.1.2.3: Determine the speedup ratio with the largest number of backtesting tasks in each speedup ratio group as the representative parameter of the speedup ratio group to form a set speedup:

It should be noted that when determining the representative speedup ratio of each group, it is necessary to record the database combination (nf, ps) corresponding to each representative speedup ratio and the backtest running time t (nf, ps) corresponding to the database combination (nf, ps) .

S4.1.2.4: According to a plurality of speedup ratio groups and the representative parameters of each speedup ratio group, a speedup ratio queue is formed.

Correspondingly, step S4.2 may include: searching in the acceleration ratio queue according to the specified number of backtesting tasks input by the user for the multi-parameter quantization strategy to obtain multiple optimal database combinations.

Further optionally, the following dynamic programming algorithm can be used to search in the speedup queue:

After searching the speedup queue through the dynamic programming algorithm, multiple optimal database combinations can be output, which can specifically be the grouping set group:

group={(nf _{good 1} , ps _{good 1} ), (nf _{good 2} , ps _{good 2} ), L, (nf _{good k} , ps _{good k} )} (6)

where k is the number of batches divided. The grouping set group indicates that the specified number of backtesting tasks is divided into k batches for backtesting. The first batch of backtesting tasks uses nf and ₁ financial database, each financial database serves ps and has _one backtesting task, and the second batch of backtesting tasks Use nf to optimize ₂ financial databases, each financial database services ps to optimize ₂ backtesting tasks, and so on, the k-th batch of backtesting tasks uses nf to optimize _k financial databases, and each financial database serves ps to optimize _k backtests Task. Among them, the total number of backtesting tasks in each batch is less than or equal to N.

In this way, by dividing the specified number of backtesting tasks into multiple batches for backtesting, the results of how many financial databases need to be opened for each batch of tasks and how many backtesting tasks each financial database serves can further shorten the backtesting of parallel multi-parameter quantitative strategies. It reduces the testing time, accelerates the backtesting efficiency of multi-parameter quantization strategies, and reduces resource costs.

S5: Back-test other stock targets or other back-test time periods of the multi-parameter quantitative strategy according to multiple optimal database combinations; among which, other stock targets are different from the preset standard stock targets, and other back-test time periods are the same as the preset standard stock targets. The backtest time period is different, and other backtest time periods have the same length as the preset backtest time period.

Optionally, a batch of tasks can also be set, and the number of the batch of tasks is the specified number of backtesting tasks n input by the user, and then according to the optimal database combination in step S4, the existing backtesting methods, such as serial The back-testing method and the single-database multi-process back-testing method perform back-testing on the multi-parameter quantization strategy, and respectively calculate the operating efficiency ratio ₁ improved by using the system resource scheduling method of the present invention compared with the serial back-testing method and using Compared with the single-database multi-process backtesting method, the system resource scheduling method of the present invention improves the operating efficiency ratio ₂ :

Wherein, serial_t is the total running time of the batch of tasks using the serial backtesting method, single_t is the total running time of the batch of tasks using the single-database multi-process backtesting method, and optimized_t is the system resource scheduling method of the present invention. The total running time of this batch of tasks. ratio ₁ represents the operating efficiency improved by using the system resource scheduling method of the present invention compared to the serial backtesting method, and ratio ₂ represents the operating efficiency improved by using the system resource scheduling method of the present invention compared to the single-database multi-process backtesting method .

This embodiment provides a system resource scheduling method based on multi-parameter quantization strategy backtesting. By setting multiple database combinations according to the amount of free memory of the server, the amount of memory occupied by the single-parameter quantization strategy, and the amount of memory occupied by a single financial database, And backtest the multi-parameter quantitative strategy according to multiple database combinations, and obtain the backtesting running time corresponding to each database combination; When performing backtesting in other backtesting time periods, you can search for multiple optimal database combinations in multiple database combinations and the backtesting running time corresponding to each database combination according to the specified number of backtesting tasks. The combination can determine how many financial databases need to be opened and the results of how many backtesting tasks each financial database serves, and use the results to backtest the multi-parameter quantization strategy, which can shorten the backtest time of the multi-parameter quantization strategy and speed up the multi-parameter quantization Backtesting efficiency of strategies while reducing resource costs.

