Agent-Based Decision Support for Managing Print Tasks
Jean-Marc Andreoli, Stefania Castellani, Uwe Borghoff, Remo Pareschi, Gunnar Teege
We consider here a typical instance of a non-trivial print tasks management problem, and we show how to
handle it within Zippin, a highly interactive constraint-based environment for scheduling and decision support.
The aim of Zippin is to provide the user with synthetic views of an ongoing scheduling process from multiple
perspectives, allowing the user to navigate through the space of possible schedules and to make informed
decisions. The navigation space is defined by a system of constraints accessed by the user through the views
and by a set of software agents, which manipulate the constraints store. Constraints identify goals that agents
have to fulfill as well as effects that their actions propagate into the environment. Thus, the agent model
implements a type of behavior which is teleological (agents have intentions that guide their actions) and
stigmergic (agents communicate through the effects that their actions produce on the environment). We also
describe how our model may be extended with agents which implement heuristic actions handling specific
In Proc. of PAAM'98, March, 1998, London, U.K.
dist.pdf.gz (149.90 kB)
dist.ps.gz (169.63 kB)