构建在联接探索和分解分布上分布估计算法的扩展算法

遗传算法（GA）在解决变量间存在较大相互作用优化问题时缺乏有效性，一种解决问题的途径是分布估计算法（EDA）。分解分布算法是一种近似高阶相互作用的EDA，它用分解Boltzmann分布来产生新的解。运用联接探测及分解分布给出一个以高概率找到最优解的新算法。该算法能解决一些分布估计算法难于处理的问题。实验证明了算法的可行性和有效性。

Algorithm extended to an estimation of distribution algorithm based on linkage detection and factorization

Genetic Algorithm (GA) has been found to be lack of effectiveness in solving optimization problems where there is a large amount of interaction between variables, one approach to solve this problem is Estimation of Distribution Algorithms (EDA). Factorized distribution algorithm is an EDA that uses approximation of higher-order interaction, and it uses a factorization of the Boltzmann distribution for the generation of new solutions. A new algorithm which finds the optimum with high probability based on the linkage detection and factorization was given. The algorithm can solve the problems which EDA may have difficulties to deal with. Experimental results prove that the new algorithm is feasible and effective.