Informs Annual Meeting 2017

TB80

INFORMS Houston – 2017

6 - Evaluating the Qualitative Aspects of Automotive Dealerships Alongside Traditional Quantitative Metrics Mark Colosimo, Urban Science Applications, Inc., 400 Renaissance Center, Suite #2900, Detroit, MI, 48315, United States, macolosimo@urbanscience.com Quantitative metrics to evaluate automotive dealerships is a practice that has existed and grown over many decades. Now, even more data is available and the number of metrics of Key Performance Indicators is overwhelming to those trying to operate a dealership business. Additionally, these indicators do not tell the whole story. There are many other activities within a dealership that are not measured or require additional research in order to determine their impact on desired outcomes. This study begins to simply what a dealership operator must regularly review to obtain a complete perspective on their operation, while incorporating new analysis via survey results of actual dealership operators. 381B Railway Operations Sponsored: Transportation Science & Logistics Sponsored Session Chair: S Viswanathan, Nanyang Business School, Nanyang Avenue, Singapore, 639798, Singapore, asviswa@ntu.edu.sg 1 - Inspecting Railways for Defects: A Game Theoretic Approach Pedro Cesar Lopes Gerum, Rutgers University, New Brunswick, NJ, pedro.gerum@gmail.com, Melike Baykal-Gursoy Train track inspections are important to repair defects that could cause delays or accidents. Even though most train companies have inspection guidelines, there is little research done on optimal track inspection scheduling policies. Our research addresses this issue by using game theoretic concepts to determine the optimal strategy for inspection crews. We define a number of zero-sum hider and searcher games, each with increased complexity. The higher the complexity, the more accurate the strategy represents the true optimal strategy for the problem; however, the more complex the game is, the more data is needed. Additionally, we provide data analysis on train track defects, exemplifying how the parameters for the games can be defined. 2 - Evolution and Development of Railway Passenger Hub Liwen Wang, Southwest jiaotong University, chengdu, China, siyu.tao@okstate.edu This article studies optimization method of railway passenger hub in the big city of contemporary in China. It reviews evolution stages and its inner dynamic on railway passenger hub in China and abroad, comparatively analyzes the different cases in order to sum up the valuable experience. It shows that railway station in urban centers of China should be both transport nodes and urban places, Urban- Station Complex is its development trend only in necessary given conditions. 3 - Exact and Heuristic Algorithms for the Real-time Train Scheduling and Routing Problem Marcella Sama, Post-Doc, Roma Tre University, 79 via della Vasca This talk deals with the real-time train scheduling and routing problem. The problem is NP-hard and finding a good quality solution in a short computation time for practical size instances is a challenging task. We propose a solution approach based on the relaxation of train routing constraints in the formulation. Working in the relaxed problem allows to quickly compute good quality lower bounds, and represents the first step toward the development of a branch-and- bound algorithm for the overall problem. Computational experiments are performed on several networks and disturbed traffic situations. 4 - Railroad Crew Balancing Model Jaydeep Kumar Chanduka, Masters’ Student, University of Illinois, Urbana-Champaign, 1618 Melrose Park Court, Urbana, IL, 61801, United States, jkc2@illinois.edu In this paper, we are trying to minimize the operational cost of crew balancing in a railroad network by optimizing crew deadhead, crew assignment, and train delay decisions. We assume that there is not always sufficient crew at each terminal for all the trains. Some of the crew must be deadheaded to different terminals while others have to be detained at other terminals to match all the trains. All the major constraints of the railway industry are considered in this paper. Some of them being: mandatory minimum rest, crew ordering, and detention rules. Crew pairing and assignment is a very intensively studied problem in the airline industry but railroad industry has evaded this problem so Liwen Wang, 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 Navale, Roma, 00146, Italy, sama@ing.uniroma3.it, Andrea D’Ariano, Marco Pranzo, Dario Pacciarelli TB79

far. In this paper, the problem is formulated by applying the concept of “virtual crew” and a multi-day time period is studied. The effect of single day model and multi-day model is analysed with one home terminal and a number of away terminals. The results show that multi-day model gives a better performance in terms of objective values and computational time. 5 - Scheduling Trains to Minimize Peak Power and Maximize Regenerative Braking Power Utilization S. Viswanathan, Professor of Operations Management, Nanyang Technological University, Nanyang Business School, Nanyang Avenue, Singapore, 639798, Singapore, asviswa@ntu.edu.sg, Mahendra Birhade, Rohit Bhatnagar We address energy-efficient train scheduling for urban Mass Rapid Rail Systems. We formulate and solve an IP model that aims to minimize both peak power and total energy consumption. Test results on real problem instances obtained from the Mumbai Suburban System show that on average our model reduces the peak power demand by 28% and increase the regenerative braking energy utilization by 7.5 % while limiting the average trip time increases to only 2%. 381C Planning and Operation of Energy and Chemical Hubs with Renewables and Storage Sponsored: Energy, Natural Res & the Environment, Energy Sponsored Session Chair: Kaveh Rajab Khalilpour, Monash University, Melbourne, Australia, kr.khalilpour@monash.edu 1 - Expansion Planning under Long-term Uncertainty for Hydrothermal Systems with Volatile Resources Álvaro Lorca, Pontificia Universidad Católica de Chile, Departamento de Ingeniería Eléctrica, Av Vicuna Mackenna 4860, Santiago, 7820436, Chile, alvarolorca@uc.cl, Benjamín Maluenda, Matias Negrete-Pincetic, Daniel Olivares The significant integration of volatile energy sources in power systems stimulates the use of greater operational details in power system expansion planning models. Motivated by this, we will present a stochastic expansion planning model for hydrothermal power systems, including uncertainty in water inflows and using representative days to capture inter-hourly phenomena such as renewable power production profiles and energy storage. Computational experiments showing the advantages of the proposed model for an actual power system will also be presented. 2 - Enhanced Representative Days and System States Modeling for Energy Storage Investment Diego Alejandro Tejada Arango, Universidad Pontificia Comillas, 23 Alberto Aguilera, Madrid, 28015, Spain, dtejada@comillas.edu, Sonja Wogrin, Efraim Centeno In this work, we analyze the impact of different options to represent the operation decisions in the study of energy storage (ES) investment for long-term planning models. We compare the representative days and the system-states approaches for the representation of these operation decisions. We proposed enhanced versions of these approaches to improve the ES investment approximation. An Spanish case study isevaluated and the results are used to identify the potential profits that energy storage investment can obtain. 3 - A Risk-averse Approach for the Planning of a Hybrid Renewable Energy System Ozlem Yilmaz, Bilkent University, Universiteler Mah, Bilkent/Cankay, Ankara, 06800, Turkey, ozlem.yilmaz@bilkent.edu.tr, Ozlem Cavus, Ayse Selin Kocaman We consider a risk-averse approach to a two stage stochastic programming problem that aims to design a hybrid renewable energy system consisting of solar panels and hydropower stations in a most cost effective way in order to satisfy the demand for energy. We use Conditional Value at Risk (CVaR) as the measure of risk and propose a scenario-wise decomposition algorithm to solve this model. The main source of the randomness is water inflow to the reservoirs. We therefore generate scenarios for inflow using a modified K-NN method developed by Prairie et.al (2006). TB80

333