Note:

Descrizione sintetica del prodotto:
One broad source of difficulty in transitioning automated planning and scheduling theories and algorithms into practical application systems is the need to integrate with user decision-making processes. Both user acceptance of system decisions in successful problem solving episodes and effective user involvement in circumstances where the system reaches a problem solving impasse require that the system be comprehensible, and this requirement, in turn, implies that the system be capable of explaining its decisions in user-understandable terms. Unfortunately, most current work in explanation instead forces users to understand the system’s underlying search models. In this paper, we consider this problem of bridging the gap between user and system models in the context of a constraint-based scheduling system. We focus specifically on the problem of explaining temporal constraint conflicts that may arise in the course of either solving a given scheduling problem or integrating new state information with a previously computed schedule. The central idea is to use domain ontology knowledge (1) to isolate those constraints that are meaningful to and can be manipulated by the user, and (2) to identify various constraint relaxation options that the user might consider to resolve the conflict at hand.

Ulteriori informazioni sul prodotto: