چكيده به لاتين
Majority of existing task schedulers in WSANs are only concerned with energy savings or meeting temporal constraints, overlooking service reliability. They do not consider the types and various features of tasks that WSANs may need to perform, making such schedulers inapplicable to critical applications such as search and rescue missions. In this study, an approach is presented to address the mentioned challenges. To this end, we have introduced a scheduling model that considers important criteria such as load balancing, reliability, and performability of actors. For schedulability analysis, a reachability model formally using process algebra is presented. Based on the proposed schedulability model, the schedulability of tasks scheduling approach, the formal availability model is presented using process algebra. Also, using the proposed scheduling model, we have proposed a task scheduling approach considering requirements of hybrid architecture. To allocate local tasks, a cost-effective and auction-based distributed scheduling protocol (DSP) is presented. Using the proposed protocol named FitCa, DSP determines a cost-effective set of actors and assigns local tasks to the best actors. In order to allocate global tasks considering the fact that they are more complex than local tasks and may have one or more dependent subtasks, using bibgeography optimization technique, allocation of global tasks is carried out. Finally, after allocation of all tasks to actors, tasks are scheduled at actor level using the proposed task scheduling protocol (TSP). To verify the correctness of the proposed approach and models, a framework based on timed automata is presented. Based on this model checking technique we analyze and prove bounded liveness, deadlock-freeness, and correctness. The results obtained from the required simulations of the proposed approach represent 14% improvement in reducing make-span, 11% improvement in increasing service reliability, 15% improvement in increasing deadline hit ratio, and 12% improvement in increasing residual energies of actors compared to the best results of the related works. It is also shown that the proposed approach can increase the life time of actor nodes.
