A Model Set Design Method Based on Idiotypic Network

Model set design method has been one of the focuses of the research on multi-model control. The immune system is a excellent biological multi-model control system. In the immune system, the antibody is the model of antigen. Many kinds of antibodies constitute the idiotypic network. This network can be regarded as the model set of antigens. Among these antibodies there exists interaction. Through this interaction the structure of idiotypic network can be adjusted to maintain the diversity of antibodies. Self-adjustment of the idiotypic network can be regarded as the optimization of the model set. Inspired by this thought, the idiotypic network was used to represent the model set of complex uncertain dynamic process. Corresponding training algorithm was given. Simulation results showed this new model set design method is effective.