基于动态参数蚁群算法的云制造服务组合-《计算机技术与发展》

文章信息/Info

Title:: Cloud Manufacturing Service Composition Based on DynamicParameters Ant Colony Algorithm

Author(s):: ZHANG Yan-kai; ZHOU Jing-quan; LI Qiang; School of Electronic Science and Engineering,Nanjing University of Posts and Telecommunications,Nanjing 210003,China

Keywords:: cloud manufacturing service composition; DPACO; QoE evaluation model; fitness function

摘要:: 为了使云制造资源更加有效地分配到各个制造任务中,提出了一种动态参数蚁群算法(Dynamic Parameter Ant Colony Optimization,DPACO)。该算法建立在 QoS(Quality of Service)评估模型之上,QoS 评估模型通过综合成本 C(Cost)、时间 T(Time)、质量函数 Q(Quality function)和满意度 S(Satisfaction)四个方面得到适应度函数 F,F 越小结果越优。 DPACO算法通过改变参数在不同阶段的值来使算法获得更快的收敛效率,加入特殊蚂蚁使得算法更好地跳出局部最优解获得全局最优解。最后通过钢铁锻造任务的云制造资源优选将 DPACO 算法与原始 ACO、PSO、DE 算法作比较,实验结果表明,DPACO 算法在求解云制造服务组合问题上能够更好地获得全局最优解,并具有较高的收敛效率。

Abstract:: In order to make the cloud manufacturing resources more effectively allocated to each manufacturing task,a Dynamic Parameter Ant Colony Optimization (DPACO) is proposed. It is on the basis of QoS (Quality of Service) evaluation model which gets a fitness function F by combining cost,time,quality function and satisfaction. The smaller the F,the better the result. DPACO can accelerate its convergence by changing the parameters in different stages and better jump out of local optimal solution for global optimal solution by adding a special ant algorithm. Finally,DPACO is compared with ACO,PSO and DE through cloud manufacturing resources optimization of steel forging task. The experiment shows that DPACO can be able to obtain the global optimal solution in solving the questions of cloud manufacturing service portfolio with higher convergence.