Data Transmission Schedule Optimization Of Green Cellular Network Based On Data Delay Characteristics

On account of global warming and depletion of fossil fuel resources,renewable energy sources(such as solar photovoltaic power and wind power)have received extensive attention in power-system field.Integration of renewable energy into mobile cellular networks tends to significantly reduce energy generation cost and carbon emission caused by conventional fuel generators.However,unpredictability and volatility of renewable energy sources make it difficult that generating electricity with renewable sources adapts to fluctuation in user’s electricity demand.Meanwhile,renewable energy fails to be used efficiently due to randomness of energy demand in data traffic.In particular,unregulated renewable power generation can either be over-or under-supplied within short time compared to load demand,which can result in energy waste or higher electricity cost due to unappropriate battery charging.This paper optimizes scheduling method for renewable energy generation and grid power generation in data transmission of green cellular networks with an aim that data communication performance with minimum power cost will be guaranteed.It is proposed to make use of both energy and data storage to tackle instantaneous mismatch between renewable energy generation and energy demand.According to the difference between user’s data processing requirements,we divide user’s data into emergent data and non-emergent data.The former needs to be transmitted immediately while the later is allowed to be transmitted after delayed for a period of time.In principle,we use renewable energy stored in energy storage to transmit emergent data and non-emergent data will delay to be sent in peak of data traffic.It reduces electricity cost since the consumption of traditional electricity with insufficient renewable resources is lowered.In consideration of statistical characteristics of data flow and renewable energy sources,we propose a threshold-based energy and data storage management method to minimize long-term energy cost of cellular system operator.In addition,when statistical characteristics of data flow and renewable energy are unknown or time-varying,we proposes to use reinforcement learning to decide data transmission scheduling problem in order to reduce energy consumption of green cellular networks in practice.