Abstract: Aiming at the problems in process of orienting edges of Bayesian networks, a novel edge-oriented approach of Bayesian networks based on tabu Genetic Algorithm(GA) is proposed, which introduces the diversification of Tabu Search(TS) into the cross arithmetic operators and the variation arithmetic operators, and is applied into orienting edges of flight departure delay skeleton. Simulation experimental results show this approach is effective and feasible.

Key words: Bayesian networks, Genetic Algorithm(GA), structural learning

摘要： 针对贝叶斯网边定向过程中存在的问题，提出一种基于遗传禁忌算法的贝叶斯网边定向方法，该方法将禁忌搜索的“多样化”引入遗传算法的交叉算子和变异算子中，生成禁忌交叉算子和禁忌变异算子，并对航班离港延误骨架模型进行定向，仿真实验结果表明，该方法是有效可行的。

关键词: 贝叶斯网, 遗传算法, 结构学习

CLC Number: 

• TP18