CN117149293B - Personalized configuration method for operating system - Google Patents
Personalized configuration method for operating system Download PDFInfo
- Publication number
- CN117149293B CN117149293B CN202311416177.5A CN202311416177A CN117149293B CN 117149293 B CN117149293 B CN 117149293B CN 202311416177 A CN202311416177 A CN 202311416177A CN 117149293 B CN117149293 B CN 117149293B
- Authority
- CN
- China
- Prior art keywords
- configuration
- parameters
- action
- value
- performance
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 67
- 238000012417 linear regression Methods 0.000 claims abstract description 24
- 238000005070 sampling Methods 0.000 claims abstract description 13
- 230000000694 effects Effects 0.000 claims abstract description 11
- 230000002787 reinforcement Effects 0.000 claims abstract description 11
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 10
- 238000007781 pre-processing Methods 0.000 claims abstract description 7
- 230000003631 expected effect Effects 0.000 claims abstract description 6
- 230000009471 action Effects 0.000 claims description 40
- 238000009826 distribution Methods 0.000 claims description 32
- 230000008569 process Effects 0.000 claims description 9
- 230000003993 interaction Effects 0.000 claims description 7
- 230000006870 function Effects 0.000 claims description 6
- 230000007774 longterm Effects 0.000 claims description 4
- 238000012545 processing Methods 0.000 claims description 4
- 238000004140 cleaning Methods 0.000 claims description 3
- 230000006872 improvement Effects 0.000 claims description 3
- 238000012544 monitoring process Methods 0.000 claims description 3
- 238000010606 normalization Methods 0.000 claims description 3
- 238000009827 uniform distribution Methods 0.000 claims description 3
- 230000002159 abnormal effect Effects 0.000 claims description 2
- 238000001514 detection method Methods 0.000 claims description 2
- 238000012795 verification Methods 0.000 claims description 2
- 238000004590 computer program Methods 0.000 description 4
- 238000003860 storage Methods 0.000 description 4
- 238000003491 array Methods 0.000 description 2
- 230000007246 mechanism Effects 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000002790 cross-validation Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 238000013450 outlier detection Methods 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/44—Arrangements for executing specific programs
- G06F9/4401—Bootstrapping
- G06F9/4406—Loading of operating system
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/092—Reinforcement learning
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Software Systems (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Biomedical Technology (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Computer Security & Cryptography (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Machine Translation (AREA)
Abstract
The invention discloses an operating system personalized configuration method, which relates to the technical field of personalized configuration and comprises the steps of collecting user demand data and preprocessing the user demand data; performing model fitting and feature selection by adopting a method combining LASSO regression and stepwise linear regression; dynamically adjusting configuration parameters by adopting a reinforcement learning algorithm; the performance profiles were compared using Wasserstein distance and L2S random sampling was used to select the configuration; and verifying whether the deployed configuration achieves the expected effect, and optimizing according to feedback of a user. The method can realize intelligent personalized configuration of the operating system, can ensure that the operating system realizes optimal or near optimal configuration effect under different use scenes and user requirements through accurate model fitting and feature selection, accurately screens key parameters with the greatest influence on the system performance from a large number of configuration parameters, and constructs a simplified and accurate configuration model according to the key parameters.
Description
Technical Field
The invention relates to the technical field of personalized configuration, in particular to a personalized configuration method of an operating system.
Background
In recent years, with the rapid development of industries such as cloud computing, big data, internet of things and 5G, government authorities, operators and internet enterprises have also put higher demands on personalized configuration at the operating system level, and the operating system is used as the core of a computer system, and the performance and configuration of the operating system directly affect the running efficiency and user experience of the system. Conventional operating system configurations are generally fixed, and it is difficult to meet the diversified demands of different users or different service scenarios. Therefore, how to implement personalized configuration of the operating system according to different usage scenarios and user requirements becomes a problem to be solved.
Disclosure of Invention
The present invention has been made in view of the above-mentioned problems occurring in the conventional operating system personalized configuration method.
Therefore, the problem to be solved by the present invention is how to provide a personalized configuration method for an operating system.
