Is self-expression useful? Evaluation by a case study

In the context of adaptive component-based systems, self-expression is the capability of changing the adaptation pattern when some changes occur in the system itself or in its environment. Even if functional requirements can be met without changing the adaptation pattern, the achievement of non-functional requirements, such as performance, can benefit from a change of adaptation pattern. The aim of this paper is to show, by means of a case study in swarm robotics, that a change of adaptation pattern can affect the performance of a system. In our case study, the robots must collect food items (foraging); we will show that while in an open arena the Reactive Stigmergy Pattern performs better, the Centralised AM Pattern better suits the presence of obstacles. Since the situation of the environment can be unknown in advance, or it can change at runtime, we claim that the change of adaptation pattern can be useful in adaptive componentbased systems.