Optimal Iterative Method for Network Utility Maximization with Intertemporal Dynamic Constraints

Ke Ma1, Peng Wang2, Thiagarajan Ramachandran2, Jianming Lian, Donald J. Hammerstrom2

  • 1Texas A&M University
  • 2Pacific Northwest National Laboratory

Details

11:10 - 11:30 | Tue 20 Aug | Lau, 5-206 | TuA5.3

Session: Network Systems

Abstract

Network utility maximization (NUM) has been often formulated to solve the optimal resource allocation problem in wireless networks and power systems. The network planner aims to optimally balance the supply and demand among producers and consumers so that the social welfare is maximized while the network constraints are respected. However, when the intertemporal dynamics of agents are taken into account, a single-period NUM formulation becomes insufficient. In this paper, the multi-period NUM with intertemporal dynamic constraints is considered and solve by a proposed iterative method. It features a price iteration scheme that achieves the optimal social welfare, and the algorithm is guaranteed to converge if proper concavity conditions hold. Simulation results verifies the effectiveness of the proposed iterative method. It is also demonstrated that there is an increase in social welfare compared to a single-period formulation.