In this letter, we first describe the motivation for such a "distributed optimal planning" functionality; then we outline the underlying requirements in the context of self managed networks and position its role within an overall autonomic control loop. Next, we present an adaptive solution to the issue of constraint-based optimal planning that is based on weighted automata calculus. Finally, we conclude with future directions for this research work.
Distributed Optimal Planning: Context and MotivationsIn self-organizing and self-managing environments, network elements should constantly readjust their configuration and behavior in order to ensure an optimal functioning of the network, while fulfilling the business objectives established by human operators. This is the essence of an autonomic behavior; it is materialized by the implementation of closed control loops [4]. This continual readjustment is achieved through network sensing, learning, knowledge sharing, and distributed cooperation among all network