KR20220153774A - Proactive adaptation approach based on statistical model checking for self-adaptive systems - Google Patents

Abstract

Disclosed is a predictive adaptation technique based on a statistical verification of an adaptation-type system. A predictive adaptation method performed in a computer system comprises: a step of monitoring a change for an operating environment of the computer system; and a step of performing a proactive adaptation process by using a statistical model checking (SMC) technique in order to select an optimal action option related to an operation of the computer system according to the change. Therefore, the present invention is capable of having an advantage of not limiting a language.

Description

Statistical verification-based predictive adaptation technique of adaptive systems {PROACTIVE ADAPTATION APPROACH BASED ON STATISTICAL MODEL CHECKING FOR SELF-ADAPTIVE SYSTEMS}

The description below is about predictive self-adapting technologies that need to respond appropriately to various situations, such as self-driving cars and smart factories.

As the complexity of the environment that affects the achievement of a system's goals increases, environmental analysis is important for reliable goal attainment. Environments such as user traffic and outdoor temperature can change over time. Complete anticipation of environmental changes in system design is difficult and often impossible. The system needs to adapt itself at runtime to change its behavior and structure as the environment changes. To achieve this, a design approach based on a mean absolute percentage error (MAPE) feedback loop is being utilized. This adaptation process includes continuous monitoring and analysis of the environment as well as planning and implementation of adaptation.

In most approaches, adaptation is triggered reactively by system failures or environmental changes. Another adaptive approach, known as proactive or predictive adaptation, has emerged, which anticipates changes in advance and is therefore more effective than reactive adaptation in a changing environment. However, predictions of environmental change are uncertain, thus affecting the outcome of prior adaptation. To address uncertainty, probabilistic model checking (PMC) techniques are utilized in some studies to verify adaptation tactics and their impact on system adaptation goals.

PMC-based approaches are the main methods used for active adaptation, but PMCs may not be suitable for validation of large and complex self-adaptive system (SAS) models due to state explosion problems. Verification of complex SAS models and adaptation tactics can fail due to time and memory constraints, as PMC requires high verification costs in time and memory to fully review a given probabilistic model. In addition, modeling languages supported by probabilistic model verifiers must be used for SAS and environment modeling. Engineers should be familiar with modeling languages such as Markov chains, Markov decision processes, or automata that can be interpreted by model verifiers.

In order to overcome the limitations of the PMC-based approach, we propose a Proactive Adaptation approach based on STATistical model checking (hereinafter referred to as 'PASTA') based on SMC.

We propose PASTA, which consumes less verification resources than PMC and requires only the simulation results of the system model without limiting the language.

We propose an approach in which SAS can generate statistically sufficient samples based on the SMC algorithm to mitigate uncertainty in the future environment faster than PMC-based approaches.

A predictive adaptation method performed in a computer system, comprising: monitoring a change in an operating environment of the computer system by at least one processor included in the computer system; and performing, by the at least one processor, a proactive adaptation process using a statistical model checking (SMC) technique to select an optimal action option related to operation of the computer system according to the change. It provides a predictive adaptation method for

According to one aspect, the monitoring may measure a value of an environmental condition via a sensor to capture the change at runtime of the computer system.

According to another aspect, the performing may include predicting a future environmental change based on environmental data accumulated through monitoring; and selecting an optimal adaptation tactic through the SMC technique based on the predicted result of the future environment change.

According to another aspect, the selecting may include generating a sample of a future environment condition based on a prediction result of the future environment change; simulating by applying the adaptive tactic according to the sample to the computer system; Evaluating performance of a corresponding adaptation tactic based on a simulation result of the adaptation tactic; and selecting and executing an optimal adaptation tactic according to a performance evaluation result of the adaptation tactic.

