A New Adaptive Elite Evolutionary Algorithm for Task Allocation of Wireless Agricultural Machinery Sensor Network

In the field of wireless agricultural machinery sensor network (WAMSN), small wireless sensors are equipped with restricted sensing capacity. To have a higher revenue of the network, task allocation scheme must be designed properly. However, the task allocation with discrete variables is an NP-hard problem in its precise formulation, so it is difficult to reach an optimum solution. In this paper, a new adaptive elite evolutionary algorithm (AEEA) is proposed for the task allocation problem in WAMSN. It is more powerful and simpler than available heuristics, and can avoid local optima while searching for a better result. In such a hybrid approach, the global exploit capacity of newly designed immune selection and local search capability of novel adaptive adjustment are integrated to yield faster and powerful convergence. In addition, it is found that the newly designed elite strategy helps evolution to avoid local optima. Simulations are carried out with Genetic Algorithm (GA) and Simulated Annealing (SA) and the experiments are compared to verify the represented scheme. Results demonstrate that the presented AEEA method significantly outperforms the GA and SA method. Besides, AEEA can significantly maximize the revenue of the network.