In order to solve the technical problems, the invention provides the following technical scheme: the personalized configuration method of the operating system comprises the steps of collecting user demand data and preprocessing the user demand data; performing model fitting and feature selection by adopting a method combining LASSO regression and stepwise linear regression; dynamically adjusting configuration parameters by adopting a reinforcement learning algorithm; the performance profiles were compared using Wasserstein distance and L2S random sampling was used to select the configuration; verifying whether the deployed configuration achieves the expected effect or not, and optimizing according to feedback of a user; the model fitting and feature selection by adopting the combination method of LASSO regression and stepwise linear regression comprises the following steps of fitting a LASSO regression model by using historical configuration data and corresponding performance data of an operating system, wherein the fitting formula is as follows,
；
in the method, in the process of the invention,refers to regression coefficient, ++>Refers to the operation performance index>Refers to the number of performance index items,/->The j-th configuration parameter, which refers to the i-th performance indicator,>means the total amount of configuration parameters of a certain performance index, +.>The normalization parameter is referred to; extracting configuration parameters with regression coefficients different from 0 in the LASSO regression model as important parameters, and recording; constructing a linear regression model by a least square method based on the important parameters; gradually adding the important parameters from the linear regression model without any configuration parameters until all the important parameters are added to the linear regression model or the p value of the F statistic of the added important parameters is greater than 0.05; from inclusion of all important parametersGradually deleting important parameters which have the smallest contribution to the linear regression model until the p value of the F statistic of the important parameters which have no important parameters to be deleted or deleted is less than 0.05; the important parameters are verified using a cross-validation method.
As a preferable scheme of the personalized configuration method of the operating system, the invention comprises the following steps: dynamically adjusting configuration parameters by adopting a reinforcement learning algorithm comprises the following steps of initializing a Q value for each possible state-action pair, wherein the state is the current configuration parameter, and the action is the selected next configuration parameter; the next action is selected using an epsilon-greedy strategy, based on the current state and the Q value of the action, as follows,
；
in the method, in the process of the invention,representing the probability of choosing a random action,/->Refers to the Q value in state s and action a, representing an estimate of the long-term return for selecting action a in state s, a refers to the selected action,/->Representing the selection of an action maximizing the Q value, < >>Refers to randomly selecting an action from all possible actions; the selected actions are performed, the rewards and new status are observed, and then the Q value is updated using the following formula,
Q′(s,a)←Q(s,a)+α(d)[r+γmax a (Q(s',a'))-Q(s,a)]
where Q (s, a) is the expected return for taking action a in state s, α (D) is the dynamic learning rate based on data type D, α (D) =1/(1+ηD), D is a measure representing the complexity of the data type, D ε [0,100], η is the influence factor based on the complexity of the data type D, η ε [0,1], γ is the discount factor, γ ε [0,1], max (Q (s ', a ')) is the Q value of the possible action taken in the next state s '; generating a configuration list containing all configuration parameters according to a regression model of the source environment; and selecting configuration parameters in the list for sampling according to the Q value function.
As a preferable scheme of the personalized configuration method of the operating system, the invention comprises the following steps: comparing performance distributions using Wasserstein distance and selecting configuration parameters using L2S random sampling includes the steps of collecting performance data of a source environment and a target environment, calculating performance distributions of the two environments using a kernel density estimation method; the wasperstein distance between the two performance distributions is calculated, as follows,
；
where m and n are two performance distributions,is the set of all joint distributions that shift m to n, where c represents one element with an edge distribution m, v represents one element with an edge distribution n, pi (c, v) is the probability of shifting the mass from c to v, inf represents the infinitum, i.e., the smallest possible value; randomly selecting configuration parameters in the configuration list using the uniform distribution; for each randomly selected configuration parameter, calculating the Wasserstein distance between the target distribution and the source distribution after adding the configuration parameter, and selecting the configuration parameter with the minimum Wasserstein distance.
As a preferable scheme of the personalized configuration method of the operating system, the invention comprises the following steps: verifying whether the deployed configuration achieves the expected effect or not, and optimizing according to feedback of a user, wherein the method comprises the following steps of logging in a management interface of an intelligent integrated platform, and uploading related service modules and demand files; in a management interface of the platform, a module and a file which are imported before are sent to a target host; the client software on the target host receives the module and the file sent from the platform and starts to execute the deployment operation; client software on the target host computer can perform personalized configuration of an operating system according to the previously imported demand file; using a performance monitoring tool to compare the service conditions of a CPU, a memory, a disk and a network of a server before and after deployment; if the efficiency improvement rate does not reach the expected value or other problems exist, the configuration strategy is adjusted according to the actual effect.
