Harnessing predictive models of durations of channel availability for enhanced opportunistic allocation of radio spectrum

摘要: A proactive adaptive radio methodology for the opportunistic allocation of spectrum is described. The methods can be used to allocate resources by employing machine learning learn models, via accruing data over time, that have ability predict context-sensitive durations availability channels. predictive models are combined with decision-theoretic cost-benefit analyses minimize disruptions service or quality associated reactive policies. Rather than reacting losses channel, policies seek switches in advance loss a channel. Beyond determining one more frequency bands statistical also employed generate price predictions order facilitate sale rental available frequencies, and these switching non-cooperating distributed allocation, centralized approaches, hybrid scenarios.