Abstract: According to the characteristics of computer distribution multi constraint conditions, in order to further improve the versatility and adaptability of computer automatic cotton, this paper put forward a kind of improved PSO (Particle Swarm Optimization) optimization method. Through establishment of the mathematical model of cotton blending, we transform it into the optimization problems with multiple constraints. On the basis of analysis of the standard PSO algorithm shortcomings, the inertia weight and learning strategy improvement factor are improved. Improved and the standard PSO algorithm solve the same cotton blending in the meantime with parameters collected from cotton spinning enterprises . The results showed that by using inertia weight and learning factor and adaptive strategy, optimizing speed, precision, the ability of local and global optimization and other indicators have been improved, reducing the cotton distribution costs of enterprises thus has a certain practical application value.

Key words: cotton assorting, improved PSO algorithm, inertia weight, learning factor