A route leads from a start location to a final destination and passes through geospatial entities that are picked according to search terms provided by the user. Each entity is pertinent to exactly on term and its degree of relevance is given in terms of a probability. In the route-search problem, the goal is to minimize the traveling distance while maximizing the success, namely, the probability that for each query term, the route visits a relevant entity. The output may include several entities for each term, in order to maximize the probability. However, when actually traveling the route, the user may find out that for some of the query terms, there are no relevant entities. This paper investigates interactive route search. Upon arrival at each entity, the user provides feedback specifying whether the entity is indeed relevant to the search. The goal is to compute the next entity to be visited, based on the feedback thus far. The challenge is to utilize the feedback in order to com...