Access Path Planning of Mobile Agent in Wireless Sensor Networks

Adopting the two-stage optimization model and hybrid optimized algorithm based on evolutionary computation, a new two-stage optimization model that more conforms to the actual demand is proposed on the basis of formal description of Mobile Agent access path planning. This new model divides the access path planning problem into two sub problems of integer linear programming --data integration sub paths and return sub paths, which can reduce search space and improve the efficiency of algorithm. Then a hybrid optimized method named GAPSO, combined with GA (Genetic Algorithm) and PSO (Particle Swarm Optimization), is advanced to solve this model, which integrates discrete PSO into the interlace operation of GA to avoid infeasible solution and improve search quality. Meanwhile convergence can be accelerated by optimizing the GA population with PSO in search of return sub paths. By means of virtual connected topology graph, the high-quality to-be-accessed candidate node set is acquired, the number of to-be-selected nodes is reduced,and the complexity of solution space is decreased, making planning algorithm performance not rely on network scale directly any more. Simulation results show that the advantages of the optimization model is obvious as the node number increases, and GASPO has a better performance than GA and BPSO in the same model.