The Portuguese Association for Artificial Intelligence hasbeen organizing Portuguese Conferences on ArtificialIntelligence, now held every second year, since 1985.This volume contains selected papers from the FifthConference on Artificial Intelligence. The conference has aninternational status: 62 contributions from 13 countrieswere received, of which 26 were from Portugal. To guaranteea high scientific standard, all the contributions werereviewed by at least three researchers,and only 20 paperswere accepted and included in these proceedings. The papersare organized into sections on constraints, searchknowledge representation, temporal reasoning, planningdiagnosis and repair, and learning.

This volume contains selected papers from the Fifth Portuguese Conference on Artificial Intelligence. Topics include constraints, search, knowledge representation, temporal reasoning, planning, diagnosis and repair, and learning.

Solving linear constraints on finite domains through parsing.- Constraint solving in finite domains under user control.- A new method for solving linear constraints on the natural numbers.- A constraint-based fuzzy inference system.- A constraint-based language for querying taxonomic systems.- Heuristic parsing and search space pruning.- Wave-shaping in multiprocessor bidirectional heuristic state space search.- The extended stable models of contradiction removal semantics.- Modeling a rational cognitive agent in SNePS.- Semantics of property inheritance in a hierarchic system with explicit negation.- Time in confluences: Dealing with delays for consistency-checking.- A temporal representation for imperatively structured plans of actions.- Maximal intervals: An approach to temporal reasoning.- Consistency driven planning.- An efficient approach to planning in assembly tasks.- Towards a theory of the repair process.- Declarative source debugging.- A neural approach to data compression and classification.- Generalization for a propositional calculus: a constraints-based approach.- Adaptive learning using a qualitative feedback loop.