According to another aspect, the performing may further include maintaining knowledge of the computer system up to date, wherein the knowledge includes historical environment data (historical environment data) for predicting the future environment change. data), a simulation executable system model and a set of adaptation tactics specifications for simulating the adaptation tactics, and an adaptation goal specification for performance evaluation of the adaptation tactics. can include

According to another aspect, in the generating step, the sample may be generated through any one SMC algorithm of Simple Monte Carlo Simulation (SMCS), Single Sampling Plan (SSP), and Sequential Probability Ratio Test (SPRT).

According to another aspect, the simulating step may include applying a given adaptation tactic to a system model with a future sample environment, a system model, and an adaptation tactic as inputs, simulating the system in the future environment sample, and determining the effect of the adaptation tactic in the future environment. It can return a log of the simulation result that represents it.

An adaptive air conditioner control method performed in a computer system, comprising: monitoring environmental data including temperature and humidity by at least one processor included in the computer system; predicting, by the at least one processor, a change to the environmental data; generating, by the at least one processor, a sample of future environment data based on a prediction result of the change; simulating by applying an adaptation tactic according to the sample to an air conditioner system; Evaluating performance of a corresponding adaptation tactic based on a simulation result of the adaptation tactic; and controlling the air conditioner system to a temperature and humidity corresponding to an optimal adaptation tactic according to a performance evaluation result of the adaptation tactic.

A computer program stored in a computer readable recording medium to execute a predictive adaptation method on a computer system, the predictive adaptation method comprising: monitoring a change in an operating environment of the computer system; and performing a proactive adaptation process using a statistical model checking (SMC) technique to select an optimal action option related to operation of the computer system according to the change.

A computer system comprising: at least one processor configured to execute computer readable instructions contained in a memory, the at least one processor comprising: monitoring a change in an operating environment of the computer system; and a computer system that processes a process of performing a proactive adaptation process using a statistical model checking (SMC) technique to select an optimal action option related to operation of the computer system according to the change.

Embodiments of the present invention have the advantage of consuming less verification resources and not restricting the language by performing predictive self-adaptation using SMC.

SMC is different from PMC in that it does not prove the model mathematically like PMC, but generates a large number of samples from a probability model and statistically verifies the model based on the generated samples.

Since SMC is based on randomly generated samples, results may vary from run to run and return verification results limited within a certain level of confidence rather than fully reliable verification results.

SMC is faster than PMC, can perform with smaller memory, and can return results similar to PMC. Therefore, by using SMC in the field of predictive self-adaptation, verification results similar to those using PMC can be obtained with less resources, cost, and shorter verification time.

Figure 1 shows the main process of PMC-based pre-adaptation.
2 is a diagram for explaining an example of self-adaptation of an air conditioner control system.
Figure 3 shows the overall PASTA process in one embodiment of the present invention.
4 illustrates a PASTA reference architecture for implementing PASTA in one embodiment of the present invention.
5 shows a class diagram of the PASTA framework in one embodiment of the present invention.
6 is a block diagram illustrating an example of a computer system according to one embodiment of the present invention.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.

Embodiments of the present invention relate to predictive self-adapting techniques.

The self-adaptive system monitors the operating environment of the system in real time and determines the optimal action based on the current situation. Since it is difficult to accurately specify the operating environment at the time of system development, it can be said to be an uncertain factor from the point of view of the system. Because of the uncertainty of how the operating environment will change in the future, the system cannot definitively predict the consequences of its actions. Therefore, the system predicts the result of its own behavior based on the limited information it has. perform self-adaptation.

In this embodiment, a pre-adaptation approach utilizing SMC is proposed to remove uncertainties of the future environment faster than PMC for adaptive tactics verification. In addition, it provides the algorithm process, reference architecture, and open source implementation skeleton of PASTA for developers who apply PASTA to SAS development. And, based on the evaluation using actual data, the pros and cons of PMC and SMC-based pre-adaptation are compared and analyzed to guide the engineer's decision-making.

First, the pre-adaptation technique is described as a related study.

