An Improved Particle Swarm Optimization Algorithm Applied to the Unified Evaluation of Circularity Error

Abstract

Minimum zone circle (MZC), minimum circumscribed circle (MCC), maximum inscribed circle (MIC) and least square circle (LSC) are four common methods used to evaluate circularity errors. A novel particle swarm optimization algorithm based on self-adapted comprehensive learning (ACL-PSO) is proposed to evaluate circularity errors with real coded strategy. In the algorithm, population learning mechanism and velocity mutation strategy are adopted. In the meantime, ACL-PSO is applied to the unified evaluation of circularity error. The experiment results evaluated by different methods indicate that the proposed algorithm not only converges to the global optimum rapidly, but also has good stability, and it is easy to generalize.