Consumer price index (CPI) was a main scientific ground for manager to put forward price policy, wage policy, and national economy development strategy. A lot of study and prediction were carried out for CPI by many scientists, and corresponding achievement was obtained. However, the simple network was applied to the past perdition methods, and the prediction results was not good. The prediction method based on the combinatorial and optimal networks of BP network, Elman network, RBF network and GRNN network was established and applied to the prediction of consumer price index. The simulation results showed that the effect of the combinatorial and optimal method was better than that of single method.

In order to make a good decision in the projection investment, and the application of the projection pursuit model of particle swarm algorithm on the investment decision was studied in depth. Firstly, the development of the projection pursuit model of particle swarm algorithm was introduced, and then the brief introduction of PP and the step of constructing the PPC model were introduced. The basic theory and calculating procession of particle swarm algorithm was analyzed. And then the projection investment decision model was established. Finally a case study was carried out and the results showed that this method was simple and effective.