The work of Cotton-blending is an important technical work to the cotton spinning enterprise, it has close relationship with the enterprise product quality, cost and economic efficiency and so on. The raw material accounts for 80% of the total cost in the cotton spinning enterprise, which directly affects the enterprise's economy efficiency, therefore Cotton-blending counts for much for the enterprise. The computer automatically Cotton-blending has the very big superiority to artificially Cotton-blending, it makes Cotton-blending get rid of the oversight which the pure dependence experience and the manual control produced, is advantageous to enterprise's better production management, thus enhanced the enterprise's management benefit and the product quality. Automatically Cotton-blending problem is a Multi-constrained combination optimization problem, its computation complex is quite high, from the computation theory angle, it belongs to the NP-hard question. Therefore research on the Cotton-blending problem has practical significance and theory value.This article first introduces the computer automatically Cotton-blending technology's background knowledge, and the paper's research background. Then gives the outline of the computer automatically Cotton-blending. The article introduces the mathematical model of automatically Cotton-blending, and theoretically analyses and compares the commonly used solution to the algorithm, had pointed out the feasibility to solve automatically Cotton-blending with scheme combination algorithm and the genetic algorithm.This article separately introduces automatically Cotton-blending algorithm: The scheme combination algorithm and the genetic algorithm, has elaborate two algorithms' principles, and demonstrates the algorithm flow, and the design to the algorithm. In the scheme combination algorithm, this article gives detailed elaboration on its difficult problem: how to enumerate all the schemes. This algorithm uses the dictionary sequence to enumerate the A level scheme, and designs one method to enumerate B level scheme, which effectively solves the problem. In the genetic algorithm the paper analyzes the basic genetic algorithm's insufficiency in solving the automatically Cotton-blending problem, through importing adaptive penalty function method, simulated annealing method , the method of buffering high fitness individuals and so on in two stages to form the hybrid genetic algorithm to make up the deficiency. Finally the paper compares the application result in automaticallyCotton-blending problem with scheme combination algorithm and hybrid genetic algorithm, and carries on the comparison of hybrid genetic algorithm MGA1 and MGA2, the experimental result demonstrated the hybrid genetic algorithm can effectively solve automatically Cotton-blending problem, the scheme combination algorithm's operation efficiency become extremely lowly along with the increase of combination scale, MGA1 and MGA2 all can highly effective carry on the operation, but they have differently in the solution multiplicity, and they suit to the different scale combination problem. |