In addition, the specified number of backtesting tasks can also be divided into multiple batches for backtesting, and how many financial databases need to be opened for each batch of tasks and the results of how many backtesting tasks each financial database serves, thereby further shortening the multi-parameter quantification. The backtest time of the strategy is accelerated, the backtest efficiency of the multi-parameter quantitative strategy is accelerated, and the resource cost is reduced at the same time.

Example 2

This embodiment provides a system resource scheduling method based on multi-parameter quantization strategy backtesting, including the following steps:

Step 1: Select Ping An Bank (000001.SZ) as the default standard stock target, and adopt the single moving average strategy as the specified single parameter quantitative strategy. The moving average length of the single moving average strategy is 5, and the amount of memory occupied by the single moving average strategy Ms is 12KB, and the number of CPU cores occupied is 1 physical core.

Step 2: Use tools or commands to obtain hardware information of the server; the hardware information includes basic CPU information and memory information.

Among them, the basic information of the CPU: the main frequency of the CPU is 2.10GHz; the number of CPU cores: 2 physical CPUs, 1 CPU has 16 cores, a total of 64 logical CPUs; memory information: the amount of free memory M is 224GB, and the memory frequency is 2666MHz.

Step 3: Open a redis database on the server as a financial database, and obtain the memory occupied Mf of the financial database. The occupied memory Mf of the financial database is 5.17 GB, that is, one physical core is occupied.

Step 4: According to the amount of free memory M of the server, the amount of memory occupied by the single moving average strategy Ms, and the amount of memory occupied by the financial database Mf, it is calculated by formula (1) that a maximum of 12 different financial databases can be set. The database serves up to 32 backtesting tasks.

Step 5: Set the number of financial databases nf from 1 to 12, and set the number of backtest tasks ps for each financial database service from 1 to 32. After backtesting the multi-parameter quantitative strategy, you can obtain the corresponding data for each database combination. The backtest running time of , as shown in Table 1 below:

Table 1 Running time distribution table

Step 6: For the running time distribution table obtained in Step 5, according to the backtest running time corresponding to each database combination, calculate the speedup ratio corresponding to each database combination, as shown in Table 2 below:

Table 2 List of speedup ratios

Step 7: Take T=100, reduce the dimension of the speedup ratio list, obtain speedup ratio groups, and determine the representative speedup ratio of each speedup ratio group to obtain the speedup ratio queue as shown in Table 3 below:

Table 3 Speedup Queue

Step 8: For the different numbers of backtesting tasks specified by the user, 24, 96, 200, and 320, the dynamic programming algorithm is used to search in the speedup queue Speedup, and the optimal database corresponding to each specified number of backtesting tasks can be obtained. The combined grouping set group, as shown in Table 4 below:

Table 4 Grouping sets of optimal database combinations

Specify the number of backtesting tasks n grouping set group twenty four {(4,6)} 96 {(8,12)} 200 {(8,12),(2,4),(8,12)} 320 {(13,12),(2,4),(13,12)}

Step 9: Back-test the multi-parameter quantification strategy according to the number of each specified back-testing task and its respective set of optimal database combinations, randomly selected stock targets; and, respectively, use existing back-testing methods, such as The serial backtesting method and the single-database multiprocess backtesting method perform backtesting on the multi-parameter quantization strategy, and use the formula (7) to calculate the improvement of the system resource scheduling method of the present invention relative to the serial backtesting method. The operating efficiency ratio ₁ , and the operating efficiency ratio ₂ that is improved by using the system resource scheduling method of the present invention relative to the single-database multi-process backtesting method is calculated by formula (8). As shown in Table 5 below, Table 5 is a comparison table of the operation efficiency of the system resource scheduling method of the present invention with respect to the existing backtesting method.

