Research And Implementation Of Model-Based Self-Adaptive Strategy Verification Mechanism

With the rapid development of internet technology and continuous accumulation of business requirements,the scale of software has shown an exponential growth trend,and the software structure has become more and more complex,resulting in increased tasks and increased costs of manual software adjustment.Therefore,the software system needs to have an independent optimization ability,that is,become self-adaptive software.This type of software can be optimized and adjusted autonomously according to the real-time operating status of the system to better achieve the system goals.Self-adaptive planning can produce the most suitable self-adaptive strategy to provide guidance for the system to adjust its own behavior,parameters,structure and other attributes.The advantages and disadvantages of the self-adaptive strategy directly affect the operating effect of the software system.Therefore,ensuring the effectiveness and reliability of the strategy is a key issue in the field of self-adaptive software.However,the current work mainly focuses on the study of the entire self-adaptive control logic,or focuses on a specific link,and lacks research on verifying the effectiveness,reliability and execution effectiveness of the self-adaptive strategy.On the one hand,the selfadaptive strategy generated by the planning module is default trustworthy and can be implemented directly in the target system in most research work.On the other hand,the existing work lacks dynamical modification of the self-adaptive planning method according to the implementation effect of it.However,the strategy generated by the self-adaptive planning method is not necessarily effective and reliable.Therefore,in order to reduce the uncertain impact of the self-adaptive strategy on the system,it is necessary to verify the effectiveness and reliability of the strategy before it is implemented.At the same time,the implementation effect of the self-adaptive strategy needs to be evaluated as the basis for feedback optimization of the planning module.At the same time,the effect of the selfadaptive strategy after it is implemented needs to be evaluated as the basis for optimization of the planning module.The main goal of this paper is to design and implement an self-adaptive strategy verification mechanism for the above problems,to ensure the effectiveness and reliability of the self- adaptive strategy before the execution of it,and to achieve feedback optimization of the selfadaptive planning process by evaluating the execution effect of the strategy.The main work of this paper includes the following five parts: Firstly,aiming at the problem that the validity of the defined self-adaptive strategy is unknown,a effective verification method based on model checking is proposed.Through formal modeling the system state model and using linear sequential logic to describe the attributes to be verified,the Nu SMV model checking tool is used for verification to ensure the effectiveness of the strategy before execution and to reduce the failure rate of the strategy.Secondly,for the problem that the reliability of the dynamically generated self-adaptive strategy during the system operation process is unknown,a Markov chain-based self-adaptive software quality attribute analysis method is proposed,which transforms the business process model of extended quality attribute into Markov.Based on the Markov model,we calculate the quality attributes of the entire business process and analyze the impact of the dynamically generated adaptive strategy on the system quality attributes to ensure the reliability of the strategy before execution.Thirdly,in view of the lack of evaluation of the execution effect of self-adaptive strategy and the lack of feedback optimization in planning method,a self-adaptive strategy evaluation method based on requirement goal model is proposed.By establishing the requirement goal model of the system and based on the established model and real-time running status of the system,the strategy execution effect is evaluated.Then,according to the evaluation results,the planning method is modified to achieve the purpose of feedback optimization for the planning module.Fourthly,a self-adaptive strategy evaluation center is designed and implemented to support the actual use of the above three strategy verification methods.Fifthly,the self-adaptive strategy verification mechanism implemented in this paper is verified by experiments,and the effectiveness of the proposed method is proved.