Informs Annual Meeting 2017

TB68

INFORMS Houston – 2017

TB68

TB69

371C Joint Session QSR/ENRE Electricity: Prediction and Optimization in Electric Power Systems Sponsored: Quality, Statistics and Reliability Sponsored Session Chair: Jinkoo Park, Korea Advanced Institute of Science and Technology (KAIST), Daejeon, Korea, Republic of, jinkyoo.park@kaist.ac.kr Co-Chair: Eunshin Byon, University of Michigan, 1205 Beal Avenue, 1205 Beal Avenue, Ann Arbor, MI, 48109, United States, ebyon@umich.edu 1 - Characterization of Lithium-ion Battery Model Uncertainty for Effective Battery SOC Estimation Zhimin Xi, University of Tennessee-Knoxville, 851 Neyland Dr, 517 JDT, Knoxville, TN, 37996, United States, zxi1@utk.edu Equivalent circuit models (ECMs) are major models employed in current battery management systems (BMS) for battery SOC estimation due to its high computational efficiency. However, fidelity of the ECMs varies for different battery operating conditions, thus often resulting in inaccurate battery SOC estimation. In this paper, approaches are developed to characterize the model uncertainty of an ECM under various battery operating conditions. Fidelity of the ECM is hence further improved without giving up the high efficiency of the model for onboard battery SOC estimation. Comparison study is conducted for the SOC estimation using the proposed approach and conventional method. 2 - Optimal Personnel Deployment Strategy for Self-perform Maintenance on Wind Farms Jason Kriel Lawley, Texas A&M.University-College Station, 3780 Copperfield Drive, Apt 325, Bryan, TX, 77802, United States, jsnlawley@tamu.edu Turbine maintenance is a major cost and key determinant of wind farm productivity. Many companies outsource critical maintenance while others perform these tasks in-house, referred to as self-perform maintenance. While expected to reduce time to profit on asset investment, self-perform requires an efficient personnel deployment strategy, proposed as the following: a balanced scorecard aligning data driven performance indicators to O&M deliverables, an optimization package delivering daily maintenance tasks paired with ideal teams of technicians, and a simulation model demonstrating implementation benefits. Development, features, and results of this strategy will be discussed. 3 - Adaptive Importance Sampling for Extreme Load Estimation in Wind Turbines Qiyun Pan, University of Michigan, 1205 Beal Avenue, Department of IOE, Ann Arbor, MI, 48109-2117, United States, qiyun@umich.edu, Eunshin Byon, Henry Lam To help wind turbine reliability analysis in a design stage, aeroelastic simulators have been developed to generate stochastic load responses imposed on a wind turbine. However, crude Monte Carlo sampling to simulate the extreme load associated with a small load exceeding probability is computationally prohibitive, and the estimation results are highly uncertain. We develop an adaptive importance sampling algorithm to iteratively refines the importance sampling density to efficiently estimate the extreme load. 4 - Game Theory Based Distributed Control Algorithm for Energy Storage Systems Jinkyoo Park, Korea Advanced Institute of Science and Technology, Daejeon, Korea, Republic of, jinkyoo.park@kaist.ac.kr, Heechang Ryu We construct a game theoretic framework that can model the strategic interactions between the grid operator and energy users. The grid operator seeks to maximize their profit and reduce the peak load by applying demand response strategies, such as changing energy price. On the other side, the users aim to minimize their energy consumption costs by optimally operating ESSs that can charge the energy when the price is low and discharge the energy when the demand and price are high. With the constructed game model, we derive the equilibrium strategy for both grid operator and users that can lead the minimize energy consumption cost for the users and the reduced peak load for the grid operator.

371D Vehicle Routing Contributed Session Chair: M’Hallah Rym, Kuwait University, P.O. Box 5969, Kuwait, 13060, Kuwait, rymmha@yahoo.com 1 - Research and Optimization of Facility Layout in Passenger Transfer Hall of Shapingba Transport Hub Based on Legion Simulation Anjun Li, National United Engineering Laboratory of Integrated and Intelligent Transportation, Chengdu, China, siyu.tao@okstate.edu, Qiyuan Peng, Qiyuan Peng, Siyu Tao, Siyu Tao, Siyu Tao Transfer hall plays a significant role in transport hub, determining the hub’s operating and evacuation efficient. Legion software is used to create the model and simulate pedestrians’ movement in transfer hall of Shapingba transport hub. According to the simulation result, the existing layout scheme of various facilities in the hall is evaluated, and some possible congestion points are identified. Ultimately, some optimization suggestions, from the view of convenient transfer, are proposed. 2 - Generating Delivery Plans in Real Time by Dynamically Evaluating Multiple Scenarios Yihuan Yang, The Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong, Kowloon, Hong Kong, yyangax@connect.ust.hk, Xiangtong Qi Many Internet sellers now commit a very short lead-time for delivery upon the confirmation of orders. This requires to generate a delivery routing plan in real time. This paper presents a solution framework that dynamically traces the buyers’ ordering process. The core is a genetic algorithm designed to consider multiple possible delivery scenarios at the same time. It generates a pool containing situations which are more likely to happen and adjusts the pool according to current customers’ purchasing progress. With consideration of possible scenarios in advance, a delivery route with low cost can be obtained immediately after the cutoff time. 3 - Branch-cut-price Algorithm for the Traveling Salesperson Problem with Hotel Selection Luiz Henrique Barbosa, UFF, Niteroi, Brazil, luizhsb@gmail.com, Eduardo Uchoa The Traveling Salesperson Problem with Hotel Selection (TSPHS), a natural extension of the classic Traveling Salesperson Problem, has only recently been introduced to the literature. Current approaches focus mainly in metaheuristic techniques. The lack of a robust a exact algorithm limits the extent to which this problem can be studied. This work proposes a sophisticated branch-cut-and-price algorithms, including recent techniques like ng-route pricing and limited memory subset row cuts. Computational results show that medium-sized instances can be solved. 4 - A Matheuristic Approach to the Orienteering Problem with Service Time Dependent Profits Qinxiao Yu, Tianjin University, 92# Weijin Road Nankai District, Tianjin, China, yuqinxiao@tju.edu.cn This paper addresses the orienteering problem with service time dependent profits (OPSTP), in which the profit collected at each vertex is characterized by a nonlinear function of service time, and the objective is to maximize the total profit by determining a subset of the vertices to be visited and assigning appropriate service time to each of them respecting to time budget. To solve this problem, we formulate a mixed integer nonlinear programming model and propose a two-phase matheuristic algorithm. Extensive numerical experiments are conducted for small and large scale instances. The results show that our proposed algorithm can be quite effective in finding good-quality solutions. 5 - A Hybrid Constraint Programming Very Large Neighborhood Search for Vehicle Routing Problems with Multiple Time Windows M’Hallah Rym, Professor, Kuwait University, P.O. Box 5969, Safat, Kuwait, 13060, Kuwait, rymmha@yahoo.com This paper models the vehicle routing problem with multiple time windows as a constraint program, and solves easy instances exactly. It approximately solves difficult instances using a hybrid iterated very large neighborhood search and a constraint programming heuristic. It obtains near-optimal feasible solutions that respect the time windows of the clients and depot as well as the capacity and load constraints of the vehicles. The computational investigation highlights the benefits of the hybridization as well as the cost saving it induces.

328