As a preferable scheme of the personalized configuration method of the operating system, the invention comprises the following steps: the preprocessing comprises data cleaning, missing value processing and abnormal value detection.
As a preferable scheme of the personalized configuration method of the operating system, the invention comprises the following steps: collecting user demand data includes designing a user interaction interface, allowing a user to input business demands and configuration preferences, and automatically generating a demand document according to input information of the user on the interaction interface.
A computer readable storage medium having stored thereon a computer program, characterized in that the computer program when executed by a processor implements the steps of a method according to the operating system personalized configuration method.
A computer device comprising a memory and a processor, said memory storing a computer program, characterized in that the processor, when executing said computer program, implements the steps of a method according to the operating system personalized configuration method.
The invention has the beneficial effects that: the intelligent personalized configuration of the operating system can be realized, and the optimal or near-optimal configuration effect of the operating system under different use scenes and user requirements can be ensured through accurate model fitting and feature selection; by adopting a method of combining LASSO regression and stepwise linear regression, key parameters with the greatest influence on the system performance can be accurately screened from a large number of configuration parameters, and a simplified and accurate configuration model can be constructed according to the key parameters.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly described below, it being obvious that the drawings in the following description are only some embodiments of the present invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
Fig. 1 is a conceptual diagram of an operating system personalized configuration method.
Detailed Description
In order that the above-recited objects, features and advantages of the present invention will become more readily apparent, a more particular description of the invention will be rendered by reference to specific embodiments thereof which are illustrated in the appended drawings.
In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present invention, but the present invention may be practiced in other ways other than those described herein, and persons skilled in the art will readily appreciate that the present invention is not limited to the specific embodiments disclosed below.
Further, reference herein to "one embodiment" or "an embodiment" means that a particular feature, structure, or characteristic can be included in at least one implementation of the invention. The appearances of the phrase "in one embodiment" in various places in the specification are not necessarily all referring to the same embodiment, nor are separate or alternative embodiments mutually exclusive of other embodiments.
Examples
Referring to fig. 1, a first embodiment of the present invention provides an operating system personalized configuration method, where the operating system personalized configuration method includes the following steps:
s1, acquiring user demand data, and preprocessing the user demand data;
s2, performing model fitting and feature selection by adopting a method of combining LASSO regression and stepwise linear regression;
s3, dynamically adjusting configuration parameters by adopting a reinforcement learning algorithm;
s4, comparing performance distribution by adopting Wasserstein distance, and selecting configuration by using L2S random sampling;
and S5, verifying whether the deployed configuration achieves the expected effect, and optimizing according to feedback of a user.
The preprocessing operation includes data cleaning, missing value processing, and outlier detection, specifically, missing values are filled in by a median or mean value when missing values are processed, outliers are detected and processed by the IQR method, outliers are defined as,
K1-1.5xIQR data less than or equal to K3+1.5xIQR
Where K1 and K3 are the first and third quartiles, respectively, and IQR is the difference between K1 and K3.
Collecting user demand data includes designing a user interaction interface, allowing a user to input business demands and configuration preferences, and automatically generating a demand document according to input information of the user on the interaction interface.
In step S2, the model fitting and feature selection using a combination of LASSO regression and stepwise linear regression includes the steps of,
s21, fitting a LASSO regression model by using the historical configuration data of the operating system and the corresponding performance data, wherein the fitting formula is as follows,
；
in the method, in the process of the invention,refers to regression coefficient, ++>Refers to the operation performance index>Refers to the number of performance index items,/->The j-th configuration parameter, which refers to the i-th performance indicator,>means the total amount of configuration parameters of a certain performance index, +.>The normalization parameter is referred to;
s22, extracting configuration parameters with regression coefficients not being 0 in the LASSO regression model as important parameters, and recording;
s23, constructing a linear regression model by a least square method based on the important parameters;
s24, starting from the linear regression model without any configuration parameters, gradually adding the important parameters until all the important parameters are added into the linear regression model or the p value of the F statistic of the added important parameters is greater than 0.05;
s25, gradually deleting the important parameters with the smallest contribution to the linear regression model from the linear regression model containing all the important parameters until the p value of the F statistic of the important parameters without any important parameters which can be deleted or deleted is less than 0.05;
s26, verifying the important parameters by using a cross verification method.
