GSPN-based Modeling of Task Allocation for Increased Utilization of Actors in Wireless Sensor Actor Networks

Wireless Sensor Actor Networks (WSANs) are a special kind of wireless network composed of wirelessly communicative stationary sensor nodes and actor nodes. Sensor nodes gather environmental information, while actor nodes respond to sensory information. Task allocation is a critical issue for properly engineering cooperative applications in embedded systems with latency and energy constraints, such as WSANs. Unresolved challenges in scheduling and task allocation to actors in WSANs have resulted in lower than expected network efficiency. Since actors in a WSAN may have different capabilities in performing allocated tasks, implementing suitable scheduling policies can enhance the network's performance. The aim of this paper is to maximize the total utilization of actors in performing tasks within a WSAN. To achieve this goal, we propose a Generalized Stochastic Petri Net (GSPN) model to illustrate the task allocation policy applied by the network sink. To maximize the utilization of actors, the arrival rate of tasks for each actor should be calculated based on its capability. By performing steady-state analysis of the GSPN model and calculating the task dispatching weights at each actor, the task arrival rates can be determined. The results of applying the proposed GSPN model to allocate tasks to actors in a typical WSAN demonstrate its applicability.