Preference Inference involves inferring additional user preferences from elicited or observed preferences, based on assumptions regarding the form of the user's preference relation. In this paper we consider a situation in which alternatives have an associated vector of costs, each component corresponding to a different criterion, and are compared using a kind of lexicographic order, similar to the way alternatives are compared in a Hierarchical Constraint Logic Programming model. It is assumed that the user has some (unknown) importance ordering on criteria, and that to compare two alternatives, firstly, the combined cost of each alternative with respect to the most important criteria are compared; only if these combined costs are equal, are the next most important criteria considered. The preference inference problem then consists of determining whether a preference statement can be inferred from a set of input preferences. We show that this problem is co-NP-complete, even if one restricts the cardinality of the equal-importance sets to have at most two elements, and one only considers non-strict preferences. However, it is polynomial if it is assumed that the user's ordering of criteria is a total ordering; it is also polynomial if the sets of equally important criteria are all equivalence classes of a given fixed equivalence relation. We give an efficient polynomial algorithm for these cases, which also throws light on the structure of the inference.
A fundamental task for reasoning with preferences is the following: given inputpreference information from a user, and outcomes α and β, should we infer that the user will prefer α to β? For CP-nets and related comparative preference formalisms, inferring a preference of α over β using the standard definition of derived preference appears to be extremely hard, and has been proved to be PSPACE-complete in general for CP-nets. Such inference is also rather conservative, only making the assumption of transitivity. This paper defines a less conservative approach to inference which can be applied for very general forms of input. It is shown to be efficient for expressive comparative preference languages, allowing comparisons between arbitrary partial tuples (including complete assignments), and with the preferences being ceteris paribus or not.
A logic of conditional preferences is defined, with a language which allows the compact representation of certain kinds of conditional preference statements, a semantics and a proof theory. CP-nets can be expressed in this language, and the semantics and proof theory generalise those of CP-nets. Despite being substantially more expressive, the formalism maintains important properties of CP-nets; there are simple sufficient conditions for consistency, and, under these conditions, optimal outcomes can be efficiently generated. It is also then easy to find a total order on outcomes which extends the conditional preference order, and an approach to constrained optimisation can be used which generalises a natural approach for CP-nets. Some results regarding the expressive power of CP-nets are also given.
Preference logics and AI preference representation languages are both concerned with reasoning about preferences on combinatorial domains, yet so far these two streams of research have had little interaction. This paper contributes to the bridging of these areas. We start by constructing a "prototypical" preference logic, which combines features of existing preference logics, and then we show that many well-known preference languages, such as CP-nets and its extensions, are natural fragments of it. After establishing useful characterizations of dominance and consistency in our logic, we study the complexity of satisfiability in the general case as well as for meaningful fragments, and we study the expressive power as well as the relative succinctness of some of these fragments.
We address the problem of learning preference relations on multi-attribute (or combinatorial) domains. We do so by making a very simple hypothesis about the dependence structure between attributes that the preference relation enjoys, namely separability (no preferential dependencies between attributes). Given a set of examples consisting of comparisons between alternatives, we want to output a separable CP-net, consisting of local preferences on each of the attributes, that fits the examples. We consider three forms of compatibility between a CP-net and a set of examples, and for each of them we give useful characterizations as well as complexity results.