The LASSO regression method can compress certain regression coefficients to 0 in the fitting process by introducing an L1 regularization term, so that automatic selection of features and processing of multiple collinearity are realized, which is important for optimizing configuration of an operating system, because the method can automatically identify and select configuration parameters which have the greatest influence on the system performance, the complexity of a model is simplified, and the interpretation and operability of the model are improved. The stepwise linear regression method further refines the feature selection process, and by gradually adding or deleting features (namely configuration parameters), the model accuracy can be ensured, meanwhile, the model is prevented from being too complex, and the calculation burden is reduced.
Dynamically adjusting the configuration parameters using a reinforcement learning algorithm includes the steps of,
s31, initializing a Q value for each possible state-action pair, wherein the state is a current configuration parameter, and the action is a selected next configuration parameter;
s32, selecting the next action by using an epsilon-greedy strategy according to the current state and the Q value of the action, wherein the formula is as follows,
；
in the method, in the process of the invention,representing the probability of choosing a random action,/->Refers to the Q value in state s and action a, representing an estimate of the long-term return for selecting action a in state s, a refers to the selected action,/->Representing the selection of an action maximizing the Q value, < >>Refers to randomly selecting an action from all possible actions;
s33, executing the selected action, observing the return and the new state, then updating the Q value by using the following formula,
Q′(s,a)←Q(s,a)+α(d)[r+γmax a (Q(s',a'))-Q(s,a)]
where Q (s, a) is the expected return for taking action a in state s, α (D) is the dynamic learning rate based on data type D, α (D) =1/(1+ηD), D is a measure representing the complexity of the data type, D ε [0,100], η is the influence factor based on the complexity of the data type D, η ε [0,1], γ is the discount factor, γ ε [0,1], max (Q (s ', a ')) is the Q value of the possible action taken in the next state s ';
s34, generating a configuration list containing all configuration parameters according to a regression model of the source environment;
s35, selecting configuration parameters in the list for sampling according to the Q value function.
The reinforcement learning algorithm can continuously learn and optimize strategies in interaction with the environment, can self-adjust even if the environment changes, adapts to a new environment, and can dynamically adjust configuration parameters by continuously exploring and utilizing reinforcement learning so as to realize continuous optimization of system performance. And when the reinforcement learning algorithm selects actions, long-term accumulated returns are considered, and the search mechanism is introduced to avoid trapping in a local optimal solution and search for a more global optimization strategy.
The comparison of performance profiles using wasperstein distances and the selection of configuration parameters using L2S random sampling includes the steps of,
s41, collecting performance data of a source environment and a target environment, and calculating performance distribution of the two environments by using a nuclear density estimation method;
s42, calculating the Wasserstein distance between the two performance distributions, wherein the formula is as follows,
；
where m and n are two performance distributions,is the set of all joint distributions that shift m to n, where c represents one element with an edge distribution m, v represents one element with an edge distribution n, pi (c, v) is the probability of shifting the mass from c to v, inf represents the infinitum, i.e., the smallest possible value;
s43, randomly selecting configuration parameters in a configuration list by using uniform distribution;
s44, calculating the Wasserstein distance between the target distribution and the source distribution after adding the configuration parameters for each randomly selected configuration parameter, and selecting the configuration parameter with the minimum Wasserstein distance.
The waserstein distance can accurately measure the difference between two distributions, provide accuracy even if the distributions have different support or in the case of sparse data, and more comprehensively reflect the performance of the system by comparing performance distributions rather than a single performance index. In addition, it can be ensured that the selected configuration enables the performance distribution of the target environment to be as close as possible to the performance distribution of the source environment, thereby achieving target-oriented configuration selection.
The L2S random sampling can introduce certain randomness, avoid sinking into local optimum, accelerate the optimizing speed, and flexibly adjust the sampling strategy to adapt to different optimizing targets and constraints.
Verifying whether the deployed configuration achieves the intended effect, and optimizing according to user feedback includes the steps of,
s51, verifying whether the deployed configuration achieves the expected effect, and optimizing according to the feedback of the user comprises the following steps,
s52, logging in a management interface of the intelligent integrated platform, and uploading related business modules and demand files;
s53, in a management interface of the platform, the previously imported modules and files are sent to a target host;
s54, the client software on the target host receives the module and the file sent from the platform and starts to execute the deployment operation;
s55, the client software on the target host computer performs personalized configuration of the operating system according to the previously imported demand file;
s56, using a performance monitoring tool to compare the service conditions of a server CPU, a memory, a disk and a network before and after deployment;
s57, if the efficiency improvement rate does not reach the expectation or other problems exist, adjusting the configuration strategy according to the actual effect.
