Informs Annual Meeting 2017

MA50

INFORMS Houston – 2017

MA50

MA51

361C Modeling and Analysis of Emerging Mobility Services II Sponsored: TSL, Urban Transportation Sponsored Session Chair: Yafeng Yin, University of Michigan, Ann Arbor, MI, 48109, United States, yafeng@umich.edu Co-Chair: Yu Nie, Northwestern University, Evanston, IL, 60208, United States, y-nie@northwestern.edu 1 - Operational Strategies for Electrical Car-sharing Systems Yang Liu, National University of Singapore, Department of Industrial & Systems Engineerin, Singapore, 117576, Singapore, iseliuy@nus.edu.sg, Dong Zhang, Shuangchi He For one-way electric vehicle (EV) car-sharing model, we proposed two operational strategies to improve its performance.The aim of two operational strategies is to utilize the battery of EVs in a more efficient manner. Two novel network flow models are proposed based on these two strategies, respectively. To satisfy the quick response requirement in the operational level, we further propose an efficient solution algorithm to solve the models. Experimental results reveal that the performance gap between the EV car-sharing system and the traditional car-sharing system can be very small if efficient operational strategies are implemented. 2 - Modeling the Operations of Electric Autonomous Taxis in New York City Liang Hu, PhD Student, Iowa State University, Ames, IA, United States, lianghu@iastate.edu, Jing Dong This paper explores the potential of replacing conventional taxis with electric autonomous vehicles (EAVs) in New York City. Customer requests, in terms of pick-up and drop-off locations with time stamps, are extracted from GPS tracked taxi trip data. Local interaction rules are designed to dispatch EAVs to serve customer requests and coordinate their charging activities. The simulation results show that using EAVs can improve taxi operational efficiency and reduce fleet size. 3 - Mobility Pattern Analysis of Free-floating Bike Sharing and Insights on System Operations Yu Zhang, University of South Florida, Tampa, FL, United States, yuzhang@usf.edu, Aritra Pal We study historical biking trajectory data of bike sharing users for understanding their mobility patterns and the correlations with environment variables and the interactions of those variables. The outcomes provide insights on system design and operations, e.g. hub regions of bike sharing program, time for conducting static or dynamic rebalancing, and time available for such rebalancing, etc. 4 - Same-day Delivery Planning with Store Fulfillment Ming Ni, University at Buffalo, The State University of New York, 326 Bell Hall, Buffalo, NY, 14260, United States, mingni@buffalo.edu, Qing He, Jose Luis Walteros, Xuan Liu, Arun Hampapur This study identifies a seasonal order fulfillment plan for delivering local online orders from nearby retailing stores that minimizes the overall panning and operational costs. It aims to develop optimization models and solution algorithms about store location selection, fleet-sizing for transportation, and inventory planning. Our methodology integrates outer search tree based on mix integer programming and local branching, optimization-based algorithm for initial lifting constraints, with accelerated Benders decomposition method to solve large-scale real supply chain planning problem. 5 - Optimal Design of Urban Taxi Fleet Electrification: A Simulation- Based Optimization Approach Fang He, Tsinghua University, Shunde Building, N502, Tsinghua University, Beijing, 100084, China, fanghe@tsinghua.edu.cn, Yinghao Jia, Zuojun Max Shen Transportation management agencies in urban cities are electrifying taxi fleets. How to allocate limited resources to optimize the benefits of taxi fleet electrification deserves thorough investigation. This research gathers the real-time vehicle trajectory data of 39,053 urban conventional taxis and 408 suburban electric taxis in Beijing to extract the customers’ travel demand distribution and electric taxi drivers’ behavior pattern. We derive a data-driven simulation model to reflect the operations of electric taxi fleet and further propose a simulation- based optimization framework to optimize the deployment of charging infrastructure and the configuration of electric taxi fleet.

361D 8:00 - 8:45 IBM/8:45 - 9:30 Palisade Invited: Vendor Tutorial Invited Session 1 - Generic Callbacks in CPLEX, A New Approach Xavier Nodet, IBM, 1681 Route des Dolines, Valbonne, 06600, France, xavier.nodet@fr.ibm.com During this tutorial, you will learn how to use a new Generic Callback framework in CPLEX Optimizer. These callbacks allow users to benefit from all the CPLEX features without restriction, and offer a simple API to either get information from CPLEX, or drive CPLEX behavior. 2 - Quantitative Risk Analysis in Excel with @RISK Rafael Hartke, Palisade Corporation, Ithaca, NY, United States, rhartke@palisade.com This tutorial will guide you in the use of @RISK for analyzing historical data, and for making forecasts and better decisions in an uncertain business environment. @RISK is part of Palisade’s DecisionTools Suite, and runs within the familiar MS Excel environment. It provides all the features you need to quantify and understand risks, along with graphical capabilities and reports to help you present the risk analysis results to a non-technical audience 361E Energy Contributed Session Chair: Yue Zhao, Stony Brook University, Stony Brook, NY, United States, yue.zhao.2@stonybrook.edu 1 - Distributed Transactive Operation of Micro-grid Cluster using a Stochastic Model Predictive Control Algorithm Yang Chen, University of Illinois at Chicago, Chicago, IL, United States, ychen429@uic.edu, Mengqi Hu To enable efficient transactive operation among multiple interconnected micro- grids (aka micro-grid cluster), a bi-level distributed decision framework is proposed. A two-stage stochastic optimization model is developed for each micro- grid at the subsystem level with consideration of solar and energy load uncertainties, while a guided particle swarm optimizer is employed at the system level to coordinate transactive operations for all the micro-girds. A stochastic model predictive control algorithm is employed to solve the two-stage stochastic model for real-time operation. A case study of a cluster of four micro-grids is developed to evaluate the effectiveness of the proposed approach. 2 - Multi-period Energy Procurement Policies with Deferrable Demand and Supplementary Uncertain Power Supplies Tian Wang, Zhongnan University of Economics and Law, Number 182, Nanhu Road, Donghu District, Wuhan, 430073, China, wangtian3261@gmail.com We analyze a multi-period energy procurement problem for an energy aggregator who is responsible for centralized controlling of energy procurement and consumption. To decide the optimal procurement amount, the aggregator makes trade-offs between two supplies, traditional energy with variable prices vs. free renewable energy with uncertain supplies. We show the optimal procurement policy is to procure traditional energy only when the price is less than a certain threshold, which depends on the statistics of the day-ahead real time prices, the wind energy distribution and the time left to the end of horizon. Numerical studies with real wind data are conducted. 3 - Self Scheduling of a Virtual Power Plant in Energy and Reserve Electricity Markets Ana Baringo, Universidad de Castilla-La Mancha, Ciudad Real, Spain, Ana.Baringo@uclm.es, Luis Baringo, José Arroyo We propose an approach for the self scheduling of a virtual power plant that participates in both energy and reserve electricity markets. The virtual power plant comprises flexible demands, a conventional power plant, and a wind-power unit, which participate in the markets as a single entity to optimize the use of their energy resources. Uncertainties in wind-power production and reserve deployment are modeled using confidence bounds, while uncertainties in market prices are modeled using scenarios. This allows us to formulate the self-scheduling problem as an adaptive robust optimization problem, which is efficiently solved using a column-and-constraint algorithm. 4 - A New Approach to Estimate Renewable Energy Potential in a Particular Location Tunde O.Aderinto, Texas A&M.University-Kingsville, Kingsville, TX, United States, tunde.aderinto@Students.tamuk.edu, Hua Li The renewable energy potential of a particular location is usually estimated through independent resource assessment of different renewable energy sources. MA52

150

Made with FlippingBook flipbook maker