Numerous studies on proactive or predictive adaptation are being conducted to address challenges related to changing environments. Unlike reacting to changes in the environment or system, anticipating and responding to predicted situations is more difficult, but can be more effective in preventing system failures and meeting requirements. A number of case studies of pre-adaptation have been conducted, demonstrating that pre-adaptation outperforms post-adaptation in terms of a system's adaptation goals. For pre-adaptation, since the prediction of the future environment is uncertain, PMC is utilized to verify the satisfaction of the characteristics of the probabilistic model in order to provide verified and reliable pre-adaptation results.

Figure 1 shows the main process of PMC-based pre-adaptation.

At the heart of the process is formal modeling of future environments, systems and adaptation tactics, and validation of the models to identify optimal adaptation tactics to achieve adaptation goals. However, PMC is not suitable for verification of large and complex models due to the state explosion problem. To validate adaptation tactics, all possible states of the SAS model must be thoroughly examined. Additionally, engineers must develop SAS models written in a modeling language that the model verifier can support.

In order to overcome the limitations of PMC, in this embodiment, an SMC-based pre-adaptation approach is proposed as an alternative to the PMC-based approach, which has been a major trend of pre-adaptation.

We are leveraging SMC to validate adaptive tactics at runtime in an uncertain environment. SMC is an efficient technique for validating probabilistic models. While PMC exhaustively examines the model, SMC simulates the model to obtain a sample and uses hypothesis testing on the sample to provide statistical evidence for satisfaction or violation of a given property. In practice, SMC only requires a set of simulation results, so it can only be applied to a viable black box model or set of simulation results. The fact that the verification result is determined by the quality of the model is the same as PMC, but as it is a simulation-based approach, it is known as an efficient alternative to PMC in terms of time and memory while performing verification under certain conditions. In this regard, SMC can be effectively used for runtime verification of SAS adaptation tactics in an uncertain environment. The following example of the SMC algorithm is widely used.

For example, the Simple Monte Carlo Simulation (SMCS) algorithm is the simplest and most intuitive SMC algorithm. Estimate the quantitative satisfaction of a property according to the proportion of samples that satisfy the property of the entire sample. The user must extract the sample.

Next, a single sampling plan (SSP) tests hypothesis H: p≥θ with fixed-size samples. where p is the probability that the system satisfies the given property and θ is the test threshold for p. The user provides two error limits α (0≤α≤1) and β (0≤β≤1) of false negative and false positive, respectively. Within the specified margin of error, the SSP assumes p and accepts or rejects H.

In addition, Sequential Probability Ratio Test (SPRT) tests hypothesis H within a given margin of error, similar to SSP, but the number of samples is automatically determined. We simulate the target system to obtain samples, and repeat the simulation to generate enough samples until we can accept or reject H within a given margin of error.

In the case of the PASTA approach according to the present invention, the SMC algorithm is selected and used to obtain statistical evidence of the performance of adaptive tactics in future environments, to evaluate possible tactics and to identify optimal tactics at runtime.

As an example, PASTA is described using an adaptive air conditioning control system. The system monitors indoor and outdoor air conditions, including temperature and humidity, and adaptively controls indoor conditions for specified target conditions. Planning for adaptive air conditioning control that instantly reacts to monitored indoor conditions can help the system achieve its goals, but as shown in Figure 2, indoor air quality may change over time due to the influence of outdoor air conditions. status can change. If adaptation plans are made without affecting environmental changes, adaptation outcomes may differ from expectations, so there may be better adaptation tactics that are not selected. The adaptive air conditioning control system developed by the PASTA approach predicts future air condition changes and selects the optimal adaptation tactic whose adaptation result is verified by the SMC at runtime.

In the following embodiment, a specific PASTA approach will be described using these cases.

In this embodiment, the PASTA approach, which is a proactive adaptation technique, is proposed using SMC.

Figure 3 shows the overall PASTA process in one embodiment of the present invention.

The purpose of PASTA is to provide efficient pre-adaptation based on the prediction of environmental changes and validation of SAS' adaptation tactics.

