The ranking cube was proposed for efficient computation of top-k (ranking) queries in relational databases. Recently, researchers have proposed another kind of query, called a skyline query. A skyline query returns all the objects pi such that pi is not dominated by any other object pj , where dominance is defined as follows. Let the value of pi on dimension d be v(pi ,d). We say pi is dominated by pj if and only if for each preference dimension d, v(pj,d) ≤ v(pi ,d), and there is at least one d where the equality does not hold.

(a) Design a ranking cube so that skyline queries can be processed efficiently.

(b) Skyline queries are sometimes too strict to be desirable to some users. One may generalize the concept of skyline into generalized skyline as follows: Given a d dimensional database and a query q, the generalized skyline is the set of the following objects: (1) the skyline objects and (2) the non skyline objects that are -neighbors of a skyline object, where r is an -neighbor of an object p if the distance between p and r is no more than. Design a ranking cube to process generalized skyline queries efficiently.