非线性混合整数规划的简化二阶震荡粒子群混合算法  

Simple hybrid particle swarm optimization based on second order oscillation of the nonlinear mixed integer programming

作　　者：赵乃刚[1] 赵佳鑫[1] 

机构地区：[1]山西大同大学数学与计算机科学学院,山西大同037009

出　　处：《兰州理工大学学报》2017年第3期102-105,共4页Journal of Lanzhou University of Technology

基　　金：国家自然科学基金(61272095);国家自然科学基金青年基金(41401521)

摘　　要：为了求解整数约束和混合整数约束优化问题,提出了简化二阶震荡粒子群优化算法.在二阶震荡粒子群算法的基础上,对更新过程进行简化,使得迭代方程由原来的二阶降成一阶,粒子的搜索过程更为简单高效,便于搜索和寻优;引入了平均个体最优位置,使得所有粒子的有效信息被充分利用;对不满足约束条件的粒子重新生成,从而加快算法的收敛速度;为了防止算法的早熟收敛现象,提出了"优胜劣汰"的更新机制.最后,为了验证算法求解整数和混合整数优化问题的可行性和有效性,将简化二阶震荡粒子群混合算法对16个测试函数进行了测试并与其他三种算法比较.实验结果表明,本文算法在精确度和成功率方面有明显的提高.In order to solve the optimization problem with integer and mixed integer constraint,a simple particle swarm optimization based on second order oscillation is proposed,where the original second-order iterative equations are reduced to first-order.So that the updating process is simplified and the particle search process is even more simple and efficient,and it is easy to search and optimize the particle swarm.The introduction of average individual optimal position makes the effective information of all the particles fully utilized and the particles,which are unsatisfied with constraints,regenerated,so that the convergence is accelerated.In order to prevent premature convergence,the 'survival of the fittest'updating mechanism is introduced.Finally,in orderto validate the Simple hybrid particle swarm algorithm's feasibility and effectiveness for solving nonlinear mixed integer programming,16 test functions are tested and compared with the other three algorithms.Experimental result shows that the algorithm presented in this article is significantly improved in terms of accuracy and success rate.

关 键 词：粒子群算法 非线性混合整数规划 平均个体最优位置 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]