Selbst-adaptive Software

Evolving Distributed Algorithms With Genetic Programming

“…The devices interacting in such networks often have scarce resources, such as small memory, low processing power, and a limited battery capacity. Additionally to these restrictions, new requirements such as self-adaptation [102] and self-organization [101] arise. A software architect designing a sensor network, for instance, must not only be concerned with functional aspects like data aggregation but also consider non-functional criteria such as energy consumption throughout the complete software engineering process.…”

“…Dazu zählen zum Beispiel kabellose Netze, Sensornetze[139], kabellose Sensornetze[42], Smart Homes, ubiquitäre Systeme[82] und experimentelle Ideen wie Amorphous Computing[2].Die Geräte, die in solchen Netzwerken miteinander interagieren, haben oftmals stark begrenzte Ressourcen wie beispielsweise wenig Speicher, geringe Rechenkraft und niedrige Batteriekapazität. Dazu kommen völlig neue Anforderungen wie Selbstadaption[102] und Selbstorganisation[101]. Der Entwickler eines Sensornetzes, zum Beispiel, muss sich nicht nur auf die funktionalen Aspekte wie Datenaggregation konzentrieren, sondern auch nichtfunktionale Aspekte wie Energieverbrauch während des gesamten Softwareentwurfsprozesses beachten.Es wäre sicher vorschnell, wegen diesen neuen Herausforderungen direkt neue Programmierparadigmen zu fordern.…”

CloudAware: Empowering context‐aware self‐adaptation for mobile applications

“…Taking into consideration the simple example of 3 parameters of which each can have 4 states, already 64 different states are possible and need to be considered by the developer. Considering real‐life examples that have way more parameters and states, as also found by Geihs, it becomes clear that compositional adaptation provides a far more flexible concept to develop self‐adaptive applications, whose 3 main enablers, according to McKinley et al, are the separation of concerns, computational reflection, and component‐based design. A well‐known concept in this context is the paradigm of aspect‐oriented programming, as presented in the work of Appeltauer et al When the separation of concerns is achieved, computational reflection then allows the program to reason about its own behavior and possibly alter it, if required .…”

“…According to the work of Geihs, self‐adaptive systems can be further differentiated by the level of anticipation they provide. If a developer predefines the systems behavior, the degree of anticipation is lower than in the case where no rules are defined, and the anticipation to the current context happens dynamically at runtime, referred to as context awareness.…”

Towards Automotive Embedded Systems with Self-X Properties

“…To fully exploit the potentials of adaptivity, it is not practical to limit the variability by calculating all possible system configurations in advance (during design time). Due to the enormous amount of possible variability in today's and future automotive software systems, it is necessary to adapt the system dynamically at runtime (Dynamic Reconfiguration) (Geihs, 2008). With respect to (Hofmeister, 1993) three different kinds of dynamic reconfiguration can be differentiated: 1.…”