基于鞍点近似的传感器网络性能优化策略

Sensor Network Optimization Strategy based on Saddle-point Approximation

为降低对温度传感器网络的性能评估成本,提出了一种更准确的传感器网络性能检测方法,避免因性能评估误差产生的负面影响。建立的传感器网络的模型,提出了基于鞍点近似估计原理的性能评估策略,详细阐述了鞍点近似算法的优化过程,并给出了其对应的误差处理方法。该方法通过采用鞍点逼近方法,并利用一组结果证明了该方法的高准确性和低复杂度。仿真测试结果表明,在某些情况下最优化联合规则等同于一个简单的多数决定原则。

For reducing the appraisal cost of temperaturesensor network performance, a more accurate examination method of sensor network performance is proposed, thus to avoid the negative influence resulted from the performance appraisal error. With the sensor network model, the performance appraisal strategy based on estimation principle of saddle-point approximation is proposed, the optimization process of saddle-point approximation method described, and its corresponding error processing method given. By adopting saddle-point approximation, high accuracy and low complexity of the method are verified with a set of results. The simulation result indicates that the optimized union rule equals, under certain conditions, a simple majority-decisive rule.