Multi-objective parameter optimization of S-MAC protocol
In the process of design of protocols for medium access control in wireless sensor networks, it is necessary to find a compromise between energy saving, time sensitivity, scalability, adaptability and complexity. In this research, as an object of optimization S-MAC protocol is chosen. The main advantage of the S-MAC protocol is a state of sleep, which significantly reduces power consumption. Although the S-MAC protocol is described with a large number of parameters, in this paper we have selected three parameters that directly affect the time frame protocol: the duty cycle, the duration of the synchronization interval, and duration of the data interval. To assess the quality of the selected parameters, three measures were selected: total electricity consumption in the network, delay in the delivery of packets to the application layer, and the total achieved bandwidth. Using ns-2 network simulator, a model for testing the influence of parameters of S-MAC protocol on network performance has been developed. The parameters values are selected using the NSGA-II algorithm implemented in MATLAB software package, and the whole process of finding of non-dominated solution set is fully automated. In addition to multi-objective optimization of parameters, an experiment with single-objective optimization using aggregated utility function and the standard GA is performed. Finally, the analysis of the obtained set of non-dominated solutions is performed and a comparison with the results of the single-objective optimization is done.