The Research On Artificial Immune Algorithm To Nonlinear Combination Forecasting Model

This essay is advanced basing on the study which I worked for the study of Combination Forecasting through a soft scientific task "the intellectual statistic analysis study on enterprise communications under the electronic business environment"made by ChongQing technological committee and task "the intellectual analysis systematizing study on ChongQing's imports and exports information"made by CQ foreign trade economic committee. Combination Forecast is one of the important part of the theories of forecast. But the traditional Linear Combination Forecasting can't help when the result of each individual forecasting model are all more or less than its real value. However, the Nonlinear Combination Forecasting method, which is in common use, needs that the objective function contains stronger analytics or structures the brand-new Nonlinear Functions. it reduces the solving availability of such a problem. The purpose of the essay is to work out a new feasible way to solve this very problem of Nonlinear Combination Forecasting. The research of AIA began from the late 80's in 20th century. It gained from imitating the biological immune system. Nevertheless, the complication of the immune system force the study of AIA hasn't attached enough importance to, consequently the study outcomes and applications over internal and external are relatively less. It can be said that, nowadays, the study of AIA combines two main features, the priori knowledge and the adaptive ability of immune system, endowing it with powerful and robust ability, which can deal with information, it is considered to be a very significant and meaningful research direction. Comparing with each AIA on the base of systematizing study on the existing AIA, it develops the memory cells'function into an important component part of algorithm through the progressive research on the biological immune systematizing mechanism,and takes fully use of the Baldwin Effect of the biological immune system, works out the improving multi-peak value AIA. At the mean time, the essay amends the arithmetic operator in the original algorithm, and applies it in the new algorithm, improves the global search ability and convergence rates in a great deal. Through this emulative case, it proves it's effectiveness and well robustness. Finally, putting multi-peak value AIA into practical to solve the Nonlinear Combination Forecasting problem, the good result demonstrates that the algorithm on solving the Nonlinear Combination Forecasting problem is a kind of feasible new method.