In summary, the intelligent personalized configuration of the operating system can be realized, and the optimal or near-optimal configuration effect of the operating system under different use scenes and user requirements can be ensured through accurate model fitting and feature selection. By adopting the method of combining LASSO regression and stepwise linear regression, key parameters with the greatest influence on the system performance can be accurately screened out from a large number of configuration parameters, and a simplified and accurate configuration model is constructed according to the key parameters, so that the operation efficiency of the system can be remarkably improved, unnecessary resource waste can be avoided, and the effect of saving resources is realized while the system performance is ensured.
Further, by means of a dynamic adjustment mechanism of the reinforcement learning algorithm, the system configuration can be optimized in real time, so that the system can automatically adapt to changes of external environments and changes of internal states in the running process. Meanwhile, by adopting Wasserstein distance to compare performance distribution differences under different configurations and combining an L2S random sampling method to further optimize configuration parameter selection, configuration accuracy can be ensured, and meanwhile, configuration diversity and flexibility can be realized, so that excellent configuration effects can be realized under different application scenes and user requirements.
Examples
A second embodiment of the invention, which is different from the previous embodiment, is:
the functions, if implemented in the form of software functional units and sold or used as a stand-alone product, may be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention may be embodied essentially or in a part contributing to the prior art or in a part of the technical solution, in the form of a software product stored in a storage medium, comprising several instructions for causing a computer device (which may be a personal computer, a server, a network device, etc.) to perform all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), a magnetic disk, or an optical disk, or other various media capable of storing program codes.
Logic and/or steps represented in the flowcharts or otherwise described herein, e.g., a ordered listing of executable instructions for implementing logical functions, can be embodied in any computer-readable medium for use by or in connection with an instruction execution system, apparatus, or device, such as a computer-based system, processor-containing system, or other system that can fetch the instructions from the instruction execution system, apparatus, or device and execute the instructions. For the purposes of this description, a "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transport the program for use by or in connection with the instruction execution system, apparatus, or device.
More specific examples (a non-exhaustive list) of the computer-readable medium would include the following: an electrical connection (electronic device) having one or more wires, a portable computer diskette (magnetic device), a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). In addition, the computer readable medium may even be paper or other suitable medium on which the program is printed, as the program may be electronically captured, via, for instance, optical scanning of the paper or other medium, then compiled, interpreted or otherwise processed in a suitable manner, if necessary, and then stored in a computer memory.
It is to be understood that portions of the present invention may be implemented in hardware, software, firmware, or a combination thereof. In the above-described embodiments, the various steps or methods may be implemented in software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, may be implemented using any one or combination of the following techniques, as is well known in the art: discrete logic circuits having logic gates for implementing logic functions on the user demand data signals, application specific integrated circuits having suitable combinational logic gates, programmable Gate Arrays (PGAs), field Programmable Gate Arrays (FPGAs), and the like.
Examples
For the third embodiment of the present invention, in order to verify the advantageous effects of the present invention, scientific demonstration was performed through experiments, and experimental data are shown in table 1.
Table 1 comparison table of experimental data
From an inspection of table 1, it can be seen that the performance of the method according to the present invention is significantly better than the prior art scheme, whether it is light, medium or heavy, and is improved by approximately 30% especially under heavy load. Compared with the prior art, the method saves configuration time under different system scales, and particularly saves 33% of time under a large-scale system. Compared with the prior art, the method has the advantages of less system resource consumption, lower error rate and better satisfaction of customers
It should be noted that the above embodiments are only for illustrating the technical solution of the present invention and not for limiting the same, and although the present invention has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that the technical solution of the present invention may be modified or substituted without departing from the spirit and scope of the technical solution of the present invention, which is intended to be covered in the scope of the claims of the present invention.
