中文摘要：

提出了一种新型进化算法即贝叶斯预测型进化算法,该算法是有效解决遗传算法中的连锁和欺骗问题的一种新方法,其主要特点是：（1）该算法基于最优解的概率分布和贝叶斯定理预测最优解所在的子空间;（2）该算法能高效利用所有先前代蕴含的信息,可以方便地引入专家知识;（3）该算法模型比较简单并且能以很快的速率收敛到最优解子空间.从理论上分析了贝叶斯预测型进化算法的收敛性、收敛速率和逆收敛算子.理论分析与在14个标准的测试函数上的仿真实验显示了该算法求解较为精确、稳定和快速.

英文摘要：

Bayesian Forecasting Evolutionary Algorithm （denoted by BFEA）,integrating with the basic principle of evolutionary computation,is proposed in this paper,which is a new technique to solve linkage problem and deceptive problem effectively.The main works and innovative points are as follows.：（1） BFEA,based on the probability distribution of promising solutions and Bayesian theorem,guides the exploration of the search space according to the prediction probability of every subspace including the optimal solutions; （2） Much more information in the generated populations is used and prior information is incorporated into the algorithm easily; （3）This algorithm has a simpler algorithm model and can converge faster to the subspaces with the optimal solutions.The convergence,convergence rate and counteraction operator of BFEA are analyzed theoretically.Both theoretical analyses and extensive experiments on 14 well known benchmark functions show that BFEA has lower computing complexity,better speed of convergence,smoother results and higher computation accuracy than several developed methods in recent years.