To meet the growing demand for wireless traffic, various types of relay and heterogeneous networks have received attention in the recent literature. This paper focuses on novel norm balancing algorithms based on cost-reward formulations on multiple access channels (MACs), and the results are extended to decode-and-forward (DF) relay systems. The achievable rates by different users in multiuser systems present an inherent tradeoff, which is formally known as Pareto efficiency. The analyses carried out in this paper show that any power allocation can be Pareto efficient in an MAC, provided certain conditions on the performance metric and decoding order. Based on the analysis, a novel method for efficient power allocation has been proposed to meet a target sum rate. The results have been explained in light of previous literature on polymatroid rate region and the requirement of rate vector equality. It is necessary to ensure and find the intersection of two polymatroids to operate the systems at a joint Pareto frontier of the relay channels. Two algorithms to test for the existence of the joint Pareto frontier have been proposed and compared.