Claims (4)
1. An operating system personalized configuration method is characterized in that: comprising the steps of (a) a step of,
collecting user demand data and preprocessing the user demand data;
performing model fitting and feature selection by adopting a method combining LASSO regression and stepwise linear regression;
dynamically adjusting configuration parameters by adopting a reinforcement learning algorithm;
the performance profiles were compared using Wasserstein distance and L2S random sampling was used to select the configuration;
verifying whether the deployed configuration achieves the expected effect or not, and optimizing according to feedback of a user;
the method for model fitting and feature selection by adopting the combination of LASSO regression and stepwise linear regression comprises the following steps,
fitting a LASSO regression model using historical configuration data and corresponding performance data for the operating system, the fitting formula being as follows,
；
in the method, in the process of the invention,refers to regression coefficient, ++>Refers to the operation performance index>Refers to the number of performance index items,/->The j-th configuration parameter, which refers to the i-th performance indicator,>means the total amount of configuration parameters of a certain performance index, +.>The normalization parameter is referred to;
extracting configuration parameters with regression coefficients different from 0 in the LASSO regression model as important parameters, and recording;
constructing a linear regression model by a least square method based on the important parameters;
gradually adding the important parameters from the linear regression model without any configuration parameters until all the important parameters are added to the linear regression model or the p value of the F statistic of the added important parameters is greater than 0.05;
gradually deleting the important parameters which have the smallest contribution to the linear regression model from the linear regression model containing all the important parameters until the p value of the F statistic of the important parameters without any important parameters which can be deleted or deleted is less than 0.05;
verifying the important parameters by using a cross verification method;
dynamically adjusting the configuration parameters using a reinforcement learning algorithm includes the steps of,
initializing a Q value for each possible state-action pair, the state being the current configuration parameter and the action being the next configuration parameter selected;
the next action is selected using an epsilon-greedy strategy, based on the current state and the Q value of the action, as follows,
；
in the method, in the process of the invention,representing the probability of choosing a random action,/->Refers to the Q value in state s and action a, representing an estimate of the long-term return for selecting action a in state s, a refers to the selected action,/->Representing the selection of an action maximizing the Q value, < >>Refers to randomly selecting an action from all possible actions;
the selected actions are performed, the rewards and new status are observed, and then the Q value is updated using the following formula,
Q′(s,a)←Q(s,a)+α(d)[r+γmax a (Q(s',a'))-Q(s,a)]；
where Q (s, a) is the expected return for taking action a in state s, α (D) is the dynamic learning rate based on data type D, α (D) =1/(1+ηD), D is a measure representing the complexity of the data type, D ε [0,100], η is the influence factor based on the complexity of the data type D, η ε [0,1], γ is the discount factor, γ ε [0,1], max (Q (s ', a ')) is the Q value of the possible action taken in the next state s ';
generating a configuration list containing all configuration parameters according to a regression model of the source environment;
selecting configuration parameters in the list for sampling according to the Q value function;
the comparison of performance profiles using wasperstein distances and the selection of configuration parameters using L2S random sampling includes the steps of,
collecting performance data of a source environment and a target environment, and calculating performance distribution of the two environments by using a nuclear density estimation method;
the wasperstein distance between the two performance distributions is calculated, as follows,
；
where m and n are two performance distributions,is the set of all joint distributions that shift m to n, where c represents one element with an edge distribution m, v represents one element with an edge distribution n, pi (c, v) is the probability of shifting the mass from c to v, inf represents the infinitum, i.e., the smallest possible value;
randomly selecting configuration parameters in the configuration list using the uniform distribution;
for each randomly selected configuration parameter, calculating the Wasserstein distance between the target distribution and the source distribution after adding the configuration parameter, and selecting the configuration parameter with the minimum Wasserstein distance.
2. The operating system personalized configuration method according to claim 1, wherein: verifying whether the deployed configuration achieves the intended effect, and optimizing according to user feedback includes the steps of,
logging in a management interface of the intelligent integrated platform, and uploading related business modules and demand files;
in a management interface of the platform, a module and a file which are imported before are sent to a target host;
the client software on the target host receives the module and the file sent from the platform and starts to execute the deployment operation;
client software on the target host computer can perform personalized configuration of an operating system according to the previously imported demand file;
using a performance monitoring tool to compare the service conditions of a CPU, a memory, a disk and a network of a server before and after deployment;
if the efficiency improvement rate does not reach the expected value or other problems exist, the configuration strategy is adjusted according to the actual effect.
3. The operating system personalized configuration method according to claim 2, wherein: the preprocessing comprises data cleaning, missing value processing and abnormal value detection.
