Optimal mobile agent routing with fusion cost in wireless sensor network

We address the problem of routing mobile agent with fusion cost in WSN under two scenarios: all source nodes, and a part of source nodes to be visited, respectively. Two objective functions in terms of minimizing transmission and fusion energy expenditure in the two scenarios are formulated, and proved to be NP-complete. Furthermore, we propose a two-level encoding based genetic algorithm (GA) to compute a suboptimal solution for the first scenario, and for the second scenario, we propose two computation paradigms to obtain a suboptimum: three-level encoding based GA, and two-step GA which consists of a two-level encoding GA and a simple GA. Simulation results show that proposed GA outperforms local closest first method (LCF) with different unit fusion costs and data reduction ratios in the first scenario. In the second scenario, simulation results with different conditions show that proposed algorithms are feasible, and three-level encoding based GA outperforms two-step GA in solution quality but the latter has superior performance in convergence speed.