Informs Annual Meeting Phoenix 2018

INFORMS Phoenix – 2018

SB47

economical agreement between the players. The resulting problem was solved using the Nash bargaining solution. 2 - Multicriteria Decision Analytic Framework for Evaluating Future Power Generation Pathways Thushara De Silva, PhD Candidate, Vanderbilt University, 105 Jefferson Square, Nashville, TN, 37215, United States, George M. Hornberger, Hiba Baroud Power generation planning objectives are reliable power system, economic efficiency, environmental sustainability, and social acceptability. Multiple alternatives, consists of different technologies and resources, must be assessed in multiple objectives. The objective of this study is to develop a multicriteria decision analysis model to select a power generation pathway for Sri Lanka by developing different alternative pathways, examining them across multiple objectives, and incorporating preferences of multiple stakeholders. A pathway; mix of renewable and fossil fuel resources aimed at achieving energy security can meet multiple criteria associated with future power generation. 3 - Calibration of Transportation Models with Scarce Cost Data: A Mathematical Program with Equilibrium Constraints Charalampos Avraam, Johns Hopkins University, 3400 N. Charles Street, Baltimore, MD, 21218, United States, Anastasia Lambrou, Wei Jiang, Sauleh Ahmad Siddiqui, Anthony So The lack of detailed cost data in critical sectors of the economic system, such as the Livestock production sector, confines the ability of economic modelers to properly account for critical details, limiting the usefulness to policymakers. In order to calibrate for unknown or uncertain parameters, our proposed calibration method is formulated as a Mathematical Program with Equilibrium Constraints, where the lower level is the market equilibrium problem and the upper level minimizes the difference of the uncertain or unknown parameters from trademark values. Furthermore, we systematize the trade - off between accurately calibrating for a set of parameters versus another. 4 - Solving Problems with Equilibrium Constraints with an Application to Energy Markets Sauleh Ahmad Siddiqui, Johns Hopkins University, 3400 N. Charles Street, Baltimore, MD, 21218, United States We provide a new set of optimality conditions for solving mathematical programs with equilibrium constraints and extend them to solve equilibrium problems with equilibrium constraints. We develop algorithms based on nonlinear programming to provide insights into energy markets with hierarchical structures. We conclude with policy insights and recommendations on how this approach can be extended. n SB47 North Bldg 229A Nature-Inspired Heuristics: Overview and Critique Emerging Topic Session Chair: Theodore P. Pavlic, Arizona State University, ASU - CIDSE, P.O. Box 878809, Tempe, AZ, 85287-8809, United States 1 - Nature-Inspired Heuristics Craig Tovey, Nature-Inspired Heuristics, Atlanta, GA, United States Evolutionary Programming, Genetic Algorithms, Simulated Annealing, Ant Colony Optimization, and Particle Swarm Optimization are among the earliest optimization heuristics inspired by animate and inanimate phenomena in the natural world. Some of the more recently invented methods have exotic names such as Roach Infestation, Shuffled Frog-Leaping, Invasive Weed Optimization, and Cuckoo Search. This tutorial is a guide to the bewildering, burgeoning menagerie of such heuristics, which now comprises more than 100 algorithms, and whose accompanying publications number in the hundreds of thousands. Their underlying principles include populations, recombination, exploration, reinforcement, encoding, selection, randomness and perturbation. They have been successful in many implementations, ofttimes winning out against classical OR/CS methods for implementation or a user’s acceptance. They have been less successful, sometimes spectacularly so, in many computational test comparisons with classical methods. I speculate as to why these success levels differ so greatly. On the one hand, I critique the insularity and mathematical naivet of this heuristic research, particularly its limited nature-versus-nature comparisons and self-contradictory stance on algorithm parameter tuning. On the other hand, I critique the OR community’s having implicitly ceded important optimization territory to others, and our failure to face what is now there. In an attempt to spur our community to thoroughly engage with nature-inspired heuristics, I conclude by observing a misalignment between individual and community incentives, identifying a few potentially powerful nature-inspired heuristic ideas for optimization, and proposing some specific research questions that may attract operations researchers.