4. The operating system personalized configuration method according to claim 3, wherein: collecting user demand data includes designing a user interaction interface, allowing a user to input business demands and configuration preferences, and automatically generating a demand document according to input information of the user on the interaction interface.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311416177.5A CN117149293B (en) | 2023-10-30 | 2023-10-30 | Personalized configuration method for operating system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311416177.5A CN117149293B (en) | 2023-10-30 | 2023-10-30 | Personalized configuration method for operating system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117149293A CN117149293A (en) | 2023-12-01 |
| CN117149293B true CN117149293B (en) | 2024-01-23 |
Family
ID=88884776
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311416177.5A Active CN117149293B (en) | 2023-10-30 | 2023-10-30 | Personalized configuration method for operating system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117149293B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117954113B (en) * | 2024-03-26 | 2024-08-13 | 北京大学 | Information borrowing method and system based on elastic priori and proportional priori principles |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112017025A (en) * | 2020-08-26 | 2020-12-01 | 天元大数据信用管理有限公司 | Enterprise credit assessment method based on fusion of deep learning and logistic regression |
| CN114816909A (en) * | 2022-04-13 | 2022-07-29 | 北京计算机技术及应用研究所 | Real-time log detection early warning method and system based on machine learning |
| US11487273B1 (en) * | 2021-04-30 | 2022-11-01 | Dalian University Of Technology | Distributed industrial energy operation optimization platform automatically constructing intelligent models and algorithms |
| CN116366453A (en) * | 2023-04-18 | 2023-06-30 | 中国电子科技集团公司第三十八研究所 | Self-adaptive dynamic deployment method for heterogeneous network element service demand characterization and virtual network element |
-
2023
- 2023-10-30 CN CN202311416177.5A patent/CN117149293B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112017025A (en) * | 2020-08-26 | 2020-12-01 | 天元大数据信用管理有限公司 | Enterprise credit assessment method based on fusion of deep learning and logistic regression |
| US11487273B1 (en) * | 2021-04-30 | 2022-11-01 | Dalian University Of Technology | Distributed industrial energy operation optimization platform automatically constructing intelligent models and algorithms |
| CN114816909A (en) * | 2022-04-13 | 2022-07-29 | 北京计算机技术及应用研究所 | Real-time log detection early warning method and system based on machine learning |
| CN116366453A (en) * | 2023-04-18 | 2023-06-30 | 中国电子科技集团公司第三十八研究所 | Self-adaptive dynamic deployment method for heterogeneous network element service demand characterization and virtual network element |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117149293A (en) | 2023-12-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| De'Ath | Boosted trees for ecological modeling and prediction | |
| US9129228B1 (en) | Robust and fast model fitting by adaptive sampling | |
| US20200401939A1 (en) | Systems and methods for preparing data for use by machine learning algorithms | |
| Carpentier et al. | Bandit theory meets compressed sensing for high dimensional stochastic linear bandit | |
| KR20230070272A (en) | Computer-based systems, computing components, and computing objects configured to implement dynamic outlier bias reduction in machine learning models | |
| US11004012B2 (en) | Assessment of machine learning performance with limited test data | |
| CN117149293B (en) | Personalized configuration method for operating system | |
| JP6821614B2 (en) | Model learning device, model learning method, program | |
| US10161269B2 (en) | Output efficiency optimization in production systems | |
| CN115983485A (en) | Prediction method and system for carbon emission of power plant | |
| Li et al. | The essential histogram | |
| US20220198260A1 (en) | Multi-level multi-objective automated machine learning | |
| Ahmed et al. | Enhancing stock portfolios for enterprise management and investment in energy industry | |
| Wu et al. | Forecasting online adaptation methods for energy domain | |
| WO2023113946A1 (en) | Hyperparameter selection using budget-aware bayesian optimization | |
| CN112463378B (en) | Server asset scanning method, system, electronic equipment and storage medium | |
| CN113469204A (en) | Data processing method, device, equipment and computer storage medium | |
| CN117669449B (en) | De-excitation circuit determining method, de-excitation circuit determining device, computer equipment and storage medium | |
| US11886516B1 (en) | Generating synthetic data based on time series predictions and plural machine learning models | |
| US20240028902A1 (en) | Learning apparatus and method | |
| US20230195842A1 (en) | Automated feature engineering for predictive modeling using deep reinforcement learning | |
| Lunde et al. | agtboost: Adaptive and automatic gradient tree boosting computations | |
| Liu et al. | BAMSProd: a step towards generalizing the adaptive optimization methods to deep binary model | |
| Fröhlich | Outlier identification and adjustment for time series | |
| WO2024112770A1 (en) | Apparatus and method for computer-implemented modeling of multi-event processes |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |