采用主成分策略的传感器网络路由评估算法
Optimal Route Evaluation Method Based on Principal Component Analysis for Wireless Sensor Networks

针对传感器网络路由在最优性评估过程中呈现的排序问题,提出了一种采用主成分策略的路由评估算法(OREPCA)。首先利用蚁群策略寻找传感器网络中的样品路由;接着再根据实际的布网环境构建出相应的路由评价指标集;然后对样品路由按指标集中各指标出现的顺序进行逐项打分获得评价指标向量;最后借助主成分策略构造一个传输路由综合评价指标函数,从而实现在不同的监测环境中对网络路由的多元化评价,规避了人为选取权重因子带来的主观随意性。仿真结果表明,与基于负载均衡策略的路由优化算法及基于多目标优化的交互式路由算法相比,OREPCA算法能将网络寿命提高14％,并能有效降低网络的通信延迟。
A novel optimal route evaluation method based on principal component analysis (OREPCA) is proposed to solve the transmission route sequencing problem appeared in the optimality evaluation process of wireless sensor networks. Sample routes information is firstly collected in a discrete time domain by using an ant colony algorithm. Then, a relevant evaluation index system of routes is constructed based on the deployment environment of the network, and a positive dimensionless method is given and is applied to the indicators of the set. The priority of all indicators is used to score each sample route and to obtain a vector of evaluation indicators. Finally, a correlation coefficient matrix among different indicators in the vector is calculated, and the principal component approach is used to obtain a route evaluation function and the weighting factor of each evaluation indicator in the function. The algorithm avoids the problem of subjective randomness caused by manually selecting weighting factors based on experience. Moreover, the evaluation function generated by the proposed method achieves the diversified evaluation and prompts dynamic server load balance in different monitoring environments. Comparisons with the multi？？objective routing algorithm and the cluster based route optimisation and load balancing protocol, OREPCA can increase the lifetime of network by 14% and put down the delay of communication notably