Dependency Network Based Planning, Understanding the True Effects of Plans and Actions

This paper describes an approach to intelligent planning which employs a dependency network model to reason with direct, n-order, cumulative and cascading effects. The dependency network is a cyclic graph of nodes (persons, organizations, locations, resources, concepts, etc.) connected by weighted arcs that define the type and strength of dependency between a node pair. The network can be employed as part of Mixed Initiative Planning (MIP) architecture or embedded as a constraint manager inside a fully automated planning architecture. The network can be employed to firstly evaluate both the intended, cumulative and cascading effects of actions and secondly to critique plans to identify their over dependence on a node or set of nodes. For example, repairing the electrical substation will restore power to the hospital and the traffic light sensors in the area. The choice of repair teams 1 and 2 to conduct the repairs means they are both dependent on two key roads to travel to the repair site. The paper provides details of both planning architectures, an overview of the dependency model and examples of its use on several scenarios.