Referring to Figure 3, the PASTA system initially PASTA continuously monitors the environment to catch changes in runtime (S301).

The PASTA system analyzes monitored (historical) environmental data and predicts future environmental changes based on a prediction algorithm (S302). A prediction or expectation of the future environment is a form of non-deterministic probability, such as a probability density function of future environmental conditions.

The PASTA system creates samples of possible future environments based on the prediction results for future environment changes and provides them to the simulation engine (S303).

In the PASTA system, adaptive tactics are applied to the system model in a given environment and simulated for sample evaluation of the performance of the tactics (S304). The simulation is repeated until the system obtains a statistically sufficient number of samples to verify the performance of the tactics against the adaptation target in expected future environmental changes (S305).

The PASTA system verifies the performance of the adaptation tactic based on the accumulated samples (S306). All adaptation tactics are repeatedly evaluated in the same way (S307), and the SAS statistically guarantees the effectiveness of the adaptation tactics.

After evaluating all possible adaptation tactics, the PASTA system selects and executes the optimal adaptation tactics (S308 to S309).

The above adaptation process (S301 to S309) is continuously repeated to respond to continuous environmental changes, and requires environmental data, an executable system model, adaptation tactical specifications, and adaptation target specifications as inputs.

The detailed description of the PASTA approach based on the above adaptation process is as follows.

Knowledge

<Principle> The PASTA approach requires SAS to accumulate monitored environmental data. It analyzes accumulated historical environmental data to predict environmental changes. Furthermore, the system has a current system model, which is an abstraction of system behavior executable by the simulator. PASTA's models are user-specific, and the modeling language and system information are chosen by the engineer, but the only requirement is that the model can be run to generate simulation logs. In addition, the system model contains a set (space) of adaptation tactics that can be verified. An adaptation tactic is an adaptation specification, such as a configuration set, that can be applied to SAS and SAS models. Adaptive goals are also specified in knowledge. Thus, the optimal tactic for the adaptation goal will be selected and executed.

<Case> The environmental factors of interest in the adaptive air conditioning control system are indoor/outdoor temperature and indoor/outdoor humidity. Therefore, the environmental data monitored at a specific time includes the values of the four factors. The simulation model mimics changes in indoor temperature and humidity that are affected by outdoor conditions and control values of the air conditioning control system. The system's possible adaptation tactics are determined by the system capabilities of each temperature and humidity control function. For example, the system can increase or decrease temperature and humidity in individual simulation time units, from 0.1°C, 0.1% up to 5°C, 5%. Tact space is the Cartesian product of possible temperature and humidity control. The adaptation goal is to adjust the indoor temperature and humidity according to the conditions desired by the user.

Environmental change monitoring (S301)

<Principle> The system continuously monitors the environment. Environment is measured as a value of environmental conditions that can be observed by sensors. The current environment data is added to the environment database. The current state of the system is also monitored and the system model is kept up to date.

<Case> An adaptive air conditioning control system continuously monitors indoor/outdoor temperature and indoor/outdoor humidity. Accumulate environmental data in the environmental database.

Prediction of future environmental changes (S302)

<Principle> PASTA predicts future environmental changes based on accumulated historical environmental data using data analysis or forecasting techniques. Given that historical environmental data consists of time series data, time series analysis and forecasting methods such as random walks, seasonal error trends, autoregressive integrated moving average models, or machine learning techniques can be applied, and the choice of forecasting method is up to the domain engineer. What is important here is that the prediction result is a non-deterministic expectation, such as a probability density function of future environmental conditions, because predictions of future environmental changes based on past data are uncertain. This uncertainty will be addressed by SMC.

<Example> The system predicts outdoor temperature and humidity changes with a distinct repeating pattern (seasonality) at 24-hour intervals. Since the environmental data of this system exhibits a distinct seasonality, seasonal differences can be used to predict them purely using a random walk model. Based on past temperature data and prediction algorithms, a probability density function can be used to predict the temperature change from now to several hours in the future. For example, if the current temperature at 2:00 PM is 24°C, the temperature at 3:00 PM may be expected to change according to a uniform distribution between 24°C and 30°C.

