Key Research Highlights > Databases & Big Data Analytics > Answering MaxBRkNN Queries
Wynne HSU Mong Li LEE |
Answering MaxBRkNN Queries
Objective To provide an improved solution to the MaxBRNN problem which involves finding a region such that setting up a new service site within this region would guarantee the maximum number of customers by proximity. Results We have proposed a solution to the generalised MaxBRkNN problem known as MaxFirst. The algorithm works by partitioning the space into quadrants and searches only in those quadrants that potentially contain an optimal region. During the partitioning process, we compute the upper and lower bounds for each quadrant that potentially contain an optimal region; these bounds are used to prune unpromising quadrants. Experiments shown MaxFirst is two to three orders of magnitude faster than the state-of-the-art algorithm. |