A Customized Particle Swarm Method to Solve Highway Alignment Optimization Problem

Yousef Shafahi, Mehdi Bagherian

Computer-Aided Civil and Infrastructure Engineering · 2012 · 149 citations · 25 references

Concepts

TL;DR

Highway alignment optimization is a complex, constrained problem requiring versatile cost functions and efficient search methods to identify the best design among infinitely many continuous solutions. The study applies a customized particle swarm optimization algorithm to find a near‑optimal highway alignment composed of circular and parabolic curves. The algorithm optimizes total cost under design constraints using GIS‑based calculations of right‑of‑way, earthwork, and other spatial costs, and its efficiency was validated in a case study on an artificial map. In a real‑world example, the algorithm produced alignments that matched or outperformed those found by traditional methods.

Abstract

Abstract: Optimizing highway alignment requires a versatile set of cost functions and an efficient search method to achieve the best design. Because of numerous highway design considerations, this issue is classified as a constrained problem. Moreover, because of the infinite number of possible solutions for the problem and the continuous search space, highway alignment optimization is a complex problem. In this study, a customized particle swarm optimization algorithm was used to search for a near‐optimal highway alignment, which is a compound of several tangents, consisting of circular (for horizontal design) and parabolic (for vertical alignment) curves. The selected highway alignment should meet the constraints of highway design while minimizing total cost as the objective function. The model uses geographical information system (GIS) maps as an efficient and fast way to calculate right‐of‐way costs, earthwork costs, and any other spatial information and constraints that should be implemented in the design process. The efficiency of the algorithm was verified through a case study using an artificial map as the study region. Finally, we applied the algorithm to a real‐world example and the results were compared with the alignment found by traditional methods.

References

25