Based on metal rubber roughcast weaving process

the dowels and paths of complicate shape metal rubber roughcast weaving were designed with developed genetic algorithm. Code

crossover and mutation were improved for dowels and paths planning problem. An evaluation factor was presented to appraise iterative procedure of genetic algorithm. Complicate shape metal rubber roughcast can be knitted well with optimized dowels distribution and laying paths. Roughcast can be woven better evidently with optimized paths than traditional experiential paths. When roughcast was knitted to ten layers

the objective function b of optimized weaving was 54% smaller than experience weaving. Knitting quality was little influenced by paths scale in each layer. With the growing laying length

contact states of roughcasts woven with different paths scale in each layer reached accordance

providing a theoretical references

for complicate shape metal rubber roughcast weaving.