This is Penn State

Representing and Reasoning About Qualitative Preferences

(Funded in part by a grant from the National Science Foundation)

Many applications, e.g., composing web services, assembling complex pieces of legislation or legal arguments, designing embedded systems, etc. rely on methods for identifying collections of objects satisfying some criteria (functional specifications). Among the collections that satisfy the functional specifications, it is often necessary to identify one or more collections that are the most preferred with respect to user preferences over non-functional attributes, e.g., reliability, cost, etc. To address the needs of such applications, we have developed:

  • A first practical solution to the problem of determining whether one outcome dominates another with respect to a given set of qualitative preferences over the attributes of the outcomes (Santhanam et al., 2010).
  • A dominance relation that allows comparison of collections of objects in terms of preferences over attributes of the objects that make up the collection the set of most preferred collections, and algorithms for dominance testing that are guaranteed to return only (sound), all (complete), or at least one (weakly complete) of the most preferred collections (Santhanam et al., 2011).
  • Applications of preference reasoning algorithms in service composition and substitution based on qualitative preferences over non-functional attributes of the services (Santhanam et al., 2008, 2009), and in minimizing credential disclosure based on qualitative preferences over sensitivity attributes of the credentials (Oster et al., 2012).

Work in progress is focused on:

  • Extensions of the framework to handle reasoning with the preferences of multiple stake holders.
  • Applications of preference reasoning in software design, testing, and cyberdefense.
  • Applications of the resulting algorithms to problems in service composition, substitution, and adaptation.