2016 INFORMS Annual Meeting Program

WD82

INFORMS Nashville – 2016

2 - Heuristics For The Covering Capacitated Vehicle Routing Problem Christopher John Wishon, Arizona State University, Tempe, AZ, United States, cwishon@asu.edu, Guilherme Sproesser Ferreira, J. Rene Villalobos The covering VRP is a variant of the traditional VRP in which a vehicle can satisfy a customer’s demand by visiting one out of a set of predetermined alternate locations within the network. This variant is motivated by many practical applications including the routing of mobile retailers, urban bus lines, and emergency vehicles. In this work, the Greedy, Sweep, Savings, and Ant Colony Heuristics are adapted to solve the covering VRP. These techniques are employed to solve nearly 200 test instances and the results are compared to any known optimal solutions. These results demonstrate the superiority of the Ant Colony and Savings methods. 3 - New Formulations For The Hub Interdiction Median Problem Prasanna Ramamoorthy, IIM Ahmedabad, Vastrapur, Ahmedabad, 380015, India, prasannar@iima.ac.in In this work, we present new formulations for the hub interdiction median problem. We present a bi-level formulation and explore reducing it to a single level using KKT(Karush - Kuhn - Tucker) conditions and closest assignment constraints. We present new closest assignment constraints for the problem in addition to the existing ones in literature (Lei 2013). The new closest assignment constraints has some interesting properties which aids in solving the problem efficiently. We also present computational results highlighting the efficiency of the closest assignment constraints and the proposed model in solving the problem.

4 - The Phantom Inventory Menace: The Effect Of Unobserved Stock-outs On Lost Sales Fredrik Eng-Larsson, Postdoctoral Associate, MIT Center for Transportation and Logistics, Boston, MA, United States, frengl@mit.edu, Daniel Waymouth Steeneck Based on retail audit data, we find retail inventory records are an unreliable indicator of out-of-stock (OOS) events. Worse yet, the records are rarely validated or corrected. As a result, many OOS event are never observed in the inventory record. We estimate the impact of the unobserved OOS events on lost sales via a novel demand estimation technique, which accounts for inventory uncertainty in the presence of scarce inventory record validation. 5 - Warehousing And Shipping Decisions For An Online Replacement Parts Retailer Kamal Lamsal, Assistant Professor, Emporia State University, Campus Box 4039, 1 Kellogg Circle, Emporia, KS, 66801, United States, klamsal@emporia.edu, Amit Kumar Verma Online replacement parts sellers offer big selection of products. These products are held in several locations. With every incoming order, the seller must decide whether to split the various items from one order or not and from where each item will ship. We work with an online OEM parts retailer which competes on customer service level. The retailer charges a flat shipping fee per product and promises a delivery time. We develop an Approximate Dynamic Programming (ADP) based algorithm that makes shipping decisions by minimizing the on hand shipping cost plus estimate of future shipping costs. We use the lessons from the exercise to decide which parts should be located in which warehouse. WD82 Broadway G- Omni Multicriteria Decision I Contributed Session Chair: Yuji Sato, Professor, Chukyo University, 101 Yagotohonmach, Showa, Nagoya, Aichi, 466-8666, Japan, ysatoh@1988.jukuin.keio.ac.jp 1 - Financial Decision-maker’s Preferences Modeling Within Goal Programming Model Belaid Aouni, Associate Dean, Qatar University, College of Business and Economics, Al Jamiaa Street, Doha, 2713, Qatar, belaid.aouni@qu.edu.qa Goal Programming (GP) model has been applied to financial portfolio selection problem where several conflicting and incommensurable attributes are simultaneously aggregated, such as return, risk and liquidity . The aggregation of the conflicting attributes requires some compromises from the Financial Decision- Maker (FDM) based on his/her preferences. The aim of this paper is to present a new typology of the FDM’s preferences modeling within the GP model for financial portfolio selection. 2 - Rethinking Sfpark’S Demand Response Pricing Tayo Fabusuyi, Research Associate, University of Michigan, 5520 Baywood Street, Floor #3, Ann Arbor, MI, 15206, United States, Fabusuyi@umich.edu, Robert Hampshire In an effort to eliminate circling and reduce parking search time and cruising, SFpark, an innovative demand-responsive pricing program was implemented by the City of San Francisco. Over a two year period, the program was piloted across seven San Francisco neighborhoods made up of 256 distinct parking blocks. The evaluation of the pilot program has however met with mixed reviews particularly with regards to the relationship between price changes and occupancy levels. These issues are addressed by employing a non-dominated sorting genetic algorithm approach from which figures of merit are generated that allow for an objective comparison between treatment and control blocks. 3 - Assets And Liabilities Management Within An Integral Risk Framework For A Microfinance Institution Tayo Fabusuyi, Numeritics, Pittsburgh, PA, Contact: Tayo.Fabusuyi@numeritics.com While the microfinance industry has recorded some success in providing financial services to the poor, it has also attracted criticism with regards to the quality of services offered and the lack of a robust risk framework observed across the industry. Using data over a period of a decade (2006-2016), our study addresses these concerns by employing a multiple-criteria decision analysis which provides management with a menu of measures of risk-return tuples that maintains fideli- ty to the bank’s non-financial constraints. The approach is enumerated using Grameen Bank as a case study.

WD80 Broadway E- Omni Retail Mgt I Contributed Session

Chair: Kamal Lamsal, Assistant Professor, Emporia State University, Campus Box 4039, 1 Kellogg Circle, Emporia, KS, 66801, United States, klamsal@emporia.edu 1 - Efficient Workforce Scheduling In Retail Stores Considering Overtime Or Part-time Workforce

Peeyush Pandey, Doctoral Student, Indian Institute of Management, Indore, IIM Indore, Indore, 453331, India, f12peeyushp@iimidr.ac.in, Hasmukh Gajjar, Bhavin J. Shah

This paper deals with the problem of identifying optimal workforce size and their schedule to satisfy hourly and daily requirements of workers at the retail store. We consider a multiple day planning horizon divided into periods of equal length for the retail store that caters to daily requirement of customer and offers a wide variety of product. An optimization model is proposed for both overtime and part- time workers considering different shift lengths and different lunch and tea breaks during the working shift. A heuristic presented in the paper guarantees to obtain optimal number of workers and their schedule. 2 - Analyzing Impact Of Cardinality And Similarity Context-effects On Assortment Optimization Uzma Mushtaque, Rensselaer Polytechnic Institute, 110 8th St, Troy, NY, 12180, United States, uzmamu@rpi.edu, Jennifer Pazour Assortment planning problem in an online retail environment explicitly modeling no-choice behavior is presented. Difficulty of selecting an item under cardinality- context effects and the attraction due to underlying utility are modeled as the fundamental driver’s behind opting for no-choice. Optimality conditions are developed for cardinality context effects and for an interaction of cardinality with similarity effects. Three different algorithms exploit the structure of the optimal solution under these two conditions. 3 - Assortment Planning And Replacement Under Attractiveness Decay Effect Huiqiang Mao, City University of Hong Kong, Tat Chee Avenue, Kowloon, Hong Kong, 000, Hong Kong, huiqiangm@gmail.com, Yanzhi Li Inspired by assortment renewal strategies of many industries, we consider the capacitated assortment planning and replacement problem, where the attractiveness of the product decays over time. Based on Locational Choice Model, we characterize the structural properties of the optimal assortment. We also consider the assortment replacement problem, where the retailer is allowed to update the assortment over the horizon. We explore conditions under which the retailer should employ this replacement strategy.

481