Due to the absence of global knowledge, elements in a self-organizing emergent system tend to make suboptimal local decisions that result in globally inefficient solutions. However, improving the solutions of such systems, which work in a bottom-up style, by the principles of self-adaptive systems, which work in a top-down style, is not a straight forward process. In this paper, we present challenges and constraints that have to be respected during this process and describe early work on an approach, how to autonomously adapt the local behavior of self-organizing elements by so-called exception rules in order to improve the performance of the global solution. In particular, we present a set of exception rules that can be employed in different situations for the improvement of environment-mediated, self-organizing emergent solutions to pickup and delivery problems.
Generated: 28/10/2010