Multi-strategy combined learning mechanism suitable for designing expert system

The complicacy of designing products induces the complicacy of designing expert system. The machine learning of this kind of system is difficult to be realized. In this paper, a kind of multi-strategy combined learning mechanism is presented, which combines the learning by analogy based on pattern matching and the self-adaptive learning based on evolution strategy as well as deductive learning based on fuzzy logic together, and overcomes the shortage of single machine learning and can be used for reference by other designing expert systems.