2016 INFORMS Annual Meeting Program

MD72

INFORMS Nashville – 2016

4 - Robust Allocation Of Testing Resources In Reliability Growth MohammadHossein Heydari, University of Arkansas, Fayetteville, AR, United States, mhheydar@uark.edu, Kelly Sullivan Reliability growth testing seeks to improve system reliability by identifying and removing failure modes. Recent models maximize system reliability by allocating limited testing resources across the system’s components, each of which exhibits reliability growth according to the AMSAA model (Crow, 1974) with known parameters. We extend this research to solve a robust version of this problem in which AMSAA parameters are uncertain but assumed to lie within a budget- restricted uncertainty set.

3 - A Multi Stage Heuristic Solution For A Consultant Assignment And Routing Problem

Andrew Junfang Yu, Associate Professor, The University of Tennessee - Knoxville, 411 B. H. Goethert Pkwy., MS 19, Tullahoma, TN, 37388-9700, United States, ajyu@utk.edu, Brett Shields

In this study, a heuristic algorithm is developed for a novel Vehicle Routing Problem (VRP), namely the Heterogeneous Fleet Vehicle Routing Problem with Multiple Depots, Fixed Visits, and Priority Matching. The heuristic is four stage, in which the first stage is a conversion of skills and priorities to one set, second a listing of consultants is constructed based on the items in step one, third a fixed visit based assignment is given considering the availability of consultants and clustered demand, and lastly an optimal routing for each consultant is determined. 4 - Two Echelon Location Routing Problem In The Presence Of Third Party Logistics Khosro Pichka, PhD Student, S. B. Lubar School of Business, 3202 N. Maryland Ave., Milwaukee, WI, 53211, United States, kpichka@uwm.edu, Amirsaman Hamzeh Bajgiran, Xiaohang Yue, Jaejin Jang, Matthew Petering This paper addresses a two echelon location routing problem (2E-LRP) considering third party logistics. A 2E-LRP is a variant of the classical LRP; however, it seeks to find a set of vehicle routs in two echelons. First, from a single depot to a set of possible satelites and second, from the opened satelites to customers. In spite of the large amount of research on LRPs, the 2E-LRP has received very little attention from researchers. A mixed integer program (MIP) formulation is proposed, and a hybrid heuristic algorithm is developed to solve its medium and large size problems. Milk-run deliveries of product to customers depend on both inventory levels and customer proximity to one another. We describe a system designed to improve delivery economies through careful assessment of the trade-offs between stopping costs and mileage costs. We use k-means clustering to great groups of “like” customer in both geography and inventory depletion dates. We use a novel “furtherst neighbor” heuristic to then sequence deliveries. We discuss the application of this approach at a US manufacturer. MD72 Bass- Omni Supply Chain Mgt VIII Contributed Session Chair: Nathalia Hernandez, Universidad del Norte, Km 5 Vía Puerto Colombia, Barranquilla, AA1569, Colombia, inhernandez@uninorte.edu.co 1 - An Econometric Analysis Of Omnichannel Retailing Using Data Analytics Serkan M. Akturk, Texas A&M University, 4217 TAMU, Wehner 320 M, College Station, TX, 77843-4217, United States, makturk@mays.tamu.edu, Michael Ketzenberg The advent of omnichannel technologies, like ship-to-store capability, enable integration of both physical and electronic marketplaces and thereby offer consumers shopping flexibility and enhanced services. For a retailer, these capabilities require significant investment in order to deliver a seemless shopping experience, yet hold the promise of greater customer loyalty, higher marketshare, and increased competitiveness. We assess this promise by analyzing the impact of ship-to-store capability on a retailer’s performance using transactional data from a national retailer. 2 - Determining The Optimal Level Of Supply Chain Forecast Coordination With Bidirectional Option Contracts Seong-Hyun Nam, Professor, University of North Dakota, 2517 Sand Hills Avenue, Grand Forks, ND, 58201, United States, snam@business.und.edu We have studied bidirectional option contracts associated to demand uncertainty. By analyzing how the buyer’s procurement and supplier’s production decisions are affected by the level of the coordination/collaboration of demand forecast under bidirectional option contracts, we have found that the optimal procurement for buyer and optimal production strategy for supplier to maximize supply chain surplus is associated with the level of forecast collaboration under bidirectional option contracts. 5 - Inventory-based Delivery Scheduling And Routing Michael F Gorman, Professor, University of Dayton, 300 College Park, Dayton, OH, 45469, United States, michael.gorman@udayton.edu

MD70 Acoustic- Omni Transportation, Rail I Contributed Session

Chair: Ginger Yi Ke, Memorial University of Newfoundland, Faculty of Business Administration, Memorial University of Newfoundland, St. John’s, NL, A1B 3X5, Canada, gke@mun.ca 1 - Passenger Centric Railway Planning Yousef Maknoon, EPFL, Lausanne, Switzerland, Yousef.maknoon@epfl.ch, Tomas Robenek, Stefan Binder, Michel Bierlaire Traditional railway systems use operations research techniques to optimize their system. Most of the time such mathematical models take only operators’ cost and benefit into account. However, passengers’ decision as the main users of such systems play an important role in the performance of such networks. We introduce a state of the art methodology that takes both sides (demand and supply) into account in timetable design as well as recovery plans for trains. 2 - The Locomotive Planning Problem With Alternative Train Speeds An Integrative Environmental Perspective Ginger Yi Ke, Memorial University of Newfoundland, Faculty of Business Administration, Memorial University of Newfoundland, St. John’s, NL, A1B 3X5, Canada, gke@mun.ca, Kan Fang, Manish Verma We develop multi-objective models for the locomotive planning problem to minimize the GHG emissions and the operational costs. The amount of GHG emitted by a locomotive depends on the train speed, train load, and the locomotive type, etc. We determine the assignment of locomotives to each train, so that sufficient power can be provided to pull the trains; we generate train schedules that meet customer demands within different time windows, and determine the number of active, deadheading, light traveling locomotives, and the train-to-train connections (e.g. consist busting) for each train. Real-life railway systems are used for numerical experiments to provide additional managerial insights. Chair: Michael F Gorman, Professor, University of Dayton, 300 College Park, Dayton, OH, 45469, United States, michael.gorman@udayton.edu 1 - An Optimization Method Of Petrol Distribution Considering Drivers’ Workload Balance Lijun Sun, Associate Professor, Dalian University of Technology, 2 Linggong Road, Ganjingzi District, Dalian City, 116023, China, slj@dlut.edu.cn, Haiyang Shi We present a multi-objective optimization model of petrol distribution considering drivers’ workload balance based on Adams’s equity theory. A heuristic algorithm is designed based on the NSGA-II algorithm. Computational results show that the proposed algorithm can generate a non-dominated solution set corresponding to distribution plans that perform well both in the total distribution cost and drivers’ workload balance. 2 - Two Level Logic-based Benders Decomposition For Optimal Service NetworkDesign Amin Hosseininasab, Carnegie Mellon University, Pittsburgh, PA, United States, aminhosseininasab@gmail.com, Fatma Gzara We present a new continuous time network and model for the service network design problem (SNDP). SNDP addresses the planning of operations for freight transportation carriers. Given a set of requests to transport commodities from specific origins to specific destinations with given availability and delivery times, SNDP determines a continuous movement of vehicles. We propose a two-stage Logic-based Benders decomposition and develop several valid cuts for SNDP. Numerical results show the benefits of the continuous time approach and the effectiveness of the solution methodology. MD71 Electric- Omni Vehicle Routing I Contributed Session

231