Informs Annual Meeting 2017

TD60

INFORMS Houston – 2017

TD58

TD59

362E TSP/VRP Generalizations Sponsored: TSL, Facility Logistics Sponsored Session Chair: Ibrahim Capar, Bowling Green, OH, United States, icapar@bgsu.edu Co-Chair: Burcu Keskin, bkeskin@cba.ua.edu 1 - A Steiner Zone Heuristic for the Close-Enough Traveling Salesman Problem Xingyin Wang, Singapore Univeristy of Technology and Design, 8 Somapah Road, #05-101, Singapore, 487372, Singapore, xingyin_wang@sutd.edu.sg, Bruce L.Golden, Edward Andrew Wasil In the close-enough traveling salesman problem (CETSP), a truck starts at a depot and serves all customers on a Euclidean plane before it returns to the depot. The truck only has to get close enough to a customer to make a delivery. The objective is to minimize the total distance traveled by the truck. The CETSP has real-world applications including meter reading. We develop a heuristic based on Steiner zones that identifies regions close enough to multiple customers and connects them by the shortest route. When compared to existing algorithms on benchmark instances, our heuristic produces results of comparable quality with much shorter computation times. 2 - The Mothership and Drone Routing Problem Stefan Poikonen, University of Maryland-College Park, 3406 Tulane Dr., Apt 22, Hyattsville, MD, 20783, United States, spoikone@math.umd.edu, Bruce L.Golden The mothership and drone problem (MADP) considers the use of two vehicles: a large “mothership” and a small “drone”. It is assumed both mothership and drone are free to move by the Euclidean metric. There exists a set of targets (“buoys”) that must be visited. The drone is launched from the mothership, visits a buoy, and returns to the mothership. The original objective is to minimize the total time required to visit all buoys and return the drone to the mothership. A branch and bound method is used and a second order cone program defines a lower bound at each node. Modified objective functions and a “close-enough” variation for signals collection will also be presented. 3 - Dynamic Discretization Discovery Approach for Deliveryman Problem with Time Windows Minh Duc Vu, Quinlan School of Business, Loyola University Chicago, 16 E.Pearson St, Chicago, IL.60611, Chicago, IL, 60611, United States, dvu3@luc.edu, Mike Hewitt In this talk, we study the Delivery Man Problem with Time Windows (DMPTW), a problem belonging to the Traveling Salesman Problem with Time Windows (TSPTW) and its variants. Unlike TSPTW and its counterpart Delivery Man Problem (DMP), research works for DMPTW are quite scarce. We present the first exact method based on Dynamic Discretization Discovery approach and partially time-expanded networks to solve the problem. Experiments on a set of 310 instances indicate that the algorithm is capable of finding good feasible solutions with 251 optimal or close-to-optimal solutions. Based on insights from the computational experiments, we conclude with future research directions. 4 - Attractive Orienteering Problem with Proximity and Timing Interactions Ibrahim Capar, Bowling Green State University, Department of ASOR, 355 Business Administration Building, Bowling Green, OH, 43403, United States, icapar@bgsu.edu, Nickolas K. Freeman, Burcu B. Keskin We consider an attractive orienteering problem for planning entertainment events that seeks to determine the profit maximizing tour among a fixed set of candidate locations over a fixed planning time. The profit contribution for each stop depends on a time-varying measure that accounts for the attraction of the event’s location and the presence of events nearby, in location and time. We formulate an MIP model and develop two heuristics; local search and scatter search. The heuristics find optimal solutions for small test instances and for large instances, they provide a 32% improvement over the best feasible solution returned by a commercial solver. 5 - Demand-resilient Electric Vehicles Sharing System Design Xin Wang, University of Wisconsin-Madison, 8916 Red Beryl Drive, Middleton, WI, 53562-4278, United States, xin.wang@wisc.edu, Yikang Hua, Dongfang Zhao, Xiaopeng Li Electric vehicle sharing advances the green city development, but faces challenges in practical operations like infrastructure deployment and vehicle rebalancing. We build a multi-stage stochastic model to reduce the customer queueing under demand-shock scenarios. Our model incorporates uncertain customer demand arrival together with the daily fleet dispatch and rebalancing decisions.

362F New Directions for Product Assortment and Recommendation Sponsored: Revenue Management & Pricing Sponsored Session

Chair: Clark C. Pixton, Massachusetts Institute of Technology, Massachusetts Institute of Technology, Cambridge, MA, 02139, United States, cpixton@mit.edu 1 - Long Tail or Blockbuster: Product Variety in Industries with Network Effects Ming Hu, University of Toronto, Rotman School of Management, 105 St. George Street, Toronto, ON, M5S.3E6, Canada, ming.hu@rotman.utoronto.ca, Yinbo Feng We provide a theory unifying the long tail theory and blockbuster phenomenon. 2 - Online Product Demand and Strategy in the Presence of Social Influence Clark C. Pixton, Massachusetts Institute of Technology, 292 Vassar Street, Apt D-5, Cambridge, MA, 02139, United States, cpixton@mit.edu We explore the demand patterns observed in product categories of online retailers, and add to the modeling literature explaining these effects. We provide managerial implications for new product introduction and demand estimation. 3 - Analytics for a Hotel Property: Demand Prediction and Optimization of Pricing and Capacity Allocation Decisions Rui Sun, Massachusetts Institute of Technology, 77 Massachusetts Avenue, MIT.Bldg. E18-436, Cambridge, MA, 02139, United States, ruisun@mit.edu, Eduardo Candela, David Simchi-Levi We present our work with a hotel company as an example of how machine learning techniques can be used to improve the demand predictions of a hotel property as well as its pricing and capacity allocation decisions. We first build a random forest model to predict the number of price-sensitive bookings and then feed these predictions into a mixed integer program to optimize prices and capacity allocations. We prove that the mixed integer program can be equivalently solved as a linear program. Finally, numerical results show that the optimized decisions are able to generate an increase in revenue compared to the historical policies. 4 - Assortment Optimization under a Mixture of Multinomial Logits with Nested and Interval Consideration Sets Jacob Feldman, Olin Business School, United States, jbfeldman@wustl.edu, Huseyin Topaloglu We study assortment problems when customers choose under the multinomial logit model with various consideration set structures. In this choice model, there are multiple customer types and a customer of a particular type is interested in purchasing only a particular subset of products. We use the term consideration set to refer to the subset of products that a customer of a particular type is interested in purchasing. We develop novel approximation schemes for the assortment problem under nested and interval consideration set structures.

TD60

370A Improving STEM Education Sponsored: INFORMEd Sponsored Session

Chair: Kingsley Anthony Reeves, University of South Florida, 4202 E Fowler Avenue, ENB118, Tampa, FL, 33620, United States, reeves@usf.edu 1 - Practices to Improve Engineering Engagement in Rising 9th Graders: A Case Study on Advanced Manufacturing Justin Yates, Francis Marion University, Florence, SC, 29501, United States, jyates@fmarion.edu, Karen Fries, Meredith Love, Tracy Meetze-Holcombe, Matthew Nelson This project is a collaboration among The School of Education, the Industrial Engineering Program, and the Center of Excellence for College and Career Readiness at Francis Marion University. Curriculum and student exercises were created for a one-day short course targeted at improving engagement in rising 9th graders with risk factors that inhibit their pursuit of college. We present pre and post data and discuss the short-term outcomes of this short course. We also discuss novel ways in which students from Industrial Engineering and the School of Education supported the planning and implementation of this short course.

397