2016 INFORMS Annual Meeting Program

WB60

INFORMS Nashville – 2016

WB61 Cumberland 3- Omni Railway Analytics Sponsored: Railway Applications Sponsored Session Chair: Qing He, SUNY Buffalo, Buffalo, NY, United States, qinghe@buffalo.edu 1 - Estimating The Probability And Impact Of Track Defects In Railroad Operations And Maintenance Planning Alexander Lovett, University of Illinois at Urbana-Champaign, alovett2@illinois.edu Slow orders and spot maintenance and are used by railroads to mitigate the impact of track defects and ensure safe train operations between capital maintenance actions. Individually, slow orders and spot maintenance do not appear to have a significant cost impact, but in aggregate, they can reduce efficiency and increase costs over the rail network. This presentation will discuss probabilistic methods to predict when slow orders and spot maintenance will be required to allow for more efficient objective track maintenance planning. 2 - Predictive Switch Health Among all the Communication and Signal (C&S) components, switches cause biggest number of train delays on line of road. In this study, we leveraged cutting- edge Big Data and Predictive Analytics techniques and developed a set of prediction models to assess future switch health. Taking information from multiple data sources, such as Computer Aided Dispatching System (CADS) event logs, switch inspection records, switch incidents records, and many others, these prediction models will enable CSX to proactively maintain our vital assets and better plan C&S workforce. 3 - Data-driven Optimization Of Railway Track Inspection And Maintenance Using Markov Decision Process Qing He, University at Buffalo, SUNY, Buffalo, NY, United States, qinghe@buffalo.edu, Siddhartha Sharma, Yu Cui, Zhiguo Li This paper develops a data-driven condition-based maintenance policy for track inspection. This paper will help in maintaining high service level of the railway tracks which is a difficult task to accomplish. Dataset is two-year track geometry inspection data which contains a variety of geometry measurements for every foot. We employ Markov Chain to model track deterioration, and build a Markov Decision Process for track maintenance decision making and optimize it using value iteration algorithm. By comparing with existing maintenance policy with Markov Chain Monte Carlo simulation, the new maintenance policy developed in this paper can save nearly 10% maintenance costs. WB62 Cumberland 4- Omni Aviation Economics Decision-making Sponsored: Aviation Applications Sponsored Session Chair: Ricard Gil, Johns Hopkins Carey Business School, 100 International Drive, Baltimore, MD, 21202, United States, ricard.gil@jhu.edu 1 - A Profit Maximizing Integrated Model Of Fleet Assignment And Aircraft Routing With Considerations Of Flight Schedule Disturbances Muhammed Sutcu, Assistant Professor, Abdullah Gul University, Airline schedule disturbances are one of the most critical problems due to the unpredictable disruptions such as technical failures and severe weather conditions. In this work, a mixed-integer mathematical model integrating fleet assignment and aircraft routing is proposed to select from among a set of flights and to assign the selected flights to appropriate aircraft for a profitable daily schedule with as minimum delay and idle time in total as possible. The uncertainties of demand, failures of the aircraft and delays arising from adverse weather conditions are also integrated to the corresponding model. To propose a solution methodology for this model is another purpose of this paper. Sumer Campus, Erkilet Bulvari, Kayseri, 38060, Turkey, muhammed.sutcu@agu.edu.tr, Baris Yildiz, Yeliz Yoldas Casey Jen, CSX, Jacksonville, FL, United States, Casey_jen@csx.com, Bob Gutman, Aihong Wen

2 - Efficient And Reliable Package Delivery Path Planning For Drones Mohannad Kabli, Mississippi State University, mrk297@msstate.edu, Sudipta Chowdhury, Mohammad Marufuzzaman The development of efficient and reliable path for drones is becoming crucial in today’s world due to its potential applicability in many commercial purposes. This research focuses on designing an efficient and reliable path planning for the delivery of packages by considering power requirement, collision, altitude, and other related factors into account. The characteristics of the optimal path are expressed in terms of a multi-objective cost function which we solved by using an Adaptive Large Neighborhood Search (ALNS) heuristic. 3 - A Continuum-approximation Approach To Optimize Routing Application of drones in various sectors is becoming common place day by day, and it has got huge potential in humanitarian logistics. This research pertains to optimization of logistics management of drones under extreme events. The key decisions investigated in this study is where to locate the transportation centers, how to assign demand points at each transportation center, and what should be the inventory policy such that the total network cost is minimized. Continuous Approximation (CA) approach is used to solve this problem. As a test bed for computational experiments, coastal region of Mississippi is selected due to its long history of getting affected by various natural disasters. WB60 Cumberland 2- Omni Routing Optimization Problems Sponsored: TSL, Freight Transportation & Logistics Sponsored Session Chair: Ahmed Ghoniem, Associate Professor, University of Massachusetts Amherst, 121 Presidents Dr., Amherst, MA, 01002, United States, aghoniem@isenberg.umass.edu 1 - A Branch-and-cut-and-price Algorithm For The Generalized Vehicle Routing Problem Mohammad Reihaneh, University of Massachusetts Amherst, Amherst, MA, 01002, United States, mreihaneh@som.umass.edu, Ahmed Ghoniem We consider the Generalized Vehicle Routing Problem in which customers are partitioned into mutually exclusive clusters, each with a specific demand. The goal is to construct cost minimizing tours such that exactly one customer is visited in every cluster, subject to vehicle capacity constraints. The proposed specialized branch-and-cut-and-price algorithm compares favorably against state-of-the-art exact algorithms in the literature and closes several open benchmark instances. 2 - A Two-level Optimization Approach For Robust Aircraft Routing And Retiming Mohamed Haouari, Qatar University, mohamed.haouari@qu.edu.qa We address the robust aircraft routing and flight retiming problem, and we propose a two-level solution strategy that embeds a simulation-optimization procedure within an evolutionary algorithm. The proposed approach requires inserting buffer times prior to the flight departure times in order to improve the robustness of both aircraft and passengers connections. We present the results of extensive computational experiments that were carried out on a set of real data. 3 - Resource Constrained Arc Routing For Snow Plowing Joris Kinable, Carnegie Mellon University, Pittsburgh, PA, United States, jkinable@cs.cmu.edu, Willem-Jan Van Hoeve, Stephen F Smith This work considers a Resource Constrained Arc Routing Problem for snow plowing, a fundamental problem faced by many cold-weather cities. In RC-ARP, routes for a heterogeneous set of vehicles must be computed such that they collectively cover a network of streets, while adhering to various resource (salt) usage and replenishment constraints. We contrast exact and heuristics approaches, as well as a decomposition method. The performance is demonstrated on real-world data from the city of Pittsburgh, PA. 4 - Vehicle Routing Problems With Drone Delivery Decisions For Drones Under Extreme Events Sudipta Chowdhury, Mississippi State University, sc2603@msstate.edu, Adindu Emelogu, Mohammad Marufuzzaman, Linkan Bian

Ahmed Ghoniem, University of Massachusetts Amherst, aghoniem@isenberg.umass.edu, Mohamed Haouari, Mohammad Reihaneh

We study a Vehicle Routing Problem with drone delivery. In this setting, a customer is directly visited by a vehicle or his/her demand is indirectly delivered from a neighboring customer using a drone. A mixed-integer formulation is presented along with a branch-and-price algorithm. Alternative solution approaches are investigated for the column generation pricing subproblem and computational results are presented.

416