Cigarette recipe maintenance method based on artificial immunity method

The invention provides a method for maintaining the cigarette blending formula based on the artificial immune algorithm. An immune neural network, an immune network, the immune algorithm and an expert system method are combined together and applied to the maintenance of the standard cigarette blending formula, and an intelligent system for the formula maintenance design is established. The methodcomprises the following steps: firstly, grouping sample data by styles based on professional expertise; then, dividing unblended cigarettes into different varieties by indices based on the idiotype immune network clustering, wherein, the tobacco leaves with the same variety are substitutable; selecting the tobacco leaves by combining the professional expertise to set up a new cigarette blending scheme, coding the cigarette blending scheme based on the immune optimal algorithm, carrying out optimization search, and using the neural network prediction model of the finished cigarettes to evaluate the cigarette blending f