求解布局分配问题的群智能聚类算法

Swarm intelligence clustering algorithm solving location-allocation problem

提出1种适用于求解布局分配问题的的群智能聚类算法,布局分配问题属于选址问题的1种,其数学表达为包含混合变量的非线性规划模型。根据问题的特点将聚类思想引入群体智能算法,种群中的个体再分出1级,称为"子个体",同一个体具有相同的适应度,而子个体根据个体适应度与自身所获得的信息进行移动,其移动有3种方式,方式1与方式2是子个体根据周围需求点的分布而移动,方式3为结合当前种群最优位置信息移动。对7组文献数据使用算法进行50次优化运算,并给出运算的最优值、均值与标准差。运算结果表明:群智能聚类算法能够达到或接近问题的最优解,与一些智能算法相比,算法在问题规模较大时运算结果更优。

To solve the location-allocation problem, a swarm intelligence algorithm based on clustering operation is preseted, which belongs to location problem expressed by a complex nonlinear programming model. According to the characteristics of the problem, clustering idea was combined into swarm intelligence algorithm, and a new level would be set up under the individual in the population,called “individual branch”. It can move independently based on their own information and the individuals fitness, and its mobile has three ways. The way 1 and 2 move according to the distribution of demand point, and the way 3 is combined with the current population optimal location information Using the 7 sets of data in the literature to test, and operation 50 times, swarm intelligence clustering algorithm can reach or close to the optimal solution Compared with some intelligent algorithms the algorithm has better results in large scale problem.