Traffic predictions bave beeD demonstrated ",jtb tbe capability
10 improve Detwon .fficieDey and QoS in broadband ATM n_rks.
Recent rese3n:b sbows tbat rurl)' logic prediction outpem.rms conventional
autoregression predictions [11 [21_ The application of fuzzy logic 0"0 bas a
potential to coDtrol traffic more effectively_ In this paper, wo propose tbe use
ortbe furl)' logic prediction on connection admission control (CAC) and congestion
control 00 bigh speed oetworks. We first modeled traffic cbaractor
isties using an oo-lioo fu:zzy logic predictor on CAC. Simulation results sbow
tbat rurl)' logic prediction improves tbe efficieney of botb conventiooal aod
measurement-based CAC.lo additioo, tbe measurement-based approacb incocporating
fuzzy logic inference and using fuzzy logic prediction is sbown
to achieve bigber netwon utilization wbile maintaining QoS. We tben applied
tbe fuzzy logic predictor to COag.StiOD control in wbicb lbe ABR queue
is estimated one round-trip in advance. Simulatioo ""ults sbow that tbe
fuzzy logic control scbeme significantly reduces convergence time and overall
buffer requirements as compared with conventional schemes.
Keyworru- fuzzy logic predictioo, asyncbronous transfer mode (ATM)
networks, quality of service (QoS), connection admission conlrol· (CAC),
congestion control.