An Adaptive Ant Colony optimization in Knowledge Graphs

Abstract

Knowledge graphs have been widely used in various fields such as question answering systems and recommendation systems. However, there are few researchers on combinatorial optimization problems based on knowledge graphs, which greatly delays the development of knowledge graphs. Also, when solving combinatorial optimization problems only by using knowledge graphs, it is impossible to obtain better results. In order to solve these problems, an ant colony optimization algorithm based on an adaptive strategy (AACO) is proposed, and the algorithm is applied to solve the path optimization model established by the knowledge graph. In the vector space based on knowledge graph embedding, the ant colony optimization algorithm has a good positive feedback mechanism and robustness to find effective paths between entity nodes. Experimental results show that this proposed AACO algorithm can accelerate the convergence speed and obtain better accuracy. At the same time, a global optimal solution can be achieved, which is suitable for solving combinatorial optimization problems.