Extension Theory and the Application in Optimization of Immune Neural Network

In the Immune Neural Network (INN), the key point is stability and convergence. The existing INN has shortages in the control of local optimization, so the paper bring forward INN algorithm which is based on extension theory. With the concepts of dependent function and matter-element, the improved algorithm firstly can optimize architecture and rule extraction of INN. And then, the new algorithm is applied in emulation experiment, which is used for computing of function. According to simulation results, the improved algorithm is compared with the existing INN. It indicates that extension theory has better advantage in optimization of INN. So the new algorithm has great reference value.