Preference and Similarity-based Behavioral Discovery of Services

We extend Constraint Automata by replacing boolean constraints with semiring-based soft constraints. The obtained general formal tool can be used to represent preference-based and similarity-based queries, which allow a user more freedom in choosing the behavior of the service to finally use, among all possible choices. A user states his preferences through a “soft” query, and obtains results that satisfy this query with different levels of preference. The soft requirements may involve a parameter data of the service operations, or the (names of the) operations themselves. Moreover, we introduce a first implementation of the search procedure by using declarative (Soft) Constraint Programming.