Online Energy Balancing Strategy Based on Lyapunov Optimization in Mobile Crowdsensing
Sha Chang, Su Deng, Yahui Wu, Wubin Ma, Haohao Zhou · IEEE Transactions on Industrial Informatics · 2022
In mobile crowdsensing (MCS), the battery of participants is often limited. When participants perform too many sensing tasks resulting in insufficient remaining energy, they will exit the MCS system. This article mainly addresses the energy balancing problem to prolong the system lifespan. By this means, it can ensure adequate participants and promote the completion of tasks. First, it formulates a discrete time optimization model, which transforms abovementioned problem into the online control of task admission and allocation. In addition, this model uses remaining energy variance of the participants to measure the degree of balance. Next, an online energy balancing strategy (OEBS) is proposed based on the Lyapunov optimization, which can realize energy balance without utility loss. Finally, an approximate optimal policy is presented based on the linear programming and genetic algorithm to solve abovementioned optimization problem. Experiments show that OEBS effectively maintains adequate participants, prolongs the lifespan of MCS system and maximizes the system utility even when there are few participants with multiple tasks. Specifically, the lifespan in OEBS is longer than that in utility optimization algorithm (UOA) and LP-relaxation algorithm significantly. The total utility in OEBS is more than that in UOA. OEBS can maximizes average utility of system by adjusting${\bm{V}}$. In addition, the energy balancing ability of OEBS is always effective as${\bm{V}}$changes.