论文部分内容阅读
针对粒子群算法在短程迭代的状况下搜索精度差、波动大、粒子状态考证不足的问题,提出了一种面向可交互式智慧鱼群的权重动态约束的粒子群算法。根据粒子状态将粒子群进行分离,对粒子群进行动态约束管理,并使用“系数收敛管理器”的概念保留了粒子间的差异化运动。设定估价函数,采用权重动态约束,完成粒子群的快速求解,并使之应用于智慧鱼群模拟。结果表明,在大规模虚拟生物集群移动中,权重动态约束效果最好;完成智慧鱼群运动时,明显优于普通粒子群算法,且速度明显加快。该方法已经很好的用在了自主开发的虚拟水族馆系统中,运行稳定可靠。
Aiming at the problem of poor precision, large fluctuation and particle state verification in the short-range iterative state of particle swarm optimization algorithm, a particle swarm optimization algorithm based on weight dynamic constraint of interactive fish swarm optimization is proposed. Particle swarm is separated according to the particle state, the particle swarm is dynamically constrained to be managed and the concept of “coefficient convergence manager” preserves the differentiated movement between particles. The evaluation function is set up, and the weight dynamic constraints are used to complete the particle swarm optimization and make it apply to the intelligent fish swarm simulation. The results show that the weight dynamic constraint is the best in large-scale virtual biological cluster movement. When the wisdom fish movement is completed, it is obviously better than the ordinary particle swarm optimization algorithm, and the speed is obviously accelerated. The method has been well used in the self-developed virtual aquarium system, stable and reliable operation.