Biological and Shortest-Path Routing Procedures for Transportation Network Design

TL;DR

This paper compares biological routing and shortest-path methods for transportation network design, finding biological routing more efficient for uniform transfer distributions but less so for skewed distributions.

Contribution

It introduces a common framework for comparing network design methods and provides insights into their relative performance under different transfer distribution scenarios.

Findings

Biological routing explores more efficient solutions for uniform transfer distributions.

Difference in performance diminishes with skewed transfer distributions.

Framework enables systematic comparison of different routing approaches.

Abstract

The design of efficient transportation networks is an important challenge in many research areas. Among the most promising recent methods, biological routing mimic local rules found in nature. However comparisons with other methods are rare. In this paper we define a common framework to compare network design method. We use it to compare biological and a shortest-path routing approaches. We find that biological routing explore a more efficient set of solution when looking to design a network for uniformly distributed transfers. However, the difference between the two approaches is not as important for a skewed distribution of transfers.