Informs Annual Meeting 2017

SD32

INFORMS Houston – 2017

4 - Real-Time Optimization of Personalized Assortments Negin Golrezaei, University of Southern California, 812 S.Marengo Avenue, Unit 6, Pasadena, CA, 91106, United States, golrezae@usc.edu Motivated by the availability of real-time data on customer characteristics, we consider the problem of personalizing the assortment of products for each arriving customer. Using actual sales data from an online retailer, we demonstrate that personalization based on each customer’s location can lead to over 10% improvements in revenue compared to a policy that treats all customers the same. We propose a family of index-based policies that effectively coordinate the real- time assortment decisions with the backend supply chain constraints. We allow the demand process to be arbitrary and prove that our algorithms achieve an optimal competitive ratio. In addition, we show that our algorithms perform even better if the demand is known to be stationary. Our approach is also flexible and can be combined with existing methods in the literature, resulting in a hybrid algorithm that brings out the advantages of other methods while maintaining the worst-case performance guarantees. 5 - Velocity-based Storage and Stowage Decisions in a Semi- automated Fulfillment System Our research focuses on a new semi-automated operating architecture of an order fulfillment system where the inventory is stored on mobile pods, moved by robotic drives, and picked manually by stationary operators. Our work is one of the first research efforts to optimize the design and operational processes of a real world large-scale semi-automated storage system. We proposed novel and effective operational policies for the storage (where to return a pod upon the completion of a pick or stow operation) and stowage (in which zone and on which pod to store the receiving inventory) decisions. We further use simulations to verify our findings with real world data provided by our industrial partner. Rong Yuan, Massachusetts Institute of Technology, 2014, 235 Albany St., Cambridge, MA, 02139, United States, rongyuan@mit.edu

4 - Decomposition Heuristics for the Last-mile Delivery Problem Andrea Arias, Pontificia Universidad Catolica de Valparaiso, Valparaiso, Chile, Andrea.arias@ttu.edu Andrea Arias, Texas Tech University, Lubbock, TX, United States, Andrea.arias@ttu.edu, Ricardo Gatica, Jimena Pascual, Timothy I. Matis We consider an extension of the Vehicle Routing Problem in which a single Unmanned Aerial Vehicle (UAV) is used to support a delivery truck in the last- mile delivery process. In this problem, the objective is to determine a coordinated tour for both the truck and the UAV, such that all the packages are delivered at a minimum cost. It is assumed that once the UAV is launched from the truck to deliver a package, the truck continues visiting other clients and meets the UAV in a later position of the tour. We propose alternative decomposition-based heuristics for the problem and show results on preliminary computational tests. 351C MAS Tutorial: Thomas R. Willemain Sponsored: Military Applications Sponsored Session Chair: Greg H. Parlier, North Carolina State University, 255 Avian Lane, Madison, AL, 35758, United States, gparlier@knology.net 1 - Three Unexploited Ideas in Simulation Analysis Thomas R. Willemain, Smart Software, Inc, 2520 Peters Lane, Niskayuna, NY, 12309-2413, United States, trw@smartcorp.com Some promising ideas have not yet been added to the simulation canon. I review three such under-appreciated ideas. Idea 1: Using bootstrap methods to generate realistic inputs from a single sample of actual system operation. Idea 2: Creating malicious simulation inputs to test design resilience. Idea 3: Using transient behavior to compute rapid estimates of steady state behavior. SD33

SD32

351B Drone Delivery Systems - IV Invited: InvitedDrone Delivery Systems (tentative title) Invited Session

SD34

351D Studies and Scheduling Sponsored: Military Applications Sponsored Session

Chair: Andrea Arias, Texas Tech University, andrea.arias@ttu.edu 1 - A Dynamic Programming Approach for the Traveling Salesman Problem with Drone

Niels Agatz, Rotterdam School of Management, Erasmus University, Burg. Oudlaan 50, Rotterdam, Netherlands, nagatz@rsm.nl, Paul Bouman, Marie Schmidt

Chair: Andrew Oscar Hall, United States Military Academy, 86 B Patridge Pl, West Point, NY, 10996, United States, andrew.hall@usma.edu 1 - Nato Systems Analysis Studies Panel

A promising new operating model is the use of a delivery truck that collaborates with a drone to make deliveries. This model gives rise to a new planning problem that we call the Traveling Salesman Problem with Drone (TSP-D). In this contribution, we propose an exact solution approach based on dynamic programming and present experimental results of different dynamic programming based heuristics. 2 - Optimal Path Construction and Routing of Drones for Inspecting Disaster-affected Regions Sudipta Chowdhury, Mississippi State University, 260 McCain Engineering Building, ISE Department, Starkville, MS, 39762, United States, sc2603@msstate.edu, Sandip Kumar, Mohammad Marufuzzaman, Linkan Bian Aerial drones are one of the most promising and powerful new technologies to improve disaster response and relief operations. This study introduces a heterogeneous fleet routing problem for drones, by considering multiple charging stations and respecting operational requirements. First, paths are constructed for the drones by obeying the restrictions set by the Federal Aviation Agency (FAA), technological, geographical, and environmental factors into account. We then propose a hybrid metaheuristic to solve the heterogeneous fleet routing problem. A case study, using three disaster-prone coastal counties in Mississippi, is developed to visualize and validate the algorithmic results. 3 - GPU Enhanced Shortest Path Finding for an Unmanned Aerial Vehicle Roksana Hossain, University of Calgary, Calgary, AB, Canada, rhossain@ucalgary.ca, Sebastian Magierowski, Geoffery G. Messier Situated robots like unmanned aerial vehicles (UAVs) need to decide the sequence of multiple destinations. The traveling salesman problem (TSP) is used for identifying an efficient sequence of goals. For real-time mission-critical operations, such decisions need to be made quickly with sufficient accuracy. In this report, to enhance the computational execution rate without incurring excessive system cost, a graphics processing unit (GPU) is used. A genetic algorithm and clustering algorithms is used to solve the TSP in an aerial network delivery context. The implemented GPU code works 4.8 times faster than serially implemented code and the algorithm can solve large problems with 4000 waypoints.

Richard F. Deckro, Professor of Operations Research, Air Force Institute of Technology, Afit/ENS; Bldg 641, 2950 Hobson Way, Wright Patterson AFB, OH, 45433-7765, United States, richard.deckro@afit.edu, Timothy Povich The NATO Science and Technology Organization’s System Analysis and Studies (SAS) Panel fosters co-operative research, exchange of information, and the advancement of science and technology among the NATO Nations in the field of operations for defense and security. This presentation provides an overview of the SAS Panel’s mission, goals and operations. Selected examples of previous studies will be provided. 2 - Kalman Filtering for Artillery Fire Shift Michael Bendersky, Holon Institute of Technology, 52 Golomb St Holon, Holon, 58102, Israel, michael.bendersky@gmail.com, Israel David Firing Shift or transfer of Fire is the shifting of artillery fire from one target to another with the application of corrections determined from the adjustment on the first target to the initial firing data for the second. In introducing a linear model for artillery fire accuracies - the parameters of which are easily obtainable from extant Firing Tables - we propose a new algorithm for making the shift, based on a static Extended Kalman Filter. It assumes concurrent fire adjustments on multiple auxiliary targets. Numerical examples are included, which attest to the potential of the approach.

112