A Genetic Framework Model for Self-adaptive Software

The recent self-adaptive software systematic literature reviews stated clearly the following insufficiencies: (1) the need for a holistic self-adaptive software model to integrate its different aspects (2) The limitation of adaptations to context changes (3) The absence of a general and complete adaptations' picture allowing its understandability, maintainability, evaluation, reuse, and variability. (4) The need for an explicit and a detailed link with resources, and (5) a usual limitation to known events.In order to metigate these insufficiencies, this paper is proposing a holistic model that integrates the operating, adaptations, and adaptations' manager aspects. The proposed model covers all possible adaptations: operating (dealing with software functions failures), lifecycle (handling adaptations required by some software lifecycle steps), and context (facing context changes events). The presented work introduces the concept of software adaptations process integrating the specifications of all the above kind of adaptations. In fact, this work shows an explicit trace to its pure bio-inspired origin.An application of the proposed approach on a "car industry case study" demonstrated its feasibility in comparison with similar works that proved its meaningful added value and its promising research perspectives.

Section: Discussionmentioning

confidence: 95%

Section: Related Workmentioning

confidence: 99%

Untitled

2021

“…A bio-inspired approach [17,18]is a combination between the biological and the artificial life, in a way to enhance the artificial life through inspired from biological life characteristics. Because they success in solving many artificial problems, there is an increasing demand for these approaches such as neural networks and genetic algorithms, and the improvements it made in hardware sections [19,20,21].…”

Section: Introductionmentioning

confidence: 99%

Requirements Variability Specification for Data Intensive Software

Muslah¹,

Ghoul²

2019

Self Cite

Nowadays, the use of feature modeling technique, in software requirements specification, increased the variation support in Data Intensive Software Product Lines (DISPLs) requirements modeling. It is considered the easiest and the most efficient way to express commonalities and variability among different products requirements. Several recent works, in DISPLs requirements, handled data variability by different models which are far from real world concepts. This,leaded to difficulties in analyzing, designing, implementing, and maintaining this variability. However, this work proposes a software requirements specification methodology based on concepts more close to the nature and which are inspired from genetics. This bio-inspiration has carried out important results in DISPLs requirements variability specification with feature modeling, which were not approached by the conventional approaches.The feature model was enriched with features and relations, facilitating the requirements variation management, not yet considered in the current relevant works.The use of genetics-based methodology seems to be promising in data intensive software requirements variability specification.

Bio-Inspired Requirements Variability Modeling with use Case

Abdel-Ghani¹,

Ghoul²

2019

Background.Feature Model (FM) is the most important technique used to manage the variability through products in Software Product Lines (SPLs). Often, the SPLs requirements variability is by using variable use case modelwhich is a real challenge inactual approaches: large gap between their concepts and those of real world leading to bad quality, poor supporting FM, and the variability does not cover all requirements modeling levels.Aims. This paper proposes a bio-inspired use case variability modeling methodology dealing with the above shortages.Method. The methodology is carried out through variable business domain use case meta modeling, variable applications family use case meta modeling, and variable specific application use case generating.Results. This methodology has leaded to integrated solutions to the above challenges: it decreases the gap between computing concepts and real world ones. It supports use case variability modeling by introducing versions and revisions features and related relations. The variability is supported at three meta levels covering business domain, applications family, and specific application requirements. Conclusion.A comparative evaluation with the closest recent works, upon some meaningful criteria in the domain, shows the conceptual and practical great value of the proposed methodology and leads to promising research perspectives. KEYWORDSSoftware requirements variability, bio-inspired variability modeling, use case variability, feature model, software requirements versions, software requirements revisions.