Joint Mode Selection and Task Assignment in Multi-Mode Sensor Network

Sensors with multiple functions and working modes constitute the nodes of the multi-mode sensor network. In this work, the mathematical model of task assignment problem in multi-mode sensor network is established. The model consists of task priorities, capability matching relationships, costs of tasks, resources of sensors and mode constraints. A two-stage scheme composed of joint mode selection and task assignment is proposed and evaluated. First, the cooperative working modes of sensors are determined with quantum-behaved particle swarm optimization (QPSO). Next, task assignment is performed with modified Hungarian algorithm under market-based assumption. Our proposed two-stage approach exhibits higher flexibility and achieves better tasks completion ratio compared with the benchmark of greedy policy, while the time complexity is much less than the conventional assignment scheme with PSO algorithm. According to the Monte Carlo simulation, the assignment of 20 sensors and 1000 tasks is completed in 0.225 seconds with the proposed strategy, while the completion percentage of tasks is 7% higher than those achieved with the greedy algorithm.