Beamforming in heterogeneous networks with Massive MIMO

Abstract

Poor TSV assignment algorithm leads to a large detour of wire-length.Some of the previous works deal with TSV assignment problem using layer-by-layer methods(LBL) only gets local optimal results.In this research work,a backward layer-by-layer reassignment(BRLBL) is developed to do further optimization.This reassignment algorithm takes global information into consideration and optimizes the outcome of layer-by-layer TSV assignment methods.Also a fast layer-by-layer reassignment(FLBL) method is proposed to shorten the runtime.Experimental results show that BRLBL achieves an average reduction over 3.7% of wire-length,while the FLBL improves the execution time of BRLBL by 10 times.