Adaptation plan using SMC (S303 to S308)

[Table 1]

<Principle> Adaptation planning of the PASTA approach involves searching for the optimal tactic among possible adaptation tactics using SMC as shown in Algorithm 1 of Table 1. Evaluating adaptation tactics using SMC consists of three steps: environmental change sampling, adaptation strategy simulation, and simulation result verification. The prediction results are non-deterministic so that the sample generator creates deterministic samples of possible future environmental conditions based on the prediction results. SMC produces statistically sufficient samples to eliminate uncertainties in non-deterministic future environments, whereas PMC probabilistically validates probabilistic models. The number of samples is determined according to the SMC algorithm as described above. The simulator takes as input a sample environment, system model and adaptation tactics. It applies the given tactic to the system model, simulates the system in a sample of the future environment, and returns a simulation result log representing the effect of the adaptive tactic in the future environment. The verifier receives a number of simulation results and evaluates the performance of the tactics against the adaptation goal represented by the verification properties. This process is performed for every adaptation tactic, and the optimal tactic is selected according to all evaluation (verification) results. Thus, the planning time required for adaptation depends on the number of tactics, the number of samples required, and the single simulation time of the model.

<Case> A sample of future outdoor temperatures (e.g. 25°C, 27°C, 29°C) is randomized by the SMC algorithm based on the range of expected temperature changes at 3:00 PM (24°C to 30°C). is selected as In the current assessment, the system model and adaptation tactics (eg lowering the room temperature by 3°C) are each simulated with a sample environment. The verifier evaluates the adaptation result of room temperature control based on the simulation results. In this example, the average distance between the target condition and the current condition is used as a verification attribute representing the adaptation target, but the maximum distance representing the presence or absence of an event occurring with a small probability in the worst case or temporal logic may be used as a verification attribute. When all possible temperature and humidity control methods are identified (evaluated), the optimal method is selected.

Adaptation execution (S309)

<Principle> The chosen optimal adaptation tactic is applied to the system managed by the system's actuators.

<Case> The adaptive air conditioner control system operates the selected optimum temperature and humidity controller. The control unit influences room conditions through actuators in the system.

In addition to the above-mentioned adaptive air conditioner control system, it can be applied to all systems that need to respond appropriately to various situations, such as an intersection dynamic signal control system, an autonomous vehicle control system, and a smart factory control system.

PASTA Implementation

4 illustrates a PASTA reference architecture for implementing PASTA in one embodiment of the present invention.

Figure 4 shows the layered architecture of SAS with the PASTA approach.

In the interaction layer (interaction layer) 410, PASTA monitors the system that manages the driving environment through a sensor 411 and influences it through an actuator 412 such as a general SAS.

In the data analysis layer 420, a forecasting engine 421 for predicting environmental changes and a knowledge management module 422 for always maintaining up-to-date knowledge of the system are provided. can be included

System knowledge includes historical environment data used to predict future environmental changes, a simulation executable system model used to simulate adaptive tactics, and a set of adaptive tactics specifications. adaptation strategies specification) and adaptation goal specifications used for performance evaluation of adaptation tactics.

The optimal adaptation tactic searching module 431 in the adaptation planner layer 430 obtains an optimal adaptation tactic through interaction with the adaptation verification layer 440. Search for

The SMC module 444 in the adaptive verification layer 440 verifies the adaptive tactics controlling the sample generator 441 , simulator 442 and verifier 443 .

Here, the sample generator 441 generates a sample of a future environment based on the prediction of the prediction engine. Simulator 442 simulates the system model with adaptive tactics in a given sample future environment. The verifier 443 analyzes the simulation results to analyze achievement of adaptation goals such as quality of service or immutable properties.

