Analytical model for multi agent - call admission in wireless networks

Abstract

Resource allocation and its planning are major issues in wireless networks. Distributed call admission control requires calls to be admitted by considering local as well as global view of the network so as to reduce overall congestion. The paper proposes MA-CAC (Multi Agent-CAC) scheme which models a D-CAC problem using multiagent system (MAS) where each cell comprises of agents. These agents interact, communicate, coordinate or negotiate with each other to determine the optimal utility function based on maximum tolerable handoff dropping probability for a cell. It also presents the analytical model for calculating dependency of handoff dropping probability of a cell on the number of calls in the adjoining cells.