Performance study of a clustering-based genetic algorithm for data gathering by a mobile robot in wireless sensor network

The problem of planning a path for minimizing the distance the mobile robot has to traverse to accomplish the task of gathering the data stored in a spatially distributed wireless sensor network is a kind of Traveling Salesman Problem with Neighborhoods (TSPN). We proposed a one-hop data-gathering scheme using clustering-based genetic algorithm (CBGA) that performs well: following the planned route, the robot can gather all data from all sensors while the travel cost of the robot decreases obviously. Further comparative simulation results of CBGA with other solutions of TSPN in a network with identical sensing radius or random sensing radius are presented in this paper to reveal the relative performance of each solution scheme and highlight the effect of clustering, the role of GA in route design. In particular, we demonstrate the scalability of CBGA to large-scale network and quantitatively reveal the significant path length reduction, showing the advantages of integration of clustering and GA.