An adaptive agent-based application for quality of service management in wireless internet network

Abstract

A conceptual adaptive QoS management model is proposed, emphasizing on managing limited resource and also maximizing overall and individual user satisfactions. This model contains QoS management functions and cycle. The functions are designed to work on both wireline and wireless networks to reduce the QoS gaps. Moreover, the cycle is stated as the processing timeline. The results from a case study of feature adaptation confirm to increase user satisfaction. Moreover, this dissertation proposes the Multicriteria-Based (MCB) scheduling policy to manage the incoming request scheduling for the web server; the initial step of the QoS management cycle. Since many researches tend to work on single criterion scheduling policies, which focused on achieving overall user satisfaction, while the individual user satisfaction remains unclear. MCB is a compromising policy, which puts effort on both overall and individual user satisfactions. The merit validation of MCB over traditional policies: FIFO, EDF, and SPT, is stated by conforming the vector calculus. Arrival time, deadline, and processing time are nominated for the scheduling criteria. The proposed model is expressed in the scheduling and performance measurement functions. The M/G/1 system is experimented based on MATLAB for indicating the merit verification of MCB. Comparing MCB to the three traditions, the simulation results indicated that MCB is an optimal policy by optimizing average, maximum, and standard deviation waiting times for ideal situation; non-deadline checking case. Both of the validation and verification results provide the evidences confirming that the MCB policy is potent enough to broaden it into the high variable environments.