Automatic Optimization of Printed Circuit Board Routing (PCB) Using Metaheuristic Algorithms.

Abstract

Automated printed circuit board (PCB) routing is an NP-hard combinatorial optimisation problem that consists of connecting all electrical nets while minimising track length and satisfying design rule constraints. This work addresses the net ordering problem in single-layer sequential routing, where the routing order directly influences the quality of the final solution. Two bio-inspired metaheuristics, the Whale Optimisation Algorithm (WOA) and the Slime Mould Algorithm (SMA), are adapted to the discrete domain using the Smallest Position Value (SPV) encoding in order to optimise this ordering. Candidate solutions are evaluated using an A* algorithm that explores eight movement directions and incorporates mechanisms ensuring valid, crossing-free routings that respect all design rules. Experiments conducted on both a synthetic and a real printed circuit board show that both approaches produce complete or near-complete routings. SMA achieves the best average performance, while WOA also delivers highly competitive results.