Maximizing QoS in heterogeneous Wireless Sensor Networks: Memory and time considerations

A Wireless Sensor Network (WSN) is a set of nodes that sense and process information from the surrounding environment. For some critical applications, WSNs are very restrictive in terms of quality of service (QoS). QoS should be guaranteed but also energy should be saved. In our work, we consider the problem of power allocation in heterogeneous Wireless Sensor Networks (WSNs), especially critical WSNs where missing the detection of an event (e.g. lost frame carrying the event alert) can provoke a human death, massive financial losses or environmental disasters. Our goal is to maximize the frame success rate of the network while preserving the scarce resource in WSNs which is the energy consumption. The maximization of the frame success rate subject to the sensor nodes powers is an NP-Hard problem. A heuristic algorithm is then needed to reach the global optimum. Since WSNs have limited memory space and computational capacity, heuristics that reduce the memory space and time usage are highly desirable. We propose, in this paper, the FLAPH which is a learning algorithm that achieves good tradeoff between memory and time consumption with comparison to another powerful algorithm referred to as Max-logit.