Informs Annual Meeting Phoenix 2018

INFORMS Phoenix – 2018

WA29

2 - Real-time Package Consolidation for Split Orders in Online Retailing

3 - Post Disaster Assessment Routing Bahar Yetis, Bilkent University, Ankara, Turkey, Buse Eylul Oruc We propose a post-disaster assessment strategy as part of response operations. The arcs and nodes to perform assessment activities on are selected based on the value they add to the consecutive response operations. The model considers motorcycles, which can be utilized under off-road conditions, and/or drones. The first objective aims to maximize the total value added by the assessment of the road segments (arcs) and the second maximizes the total profit generated by assessing points of interests (nodes). An epsilon-constraint method and a heuristic is proposed. To test the mathematical model and the heuristic method, a data set belonging to Kartal district of Istanbul is used. n WA30 North Bldg 221C Computational Problems in Warehousing and Order Picking Sponsored: TSL/Facility Logistics Sponsored Session Chair: Sabahattin Gokhan Ozden, Penn State University, Abington, PA, United States 1 - Wait or Start? A Dynamic Warehouse Order-picking Process Phenomenon Atieh Madani, University at Buffalo (SUNY), Buffalo, NY, 14260, United States, Rajan Batta, Mark Henry Karwan This study focuses on optimizing warehouse order picking process in a dynamic environment. The logic behind this process determines the timing for any picker to start the tour to pick up the items of orders. The number of items in the picking list has the opposite effect on transportation cost and customer waiting time and this phenomenon brings us to one of the challenging problem in warehousing. Our approach to solve this problem involves probabilistic methods such as dynamic programming along with a heuristic on TSP to calculate the expected length of tour for the future periods. 2 - Optimization of a Fast-pick Area in a Cosmetics Distribution Center Alice E. Smith, Auburn University, 3301 Shelby Center, Dept of Industrial/Sys Engineering, Auburn, AL, 36849, United States, Mario Velez Gallego Fast-pick areas are used in warehouses to improve labor efficiency by concentrating picking activities within a compact area, minimizing the distance traveled by the pickers. One problem that must be solved when a fast-pick area is to be implemented is the assignment-allocation problem. This deals with deciding which products should be assigned to the fast-pick area, and how much space should be allocated to these products. This research was motivated by the picking operation of a cosmetics distribution center where several fast-pick areas are in place. A mixed integer linear programming formulation is proposed for solving the variant of the assignment-allocation problem found in this company. 3 - The Aisle Design Problem for Order Picking Warehouses Kevin Gue, University of Louisville, Louisville, KY, 40292, United States, Alice E. Smith, Sabahattin Gokhan Ozden The aisle design problem is to arrange picking and cross aisles in a warehouse such that expected distance to store or retrieve items is minimized. This problem has been addressed for unit-load warehouses, in which workers must visit only one or two locations per tour, but until now not for order picking warehouses, in which workers must visit many locations. We present results of a multi-year quest for the best aisle designs for order picking. Our computational system produced many interesting designs unknown to theory or practice, but surprisingly, none is significantly better than traditional designs. We conclude that current practice is likely the best practice. 4 - A Computational System to Solve Warehouse Aisle Design Problem Sabahattin Gokhan Ozden, Penn State Abington, 1600 Woodland Rd., Abington, PA, 19001, United States, Alice E. Smith, Kevin Gue We develop a warehouse layout optimization system that models layouts, allocates products to storage locations, calculates routing distances, and performs heuristic optimization over a comprehensive set of layout design parameters. The system searches over 19 different design classes simultaneously by using a layout encoding. The system is scalable which means that certain calculations such as routing can be distributed to run on multiple computers to decrease the overall computational time. It can solve optimization experiments or single design assessments in batch. In this way, researchers can create design of experiments in Excel and import it to the system to get the results.

Yuankai Zhang, Dalian University of Technology, Office Room A1214 Innovation Park Building, Dalian, 116024, China Yuankai Zhang, University of Arizona, Tucson, AZ, 85719, United States, Xiangpei Hu, Wei-Hua Lin The order splitting has been one of the main challenging problems for multi-item order fulfillment in online retailing. The key issue to fulfill split orders with fast- delivery options in lower costs is: how to make package consolidation (consolidating and packing items of several sub-orders together) decisions in real- time using capacitated warehouses. We formulate an analytic model and propose a rolling horizon based heuristic algorithm to generate efficient real-time package consolidation schemes for online retailers. 3 - A Benders Decomposition Algorithm for the Fulfilment Planning Problem of an Online Retailer with a Self Owned Logistics System Shuqin Li, Shanghai Jiao Tong University, No.800 Dongchuan Rd., Shanghai, China, Shuai Jia We consider an order fulfilment planning problem in an e-tailing environment. For each planning period, the e-tailer makes decisions for assigning orders to the fulfilment centers and the delivery stations, as well as the decision for shipping orders under a time window constraint. We develop a mixed integer program to minimize the order processing cost and shipping cost. We show that the problem is strongly NP-hard, and propose a benders decomposition algorithm for solving the problem. We evaluate the computational performance of the algorithm on problem instances that are generated based on the logistics network of JD.com, a leading e-tailer in China who operates a self-owned logistics system. 4 - How Survival of Micro Retail Companies Affects Logistics Costs in Large Consumer Packaged Goods Companies Josue Velazquez Martinez, Research Scientist, Massachusetts Institute of Technology, Cambridge, MA, 02139, United States, Ximena Castanon Choque, Christopher Mejia Argueta, Jan C. Fransoo In developing economies, many mom-and-pop stores disappeared per year due to a lack of productivity, while many others appeared because of low barrier of entry. Therefore, the overall micro-retailing market grows per year and represents from 40-70% of demand of large CPG companies. In this paper, we study the impact of these dynamics on logistics costs. We discuss a practical case for a distributor in Mexico City, and by using cost-to-serve estimations and continuous approximation models for routing; we show that by improving survival rate, we may avoid losses in logistics costs up to 30%. n WA29 North Bldg 221B Joint Session TSL/Practice Curated: Patrolling and Service Delivery Problems Sponsored: TSL/Urban Transportation Sponsored Session Chair: Rajan Batta, University at Buffalo (SUNY), Buffalo, NY, 14260, United States 1 - Taking Out the Guesswork: An Analytics Approach to Police Contraband Searches Anthony Bonifonte, Denison University, 3267 Raccoon Valley Rd, Granville, OH, 30106, United States, Tong Zhou Police searches of vehicles stopped for traffic violations can seize contraband and improve public safety, but unnecessary searches are time consuming and can lead to community ill will. Using publicly available data, we develop an analytics approach to assisting officer decisions in whether to search a vehicle or not. Using variables such as time and location of stop, we create an objective estimate of whether a vehicle search will discover contraband. We adopt a reject inference approach to account for having outcome data only for vehicles searched. These findings can improve the use of police time and safety, while simultaneously reducing bias in decisions and improving community relations. 2 - Using the Hypercube Queueing Model for Response to Jobs with Degradation Rajan Batta, University at Buffalo (SUNY), 410 Bell Hall, Buffalo, NY, 14260, United States, Fatemeh Aarabi The problem of scheduling dynamically arriving jobs over a region is considered, with the features of job degradation, travel time, and multiple servers. A modification of the Hypercube queueing model is studied to address job degradation considerations. A 3 server example is presented along with optimization of response areas, server dispatch preferences and idle server management. Numerical results will be provided.

420