n SB45 North Bldg 228A Real-time Optimization in Power Networks Sponsored: Energy, Natural Res & the Environment/Electricity Sponsored Session Chair: Enrique Mallada, Johns Hopkins University, Baltimore, MD, 21218, United States Co-Chair: John Simpson-Porco, University of Waterloo-ECE Department, Waterloo, ON, N2L 3G1, Canada 1 - A System Level Approach to Frequency Control Nikolai Matni, UC Berkeley, Berkeley, CA, United States We show how the System Level Approach to controller design can be used to explore design tradeoffs in primary and secondary frequency control. In particular, we show that localized distributed optimal controllers can be computed and implemented at scale, and explore tradeoffs in system performance, robustness, actuation density, sampling time, communication speed, and coordination between subcontrollers. We end with a brief discussion of progress made towards co-designing such low-level feedback, or reflex, layers and higher- level planning layers, such as those solving optimal power-flow problems. 2 - Optimal Steady State Control for Frequency Regulation John Simpson-Porco, University of Waterloo, 200 University Avenue West, Waterloo, ON, N2L 3G1, Canada, Liam S. Lawrence, Zachary Nelson, Enrique Mallada We consider the problem of designing a feedback controller that guides the input and output of a linear time-invariant system to a minimizer of a convex program in the presence of unknown disturbances. The proposed controller combines proportional-integral control with gradient feedback, and enforces the KKT optimality conditions in steady-state, without incorporating dual variables into the controller. The results are applied to derive different centralized and distributed secondary frequency regulation strategies for power systems. 3 - Online Optimization with Feedback for Power Grids Emiliano Dall’Anese, University of Colorado Boulder, CO, United States This talk addresses the design of feedback-based online algorithms for distribution systems. The time-varying optimization formalism is leveraged to model optimal operational trajectories of a distribution system, as well as explicit local and network-level operational constraints. The design of the algorithms capitalizes on an online implementation of primal-dual projected-gradient methods; the gradient steps are modified to accommodate measurements from the system. The algorithms can cope with model mismatches, avoid pervasive measurements, and lend themselves to distributed implementations. Convergence claims are established in terms of dynamic regret and Q-linear convergence. 4 - The Effect of Power System Dynamics on Feedback Steady-state Optimization Adrian Robert Hauswirth, ETH Zurich, ETL I. 13, Physikstrasse 3, Zurich, 8092, Switzerland We consider the problem of optimizing the steady-state of a power system in closed loop. While the design of the feedback steady-state optimization law is based on a time-scale separation argument, in reality the dynamics of the (slow) iterative optimization routines can interfere with the (fast) natural system dynamics (e.g. primary frequency control). We provide a study of the stability and convergence of these optimization routines when applied to power systems, and we suggest solutions to maximize the robustness of these methods with respect to the underlying power system dynamics. n SB46 North Bldg 228B Decisions in Energy Markets using Extended Concepts of Equilibrium Sponsored: Energy, Natural Res & the Environment/Energy Sponsored Session Chair: Sauleh Ahmad Siddiqui, Johns Hopkins University, Baltimore, MD, 21218, United States 1 - Demand Response by Load and EV Aggregators Linked by an Option Contract for Load Sharing Kevin Melendez, University of South Florida, Tampa, FL, 33613, United States, Tapas K. Das, Changhyun Kwon With the increasing penetration of electric vehicles (EVs) on the market, the coordinated control of a large feet of EVs by an aggregator has obtained special attention in the last decade. The cooperation among the EV aggregators and other demand response agents must be addressed to adequately incorporate this new player into the electricity market. In this work, we proposed an option contract among the EV aggregator and a demand response aggregator to define the

47