No abstract

Design of Radio Resource Management (RRM) protocols in wireless networks plays a key role in the system's performance. By handling different service classes, RRM protocols must find out a trade off between the Quality of Service (QoS) requirements of real time applications and non-real time applications as well as between users and Service Providers expectations. In the current paper, Traffic Management (TM) mechanisms are combined to deal with the problem of designing multiservice RRM protocol . Particulary, we use Call Admission Control (CAC), Scheduling, and Buffer Management to ensure the QoS provisioning of the different service classes accordingly their QoS demands. Analysis of the results is outlined and the conclusions are drawn about their performance.

Abstract-In this paper we consider the problem of Radio Resource Management (RRM) in mobile cellular networks offering not only real time services (voice services), but also data services. In particular, we investigate integrated Call Admission Control (CAC) and Scheduling policies that can provide a good tradeoff between the QoS of voice and data services, considering two alternatives, and developing analytical techniques for the assessment of their performance metrics. In all schemes the admission criteria is based on network load (channels and buffer occupancies); and the CAC also performs service classes differentiation reserving some radio channels in order to favor handoff and data packet traffics over new call connection requests. The scheduling policies are FIFO and Priority Queueing (PQ). Analysis of these schemes are outline and the results show that: PQ may be perfectly employed when it is needed to employ service classes differentiation between the non-real time traffics; and that FIFO achieves acceptable results when all data packet traffic have to be queued in only one buffer. The results also show that the use of the reserved radio channels to carry out the data packet traffic can enhance the QoS perceived by the lower priority data packet traffic.

O presente trabalho compara a aplicação de agentes inteligentes aplicados ao mundo do Wumpus. Os agentes estudados são baseados em regras e algoritmos genéticos (AG). Os experimentos mostram a possibilidade de utilização de ambas técnicas para problemas de busca, bem como um melhor resultado para aquele agente baseado em AG. Além disso, compara variações dos AGs, investigando a utilização de elitismo no processo de seleção da próxima geração.

No abstract