Research And Implementation Of Lightweight Semantic Reasoning Engine Based On Genetic Algorithm Optimization

Semantic reasoning is the key technology to realize knowledge discovery and is also the core link in the Semantic Web.To knowledge the data in the Internet of Things(Io T)in order to realize the efficient sharing of data,the semantic reasoning engine plays a vital role.However,the current mainstream semantic inference engines occupy too much computing resources and storage resources of computing devices in the Io T,so mainstream inference engines are not suitable for those edge computing devices of the Io T with less computing and storage resources.At the same time,the current reasoning engine cannot reason about the uncertain knowledge in the Io T.Many scholars mainly study the expression of uncertain knowledge.It is a static knowledge representation method and is not suitable for the Io T environment with dynamic data changes.Combining the above two problems and the rule-based reasoning algorithm-RETE,this thesis proposes a lightweight semantic reasoning engine optimized based on genetic algorithm.This reasoning engine has the advantage of occupying fewer computing resources and storage resources and implementation the dynamic reasoning of uncertain knowledge.Aiming at the sequence optimization problem of the alpha network sequence in the RETE reasoning algorithm,genetic algorithm(GA)is used to optimize the Alpha network of the RETE algorithm;for the problem of excessive resource occupation of the traditional reasoning engine,a lightweight reasoning engine implementation is designed;in view of the problem that the current reasoning engine cannot reason about uncertain knowledge online,a quaternion knowledge expression method with probability factors is proposed to represent the uncertain knowledge,and the construction of the ontology and the labeling of the data are extended at the same time,the traditional RETE network is improved to support reasoning based on quaternion knowledge;for the calculation of uncertain knowledge,combined with Bayesian model and full probability model,probabilistic reasoning rules based on quaternion are formulated.Experimental results show that genetic algorithm has a certain optimization effect on the alpha network in the RETE algorithm.The reasoning time of the lightweight semantic reasoning engine in this study is less than that of the traditional reasoning engine based on the RETE algorithm;it consumes resources in memory and non-volatile storage.On the one hand,the lightweight reasoning engine in this research occupies fewer computing resources and storage resources than the current mainstream reasoning engines;at the same time,the reasoning engine in this research realizes the function of online reasoning for uncertain knowledge in the Internet of Things environment.