Solving operational models of interdependent infrastructure systems

Abstract

We formulate and solve a model of interdependent fuel and electric power infrastructure systems with explicit representation of the fuel required to run some electric power generators and the power required to heat and pump fuel. Our model determines a set of fuel and power flows that result in the minimum-cost of operating both systems, including penalty costs for failing to deliver each material to each of several external customers. We then formulate models of each system separate from the other, and, for each system, represent each interdependence relationship as a demand node with associated penalties. We implement an iterative algorithm for solving various instances of the problem; the algorithm alternates between solving each system separately, and passing material requirements to the other model. We then evaluate how well our algorithm performs in comparison to the monolithic formulation. We conclude with suggestions for improvements to the algorithm.