The topic of preferences is gaining more and more attention in diverse areas of AI such as nonmonotonic reasoning, qualitative decision theory, soft constraints, configuration, and AI planning.
As described in the article of Jon Doyle and Richmond Thomason about qualitative decision theory (AI Magazine, 1999), AI provides qualitative methods for treating preferences that can improve or complement numerical methods for treating preferences from classical decision theory. (Qualitative) preferences have also been essential in treating conflicting information in nonmonotonic reasoning, inheritance of defaults, temporal reasoning, diagnosis, and other areas in knowledge representation and reasoning. More recently, preferences have also been used in constraint satisfaction and constraint programming for treating soft constraint, for describing search heuristics, and for reducing search effort.
It appears that preference is a concept that complements the concept of a constraint and that provides an AI counterpart to the notion of an objective used in operations research. Preferences allow to treat conflicting information and to choose the interesting alternatives such as
In contrast to hard constraints, preferences are not eliminating the non-selected alternatives, which may become interesting if additional information is added.
AI permits complex preference representations (e.g. based on logic or constraints) and allows to reason with and about preferences. Thus, it gives a perspective for formalizing information that was never adequately formalized before, but which is essential for all domains of our social life ranging from daily decision making (e.g. which products to buy? which articles to read?) up to legal reasoning (which law takes priority over which other) and scientific debates (which hypotheses best explains the given phenomena). Hence, preferences are of crucial importance in the design and development of intelligent systems such as such as web-based configurator, configuration of alerting and filtering systems, temporal reasoning and scheduling systems, and robot planners and behavior systems.