Stepwise Optimisation Method for k-CNN Search
Jun Hong Feng, Naoto Mukai, Toyohide Watanabe · 2004
The problem of k-CNN search along a specific route on road network is to find out k nearest neighbor (k-NN) objects for any place on the route. k nearest neighbors are selected based on the path length from the route to the objects, and the continuous search for all the points on the route should be considered. A k-CNN search method is proposed by using an incremental k-NN search based on road network. The method is an extension of CNN search method proposed by adding new data structur e-afi xedlength queue for recording up-to-now intermediate results. Because the search regions can be reduced stepwise by the intermediate results, our k-CNN search is efficient.