Enhanced DNA sequence design with learning PSO

DNA computing is a novel computing technology with high storage density and parallelism and promising applications. The reliability of DNA computing relies on data encoded DNA sequences, but designing qualified DNA sequence is a very complex task. To accomplish this task more effectively, a learning particle swarm optimization (LPSO) is proposed. LPSO uses improved refraction opposition-based learning (ROBL) to enhance the local exploitation, improved salation learning (SL) to consolidate the global exploration,Then, Gaussian Mutation (GM) is introduced to improve the population diversity and bring the population closer to the global optimal solution. The feasibility of LPSO is demonstrated by results of DNA encoding sequence optimization experiments, Then the help of three strategies to improve PSO algorithm is discussed through ablation study.