Search applications in which queries are dependent on their context are becoming increasingly relevant in today's online applications. For example, the context may be the location of the user in location-aware search or the social network of the query initiator in social-aware search. Processing such queries efficiently is inherently difficult, and requires techniques that go beyond the existing, context-agnostic ones. A promising direction for efficient, online answering -- especially in the case of top-k queries -- is the one of materializing and exploiting results of previous queries (views).

We consider in this paper such an approach for query optimization based on cached results, focusing on two important sub-problems. First, the use of precomputed results in context-aware search requires adapted algorithms for top-$k$ query processing, in which only uncertain scores or score ranges are given, instead of exact ones. This is also the case when, for the sake of efficiency, classic top-k algorithms (e.g., NRA) are allowed to stop early - when the final result is reached -- without giving exact scores and ranks within this result. Second, in the presence of cached queries, we need adapted list selection techniques, which can leverage both the descriptions of queries and statistics over their results. We present novel algorithms that address these two problems, and exemplify their use in two important application scenarios: location-aware search and social-aware search. We validate our approaches via extensive experiments, using both synthetic and real-world datasets.