2016 INFORMS Annual Meeting Program

WC60

INFORMS Nashville – 2016

WC60 Cumberland 2- Omni Prescriptive Analytics for Transportation and Inventory Sponsored: TSL, Freight Transportation & Logistics Sponsored Session Chair: Wesley Dyk, Revonos, Lone Tree, CO, United States, wesley.dyk@revonos.com Co-Chair: Alexander Engau, University of Colorado Denver, NA, Denver, CO, United States, alexander.engau@ucdenver.edu 1 - Stochastic Liquid Inventory Management For Transportation And Logistics Alexander Engau, Associate Professor, University of Colorado Denver, Denver, CO, United States, aengau@alumni.clemson.edu, Wesley Dyk This talk presents a recent case study with an independent energy provider in the Denver-Julesburg Basin in Northeastern Colorado, one of the largest natural gas deposits in the United States. In contrast to related exploration and production operations in other geographic locations, for this present case, the inherent logistics of liquid products including crude oil and waste water is limited to distribution and transportation primarily using trucks which require the careful management of their generally uncertain load inventories and qualities. These challenges will be addressed in some detail and their current solutions using stochastic optimization outlined for questions and discussion. 2 - Network Flow Relaxations For Batched Transportation Models Wesley Dyk, Revonos, wesley.dyk@revonos.com, Alexander Engau We present a new approach to solve mixed-integer batched transportation models using a new class of network flow relaxation for actual use in real-time applications. This approach has been implemented in actual practice for inventory management and liquid hauling of crude oil and waste water using trucked transportation in the petroleum exploration and production sector to give Anirudh Subramanyam, Carnegie Mellon University, Pittsburgh, PA, United States, asubramanyam@cmu.edu, Frank Mufalli, Jose M Pinto, Chrysanthos E Gounaris It is fairly common during the tactical planning of multi-period vehicle routing operations that customer requests arrive dynamically over a planning horizon. This is the case in settings where customer service is provided by appointment. In such settings, the dispatcher must incorporate enough flexibility in the routing plan to accommodate potential customers who have not yet called in to place their order. Failure to do so may result in insufficient fleet capacity to serve future requests. We shall describe an adaptive robust optimization approach to solve this problem and illustrate the tradeoffs in using an adaptive strategy over a static strategy in terms of computational time and cost savings. 4 - Analytics For Logistics Supply Chain: An Overview And Research Outlook M. Ali Ulku, Rowe School of Business, Dalhousie University, Halifax, NS, Canada, ulku@dal.ca There are myriad of opportunities, and thereby challenges, for research and education in supply chain management. This talk comprises of an overview of the current literature and practices and an outlook as to how companies can employ analytics for logistics supply chains so as to gain a competitive edge. WC61 Cumberland 3- Omni Yard and Terminal Operations Sponsored: Railway Applications Sponsored Session Chair: Tyler Dick, University of Illinois, Urbana, IL, United States, ctdick@illinois.edu 1 - Switch-it; A New Yard Modeling Tool Chip Kraft, Transportation Economics & Management Systems, Inc., ckraft@temsinc.com SWITCH-IT simulates any type of rail yard, tracking each car movement as trains arrive and depart and cars are sorted, inspected, repaired and flat switched. Command scripts give the user total control over the simulation. The model directive insights to decision makers in short timeline scenarios. 3 - Adaptive Robust Optimization For Tactical Vehicle Routing Problems

supports fixed block-to-track assignment, dynamic block “swinging” and matrix sorting; pull and cut, shove and couple, and push back to clear from either end of the yard. It simulates lead confliction, crossovers and ladder tracks. Animation uses a “console” presentation which accepts, but does not require detailed track network coding. Its simple yet powerful approach to modeling complex yard operations provides quick feedback and a simple user interface. 2 - Animation Of Switch-it Analysis Using Anylogic Roger William Baugher, President, TrAnalytics, LLC, 100 Villamoura Way, Johns Creek, GA, 30097, United States, rwbaugher@aol.com In an earlier presentation, the features and capabilities of the SWITCH-IT program were described and demonstrated. While the program provides an animation capability, a richer animation environment was desired. Using a newly developed interface, data were fed to AnyLogic’s Rail Library, resulting in new and enhanced animation capabilities, which will be described and demonstrated. 3 - Blocking Capacity, Throughput Volume And Level Of Service In Hump Classification Yards Tyler Dick, University of Illinois, ctdick@illinois.edu This presentation summarizes simulation research on fundamental hump classification yard capacity relationships. In particular, simulation experiments have been conducted to quantify the relationship between level of service and throughput volume of railcars as a function of the number of blocks assembled in the yard. 4 - Boston North Terminal District: Infrastructure Demand Modeling And Improvements Bradford Kippen, University of Illinois, Urbana, IL, 61801, United States, kippen2@illinois.edu Boston’s Tower A Interlocking connects all of the MBTA’s North Side commuter rail lines with North Station terminal and has over 260 daily scheduled trains. Proposed replacement of the Charles River Drawbridges, critical interlocking appliances, catalyzed development of Infrastructure Demand Models which quantify minute to minute demand for infrastructure at key locations given a proposed plan of operations and conditional infrastructure constraints. This presentation outlines the project background, modeling methodology, proposed improvements, and potential applications. WC63 Cumberland 5- Omni New Objectives and Solution Concepts in Location Analysis Sponsored: Location Analysis Sponsored Session Chair: Dmitry Krass, University of Toronto, 105 St. George St., Toronto, ON, M5S 3E6, Canada, krass@rotman.utoronto.ca 1 - Correct Discounted Costs In Hub Location Models Vladimir Marianov, Pontificia Universidad Catolica de Chile, marianov@ing.puc.cl, H.A. Eiselt, Armin Luer-Villagra Frequently, traditional single allocation hub location models incorrectly apply discounted costs to trip legs without enough traffic to economies of scale to apply. Conversely, discounts are not applied to some legs with high traffic. We develop a model and a heuristic to correct this flaw. The assumptions are the same as in the usual models. Computational experience is presented. 2 - Multi-objective Single Median Models Joyendu Bhadury, University of North Carolina-Greensboro, NC, j_bhadur@uncg.edu, HA Eiselt This talk will focus on one median problems in location theory with multiple objectives. The goal is to find Pareto-Optimal Set of solutions in each case. Solved models will be discussed and new ones will be presented along with their applications in practice.

446