Advisor: A knowledge-based planning system

This work investigates the ability of a planning system called Advisor to satisfy the requirements posed by complex application domains. This paper concentrates on a complex domain that concerns the planning of cargo handling operations for chemical carriers, and identifies the key issues of this domain in correspondence to the requirements of complex domains. Advanced expressiveness, representation and handling of temporal information in a symbolic and quantitative way, user interaction and utilization of appropriate knowledge structures for increasing efficiency, are among the issues under consideration. Advisor has been constructed in a three level modular architecture, enabling users to interact directly with each level. The main module is the TRL-Planner which is supported by the Symbolic Constraint Solver (SCS) and the Temporal Reference Language interpreter (TRLi).