A Proposal of Road Network Hierarchization Methods Using Betweenness Centrality for Route Finding Problems

Abstract

In vehicle routing problems, it is required to compute the necessary customer-to-customer routes within a settime frame, and to perform route searches that take into account the preferences of individual drivers at the time ofroute search. In this paper, we propose a road network hierarchization method aimed at reducing the total computationtime for route search and performing route searches that consider the preferences of drivers. The proposed methodconstructs a hierarchical network based on betweenness centrality, one of the centrality measures in network analysis.To verify the effectiveness of the proposed method, computer experiments were conducted on the road network of a14 km square area in the central part of Sapporo city, with multiple drivers’ preferences for narrow roads prepared.The total computation time for route search, the cost of routes, and the average route length were evaluated. As aresult, by using the proposed method, the total computation time for route search was reduced to 4 ∼ 7% of the exactsolution, and the cost of routes worsened by about 5 ∼ 10%. Additionally, regarding the average route length, whereasthe Euclidean distance, frequently used as a measure for the length of routes between a delivery base and customers invehicle problems, tends to result in an error of about 30 ∼ 40% when compared to the exact solution, the applicationof the proposed method has been able to contain this error to less than 10%, thereby confirming the effectiveness ofthe proposed method.