For some complex specified domain retrieval requirements

many researches use semantic related technologies to resolve such problems.In terms of rule inference

it always leverages the inference algorithm of Semantic Web directly.However

the efficiency is not good.On the respects of searching results ordering

the ordering algorithm for search engine

which is based on keywords

is not suitable for ranking on search results generated by semantic retrieving.Focusing on the above issues

this paper proposes a semantic rule modeling method

and gives a new rule reasoning algorithm based on closed world assumption backwards reasoning chain to get higher inference efficiency compared to most semantic inference engines.Moreover

this paper proposes a new ordering algorithm based on feature similarity.Taking advantage of the above methods this paper describes

the search engine-Material Hub has been built up.Experiments show this semantic search engine improves the searching precision rate

recall rate

and rational responding time.So far

this system has been applied in industry.