Software behavior adjusting method based on multivariable control

Abstract

The invention discloses a software behavior adjusting method based on multivariable control. The method comprises the following steps: analyzing the functional logic of target software, determining adjustable configuration parameters influencing the resource consumption of the target software in the working process of the target software, and establishing a mathematical model, namely a transfer function matrix, between the adjustable configuration parameters and software behaviors; designing a decoupler and a control loop for the transfer function matrix by applying a multivariable control theory, and constructing an independent adjusting loop for each software behavior needing to be dynamically adjusted; and attaching the designed decoupler and control loop to the original software logicin an asynchronous execution mode, and adjusting the software behavior in a set adjustment period. According to the method, theoretical knowledge of control engineering and software engineering is combined, the online adjustment problem of actual software under various software behavior scenes needing to be regulated and controlled is effectively solved, and the whole method is simple and convenient to design and apply.

Description

technical field [0001] The invention belongs to the field of software engineering, in particular to a method for adjusting software behavior based on multivariable control. Background technique [0002] As computing systems become increasingly complex, it places a tremendous burden on application developers. This is exacerbated by the dynamic nature of modern systems, which can experience sudden and unpredictable changes in software operations; for example, application workload fluctuations and system component failures. Software engineers are under increasing pressure to manage this complexity and ensure applications run successfully in dynamic environments. In response to this situation, there have been proposals to use autonomous or adaptive technologies to help engineers manage this burden. Adaptive systems respond to unpredictable changes by adjusting the system's own behavior to maintain the stability of the control objective. While tuning of the functional aspects ...

AI Technical Summary

Problems solved by technology

With the improvement of software self-adaptive requirements, it is difficult for these two methods to achieve satisfactory results

The former method essentially adopts the method of mode switching, which can only be used in artificially imagined scenarios, and the control quantity is discretized, it is difficult to guarantee its control accuracy, and the design of its mode will increase the work of software engineers burden; the latter method is only applicable to a small number of code blocks. For large-scale software systems, due to modeling errors and other issues, it is difficult to apply this idea to software adaptive logic design, and it is only adjusted for some single variables. It does not meet the actual needs of multiple adjustable parameters and multiple controlled objects in actual engineering development

Images