Table 5 Operational efficiency comparison table

According to Table 5, it can be seen that when the number of specified backtesting tasks is large, using the system resource scheduling method of the present invention improves the efficiency by more than 20 times compared to the serial backtesting method, while using the system resource scheduling method of the present invention is relatively The efficiency improved by the backtesting method of single database and multiple processes is more than 4 times.

It can be seen that by using the system resource scheduling method based on the multi-parameter quantization strategy back-test of the present invention, the operation efficiency of the multi-parameter quantization strategy back-test can be greatly improved, and the resource cost can be reduced at the same time.

Example 3

As shown in FIG. 2 , this embodiment provides a system resource scheduling apparatus based on multi-parameter quantization strategy backtesting, including a standard test unit 201 , a setting unit 202 , a backtest test unit 203 , an optimization unit 204 and a backtest execution unit 205, of which:

The standard testing unit 201 is configured to perform back-testing on a specified single-parameter quantitative strategy according to a preset back-testing time period and a preset standard stock target, and record the amount of memory occupied by the single-parameter quantitative strategy during the back-testing process;

The setting unit 202 is configured to set multiple database combinations according to the amount of free memory of the server, the amount of memory occupied by the single-parameter quantization strategy, and the amount of memory occupied by a single financial database; wherein, each database combination includes a certain number of financial databases and each The number of backtesting tasks served by each financial database, and the number of financial databases in each database combination is different;

The back-testing unit 203 is configured to back-test the multi-parameter quantitative strategy according to a plurality of database combinations, a preset back-test time period and a preset stock target, so as to obtain the back-test running time corresponding to each database combination;

The optimization unit 204 is configured to search among multiple database combinations and the backtesting running time corresponding to each database combination according to the specified number of backtesting tasks input by the user for the multi-parameter quantification strategy, so as to obtain multiple optimal database combinations ; wherein, the multiple optimal database combinations include the optimal number of financial databases that need to be opened when backtesting the multi-parameter quantitative strategy according to the specified number of backtesting tasks, and the optimal number of backtesting tasks served by each financial database;

The backtest execution unit 205 is configured to perform backtests on other stock targets or other backtest time periods of the multi-parameter quantitative strategy according to multiple optimal database combinations. Other stock targets are different from the preset standard stock targets, and other backtest time The period is different from the preset backtest period, and other backtest periods have the same length as the preset backtest period.

This embodiment discloses a system resource scheduling device based on multi-parameter quantization strategy back-testing. By setting multiple database combinations according to the amount of free memory of the server, the amount of memory occupied by the single-parameter quantization strategy, and the amount of memory occupied by a single financial database, And backtest the multi-parameter quantitative strategy according to multiple database combinations, and obtain the backtesting running time corresponding to each database combination; When performing backtesting in other backtesting time periods, you can search for multiple optimal database combinations in multiple database combinations and the backtesting running time corresponding to each database combination according to the specified number of backtesting tasks. The combination can determine how many financial databases need to be opened and the results of how many backtesting tasks each financial database serves, and use the results to backtest the multi-parameter quantization strategy, which can shorten the backtest time of the multi-parameter quantization strategy and speed up the multi-parameter quantization Backtesting efficiency of strategies while reducing resource costs.

Obviously, the above-mentioned embodiments of the present invention are only examples for clearly illustrating the present invention, rather than limiting the embodiments of the present invention. For those of ordinary skill in the art, changes or modifications in other different forms can also be made on the basis of the above description. There is no need and cannot be exhaustive of all implementations here. Any modification, equivalent replacement and improvement made within the spirit and principle of the present invention shall be included within the protection scope of the claims of the present invention.

