Abstract: Particle Swarm Optimization（PSO） and Differential Evolution（DE） are two latest optimization techniques.These algorithms have been very successful in solving the global continuous optimization，but their applications to combinatorial optimization have been rather limited and are not as effective as in global continuous optimization.Recently，a Differential Barebones PSO（DBPSO） is also proposed for global continuous optimization.Firstly，a discrete DBPSO is proposed for combinatorial optimization，and then the Estimation of Distribution Algorithm（EDA） is incorporated into the discrete DBPSO to improve its performance.The proposed discrete DBPSO algorithm based on EDA combines global statistical information extracted by EDA with local evolution information obtained by discrete DBPSO to create promising solutions.The results of experiment show that the EDA can significantly improve the performance of the discrete DBPSO.

Key words: discrete differential barebones Particle Swarm Optimization（PSO）, estimation of distribution, unconstrained binary quadratic programming problem, combinatorial optimization

摘要： 粒子群优化（PSO）和差分演化（DE）是两种新兴的优化技术，已经成功地应用于连续优化问题，但是它们至今尚不能像解决连续优化问题那样有效地处理组合优化问题。最近，有人提出差分骨干PSO（DBPSO）用于解决连续优化问题。首先提出离散DBPSO用于组合优化问题，然后在离散DBPSO中引入分布估计算法（EDA）来提高性能，把EDA抽样得到的全局统计信息和DBPSO获得的局部演化信息相结合来产生新解，形成基于EDA的离散DBPSO。实验结果表明EDA能大大提高离散DBPSO的性能。

关键词: 离散差分骨干粒子群优化, 分布估计, 无约束二进制二次规划问题, 组合优化