不确定环境下基于鲁棒粒子群优化的物流射频识别网络优化

针对电子标签位置不确定的物流射频识别(radio frequency identification,RFID)网络优化问题,综合考虑覆盖率、负载平衡程度、成本,建立了鲁棒优化模型.为求解负载平衡程度,采用基于Korobov点阵的蒙特卡洛方法.为减少计算量,提高算法寻优能力,提出一种基于不对称时变S–形(Sigmoid)函数的鲁棒粒子群算法(PSO).样本规模仅取部分较小整数、部分较大整数.仅在算法迭代后期,样本规模期望值大,保证算法开发精度;在较多迭代次数中,样本规模期望值小,加快算法探索速度.仿真实验表明,该方法具有较佳的搜索性能.

Logistics RFID Network Optimization Based on Robust PSO under Uncertain Conditions

To deal with the logistics radio-frequency-identification (RFID) network optimization problem when the position of the electronic tag is uncertain, we build a robust optimization model in which the coverage rate, the load balance and the cost is considered. The Monte Carlo method based on Korobov Lattice is applied to calculate the load balance. A sort of robust particle swarm optimization (PSO) algorithm based on asymmetrical time-varying sigmoid function is put forward to reduce the computation complexity and enhance the searching ability. Only some small integers and large integers are employed for the sample size. In the anaphase of the algorithm, the expected value of sample size is large, thus the exploitation precision is ensured. In most other iterations, the expected value of sample size is small, thus the exploration speed is accelerated. Simulation results show that this method possesses better searching ability.