Abstract
The proportional fair allocation depends firstly upon the priority weights given to users and secondly upon the proportional distribution of resources according to these weights. This distribution is trivial in single-technology systems but remains unresolved for heterogeneous wireless systems. This letter addresses this allocation problem not paying attention to the weights calculation, because this issue has been widely studied in the past. The novel proportionally fair scheduler for heterogeneous wireless systems (PFS-H) is based on linear programming and allocates any arbitrary resource quantity to users that are simultaneously served by several technologies. Given a priority list, the solution of the optimisation problem divides users and technologies into clusters to perform the fairest allocation of resources among users while maximising its usage. Results show that the PFS-H improves fairness up to 60% as compared with a single-technology policy. Copyright © 2011 John Wiley & Sons, Ltd.
