| We study dependent task offloading in renewable-powered multi-access edge computing (MEC). We propose G-DTO-SC, a novel problem to jointly (i) offload tasks across edge servers and a cloud, (ii) decide which services to cache on edge servers that have limited resources, and (iii) utilize energy transfers of harvested green energy among edge servers under application-delay constraints. Its objective is to maximize green energy usage. We formulate G-DTO/SC as a mixed-integer linear program (MILP). We analyse the complexity of the MILP in terms of the total number of decision variables and constraints. We aim to present an empirical evaluation of the MILP in the extended version of this paper, which reports optimal results across multiple topologies and sensitivity to energy and edge server cache capacity limits with baseline comparisons. We envisage the MILP would be able to solve only small instances of the problem due to its complexity. Thus, next, we will design a heuristic solution to solve large instances of the problem. |