A new "data size" based algorithm for itinerary planning among mobile agents in wireless sensor networks

Abstract

The integration of mobile agent systems in the wireless sensor networks (WSN) has a positive impact on this type of networks. The itineraries followed by mobile agents have a influences significantly on the energy consumption. Finding an optimal sequence of visited source nodes (data sensors) is an NP-hard problem. Many researches have dealt with itinerary planning in WSNs through the use of a single agent (SIP: Single agent Itinerary Planning) or multiple mobile agents (MIP: Multiple agents Itinerary Planning). The itinerary planning, for multiple agents in WSN (MIP), must consider the following three issues: (i) the appropriate number of Mobile Agents, (ii) the suitable grouping of source-nodes, (iii) and finally, the optimal itinerary for each Mobile Agent to visit all its associated nodes. In the current MIP solutions, the geographical distance is the unique factor motivating to plan the itinerary of the agents. These solutions don't consider the data size provided by each node which is another factor affecting energy consumption too. In this paper, we provide a new MIP solution which is based on the both basic factors (geographical distance and data size) affecting energy consumption. Using these two factors, this study provides a new way for determining the number of mobile agents and for grouping source-nodes. Once, the number of mobile agents and the node groups are defined, we identify the itinerary (a path passing through the set of source nodes that must be visited by the mobile agent). The itinerary is defined using the GRASP (Greedy randomized adaptive search procedure) heuristic algorithm. The simulation results prove that our proposal is more efficient than other existing approaches in terms of task duration and the amount of consumed energy.