The knowledge layer (450) includes an environment database (451), a system model manager (452), an adaptation tactic repository (453) and an adaptation goal manager (453). manager) 454 may be included. Knowledge layer 450 interacts with other layers while providing and updating SAS knowledge.

Since the above architecture is a reference, it can be extended to further include essential components of SAS.

In addition, to support engineers developing SAS based on the PASTA approach described above, we implement a PASTA skeleton based on a reference architecture with guiding comments and release the source code in an open source repository. The skeleton is available in Java and Python. Engineers must write application-specific code after comments tagged with 'todo'.

5 shows a class diagram of the PASTA framework. Adaptation is activated by the 'adaptManagedSystem' operator. Facilitates easier PASTA implementation allowing the use of third-party libraries or tools for some components, such as prediction engine 421 or SMC module 444.

The present invention is implemented in a modular form and is easily applicable to various systems, and predictive self-adaptation can be performed using structured and unstructured data, and through this, situation understanding and scheduling, decision-making and prediction, recommendation and It can be applied in a variety of situations, such as actions.

6 is a block diagram illustrating an example of a computer system according to one embodiment of the present invention. The PASTA system described above may be implemented by the computer system 600 configured as shown in FIG. 6 .

As shown in FIG. 6, a computer system 600 is a component for executing the PASTA method according to embodiments of the present invention, and includes a memory 610, a processor 620, a communication interface 630, and an input/output interface. (640).

The memory 610 is a computer-readable recording medium and may include a random access memory (RAM), a read only memory (ROM), and a permanent mass storage device such as a disk drive. Here, a non-perishable mass storage device such as a ROM and a disk drive may be included in the computer system 600 as a separate permanent storage device distinct from the memory 610. Also, an operating system and at least one program code may be stored in the memory 610 . These software components may be loaded into the memory 610 from a recording medium readable by a separate computer from the memory 610 . The separate computer-readable recording medium may include a computer-readable recording medium such as a floppy drive, a disk, a tape, a DVD/CD-ROM drive, and a memory card. In another embodiment, software components may be loaded into the memory 610 through the communication interface 630 rather than a computer-readable recording medium. For example, software components may be loaded into memory 610 of computer system 600 based on a computer program installed by files received over network 660 .

The processor 620 may be configured to process commands of a computer program by performing basic arithmetic, logic, and input/output operations. Instructions may be provided to processor 620 by memory 610 or communication interface 630 . For example, processor 620 may be configured to execute received instructions according to program codes stored in a recording device such as memory 610 .

The communication interface 630 may provide functionality for the computer system 600 to communicate with other devices and each other via the network 660 . For example, a request, command, data, file, etc. generated by the processor 620 of the computer system 600 according to a program code stored in a recording device such as the memory 610 is transmitted to a network ( 660) to other devices. Conversely, signals, commands, data, files, etc. from other devices may be received into computer system 600 via communication interface 630 of computer system 600 via network 660 . Signals, commands, data, etc. received through the communication interface 630 may be transferred to the processor 620 or the memory 610, and files, etc. may be stored as storage media that the computer system 600 may further include (described above). permanent storage).

The communication method is not limited, and may include not only a communication method utilizing a communication network (eg, a mobile communication network, wired Internet, wireless Internet, and broadcasting network) that the network 660 may include, but also short-distance wired/wireless communication between devices. have. For example, the network 660 may include a personal area network (PAN), a local area network (LAN), a campus area network (CAN), a metropolitan area network (MAN), a wide area network (WAN), and a broadband network (BBN). , one or more arbitrary networks such as the Internet. In addition, the network 660 may include any one or more of network topologies including a bus network, a star network, a ring network, a mesh network, a star-bus network, a tree or a hierarchical network, and the like. Not limited.

The input/output interface 640 may be a means for interface with the input/output device 650 . For example, the input device may include devices such as a microphone, keyboard, camera, or mouse, and the output device may include devices such as a display and a speaker. As another example, the input/output interface 640 may be a means for interface with a device in which functions for input and output are integrated into one, such as a touch screen. The input/output device 650 may be configured as one device with the computer system 600 .

