作者： Jinwu Gao , Xueqin Feng
DOI:
关键词: Curse of dimensionality 、 State space 、 Function (mathematics) 、 Set (abstract data type) 、 Artificial neural network 、 Artificial intelligence 、 Computer science 、 Mathematical optimization 、 Space (commercial competition) 、 Algorithm 、 Fuzzy logic 、 Sample (statistics)
摘要: In this paper, a fuzzy inventory problem with multiple commodities is casted into dynamic pro- gramming model continuous state space and decision space. order to solve the programming model, genetic algorithms are used get samples of optimal cost functions, then neural networks trained approximate function on randomly generated sample set, which may bypass "the curse dimensionality". A hybrid intelligent algorithm thus produced functions that represented by networks. Lastly, numerical example given for illustrating purpose