Regresar

Multi-task versus consecutive task allocation with tasks clustering for Mobile Crowd Sensing Systems

Abstract:

In this paper, we address the problem of multi-task allocation for dense mobile crowd sensing systems (MCS). The multi-task allocation problem is well-known to be time-consuming as the number of tasks or workers increases. However, the problem can be decomposed into small subproblems based on the location of both tasks and workers. For this reason, we propose to investigate two ways of reducing the complexity of the multi-task allocation: i) grouping tasks based on their proximity and their expected QoI such that the quality of information is maximized and ii) a meta-heuristic algorithm based on Particle Swarm Optimization (PSO) to solve the selection of workers for several tasks or sequentially solve the workers selection for single task which gives a consecutive allocation of the tasks for each worker. Simulation results shows that the consecutive allocation provides better quality of information for the allocated tasks in comparison with the multi-tasks allocation for a given cluster configuration while guaranteeing the delivery of the requested sensed data on time.