Abstract
A wide area telecommunication network consists of a set of geographically distributed sites that are connected by communication links. An important and frequently occurring problem in such networks is circuit disruption. Circuits may be disrupted by line, node or power failures. Such situations may arise after natural disasters such as hurricanes, earthquakes, floods, or snow storms. Emergency situations such as these disrupt circuits and cause changes in network parameters affecting otherwise routine network activities. In such environments, it is important that the network system has a predefined methodology to plan the restoration process. When a circuit is disrupted, it is desirable to find an optimal or near-optimal alternative path that restores the circuit.
The objective of this paper is to apply a Distributed Artificial Intelligence (DAI) methodology to the network restoral planning problem. This approach uses a bidirectional search methodology to identify a near-optimal restoral circuit. The goal is to use dynamic information about a network and domain-specific heuristics to evaluate alternate restoration plans and recommend the best (least-cost) plan under the circumstances.