Robust optimization of forecast combinations

Abstract

In this paper, we introduce RF energy harvesting paradigm into WNCS framework to study the optimal power control, energy harvesting and scheduling problem with the objective of providing maximum level of adaptivity under strict timing and reliability requirements employing the constant rate transmission model. We formulate the problem as a Mixed Integer Linear Programming Problem (MILP). We show the power allocation can be separated from the scheduling and harvesting at optimality. Then, we introduce a heuristic algorithm for the scheduling problem, periodic list scheduling (PLS), inspired from list scheduling of jobs with sequence dependent setup times on identical machines. We then demonstrate via extensive simulations the superiority of the proposed algorithm in terms of closeness to the optimal, adaptivity and runtime.