The multi-parameter combination mind-evolutionary-based machine learning and its application
Keming Xie, Changhua Mou, Gang Xie · 2002
Based on mind-evolutionary-based machine learning (MEBML), four conceptions: degree of interaction, individual similartaxis, group similartaxis and dead band space are defined in this paper. Consequently, the multi-parameter combination mind-evolutionary-based machine learning (MPCMEBML) is established. MPCMEBML can solve efficiently the searching problem of the multi-dimension parameter space. In MDPMEML, the whole parameter space is separated and combined into several sub parameter spaces and the global billboard is endowed with the interactive function. Not every single group searches the whole parameter spaces. Information exchange will supplement the insufficient knowledge that not every group can get by itself. The new fuzzy logic controller is constructed by using MPCMEBML. Simulation results show that MPCMEBML is possessed of a strong ability of searching in the multi-dimension parameter space and the new FLC has good performance.