自适应交叉与组合变异的多任务GP进行本体匹配OA

Ontology matching is an effective means to solve the problem of ontology heterogeneity.In order to improve the quality of ontology matching and suppress the bloating in genetic programming(GP),a multi-task genetic programming algorithm with adaptive crossover and combinatorial mutation is proposed to achieve the interaction of the knowledge between the two task populations.The small-sized tree is introduced to suppress the bloating,and the additional task populations are used to guide the target task population to jump out of the local optimum.In this algorithm,a new adaptive inter-task crossover operator is used to select different crossover strategies according to the performance of individuals and their parents,allowing the algorithm to fully explore the search space.A variation operator based on combinatorial probability is proposed to guide the target task population to realize better quality variation,and a new fitness function is designed to suppress the tree size and optimize the matching performance while reducing the tree size.The experiments were conducted on the Benchmark test set of ontology alignment evaluation initiative(OAEI).The results show that the proposed method can realize excellent matching performance on all the test sets,with better results compared to other cutting-edge methods.