Performance Evaluation of Routing Protocols for Mobile Ad Hoc Networks Using Statistical Taguchi’s Experimental Design

This paper presents the application of Taguchi approach in investigating the performance of mobile ad hoc networks (MANETs) with respect to drop rates. The performance of MANETs was influenced very much by the routing protocol designs and quantitative factors such as network size, terrain, pause-time, node speed, transmission range, traffic load, and packet rates. The NS-2 simulator was used for the simulation of the MANETs under a specific scenario. The tools of Taguchi approach such as L12 orthogonal array, analysis of signal-to-noise ratio, and analysis of variance (ANOVA) were employed. Results showed that the effects of network size, terrain, node speed, transmission range, traffic load, packet rates, and routing were found to be statistically significant, and the most significant factor was transmission range. Using the combination of optimal levels of the significant factors, the specific values of drop rates was predicted. Good agreement between the predicted drop rates and actual drop rates was observed in this study. Key–Words: Experimental design, Taguchi technique, analysis of variance, performance evaluation, mobile ad hoc networks, routing protocols, SAS JMP.