The prevailing research on privacy-preserving coded computation neglects the heterogeneity and time-varying nature of edge network,leading to compromised computational efficiency.To address this gap,a method for privacy-preserving coding computation offloading based on fountain codes is in-troduced.Firstly,a polynomial-fountain code encoder tailored for privacy-preserving distributed matrix multiplication is designed.This encoder supports ongoing task outputs until successful decoding,offer-ing an adaptive coding rate adept at accommodating fluctuating network conditions.Secondly,the tem-poral attributes of node computation is analyzed,and three distinct task allocation models are pro-posed,which enriches the understanding of allocation strategy performance.Finally,a dynamic task distribution method is subsequently presented,emphasizing node heterogeneity by allocating a larger share of tasks to high-performance nodes.This method adaptively fine-tunes the allocation strategy based on feedback,catering to the evolving system dynamics.Simulation outcomes validate the effi-cacy of this method,concurrently ensuring rigorous privacy protection as per information theory stan-dards.
维普
万方数据
