Mobile agent routing using variable-dimension PSO algorithm based on chord-length parameterization

The Fixed-Dimension particle swarm optimization (PSO) can´t be used to solve the mobile agent (MA) routing problem where the number of sensor nodes which can be viewed as the dimension of particle is not always fixed, since the updating formulae are only fit for particles of same dimension to update their velocity and position. To solve MA routing problem, a variable-dimension PSO based on chord-length parameterization (CLP-VDPSO) is proposed in this paper. Moreover, considering the communication radius of a sensor node, we also present relay node selection criteria. Compared with the MA routing algorithms based on genetic algorithm (GA), the proposed scheme can achieve better performance, such as less cross-over points, high routing success rate, less number of iterations, high received signal strength and less cost of per signal strength.