The invention discloses a hierarchical collaborative resource optimization method of an MEC-assisted VR video
stream XL-
MIMO transmission system. According to the method, the interference between the users is managed by optimizing the link selection between the users and the BS subarrays, and the transmission capability of the
system is improved. Meanwhile, the invention also relates to optimization of block QP selection, calculation unloading decision, beam forming vector and calculation
resource allocation, and aims to minimize weighted
energy consumption under the constraint of meeting the requirements of time
delay and field-of-view
distortion. In order to improve the optimization efficiency of a high-dimensional mixed integer nonlinear
programming problem, the invention provides a hierarchical collaborative optimization joint calculation unloading and
resource allocation module, and an upper layer solves QP selection, beam forming design and calculation
resource allocation by using a traditional convex optimization method. And the lower layer optimizes link selection and computes an unloading decision based on a decision-based dual-depth Q network and
imitation learning. The method can overcome the defects that a traditional optimization method is long in calculation time consumption and prone to falling into
local optimum, so that the resource allocation efficiency is remarkably improved, and the
system environment change is quickly responded.