A conservative constrained clustering-merging algorithm for particle-in-cell codes

• Published in: Computer physics communications Vol. 313; p. 109621
• Main Authors: Cai, Dong-sheng, Wang, Ping-yang
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.08.2025
• Subjects: Cluster size; Constraint k-means; Particle merging; Particle-in-cell; Cluster size; Particle merging; Constraint k-means; Particle-in-cell
• ISSN: 0010-4655
• DOI: 10.1016/j.cpc.2025.109621

The particle merging algorithm enables particle-in-cell codes to simulate the process of rapidly increasing particle numbers. Dividing particles that are close in phase space into a subset for merging is beneficial for preserving the particle distribution function (PDF). However, larger subsets can cause particles with significant differences to be grouped together. To address this issue, we proposed a conservative constrained clustering-merging algorithm which employs the constrained k-means method to keep the number of particles within each subset at a low level while meeting the requirement of conserving physical quantities. Subsequently, the particles in each subset are merged by probabilistically adjusting their weights. The impact of subset size on the merging results and computational performance is also discussed.