Using Mran for Call Admission Control of Atm Networks in Opnet

This paper explores the use of OPNET to introduce a neural network Call Admission Control scheme in ATM networks to improve traffic management. The advanced capability of modeling traffic sources in OPNET was employed to depict a realistic ATM network model. The conventional CAC algorithm was explored under various traffic scenarios built using profiles based on varied traffic statistical distributions. The architectural hierarchy in OPNET was then exploited to introduce MRAN Minimal Resource Allocation Network a minimal RBF neural network. MRAN was used to define a dynamic bandwidth allocation approach to the CAC algorithm. The performance of the network was scrutinized under the same traffic conditions. In conclusion, collaborating statistics such as Utilization and Call Blocking Ratio from a set of varied traffic simulations indicate that the MRAN CAC algorithm results in superior network performance and traffic management when compared to the conventional schemes.