Also, in other embodiments, computer system 600 may include fewer or more components than those of FIG. 6 . However, there is no need to clearly show most of the prior art components. For example, the computer system 600 may be implemented to include at least some of the aforementioned input/output devices 650 or may further include other components such as transceivers and various databases.

As described above, according to embodiments of the present invention, by implementing predictive self-adaptation technology using SMC, verification results similar to those using PMC can be obtained with less resources, cost, and shorter verification time.

The devices described above may be implemented as hardware components, software components, and/or a combination of hardware components and software components. For example, devices and components described in the embodiments include a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable gate array (FPGA), and a programmable PLU (programmable logic unit). logic unit), microprocessor, or any other device capable of executing and responding to instructions. The processing device may run an operating system (OS) and one or more software applications running on the operating system. A processing device may also access, store, manipulate, process, and generate data in response to execution of software. For convenience of understanding, there are cases in which one processing device is used, but those skilled in the art will understand that the processing device includes a plurality of processing elements and/or a plurality of types of processing elements. It can be seen that it can include. For example, a processing device may include a plurality of processors or a processor and a controller. Other processing configurations are also possible, such as parallel processors.

Software may include a computer program, code, instructions, or a combination of one or more of the foregoing, which configures a processing device to operate as desired or processes independently or collectively. You can command the device. The software and/or data may be embodied in any tangible machine, component, physical device, computer storage medium or device to be interpreted by or to provide instructions or data to a processing device. have. Software may be distributed on networked computer systems and stored or executed in a distributed manner. Software and data may be stored on one or more computer readable media.

The method according to the embodiment may be implemented in the form of program instructions that can be executed through various computer means and recorded on a computer readable medium. In this case, the medium may continuously store a program executable by a computer or temporarily store the program for execution or download. In addition, the medium may be various recording means or storage means in the form of a single or combined hardware, but is not limited to a medium directly connected to a certain computer system, and may be distributed on a network. Examples of the medium include magnetic media such as hard disks, floppy disks and magnetic tapes, optical recording media such as CD-ROM and DVD, magneto-optical media such as floptical disks, and ROM, RAM, flash memory, etc. configured to store program instructions. In addition, examples of other media include recording media or storage media managed by an app store that distributes applications, a site that supplies or distributes various other software, and a server.

As described above, although the embodiments have been described with limited examples and drawings, those skilled in the art can make various modifications and variations from the above description. For example, the described techniques may be performed in an order different from the method described, and/or the components of the described system, structure, device, circuit, etc. may be combined or combined in a different form than the method described, or other components may be used. Or even if it is replaced or substituted by equivalents, appropriate results can be achieved.

Therefore, other implementations, other embodiments, and equivalents of the claims are within the scope of the following claims.

Claims (17)