Claims (7)

1. A system resource scheduling method based on multi-parameter quantization strategy feedback is characterized by comprising the following steps:

s1: according to a preset backtesting time period and a preset standard stock mark, performing backtesting on a specified single-parameter quantization strategy, and recording the memory usage amount of the single-parameter quantization strategy in the backtesting process;

s2: setting a plurality of database combinations according to the idle memory amount of the server, the memory occupied amount of the single-parameter quantization strategy and the memory occupied amount of a single financial database; each database combination comprises a certain number of financial databases and the number of the return testing tasks served by each financial database, and the number of the financial databases in each database combination is different;

setting a plurality of database combinations to satisfy the following inequality according to the idle memory amount of the server, the memory occupied amount of the single parameter quantization strategy and the memory occupied amount of a single financial database:

Ms*ps*nf+Mf*nf＜M

wherein, M represents the free memory amount of the server, Ms represents the memory amount occupied by the single-parameter quantization strategy, and Mf represents the memory amount occupied by a single financial database;

s3: carrying out retest on the multi-parameter quantization strategy according to the plurality of database combinations, the preset retest time period and the preset stock labels to obtain the retest time period corresponding to each database combination;

s4: searching the plurality of database combinations and the corresponding backtesting running time of each database combination according to the specified number of the backtesting tasks input by the user aiming at the multi-parameter quantization strategy so as to obtain a plurality of optimal database combinations; the optimal database combinations comprise the optimal number of financial databases to be started when the multi-parameter quantization strategy is subjected to the recovery according to the specified recovery test task number and the optimal number of recovery test tasks served by each financial database;

the plurality of database combinations are specifically an ordered even pair (ps, nf), and the number N of the plurality of database combinations is represented by the following formula:

N＝nf*ps

wherein nf represents the number of financial databases, ps represents the number of the backtesting tasks served by each financial database;

the plurality of optimal database combinations are specifically one or more optimal ordered even pairs (ps)_{Youyou (an instant noodle)}，nf_{Youyou (an instant noodle)}) The number N of optimal database combinations included in each of the optimal ordered pairings_{Youyou (an instant noodle)}Expressed by the following formula:

N_{superior food}＝nf_{Superior food}*ps_{Youyou (an instant noodle)}

Wherein, nf_{Superior food}Representing the optimal quantity, ps, of financial databases_{Superior food}Representing the optimal number of the return testing tasks served by each financial database; and N is_{Superior food}≤N；

S5: and according to the optimal database combinations, carrying out retesting on other stock labels or other retesting time periods of the multi-parameter quantization strategy, wherein the other stock labels are different from the preset standard stock labels, the other retesting time periods are different from the preset retesting time period, and the other retesting time periods are the same as the preset retesting time period in length.

2. The method for scheduling system resources based on multi-parameter quantization policy backtesting as claimed in claim 1, wherein step S4 comprises the following steps:

s4.1: compressing the plurality of database combinations and the corresponding backtesting running time of each database combination in an acceleration ratio grouping mode to obtain an acceleration ratio queue;

s4.2: and searching in the acceleration ratio queue according to the number of specified backtesting tasks input by a user aiming at the multi-parameter quantization strategy so as to obtain a plurality of optimal database combinations.

3. The method for scheduling system resources based on multi-parameter quantization strategy feedback according to claim 2, wherein step S4.1 comprises:

s4.1.1: calculating the acceleration ratio corresponding to each database combination according to the retest running time corresponding to each database combination to obtain a plurality of acceleration ratios;

s4.1.2: and grouping the plurality of speed-up ratios to form a speed-up ratio queue.

4. The method of claim 3, wherein the step S4.1.2 comprises the following steps:

s4.1.2.1: according to the maximum acceleration ratio, the minimum acceleration ratio and the preset number of groups in the multiple acceleration ratios, calculating each group interval delta speed by the following formula:

wherein T represents a predetermined number of groups, speedup_maxIndicating the maximum acceleration ratio, speedup_minRepresents an acceleration ratio minimum;

s4.1.2.2: grouping a plurality of acceleration ratios according to each group interval by the following formula to obtain a plurality of acceleration ratio groups S_i：

S_i＝{s|[speedup_min+(i-1)*Δspeedup]≤s≤[speedup_min+i*Δspeedup],s∈speedup},1≤i≤T

Each acceleration ratio group comprises a plurality of acceleration ratios, a database combination corresponding to each acceleration ratio and a retest running time corresponding to the database combination;

s4.1.2.3: and determining the acceleration ratio with the maximum retest task number in each acceleration ratio group as a representative parameter of the acceleration ratio group to form a set speed:

wherein S represents an acceleration ratio, S_iRepresenting the acceleration ratio set, nf representing the number of financial databases, ps representing the number of the backtesting tasks served by each financial database;

s4.1.2.4: and forming an acceleration ratio queue according to the acceleration ratio groups and the representative parameters of each acceleration ratio group.

5. The method as claimed in any one of claims 1 to 4, wherein the multi-parameter quantization strategy is obtained by setting the mean line length of the single-parameter quantization strategy as a variable parameter.

6. The method as claimed in claim 5, wherein the single-parameter quantization strategy is a single-average line strategy.

7. A system resource scheduling device based on multi-parameter quantization strategy feedback is characterized by comprising:

the standard test unit is used for carrying out retest on the appointed single-parameter quantization strategy according to the preset retest time period and the preset standard stock mark and recording the memory occupied amount of the single-parameter quantization strategy in the retest process;

a setting unit, configured to set a plurality of database combinations according to the idle memory amount of the server, the memory occupied amount of the single-parameter quantization policy, and the memory occupied amount of a single financial database; each database combination comprises a certain number of financial databases and the number of the return testing tasks served by each financial database, and the number of the financial databases in each database combination is different;

setting a plurality of database combinations to satisfy the following inequality according to the idle memory amount of the server, the memory occupied amount of the single parameter quantization strategy and the memory occupied amount of a single financial database:

Ms*ps*nf+Mf*nf＜M

wherein, M represents the free memory amount of the server, Ms represents the memory amount occupied by the single-parameter quantization strategy, and Mf represents the memory amount occupied by a single financial database;

the retest test unit is used for retesting the multi-parameter quantization strategy according to the multiple database combinations, the preset retest time period and the preset stock labels so as to obtain the retest running time corresponding to each database combination;

the optimizing unit is used for searching the plurality of database combinations and the corresponding backtesting running time of each database combination according to the specified number of backtesting tasks input by a user aiming at the multi-parameter quantization strategy so as to obtain a plurality of optimal database combinations; the optimal database combinations comprise the optimal number of financial databases required to be started when the multi-parameter quantization strategy is subjected to the recovery according to the specified recovery test task number and the optimal number of recovery test tasks served by each financial database;

the plurality of database combinations are specifically an ordered even pair (ps, nf), and the number N of the plurality of database combinations is represented by the following formula:

N＝nf*ps

wherein nf represents the number of financial databases, ps represents the number of the backtesting tasks served by each financial database;

the plurality of optimal database combinations are specifically one or more optimal ordered even pairs (ps)_{Youyou (an instant noodle)}，nf_{Superior food}) The number N of optimal database combinations included in each of the optimal ordered pairings_{Superior food}Expressed by the following formula:

N_{superior food}＝nf_{Superior food}*ps_{Superior food}

Wherein, nf_{Superior food}Representing the optimal quantity, ps, of the financial database_{Superior food}Representing the optimal number of the return testing tasks served by each financial database; and N is_{Superior food}≤N；

And the backtesting execution unit is used for backtesting other stock marks or other backtesting time periods of the multi-parameter quantization strategy according to the optimal database combinations, wherein the other stock marks are different from the preset standard stock marks, the other backtesting time periods are different from the preset backtesting time period, and the other backtesting time periods are the same as the preset backtesting time period in length.

Images