approach and may miss some of the pruning opportunities.For example, the candidate shown as an x with a square around it could be pruned with a more complex pruning strategy. However, our approach is relatively inexpensive and can be shown to require time proportional to . Choosing a more expensive pruning algorithm may decrease the overall number of distance calculations. This may, however, be at the expense of higher overall computation time due to an offsetting increase in cost of pruning.