In a predictive adaptation method performed in a computer system,
monitoring a change in an operating environment of the computer system by at least one processor included in the computer system; and
performing, by the at least one processor, a proactive adaptation process using a statistical model checking (SMC) technique to select an optimal action option related to operation of the computer system according to the change;
A predictive adaptation method comprising a. According to claim 1,
The monitoring step is
Measuring a value of an environmental condition via a sensor to capture the change at runtime of the computer system.
A predictive adaptation method characterized by. According to claim 1,
The above steps are
Predicting future environmental changes based on environmental data accumulated through monitoring; and
Selecting an optimal adaptation tactic through the SMC technique based on the predicted result of the future environmental change
A predictive adaptation method comprising a. According to claim 3,
The selection step is
generating a sample of a future environmental condition based on a prediction result of the future environmental change;
simulating by applying the adaptive tactic according to the sample to the computer system;
Evaluating performance of a corresponding adaptation tactic based on a simulation result of the adaptation tactic; and
Selecting and executing an optimal adaptation tactic according to a performance evaluation result of the adaptation tactic
A predictive adaptation method comprising a. According to claim 4,
The above steps are
Keeping knowledge of the computer system up to date
Including more,
The knowledge includes historical environment data for predicting the future environment change, a simulation executable system model for simulating the adaptation tactics, and a set of adaptation tactics specifications. , including an adaptation goal specification for performance evaluation of the adaptation tactic.
A predictive adaptation method characterized by. According to claim 4,
The generating step is
Generating the sample through any one SMC algorithm of Simple Monte Carlo Simulation (SMCS), Single Sampling Plan (SSP), and Sequential Probability Ratio Test (SPRT)
A predictive adaptation method characterized by. According to claim 4,
The simulating step is
Applying a given adaptation tactic to the system model with a future sample environment, system model, and adaptation tactic as input, simulating the system in a sample of the future environment, and returning a simulation result log representing the effect of the adaptation tactic in the future environment.
A predictive adaptation method characterized by. In the adaptive air conditioner control method performed in a computer system,
monitoring environmental data including temperature and humidity by at least one processor included in the computer system;
predicting, by the at least one processor, a change to the environmental data;
generating, by the at least one processor, a sample of future environment data based on a prediction result of the change;
simulating by applying an adaptation tactic according to the sample to an air conditioner system;
Evaluating performance of a corresponding adaptation tactic based on a simulation result of the adaptation tactic; and
Controlling the air conditioner system to a temperature and humidity corresponding to an optimal adaptation tactic according to a performance evaluation result of the adaptation tactic
Adaptive air conditioning control method comprising a. A computer program stored in a computer readable recording medium for executing a predictive adaptation method on a computer system,
The predictive adaptation method,
monitoring changes in the operating environment of the computer system; and
Performing a proactive adaptation process using a statistical model checking (SMC) technique to select an optimal action option related to the operation of the computer system according to the change.
Including, a computer program. In a computer system,
at least one processor configured to execute computer readable instructions contained in memory;
including,
The at least one processor,
monitoring changes in the operating environment of the computer system; and
A process of performing a proactive adaptation process using a statistical model checking (SMC) technique to select an optimal action option related to the operation of the computer system according to the change
A computer system that processes According to claim 10,
The at least one processor,
Measuring a value of an environmental condition via a sensor to capture the change at runtime of the computer system.
Characterized by a computer system. According to claim 10,
The at least one processor,
Predict future environmental changes based on environmental data accumulated through monitoring,
Selecting an optimal adaptation tactic through the SMC technique based on the prediction result for the future environmental change
Characterized by a computer system. According to claim 12,
The at least one processor,
Based on the prediction result of the future environmental change, a sample for future environmental conditions is generated,
Applying and simulating an adaptation tactic according to the sample to the computer system;
Evaluate the performance of the adaptive tactic based on the simulation results for the adaptive tactic;
Selecting and executing an optimal adaptation tactic according to the performance evaluation result of the adaptation tactic
Characterized by a computer system. According to claim 13,
The at least one processor,
Keep your knowledge of the computer system up to date;
The knowledge includes historical environment data for predicting the future environment change, a simulation executable system model for simulating the adaptation tactics, and a set of adaptation tactics specifications. , including an adaptation goal specification for performance evaluation of the adaptation tactic.
Characterized by a computer system. According to claim 13,
The at least one processor,
Generating the sample through any one SMC algorithm of Simple Monte Carlo Simulation (SMCS), Single Sampling Plan (SSP), and Sequential Probability Ratio Test (SPRT)
Characterized by a computer system. According to claim 13,
The at least one processor,
Applying a given adaptation tactic to the system model with a future sample environment, system model, and adaptation tactic as input, simulating the system in a sample of the future environment, and returning a simulation result log representing the effect of the adaptation tactic in the future environment.
Characterized by a computer system. According to claim 10,
The computer system is an adaptive system of any one of a computer-implemented air conditioner control system, an intersection signal control system, an autonomous vehicle control system, and a smart factory control system.
Characterized by a computer system.

Images