Hybrid strategy for reducing transport cost in spatially embedded networks

The dynamics of transportation in networks is a universal phenomenon in many real complex systems. Previous studies on transportation dynamics in spatially embedded networks have mainly focused on the average transport time from source to target. However, previous studies have rarely considered the transport cost from source to target in spatially embedded networks, a factor that is essential if geographical distance is to be accounted for. In this paper, we aim to minimize transport cost, a problem essentially different from that of minimizing transport time. We propose a novel hybrid algorithm that integrates information of both geographical distance and the direction of movement. The proposed hybrid algorithm can reduce the chance of zigzag motion in the network, and has the potential to yield lower transport cost than Kleinberg’s algorithm. Ultimately, we find that average transport cost scales linearly with system size for both Kleinberg’s algorithm and our hybrid algorithm, but grows more slowly for our algorithm.