Artificial Intelligence - Special issue: Qualitative reasoning about physical systems II
Consistency techniques for numeric CSPs
IJCAI'93 Proceedings of the 13th international joint conference on Artifical intelligence - Volume 1
Using Qualitative Reasoning for a Recommender System
Proceedings of the 2010 conference on Artificial Intelligence Research and Development: Proceedings of the 13th International Conference of the Catalan Association for Artificial Intelligence
Refining reasoning in qualitative probabilistic networks
UAI'95 Proceedings of the Eleventh conference on Uncertainty in artificial intelligence
Improving the computational efficiency in symmetrical numeric constraint satisfaction problems
CAEPIA'05 Proceedings of the 11th Spanish association conference on Current Topics in Artificial Intelligence
Order of magnitude qualitative reasoning with bidirectional negligibility
CAEPIA'05 Proceedings of the 11th Spanish association conference on Current Topics in Artificial Intelligence
Applied Ontology
Applied Ontology
Using L-fuzzy sets to introduce information theory into qualitative reasoning
Fuzzy Sets and Systems
Hi-index | 0.00 |
In [Dague, 1993], a formal system ROM(K) involving four relations has been defmed to reason with relative orders of magnitude. In this paper, problems of introducing guantitative informahon and of ensuring validIty of the results in IR are tackled. Correspondent overlapping relations are defmed in R and all rules of ROM(K) are transposed to R. Unlike other proposed systems, the obtained system ROM(R) ensures a sound calculus in R, while keeping the ability to provide commonsense explanations of the results. If needed, these results can be refmed by using additional and complementary techniques: k-bound-consistency, which generalizes interval propagation; symbolic computation, which considerably improves the results by delaying numeric evaluation; symbolic algebra calculus of the roots of partial derivatives, which allows the exact extrema to be obtained; transformation of rational functions, when possible, so that each variable occurs only once, which allows interval propagation to give the exact results. ROM(R), possibly supplemented by these various techniques, constitutes a rich, powerful and flexible tool for performing mixed qualitative and numeric reasoning, essential for engineering tasks.