Is Euclidean Distance Really that Bad with Road Networks?

Spatial queries play an important role in many transportation services. Existing solutions to spatial queries commonly rely on the measurement of road network distance, which is the length of the shortest path from one point to another in a road network. Due to the high computation cost of measuring road network distance, a service provider may not be able to handle all the queries in a timely manner. We are interested in Euclidean distance-based solutions to spatial queries as Euclidean distance is significantly cheaper to compute than road network distance. A common view is that Euclidean distance is not suitable for solving any spatial query in road networks as road network distance can be significantly different to Euclidean distance. We challenge this view by evaluating the performance of an Euclidean distance-based approach in solving Group Nearest Neighbor queries, which can be used in transportation applications. Our study shows that Euclidean distance can help to achieve an excellent level of accuracy in solving the associated query type. In return, this opens the door for other studies to check whether similar results could be found in other query types and that per query type a judgement should be made.